Victorian Boombox Mk III. A brand new start.

[J. Wilhelm]

While I figure out how to build a pump for the COVID mask, I decided to revisit the boombox project. The though occurred to me to play with the main board of the Mk I boombox to see if I could connect it to the new horn speakers. Otherwise comparing the old boombox to the new requires I compare amplifiers as well.

The one catch is that the old boombox was not working well. I needed to shake some connectors between the tone control and the amplifier to make it work. On Sunday, I took the extraordinary step of taking apart the old boombox to see if I could bypass the old cables (11 conductor bundle of wires). I've been avoiding that chore for years. The difficulty was on obtaining a replacement 11 pole plug and resoldering a rusted out bundle of hair thin wires. But the issue turned out to be not the plug or the wires. The jack on the main board was broken away from the motherboard, and some solder joints broke I didn't find out until I opened the large Subwoofer box.

Waking up an old giant

You see, I had purchased a second Altec Lansing Subwoofer at the second hand shop. That was my "nuclear option" for this project, in case I could not make the horns work. Sadly, the new Altec Lansing box has a blown amplifier. Naturally, since I only paid $6 for it. But taking it apart revealed how simple the circuit is inside. And it gave me the idea to go ahead and take apart my old boombox.

It's basically a small motherboard built around a TDA 7375 hi-fi amplifier. It's got a two-stage active low pass filter built with Op-amps. Pretty basic stuff. The box is empty otherwise, besides the Subwoofer unit and it's a ported design (Bass Reflex). It's only about 35 watts or so, basically the same as the tiny Sony stereo I show at the start of the thread, except this is an analog, not digital amplifier, which means that the power source of the Altec Lansing is much bigger and much heavier (but probably sounds better). In this case old school is better than new school. On the other hand, the Altec Lansing will not be a match for my Sony tabletop 100 W amp I'm using as a testbed.

In the meantime, I will order the inductors from Parts Express for use as a passive low filter. I'm just tired of trying to make my own inductors. It's so difficult to get the right materials now. What I may do, however is to follow Altec Lansing's example and just build an active Subwoofer amplifier. I'm so impressed to see how small they made that amp.

[J. Wilhelm]

So the year has turned. Having reached a certain point with the COVID respirator, I'm now ready to revisit this project. It has been patiently waiting for me.

I'm about to take a turn on this project. The difficulty in getting the right parts for the 4th order crossover for the subwooofers made me think about exploring a new direction. I can continue finishing the passive crossover, but last year, I stumbled upon an offering by Parts Express for ready-made amplifiers with a subwooofer's stereo output. No doubt, the subwooofer output will be some sort of high-order low pass filter, and it will feature a separate volume control. Now the offering in question revolves around a handful of relatively new on-chip digital Class D amplifiers.

I've already expressed my misgivings on digital amplifiers. They're not the best or the cleanest sound makers you can get. But what they will do is drastically reduce the weight and power consumption of the system. Since I had set out to use a digital Sony tabletop stereo from the start, this solution below is not that different:

40 W per channel + 68W subwooofer amp. Based on ST mosfet power amplifiers hooked up to
a stereo TDA7266S 5.5 W pre-amp, and a mono TDA7266S for the subwooofer

This is an unknown brand, made in China, with all of the risks and benefits that entails. At $40 USD, it gives you 40w per channel plus 68W for the subwooofer output (monaural with connections for two speakers. What makes this an attractive option is that the crossover frequency is adjustable from 50 Hz to 200 Hz. And I'm thinking that it makes it a snap to bring the crossover frequency to the 100 Hz range or lower which would let me fine-tune the subwooofers to component the main Sony drivers.

There are other options, like a version with Bluetooth + FM Tuner for $43 USD, but I would have to finish the 4th order passive crossover as it doesn't have the tuned sub output. Plus the one shown above is on clearance - without the power supply, for a mere. $20 USD.

Me thinks I've found a good answer, at least for testing. I will have to judge the quality in person, but if it works, it might be fun to compare to the venerable Mk. I (below), whom I may say, looks as chipper as ever. Renewed, the old man is donning boxing gloves and ready to prove it can beat anyone.

This might be fun!

Victorian Boombox Mk. I
The old man is bullying his counterpart. "What's taking so long?!"  
he asks with a cigar between his teeth on the ring.. "Are you afraid to lose some teeth"?

The underdog, newcomer Victorian Boombox Mk. III

[Banfili]

Ooo, er! Nice, very nice!

[J. Wilhelm]

Ooo, er! Nice, very nice!

I'll give it a try. I have a 3 amp 12v supply that I can use for it, so it's only about $20 - total $28 with tax and shipping.

I've been reading the reviews on it. Not fantastic, there are some notable shortcomings. First, the volume knob for the subwooofer output is entirely independent from the volume of the stereo output. Also, you need a 14 volt supply (which they don't sell, to allow you to extract the maximum power without clipping. So the output will be a bit less than the 40W.

Many reviewers complained that the subwooofer volume knob was very "non linear" slamming from zero volume to max volume in about 1/10th of a full turn. And that the rolloff of the frequency response is too shallow (-2dB per octave to be of any use.

So it's no surprise that the volume controls are separate. I'm struck by how the reviewers online missed the fact that there is a built in blue LED next to the knob that flashes every time the amp doesn't have enough voltage and clips the waves. If they had a little imagination they'd know that the LED is used to trim the subwooofer volume down using that fidgety control knob. So that control knob is not a volume at all, but rather a fine tune limiter to keep the base from clipping. Not one reviewer online figured that out, and they just complained about the clipping and the non-linear volume knob . That should be a screwdriver type adjustment, not a volume knob. The other volume knob is adjusted exactly the same way. This was originally an automotive amplifier, of the type you hide in the boot (trunk) of your car.

So I'm still going to buy it despite its shortcomings. The real volume knob should be at the source of sound, not the amplifier. The reason I like it is because the power supply is very light (a computer monitor power supply) and the amplifier has 3 separate channels.

One thing that I discovered today is that having a single channel for the subwooofer allows me to reduce the size of the inductors. I've had a terrible time trying to make or buy a pair 11mH inductors, with my latest price tag of about $50 to buy the whole set 5mH + 11mH from Parts Express, or make my own for $30 or so.

It turns out that with the two subwooofer speakers in parallel, the impedance (resistance) of the load on the 3rd channel goes down to 2.2Ω which is listed on the amplifier specs as being acceptable. When you do that, basically you chop the inductance values roughly in half. That means a single set of 11mH and 5 mH are half the size and half the number of inductors I need for a two channel subwooofer. So less than 1/2 maybe 1/3 of the cost (cross fingers) of the stereo subwooofer components (L1 and L2 below). The value of the capacitors is doubled, but that's fine because I have two sets of capacitors which can be connected in parallel, no problem. It's worth a try.

[J. Wilhelm]

Aaaargh! They always have a way to extract more money from you, don't they? Ive been looking at the cost for the components to finish ONE subwoofer channel. It's basically $25 + shipping, just as I suspected.The reason is that I will have to buy at least one more capacitor, as I just found out that the capacitors they sent me had the wrong value. As it stands the circuit for the 3rd channel would look like this:

and I need to buy this:

Adding a resistor to the two speakers doubles the resistive impedance (or equivalently I can use just one speaker), which reduces the value of the capacitor that I need to 500μF -- in other words I only need to buy one capacitor because that's the largest value they sell   I could buy two capacitors instead, but at $8 a pop, it adds up quickly. Another advantage of increasing the resistance of the drivers besides saving me another $8 is so that the amp won't get blown by accident. 2Ω is too low already, though the amp should be able to handle two 4Ω woofers in parallel (as that is the usual rating for most woofer drivers), and the single channel is made from connecting in parallel two channels).  I can make a resistor from any piece of round steel bar I have in the garage.

I will have to address the "non-linear" volume knob for the sub woofer output, as this is a major complaint in the revues for the amp. One idea is to open the amp and replace what I assume is a variable resistor pot with a linear slider resistor, like that used in old school equalizers. Otherwise, there are ways to mechanically extend the travel of the little pot by using a tiny lever, or a gear train attached to a rack and  pinion mechanism as well.  Oh, Lego, where are thou? I think that's be fun to add some slider mechanisms to the speaker enclosure

The project is getting expensive again    Oh well. I think my roommate owes me some cash from some hardware and paper towels. The amp is already on its way. I tried looking for pre-assembled low pass filters or crossovers to see if that would be cheaper, but it'ss very rare to find crossovers stronger than -12 dB/octave and I want at least -12 dB so the little woofers are well isolated.

[J. Wilhelm]

Sometimes it feels like you're taking two steps forward and one step back. After receiving said amplifier, and knowing in advance it had serious shortcomings I made an assessment of what I could do with it.

The shortcomings/characteristics are as follows

1. The amplifier indeed needs more power to prevent clipping. Simple math (Ohm's Law) will show that there is no way the amp can produce the 40+40W + 60W sub output. Simply no way with a 12V PC monitor power supply. The testing reveals that the amp for the subwooofer will start clipping right away if you increase the volume past "barely audible."

To their credit the blue overload indicator is very helpful in determining how much volume you can produce before the amp starts clipping. The cause for the clipping is very simple: there is not enough power coming from the power supply. It needs more than a 3A source of power and more than 12V, which is what they're providing when they sell you the amp. Data sheets from the manufacturer indicate that the power source should provide at least 5 amperes of current at 14V. But simple math again reveals that even those specs are wrong. By my estimate, you need a 10 or 12 ampere 13.5 volt source. In other words, a car battery. Makes perfect sense now. This is an automotive amplifier disguised as a hi fi amplifier. The online reviewers failed to see that (most of the ones I saw).

2. After some research, I found some DIY folk who had opened the carcass of the amp to see "what was wrong." Most online revuers had noticed that the volume for the sub was very non-linear, and also that the crossover frequency adjustment knob seemed to do nothing. They couldn't tell if the crossover was at 100 or 500 Hz. Some bright people figured out that the amplifier was connected to a low pass active preamplifier built around an Op-Amp, and they matched a generic circuit diagram to the printed circuit board from China.

For those of you who don't know what an "Op-Amp" is (Operational Amplifier), it's basically a voltage-only amplifier made from a large number of transistors on a tiny chip. The circuit is called op-amp because in the mid-late 20th century you could make analog computers that could add and subtract voltages with these chips, to reproduce mathematical equations. They have a very linear response and thus can be used to make anything from low noise pre-amps, to oscillators and many other useful devices. They don't provide any significant power, but because they're very linear and work with very small currents, they can "operate" on the voltage very cleanly, amplifying any signal (with basically zero noise) to the voltage you want, before you pump it to power transistors/vacuum tubes/MOSFETs, or all in one power amplifiers chips that produce the actual power to drive the speakers. They're awesome little gizmos from yesteryear.

What these DIYs found out is that probably some capacitors which form the low pass filter had the wrong value. This caused the crossover to have a very high cutoff frequency, which meant the subwooofer output was not a subwoofer output at all, passing most frequencies untouched to the subwoofer drivers, and early versions of the amplifier sold in the US had that problem... Luckily for my unit, it seems they fixed the problem. I do have a fairly significant low frequency output for the subwooofer. So I guess they fixed their capacitor supply issue   That still doesn't explain why the amp is clipping, though. The DIYs had assumed that the lack of a real low pass filter was overloading the subwoofer amp. They were actually wrong on that. They were right about the crossover, but my amp still clips at partial power. I will assume that's entirely because of the inadequate power supply, since that is what the math tells me.

So where does that leave me? I have a 3 channel amplifier that clips the 3rd channel like Edward Scissohands, and only gives me a clean output at very low volume. The front channels seem a lot better, though.

In looking for a solution, I could perhaps build an appropriate crossover for the speakers, and spend about $50 for a set capacitors and inductors to make a **passive** subwoofer, part of the speakers themselves, as I planned in the previous pages.

The dark circles at the ends of the horns are the subwooofer drivers

Or I can build an active low pass filter which will have much smaller components because the Op Amp in the circuit has a feedback loop that can "multiply" the effect of electronic components, such as capacitors. Better yet, the only thing I need is capacitors and zero inductors. The downside is that I will have to always use a separate amplifier to drive the subwooofer drivers.

So it occurs to me that I can use the two channels in the small amp that seem to work fine without clipping. The idea is to build a second order low pass filter pre-amp and feed that to the amplifier input. A downside of this method is that the Op Amps I have use a positive and a negative voltage source on the form of two 9v batteries. Not ideal, but in the past I've built very high quality surround sound decoders (analog) and pre-amplifiers using this method.

A second order (-12dB /octave) active low pass filter

Scrounging around in my old toolbox, I found the necessary components to make a second order active crossover with an 80 Hz Cut-off. Based on the an LM 1480 dual Op Amp chip, I need 4 capacitors and 8 resistors of much lower value than before I considered this option (0.002uF, and 1MOhm). I can't change the crossover frequency to something higher, because I only have the resistor and capacitor values to build an 80 Hz crossover. But if you remember, that was about the right frequency for the cutoff I needed in the first place, before I started playing with giant inductor coils and having a hard time buying "audio grade" capacitors I needed to order online.

These components have been waiting patiently in my toolbox for about 30 years

This time I have all the components at hand, and the know-how to put them together. In theory, I should even be able to hook the pre-amp to the speaker wires coming out the stereo source. So the volume control knob on my stereo controls both the subwoofer and the main speaker output. That would be an ideal solution. So I may be able to have say at least a usable 15 + 15 W or so **stereo** subwoofer output.

Let's see how it goes!

[J. Wilhelm]

Eureka!

I've assembled the circuit and tested it. It works flawlessly. Finally I have a crossover. It still is not a 4th order (-24 dB / Octave) low pass filter, but it is a 2nd Order filter and if I ever want a 4th order filter, you can just daisy chain two of these circuits together. The advantage for me is that I can divorce myself from having to buy $50 King Kong size electronic components every time I want to build a speaker set.

C1=C2=C3=C4= 0.002μF, R1=R2=R3=R4=1MΩ, very high resistor value (read text), f_c= 89Hz, Second Order -12dB

The active filter has a fixed gain of two, which is rather irrelevant but helps boost the input signal a bit when you use line level inputs. I ended up finagling a breadboard out of a piece of craft wood, because I just don't have the patience to wait for another month to get something so basic (the "Internet Era   just like mail ordering from the Sears catalog in the old Wild West and waiting for Wells Fargo carriage to deliver!). The circuit has extremely high resistance values. I need to make them high to make the capacitor values go low enough to use ceramic capacitors (less than a micro Farad), which I had in my tool box. The cutoff is a very appropriate 89 Hz for the sub output.

One thing about using high value resistors is that they severely limit the current going through the circuit. That's actually good for two reasons:

1 High Rs keep the noise low in Op Amp applications, and even though it doesn't matter here, it extends the flat frequency response of the amplifier so you can get a truly flat curve over the audible range.

In the past, I designed a tiny pair of Binaural microphones that you fit into the entrance of your ear canal, to make 3D sound recordings, using your external ear (pinna) to impart the proper reverberations to sound that allow animals to discern the 3D direction of sound (it's explained by a science called Psychoacoustics, and the reason I know about that, is that it was developed in the aerospace industry before hitting the Hi Fi and gaming markets when I was just coming out of high school in the late 80's / Early 90s).

For the binaural microphones, you need a truly flat frequency response, and to do that I took a similar circuit and used 100 kΩ, which for the LM741 and  LM 1458 gave me a response curve flat way beyond the audible range to less than -3B at over 100 kHz, so from 20 Hz to 20kHz is was truly flat. The microphones were excellent, for Binaural recordings, or otherwise. The sound was ridiculously clean. So this circuit benefits from that cleanliness and flatness in the sound.

2. An Op Amp already has a very high input resistance by design. It's purpose is to *sample* the input voltage, but not necessarily pass much current across the circuit. The output current is provided by the battery/power source, so its a good way to insulate one device from another . The high resistance values further help to insulate circuits belonging to whatever you plug to the input/output of the filter, as well as maintain noise isolation between left and right channels in the front end of the circuit.

The low pass filter/active crossover works with a dual power source with two 9V batteries. These batteries can be replaced with two tiny transformers besides the power source of the amplifier used. The reason for the dual system is because of the way the transistors amplify the signal. If you have only one power source from say 0V to 9V, then the output of the circuit will have to be from 0 to 9V, and if the input sound oscillates between say -3 V to + 3v, the output signal will be "elevated" between 0V and 6V. Not necessarily a problem but it complicates the circuits if you want to take that signal as use it as input for some other generic device.

The other issue is that using two 9V sources (from -9V to +9V) gives you a maximum amplitude of 18 V for you output sound wave, which is plenty wide to avoid clipping; I keep thinking that maybe the clipping of the Lepai sub-woofer output is because of a poorly designed Op-Amp preamplifier in their low pass filters. It could be. I'm not necessarily wanting to open the box and corroborate, but I'm just having a hard time believing that someone would be so dumb as to mess up a simple Op-Amp circuit, which is what the DIYs online were suggesting - It's either that of the power source is too weak as I explained in a previous post (actually both phenomena are technically the same problem if you see it correctly - peak to peak voltage in the input signal) - but the evidence is that the designers of the Lepai messed up a pretty basic circuit, worthy of Radio Shack's Forrest M. Mims III "Mini Notebook." And I know, because I just built my low pass filter, using Forrest M. Mims specs, and it works **perfectly** WITHOUT clipping or distorted noise!!!

So what's next for this circuit? Testing. Initial tests are revealing some unwanted oscillations in my wooden speaker cabinets. The subwoofers really shake things up and I need to see if those subwoofer drivers (the black ones) I scrounged from Goodwill are up to the task. They may not be. Heck, the stupid Lepai amplifier, may not be up to the task either, but at least I know that my active crossover idea works. If I need to, I can go hunt for better drivers, this time order new ones from Parts Express. With all the money I'm saving for not buying the King Kong size inductor coils, I can actually afford a pair of decent subcompact subwoofer units!! Or another amplifier!!

[MWBailey]

Glad to hear you're getting it all figured -  Can't help wondering what you'll do to steam the new one up...

[J. Wilhelm]

Glad to hear you're getting it all figured -  Can't help wondering what you'll do to steam the new one up...

Progressively as I encountered more issues the aesthetic side of the project vanished away, and I just let the acoustics dictate the look of the horn. The last idea I had was to copy the style of those Alphorns. And then later I considered carving some design on the horn.

I could in theory go back to the phone case and tablet days, and use Walnut or Mahogany stain and a Copperplate decoration with the pipe holders and copper wire, etc. But somehow I'm not feeling it. To begin with, the surface area to be covered is gigantic this time. And the weight added by such a style would make it much heavier. Plus I want to avoid the "glue a cog and call it Steampunk" phenomenon. Sound vibration is a major issue now with the woofer and sub woofers (not that it kept me from making the Mk. I, though).

The other thing I could do is try to use inlays. Like, different type of wood and metal wire inlays. That could have the effect of adding the ethnic Alphorn look, and might be easier than trying complex carving into the surface of the pine planks. I'm currently exploring those issues.

I still have a lot to do functionally, though.  I have a low-pass filter that I can build up in modules to tweak the sound of the subwooofers. I'm not convinced that the little Samsung subwooofer drivers I got from Goodwill are up to the task. I'm looking at increasing the subwooofer output - it's low right now with the Lepai amp, which produces less power than advertised even if I say the wattage has to be less than the margin of 12V X 3A = 36W = 18W + 18W. Assuming a 90% efficiency, typical of digital amplifiers, you're talking no more than 16W + 16W believable power for the subwooofer units alone. Not a problem exactly for listening to classical music at mild volumes, but it falls short of the old Victorian Boombox, and worse, I believe the wattage rating of the Samsung subwooofers is even lower than that, probably 8W or so if I look at the tiny circuits that powered them.

So I'm looking at tweeking the amplifier, or even building my own, and buying a pair of tiny subwooofers from Parts Express for $21 each with much better specs and closer to 50W rating.

[Miranda.T]

I really like your inlay idea; I think it will look fabulous on the boombox and be perfect for both the scale of the build and enhancing the vintage feel of the design. So is the compass motif one you are thinking of adopting? I must say it is very striking. For the electronics, I can see the attraction of the custom build, giving you precise the control over the output. Also good fun designing & building the circuit and very satisfying when it bursts into life.

Yours,
Miranda.

[J. Wilhelm]

I really like your inlay idea; I think it will look fabulous on the boombox and be perfect for both the scale of the build and enhancing the vintage feel of the design. So is the compass motif one you are thinking of adopting? I must say it is very striking. For the electronics, I can see the attraction of the custom build, giving you precise the control over the output. Also good fun designing & building the circuit and very satisfying when it bursts into life.

Yours,
Miranda.

Thank you dear Miranda. I think your crystal ball is right on the money. I don't have to do an inlay, per se, but I can do a veneer work to achieve the same effect. I'm inclined to get rid of the hinges and assemble it like a sandwich, with straight boards on top and bottom, keeping the system as a single piece. The front would be subject to inlays or veneer with craft wood.

As far as the amplifier, the Sony is feeding the blue woofers and tweeters and the Lepai is just driving the subwooofers. There are some serious issues with the Lepai. It still distorts a lot - not clipping this time, but inexplicably produces low volume and drives the Samsung subwooofers to severe distortion. Since I was using one the Samsung subwooofers at moderate and even high volumes before without issues, then it stands to reason that the Lepai is the primary source of trouble.

I think the Lepai amplifier produces very high voltages and the average voltage must be positive, above zero volts at all times. This causes the Samsung coils to be displaced constantly (not in neutral position at zero signal) which leaves very little space for the paper cones to move before they start flexing and distorting the sound. This is a very serious design error, and considering the other design error in their active low pass filter (Op Amp), which I think is unforgivable, I'm wondering exactly *who* or *what* designed that amplifier. After re-reading the description given to me by the DIY folks who opened the Lepai, apparently these were two first order filters, each built around a single Op Amp cascaded together before the output went into the amplifier. Then they used a hi-fi TV grade low power amp (TDA), before connecting the output and building more watts with MOSFETs. Nothing wrong with the latter stage, that is common with power amplifiers, and designers prefer to build power with MOSFETs at the end rather than use a single chip power amp. All good, but when you complicate the circuit with too many stages you also introduce many sources of error, from design errors (which I think this is) through to problems with tolerances of components, etc.

My design approach is modular and keeping each stage as simple as possible. Chips help to prevent errors. The last power stage should be as monolithic as possible, so I don't introduce fatal errors and noise, and things like that. Cleanliness is godliness in audio circuits, IMHO. So what I envision is a front end with one LM 1458 chip - dual Op Amps for the left-right low pass filter connecting to a monolithic power amplifier chip, and another 1458 Op Amp for tonal control (if not included in power amp chip - otherwise just use the 1458 as buffer to equalize input impedance with subwooofer section). Perfect symmetry. Each Op Amp chip will feed to its own power amplifier. Both power amplifiers being identical. So half the power goes to the Sony drivers and the other half goes to the Timphany subwooofers. Equal power, equal pre processing.  50 Watts to the Sony, 50 Watts to the Timphany. And I'll toss the Lepai out to the he garbage bin where it belongs.

[Miranda.T]

Sounds (no pun intended) great on both counts; form and function equally perfected. I'd definitely agree on the modular approach, much easier to debug. Apologies if I've missed this, but do you have access to an oscilloscope? If by any chance not not a company called PICAXE are offering their digital scope (outputs to a PC, free software) for currently under £15 (although shipping from the UK would add to this of course). I think I might order one for myself...

Yours,
Miranda.

P.S. if the oscilloscope was of interest it's at https://www.picaxestore.com/picaxe/project-kits/project-kits/kit120

[J. Wilhelm]

Sounds (no pun intended) great on both counts; form and function equally perfected. I'd definitely agree on the modular approach, much easier to debug. Apologies if I've missed this, but do you have access to an oscilloscope? If by any chance not not a company called PICAXE are offering their digital scope (outputs to a PC, free software) for currently under £15 (although shipping from the UK would add to this of course). I think I might order one for myself...

Yours,
Miranda.

P.S. if the oscilloscope was of interest it's at https://www.picaxestore.com/picaxe/project-kits/project-kits/kit120

Thanks for the tip. No! I don't have one! I've been only doing spectrum analysis and tone generation with my phone! I'll look into it!

No question. I have blind faith on the Op Amp circuits, because I've built so many in the past, so I can just order the parts, but the amp from scratch is a bit trickier. What I may do is find a pre-assembled amplifier board. Most boards nowadays are based on the digital variety of chips. If I get lazy and decide to go that way, I'd be looking at a 25 or 50W/channel (two boards) built around the TDA7492 amp chip. It's got good reviews.

The one thing that I must absolutely do, either way is to get mea good power supply, a 4A laptop power supply at least. These amps seem to work better above 20V I just found a 19V, 4.74 (90W) laptop power supply, and it works with the Lepai. I'll be re-testing the Lepai to see if the two "front" channels sound better. I'm still probably going to build a 4 channel TDA7492 X 2 amplifier, though.

[J. Wilhelm]

I'm laughing because this guy is jury rigging his power supply, well at least he has an oscilloscope.

Class D TDA7492 stereo amplifier board test & review

But this review is kind of helpful. The consensus is that the TDA 7492 is a halfway decent chip for an inexpensive amplifier. Naturally this board is on backorder at Parts Express. If I can get a decent pair of boards this might make a nice amplifier. Otherwise it'd be making it from scratch and re-inventing the wheel.

https://store.sure-electronics.com/product/AA-AB32165

And this review tells you exactly what kind of problems you can expect with cheap boards bought online. Luckily there is a quick fix. This review not only involved the same amplifier chip, but the Peerless speakers as well!! Wow. Really good tips. I also remember using SPICE a really really long time ago (which makes me a fossil of sorts).

http://www.trevormarshall.com/class-d-tutorial/

Let's see what I can do.

[J. Wilhelm]

Exciting times. I've been extracting some really beautiful sounds from the horn at low volumes. I found this laptop transformer that I had supplying 90W of power at 19V and 4.74A. At 90% efficiency in these digital amps, that translates to a margin of 81W, or 40.5W per subwooofer driver. The pair of Samsung subwooofers is a stereo subwooofer system, which is basically unheard of in conventional systems, but it was fun to corroborate that you can indeed hear a difference between left and right channels at around 70Hz or so (drums in Ravel's "Symphonie Espagnole" ). I talked a bit about stereo subwooofers at the beginning of the thread.

The new power supply totally got rid of the clipping from the Lepai amplifier. So it's 19V peak to peak amplitude for the Lepai and 18V peak to peak for the low pass filter that feeds it. It gives me hope I could somehow use the Lepai for the subwooofers, as long as it's only used for the subwooofers and not for anything else. I might perform surgery on the Lepai to remove the pesky 3rd "subwooofer" channel, to make sure all the power goes to the right/left "front" amps. If I do that, I could revisit the idea of using the Sony digital stereo for the Sony blue drivers and tweeters. Oddly that is a return to the original concept of the digital Sony driven Mk. III. Wouldn't that be something? Otherwise I'll get the TDA boards in my previuous post. I'm just trying to protect my money, because honestly, even though I got a decent tax return this year, I'm too afraid to spend it.

After getting rid of the Lepai's distortion, it's clear that the 8W tiny subwooofer units hit their maximum pretty easily just by looking at the cone travel and hearing as the driver's cardboard panel flexes. They were obviously designed for low power tabletop PC sound, but so was the Altec Lansing subwooofer I used in the Mk. I, which totally puts them to shame in terms of maximum volume. So out they go and I'll order a pair of Tymphany subwooofers tonight.

Now, for other noises coming out from the horns: Even at low volume you're getting mechanical noises from the joint between the woofers and the wood panels. You can't tighten those screws enough, because you're talking about a joint space of less than 1/10th of a millimeter. So I will need to start using 2mm neoprene sheet to isolate all mechanical joints in the horns, now that the horns are coming to life. These mechanical vibrations can very strong at moderate volumes, and the wood is absorbing a fairly large amount of mechanical energy. Joints will need extra careful attention, as the manner in which you hold wires and cables inside the horns.

[Banfili]

That's looking really impressive, J. Wilhelm!

[J. Wilhelm]

That's looking really impressive, J. Wilhelm!

Thank you! I've been having a lot of fun trying out the horns with the subwoofers working, even though they're severely limited in volume because the Samsung subwoofers can only take 8 Watts of power. So at least they're great at low to moderate volume for classical music, with the large Sony amplifier (bottom of picture - please excuse the mess in my room  ).

But now I'm having my fair share of problems making it work with the smaller Sony amp though. That Lepai amplifier is giving me a lot of headaches.

I had already managed to get some use from the Lepai amp by taking advantage of the gain of 2 in the active crossover. I was even getting some nice sound from the big analog Sony amp, and I had it set up so that I was only dealing with the shortcomings of the quaint 8W subwoofers. Surely. all I needed was a better type of subwoofer, and I should be able to hook the Lepai and the horn to the smaller digital Sony amplifier, right?

Well, sort of, yes, it did work and I started getting some nice sound  out of that setup too. My hope was that I could switch to the smaller Sony digital amplifier for the front speakers, and maybe use the Lepai for the sub speakers on the sides. It sounded like a good idea. The Sony digital amp doesn't have a line out output though. I had to solve that problem. By trial and error, I had figured out how to tap from the Sony amp output so that I could sample the sound with the Op Amp low pass filter and then pass it on to the Lepai. At about a gain of 1/2 from the output voltage to the front speakers, you can bring the level of the  Op Amp, and thus the volume of the the subwoofers at about the same level as I got when hooking up to the big Sonyt amp. It all works fine, actually... Except for one little problem. The Lepai amplifier is not playing nice with the smaller Sony amp.

I discovered a nasty property of the Lepai. It goes into wild oscillations if you turn on and off the amp, or disconnect it while the Lepai is on, and worse, even while changing radio stations    At first, I thought that it was my little Op Amp circuit's fault. After all, it has no other safeguards against oscillations other than it's ridiculously high input impedance of 1 MΩ  (which should shield it from almost anything) and a paltry low gain of 2. So I went ahead and I added capacitors between the power inputs to the Op Amp and the ground, then added a second ground line to go directly from the Sony amp to the Lepai, and made sure that gain was low on both amplifiers. Turn Lepai on, connect-disconnect - all good, sounding nice. Then press the radio station tuning button and !!!!!karrragh!!!!! A sound somewhere between a fog horn and the mating call of a kraken shook the house.

Added two 0.68μF capacitors to the voltage supply to stabilize the Op Amp against oscillations

Tried other permutations, Adding ground contacts everywhere I could tried the sequence again, !!!!!karrragh!!!!
Decided to turn off the Sony amp !!!!!karrragh!!!!
With the Sony off, I unhooked the Op Amp from the Sony !!!!!karrragh!!!!
Then disconnected the Op Amp from the Lepai, !!!!!karrragh!!!!
Finally switched off the Lepai off !!!!!karrraaaaaaagh!!!!

It actually took a few seconds before the Lepai shut down completely disconnected from everything.

So the conclusion is that the Lepai can be set easily into wild oscillations if you hit it just the right way. And the Sony digital amp knows exactly how to do that.
Later I confirmed that if I disconnect and connect the large Sony amplifier while the Lepai is on, for a fraction of a second the Lepai will do the same oscillations, but it will shut up as soon as you secure a ground connection between the two amps (by way of the Line Out connection I use for the Op Amp circuit).

So yeah. Another reason to hate the Lepai. It's unstable as heck. I can't use it with the digital Sony unless I figure how to buffer it properly. So far I can only use it with the large analog Sony amp. Not only is the Lepai bad at amplifying with a low output volume, and has very low input sensitivity , but on top of that, it's inexplicably sensitive to the mains 60Hz signal and can go bezerk if you hit it the right way. *sigh* Something is terribly wrong with the front end of that amplifier. Exactly what kind of pre-amplifier did these folk build? It has to be the fault of the Op Amp circuits they used for input buffer and tonal control. Similar to the low pass filter they used for the subwoofer which produces excessive clipping. It seems to me the entire front end of the Lepai is badly designed, or trying to do something it shouldn't like an out of control feedback loop in the Op Amps. You can turn an Op Amp into an oscillator quite easily if you push it too hard. And that is exactly what it sounds like. A square wave oscillator, perhaps interacting with the digital amp chip which itself is a square wave oscillator (pulse frequency modulation).

~ ~ ~

OK. So how did the small digital Sony amp do? Actually quite well while connected to the horn. Not quite the same frequency range and depth as the large Sony amp, but I'd say 95-97% of the way. Definitely acceptable sound quality, and likely better frequency response that the Altec Lansing in the Mk.I. The only caveats are:

1. The stereo separation and dynamic range of the digital Sony is a bit flat, unless you use a special feature of the amplifier called "DSGX" (Dynamic Sound Generator), which to me sounds like Sony said "yeah, we know the sound is a bit flat, but this is what we did to fix it." Then it sounds just as bright as the large Sony amplifier, and it goes all the way to 100% sound quality.

2. The volume of the Sony digital amp is relatively low. It starts sounding nice at about 22 out of maximum 30 clicks maximum on the volume knob, even with the 40W per channel subwoofers thrown in (limited to 8W by the Samsung units). In theory that should be 110 W overall nominal or say 76W overall even if limited by the Samsung speakers, compared to the 40 W nominal of the Altec Lansing, but I just can't hear it being that loud. The Altec Lansing can be a bit louder. This means that a 25W + 25W digital amp of any brand will probably *not* be enough to push the speakers all the way they should, and the Lepai is not being efficient at all. These speakers are begging for at least 50 W per channel to extract all the potential of the horn. This is a strong consideration point for a future choice of amplifiers.

[Miranda.T]

Looking great, but I'm sorry to hear the electronics is still giving problems. If only there was some way to earth the Lepai's input whilst the input is transitioning (on/off, tunings etc.) Tricky one indeed.

Yours,
Miranda.

[J. Wilhelm]

Looking great, but I'm sorry to hear the electronics is still giving problems. If only there was some way to earth the Lepai's input whilst the input is transitioning (on/off, tunings etc.) Tricky one indeed.

Yours,
Miranda.

I think it's a combination of factors. Digital amplifiers tend to have closed loop circuits for the output, so you don't have a real ground. The digital Sony does not have a real ground. So as soon as the output goes to zero, the reference voltage "ground"  for each channel floats, and that gets the oscillations started. The oscillations actually happen in the Lepai, because even if you disconnect everything the Lepai continues screaming away until it runs out of current when you shut it down. There's no fixing that, other than to use a buffer with a real ground between the Lepai and the Sony digital. The Analog Sony doesn't have that situation, because it has a real ground and most likely a real transistor or Op Amp buffer handling the Line In /Line Out connections at line level. Yay for old school technology.

Speaking of which, I got an awful lot to talk about an idea that's emerging from my use of Op Amps in this circuit. I'm starting to remember all the circuits I used to build in high-school and my first years in college, and I'm thinking I'd like to implement some of these really old school sound processing ideas into the boombox, just for anachronistic fun. I'm debating whether to include that in this thread (certainly if implemented these ideas, it'd be part of the Mk III project.

Does anyone recognize or remember the name "Matrix Stereo"? This tech will take you back to 1969/70. My grandfather actually owned one of these (below) and I got to play with it throughout the 80s.

Sony MR-9300WA Portable AM/FM radio with 3 channel Matrix Stereo

And my grandmother had one almost identical to one of these below, except with a compact cassette deck, instead of an 8 track deck.

Sanyo stereo receiver/turntable/cassette with 4 channel matrix decoder. Circa 1973.
I don't know the model number, but I may have the instruction manual

[J. Wilhelm]

OmyGod this is what I need  

A good, old fashioned thick film AB type amplifier chip (this is H type, a high tech version of AB).
https://m.aliexpress.com/item/4000981509316.html?trace=wwwdetail2mobilesitedetail

80s high tech, baby   My 1985 Garrard (which I have in storage) uses one of these thick film type monsters. When they say beautiful sound, they mean it. But 140W per channel is overkill. I'd be happy with 70W per channel and here's another good choice:

Sanyo STK401-140, 70W per channel. Type H amplifier (Type AB with variable voltage rail)
$25.38 per module. (need 2 for4 channels), assembled or $20.68 unassembled kit
https://www.ebay.com/itm/353235139458

The only caveat is that I need to get the right alternating current power supply, a dual voltage alternating current supply over 20V.  But I'm so close to taking the plunge. I don't mind paying a couple of extra bucks to have it assembled. I don't see the point when they can solder better than I can.

This would have to be a one-shot project, because you don't know if you can get any more of these discontinued chips. Type H amplifiers are basically the last analog amplifiers that came out using a type AB architecture, and they disappeared after Type D amplifiers came out.

[J. Wilhelm]

I don't know whether to laugh or cry. Be happy or sad. I have a lot of things happening.

On the good side:

I have finally reached my goal of transporting myself to the storage unit I rent, without a car and solely by foot and bus. The storage unit is very much inaccessible by pavement, on account that it lies along a busy freeway between major avenues set a couple of miles apart. It takes me riding two buses and walking at least a couple of miles just to get to it. A major expedition.

Exciting times. I found all my belonging relatively intact amid a floor full of mice droppings. I remain hopeful that the critters haven't made nests in my important papers, text books, family photos and 8mm films    I have rescued the mythical and very rare 1985 Garrard (Gradiente) RDS-20, I often talk about, and I've brought it back home, still working. The speakers I did not bring, and they're a bit of a mess. These are 45W 3 way speakers with beautiful white paper cones, which now look brown (I don't know if I can rescue those). But the receiver/amp is working perfectly. It's even rejecting a 6 ohm load automatically (protection mechanism) so I have to bridge 4 outputs with the section button  to drive the little subwoofers - with no audible native distortion or clipping artifacts of any kind (except something else coming in from the Sony amplifier or low pass filter - read below).

Rock solid stability, I can connect or disconnect whatever I want without sending the amplifier into a fit. I remember that the Garrard produced beautiful sound and it doesn't disappoint. Excellent sensitivity. If I connect it to the Line Out of the Sony reference amplifier, the volume is exactly the same as if I disconnect the Line In and tune the radio to the same station. In other words, the Sony and the Garrard have exactly the same line level standard. Something the Lepai doesn't have. I knew these should be non-issues. You guess why now 30 years later it's so much of an issue among DIY's and audio experts  

Also really good news, the RDS-20 is extremely flexible. This is a 2 channel amplifier that can be split into 4 outputs, without having to work through a digital surround decoder like the Sony, which means I could drive the two subwoofers directly if I used passive crossovers (low pass filters) instead of the active Op-Amp amplifier (at this stage I'm trying to figure what's more expensive. The elphantine passive crossover, or the Op-Amp active crossover plus an extra amplifier. It'd be very useful to only have to use one amplifier).


The bad news:

The Garrard RDS-20 is based on the STK-4121 II, a precursor of the chips above. And while it is a good Type AB amplifier module with great sound, this particular chip only produces 15W!!! In the series STL-21X1 II, the highest power you can have is 50W for the STK 4191II module! I can't fathom how it can drive the three way 45W speakers! I'm going to have to open the RDS-20, to see if they're complimenting the power of the STK module with FETs or what!!!

More bad news: I found a fatal flaw in my Sony reference analog amplifier. It looks like the Line Out/Line In buffers in the receiver are busted. I will lose one channel temporarily if I try to connect the Sony's line in to the line out of another source, like my Extigy sound card. For days I had been listening to the subwoofer output randomly lower in volume, and then produce a lot of noise, like radio interference. I assumed that was part of the clipping in the Lepai amp, so I ignored it. But connecting the Line Out of the Sony directly to the Line In of the Garrard revealed the same noise being pumped into the Garrard. Obviously the Garrard can't have the same fault as the Lepai.

Now this could be a fatal flaw on the Op Amp crossover, because I used very high resistor values I've never used, but the other possibility is that the noise is related to the sudden random loss of one channel in the input/output buffer in the Sony. This means that I can't elucidate whether the noise is due to a bad Op Amp circuit in my active crossover, or a bad Sony buffer. My bets are on this being the Sony's fault; it is, after all a Goodwill second-hand find with a busted display - my feeling is that some electrolytic capacitors in the line in/out buffers are burned out and since electrolytic capacitors have a "partial self-healing ability" as soon as you turn off the Sony for 5 minutes or so, the capacitors regain their ability and they start working again for another half an hour or so before going bad again. That would explain the random noise happenings and the loss of one channel when connected to the sound card.

*Facepalm*

All my Goodwill stuff is bad in some way! The Sony amp's buffers could have been burned by the previous owner doing something dumb like connecting a speaker directly to the Line Out (shorting the Op Amp) or pumping amplified power to the Line In of the Sony (quite likely). I think that's a very likely explanation. I'll check the Op Amp circuit I made last. I have more confidence in it.

The "more good news" to balance the "more bad news" is that I have my good Sony amplifier in the storage unit. This is a very similar amp to the (large) silver one in the pictures above (the one at the *bottom* of the entertainment center furniture), but it's a tad better model, 5 years younger, and as far as I know with no faults whatsoever. Tomorrow, I mount an expedition to Mt. Kilimanjaro and I will bring the elephants to carry it back to my residence. So expect a "new" black Sony reference Amp tomorrow in this thread!! I may give the silver one away - it's no good to me without a line level input.

Also: I may find the other Creative Extigy sound card I own, which also is "new" (i.e owned only by me and treated well). I will give away my Goodwill purchased one (faulty optical line out...   )

Also! I rescued my Binaural Microphone. It's high fidelity secrets lie inside this box below with an attached schematic diagram I glued to the interior side of the aluminium panel!

My focus now is to get rid of all the stuff of questionable quality - all my Goodwill purchases, and populate my test bench with stuff I know works well, since I have it. Then calmly determine which way I want to go. I'm definitely not looking at "Digital" (which by the way are not digital) Class D amplifiers. Garbage be gone!!

I may take ONE last look at the Lepai to see if I can perform surgery, and remove the entire front end. I may open the Sony "digital" and also make some exploratory surgery, to see if I can build a proper line level output. But the thing that worries me the most is the lack of line level standards in Class D amps. They're all over the place. The input sensitivity of the Lepai is low, so you have to crank up the volume knob all the way up to hear anything, thus likely producing all the sound problems I've been talking about. As soon as Class D amplifiers were brought in to replace Class AB and Class H, and portable players adopted Class D, long-held industry standards were abandoned, and along with the sound quality, the line level input standards were gone with the wind too. This is a real problem with "digital" amps. Nothing can take away my impression that they're just garbage now. More trouble than they're worth.

[J. Wilhelm]

The big boy is here. And at 20 lbs, it is a big one to carry for miles and on the bus. When I was a couple of houses away from my arrival, my leg muscles began to cramp. I had to stop for 10 minutes while standing until my thigh muscles stopped "dancing."

20 lbs of exceptional sound. 400 W Sony receiver with 5 independent channels.
The magnificent, the one and only, Sony STR-DE445 


I had to get creative to bring it over. My Sherpa Harness 2000™ helped with that.

He also brings a surprise,which is an absolute game changer: line level 5.1 channel input. It featuresa manual input for a 5 channel decoder with choice of sub out or center channel out, at 80W per channel. That's 400W overall. The reason this isa game changer, is that you can build a decoder with 5 separate channels. This means that I can two of the channels to connect the active stereo subwooofer. I don't need two separate amplifiers. I only need this one for the horns.

The caveat is that I can't use this amp as the source of sound. The sound must come from an external source, like a sound card or handheld player. The only thing I need is to build a line level decoder, which could be a buffered front left/right pair of channels, and an active filter (like the one I built for the subwooofers) for the rear front left/right channels, plus one extra center channel -- useful for a Sony Matrix Sound decoder with +-( R-L) sides and an (R+L) center channel, or a 4 channel decoder like the Sanyo Matrix surround decoder!

The possibilities are endless if I dedicate myself to building a line level decoder!! This is right up my alley, since I've built all three of those circuits before!

[J. Wilhelm]

I just had a nightmarish two days with my black Sony amplifier. I noticed that the amplifier would go into self-protect mode at random intervals and then it just went into self protect mode permanently. I feared the worst: a bad component, or worst, à blown power amplifier Field Effect Transistor (FET). Since you can't defeat the Self Protect mode, this means a trip to the electronics repairman, or chucking the amplifier to the garbage bin -  I thought.

Luckily, the interwebs provides all sorts of answers and I scoured the internet for 2 days, seeking an answer. One website suggested the issue was due to a blown FET (the worst case scenario outside of a blown computer chip). This amplifier has 5 channels, and therefore 10 FET power transistors transistors. One pnp and one npn pair of transistors form a push - pull circuit to pump 80 watts per channel. If one or two of each pair of transistors is blown, then the voltage would be spiked to about 50 volts, or dead, in one channel, instead of a few volts like the others. So I followed the instructions and measured the output of the 10 transistors, not finding a problem. *whew*

Next I read from many websites, that the Sony STR series (among others) from Sony had a particular problem with excessive heat on the main board. Many users of different Sony models have complained about inexplicable self protect instances where the problem seems to come and go. Sometimes disappearing if they screw and unscrew the boards to the chassis. Many people think that a false ground is to blame (corrosion) at the pads that connect the main board to the chassis. But electronic repairmen point to fractured solder joints as the problem. Excessive heat on the main board causes the copper leads of components and solder joints to expand at different rates, creating cracks at the solder joints that grow in time.

So I went through the trouble of disassembling the whole unit and flipping the main board upside down so I could look at the (seemingly) few thousand solder joints underneath. And I did find a number of stressed joints, right where the repairmen said they would be. Before passing signals to the bank of 10 FETs, Sony uses one pair of NEC μPC2581V stereo power amplifiers (I reckon they would have 3 channels per pair, or three channels for one chip and two for the other) as a pre-amplifier. The problem with Sony, is that these audio chips are not using heat sinks, so the chips get very very hot, and thus that area of the main board gets extremely hot too. This is a major issue. The joints I found were not disconnected, however, they showed stress cracks that will *soon* need to be re-soldered, and would explain the intermittent problem in the near future, but I found no broken contacts.

There was one section of the board under a bank of resistors coming out of the amplifier chips that had capacitors soldered on the underside of the board. The were insulated from the main board by two pieces of electrical tape. This seemed rather unprofessional coming from them, but noticing how the main board is being baked, perhaps protecting the electrolytic capacitors from the tremendous heat was a smart idea (still amateurish thing to do, though).

Guess what I found then? Apparently there was a tiny corpse caught in the electrical tape. Apparently, one ant had crawled in between the capacitors and the main board and got fried while caught in the stickiness of the electrical tape. I literally found a bug.

You need to zoom in. This one is small!

Anxious to see if the bug was the problem, I reassembled the Sony and ... it works!

Now, this does not mean I fixed the problem for good. So far the amplifier is working again, but those stressed solder joints I was talking about will eventually break (see picture below) and I have a lot of those joints around said chips and the bank of resistors I mentioned. That last freeze that we had in Texas surely did not help. Thermal stress works in both directions, and this amp was inside a metal storage box, surely at very cold temperatures. It's a shame I did not raid my storage unit last year. That main board needs a cooling mechanism now and I need to re-wet (re-solder) all those joints. But I neither have the right illumination, magnifying glass (my eyesight is poor) nor the low-power soldering iron to do that safely around those NEC chips. So for the moment I've temporarily fixed it...  

Examples of "cold" solder joints.
A joint showing strain due to cyclic thermal stress. I have about 10 or 15 at least of those joints similar to
the ones in the picture, all around the NEC chips and the bank of resistors attached to the chips



A joint with a full fracture developed eventually from cyclic thermal loading

I think that for the moment, I will need to go easy on the amp and I might have to delay high power testing for prolonged periods of time, while I find the right tools or find some kind technician who would do me the favor of re-wetting those solder joints. I'm just tired of fishing for quality amplifiers and I need to get my finances in check. This amplifier does everything I need in one simple box while I bring those subwoofers to finish the horn.

Provided the black Sony doesn't go bad on me again, the goal would be to build a sound processor using Op-Amps as quickly as possible, and test it with a new set of subwoofers. This time I don't have all the components I need. I need to get another LM 1458 chip (or I can use two LM 741's) and a socket to build a buffer.  But I want to build not just a buffer, but a sound decoder with the Sony 4-Channel version of Matrix Sound. Since that is so close to the 4-channel Sanyo Quadraphinic decoder, I'm thinking I might do that too. The Sanyo quadraphonic sound decoder, is in fact nearly identical to the analog Dolby Surround (1981 version), so that's three sound decoders in one, plus the subwoofer processor.

How's that for a project?   What I want now is to stop bringing and spending more on amplifiers. If I do build an amplifier for the horn, it will have to be an AB type amp with 4 channels, similar to Sanyo STK boards above. I just don't know if I can work at all with the Class D digital amps. They're just a mess!

If I make the Sony "Digital" amplifier work, it's because I performed surgery in it to extract a line level signal from the entrance of the amplifier, with the aid of an Op-Amp buffer. In theory, I can use the Auxiliary input of the Sony to make a pre-amplified output for the subwoofer, but then the subwoofer will not be able to use the on-board FM radio or CD player / iPod dock. Another strategy is to borrow the signal from the iPod dock as a line level input. That's what I did in the Mk. I Boombox, because I had a stand alone iPod dock.

The Lepai amp is sounding nice, but has absolutely no headroom to increase volume- unless I also perform surgery on it), and I don't know if it will go bezerk again on me if I connect it to the Sony Digital. I just don't know what is going on, because as Miranda says, I need an oscilloscope to figure it out. With all these simultaneous projects, I should open an electronics repair shop!!

[J. Wilhelm]

As now I can start forgetting about messing with amplifiers, this is a great opportunity to think about the front end of the boombox. In any amplifier, the front end will have something to do with buffer circuits, and tone control, volume, etc. But because that is a type of sound processing, then if you have anything else special you want to do to the sound then that is the stage to build a sound processor.

As you can see in the posts above, I resolved to build a low pass filter as an active crossover for the speaker. This means that this low pass filter is in fact a sound processor located in the front end of the boombox. Since I'm already in the business of processing sound, and I need a buffer anyway, I though that adding extra sound processing power would be a natural two for one deal, so to speak. It will not involve any higher costs than building a buffer which is necessary to make sure that the stereo input impedance to the "front" speakers is the same as the impedance of the subwoofer speakers. In my perfect symmetry philosophy, the "front" amplifier should have the exact same buffer as the "rear" subwoofer amplifier. I need to make sure that the inputs accept Line Level voltages anyway.

There is a very large number of things I can do with another LM 1458 Op Amp, all having a similar circuit complexities, save for a couple of the methods shown below. I discussed in a post above the idea of incorporating a stereo expander, similar to the one that Sony used for a 3 speaker portable radio. It turns out that in 1975, Sony came up with one of the earliest designs of the classical boombox, with a 4 channel "Matrix Stereo."

1975 Sony CF-580 boombox with 4 channel Matrix Stereo expander. The "surround speakers" are on the sides.

The idea would be to build something similar to that Sony boombox. Is it Surround? Quadraphonic? Stereo expander? Nobody knows. It a bit of a gimmick for sure, but the truth is that it sounds fantastic in the 3-speaker version that my grandfather had. I would call it a stereo expander in that mode, not a surround. The actual circuit, of course is a trade secret, and not likely to be divulged. But what I do know is that the circuit is a variation on one of the sound field decoder methods below, some of which I have built.

Some of the circuits I built and heard about 30 years ago  with the tiny LM741 or LM 1458 Op Amps
(except the Sony 3 channel which I didn't build, only played with). I'd like to try my hand at building a 4 channel Sony style expander this time.
The additional speakers, rear or otherwise usually derive their signal from the difference between right and left channels
that is you mathematically subtract the right signal from the left signal and amplify or filter that in some way.

In the chart below "n" is a multiplier, that is a gain over the channel before adding or subtracting from the other one

These methods above are actually based on 1970s Quadraphonic era systems, of which there are many. The two big players at the time were CBS records and Sansui in Japan who basically fought over the format to be used in consumer electronics and also FM radio transmission in the United States. In the end the Quadraphonic technology died, because there were too many standards being pushed to the public, and all standards had serious issues with them, like poor speaker separation, and incompatibility with one another.

The CBS method "SQ" has the most financial backing, and they were very powerful in influencing record companies to produce in their format. SQ was almost fully compatible with FM stereo, and any stereo player could play an SQ tape or record with only minor issues (vanishing center channel voice). CBS tried to have the FCC consider their system as the proposed default for quadraphonic FM radio transmissions in the 1970s. The QS method developed by Sansui in Japan was to be the direct competitor to SQ, and many Japanese brands like Sansui's Sanyo brand, which carried the 4 channel "regular" matrix, as it became known among users. The latter was the system used for the phonograph I show abovem which my grandmother wouldn't let me touch in the 1970s:

Some Quadraphonic Matrix decoders from the 1970s.

The 'j" letter indicates that the signal from a right or left speaker is phase-shifted by +90 degrees.

As you can see all of the Quadraphonic systems are very similar to the stereo expanders and the Dolby Surround system. They all involve adding or substracting left and right signals, multiplying them by a factor and phase shifting them as well, in the case of SQ and QS. Of all the circuits above SQ and QS are the most difficult to buid because they need an extra phase shifting circuit, and the quality of the decoding will depend on how constant that phase shift (indicated by the letter "j" above - it's complex number notation) can be across the human hearing spectrum. Because of that the quality of quadraphonic decoders could vary from set to set, and in fact some specialized chips were produced to de code SQ encoded signals, some of which you can still find online.

I could venture into the Quadraphonic sound (Sansui QS looks most interesting and has much better reviews) - and I even know how to phase shit a signal with an Op-Amp, at lest for the range of 100 Hz to 10 kHz, but the quality of the phase shifter varies with frequency:

This is a +90 degree phase shifter built with an Op Amp centered around 1 kHz (human voice), and it's usable range is 100Hz to 10 kHz
the phase shifts from zero at 100 Hz to 180 degrees at 10 kHz, so the phase shoft is not constant. That affects the quality of the matrix decoder

But I don't know how good the decoding would be, and it requires the speakers to be separated (not really an issue, you can have jacks for external amps/speakers). But I think I'd prefer to tackle the Sony styled 4 channel Matrix as a stereo expander instead, and pump the "Matrix Sound" from the subwofers (which can carry higher frequencies as well).  Another reason for avoiding phase shifting circuits, is I don't own a collection of matrix encoded material, so it's easier to build a stereo expander as an experiment that just creates special effects for fun. Because the Sony Matrix Stereo systems are secret, it will be guesswork on my part.

So now I ponder the possibilities. I have a circuit in mind that I'll explain in my next post!  

[J. Wilhelm]

My project has stalled temporarily due to some unforseen problems with insects at my house (réf. Gaah thread) , and a good chunk of cash, more money than I was spending on the boombox project, has gone to purchase insecticides and make a collapsible bed. Oh well I'll be able to recover at least half of the money by working an extra day this week.

In the meantime, I've had an idea related to my interests in more advanced surround sound techniques, back in my early years of college. As it happened, in high-school and my early college years, I was experimenting with Dolby surround and Matrix Quadraphonic as I wrote above. But while I'm college, I had access to a higher science and library materials. I started a project around 1990 which I called Psychoacoustic Surround Sound System (or PSSS for short) which explored knowledge on human sound processing more or less historically, developing my work in stages (Surround Sound was Stage II of the project). The academic research during the late 70s I could find at the library were the peak of human knowledge regarding the way humans locate sound in 3-dimensions. As it happens, the science of Psychoacoustics is a cross between physics, anatomy and psychology and its a rather ethereal subject that eventually gave us 3D sound in computer games and virtual surround sound coming from only two speakers in television systems. It's not a science that started researching exactly how humans perceive sound, but more like an aggregate of research that accumulated throughout the centuries, as far back as the middle ages, and then by the late 19th century began to coalesce into a field of study. It came of age during the 20th century when military technology pushed the science further.

The reason I found out about the science was an article on Popular Science / Popular Mechanics / OMNI (?) on a contract awarded by the United States Air Force to NASA and the late Hughes Aerospace, circa 1986, for a system which could simulate 3 dimensional sounds inside fighter jet cockpits. The idea was to take away the heavy visual information load pilots had to endure during combat, and place the burden of processing that information on the pilot's hearing system instead. In theory that should allow pilots to focus on the airspace around them, instead of having to look at dials and readouts on their front panel or heads up display.

The key lay hidden in findings by scientists around the world for centuries, but the pinnacle of science's understanding on how humans process sound did not happen until the late 1970s and early 1980s, when scientists could use computers to measure sounds captured by tiny microphones embedded in the ear canals of test subjects just outside the ear drum. Coupled with physical acoustics, biology, anatomy, and psychology, scientists could evaluate people's perceptions of sound, and quantify the ability of human beings to detect a source of sound in 3 dimensions. Only a handful of scientists around the world were sufficiently trained to cover all of the subjects required during the mid-late 20th century, so it was rather easy to follow their progress.

What scientists discovered, besides the obvious, such as a sound intensity difference between ears, is that sound phase differences between right and left ears could actually be perceived by the human brain. Things like sound wave front time of arrival could also be measured by the brain. And the latest findings from the 1970s were that the outer ears (pinnae), your head shape and even your upper body, comprise a complex surface that diffracts, and reverberates sound in very complex ways, changing the way your eardrums register sound, depending on the location of the sound source. The last piece of the puzzle was finding that as a sound source location changes, the frequency content of sound is changed drastically, depending on the direction of sound. The outer ear and head, in particular, create a deep notch on mid and high frequency sounds, depending on sound source location.

Using computers to operate in Laplace transform space (frequency and phase), scientists could now make a map of frequency and phase changes in sound perceived by the ear drums, depending on which direction the sound waves came at you. Coupled with previous knowledge about sound intensity and phase changes, scientists could now, theoretically, program a computer to vary the frequency content, phase, and time delay between right and left ears to create the illusion that a sound source was moving about your head in 3 dimensional space. There needed to be a "key" or mathematical operator that could take regular monaural sound and map it to 3D space; the Fourier Transform of the 3 dimensional sound field around your head is called the "Head Related Transfer Function" (HRTF). And with that map, coupled with a similar map for phase changes, in theory you can take a monaural sound, or a two channel stereo sound and turn that sound into source located anywhere in a 3 dimensional sphere around your head.

In the US Air Force contract proposal, NASA came to the conclusion that a computer was needed to vary the sound content on phase and frequency, and you absolutely needed to play back the sound via headphones, in order to get the right 3D effect. The issue was cross-talk between right and left speakers, they noted, and the 3D sound cues were too subtle to survive that low stereo separation of regular speakers. They argued also that the HRTF needed to measured for every single person intended to become a test subject or user, as there is a great variety of shapes and sizes of heads, noses, and ears in the population. The 3D sound needed to be customized to the individual. Using a computer system called "Convolvotron" they set out to make their case to the Air Force.

Hughes Aerospace came back with a different answer. They were of the opinion that a sufficiently generalized HRTF could be found to work on most humans, so a regular set of stereo speakers could do the job. The amount of cross-talk between right and left speakers, they opined, wasn't that critical. And if you didn't mind only having a 180 sound sphere in front of you, as opposed to a full 360 degree sphere, and further you didn't need arc 1-5 degree accuracy on the sound source location, such as humans have, then a pair of stereo speakers could be made to work. Hughes went to develop a stereo sound processor called "Sound Retrieval System" (SRS), which eventually was sold and used for Hi Fi equipment and flat panel televisions that still have the system today. A new company called SRS Corporation (?)  was formed to market the system, but today you have 1001 other similar methods used in computer games and such. The black Sony amplifier has about 3 different variations on the method for generating a "virtual surround sound" stage.

So... I was actually going through my old college records, and I found a spreadsheet with the main data on the HRTF published by some of those high brow scientists in the late 1970s. The data set I found comes from a German research team published in 1977, in the Journal of the Acoustical Society of America, and is the most complete set of HRTF measurements I ever found. Their data gives the HRTF field from 500 Hz to 15 kHz, and I'm sure there are newer, better measurements out there, but this data set is pretty compete. I was thinking I might explore creating a simple analog processor to try to approximate this HRTF, similar to what Hughes did for their SRS system.

Data adapted from Mehrgardt, S., Mellert, V. (1977). Transformation characteristics of the external human ear.
J. Acoust. Soc. Am. 61, 1567–1576.

Now, to be honest, I don't even know if there's a simple way to do it without resorting to a digital computer. But back between 1988 and 1997, in between studying various aeronautical engineering topics, I was dead serious about developing a stereo Matrix system coupled with an HRTF operator, similar to the Hughes SRS method. Basically using the "right minus left" difference between stereo channels to pass through the HRTF operator and generate *some* form of virtual surround sound. Hopefully without the use of computers, just a simple system using operational amplifiers and phase changers... I'd be shooting for something relatively simple using a generalized HRTF based on the figure above.

Maybe it's time I restart Project PSSS III... I'll try to think about this while I'm not fighting bugs and building beds.