Soundstream Class D controller IC

I have a 2021 Soundstream AR1.4500D, 4.5kW Arachnid and I'm trying to get directly at the Class D driver (controller) IC input pin(s), bypassing all the preceding filters except the IC's input coupling capacitor(s), which I'll change. I need a single pole High Pass 0.1 Hz response.

I don't have a schematic of the small, vertical control(?) board. At the top left of the board, what looks suspiciously like a 14 pin Class D driver chip is an STMicro(?), marked STMZ3U748. It may be an old or proprietary device, because I can't find a data sheet at the STMicro website. It could also be my aging eyes.

I may have completely missed the boat and the Class D controller is something or somewhere else. Please don't tell me it's on the backside of the main board.

Does anyone have a link, a schematic of the board and/or a data sheet for the Class D controller device? Anything from that same or a similar Soundstream Class D amplifier series will be immensely helpful.

Thanks,

Ron
 
Post a good quality photo of the component side of the entire main board and the component side of any driver boards if no one can help.
Thanks Perry,

The "driver?" board was mounted vertically, about 2 inches from the toroidal output inductor.

Having blown all the output transistors in my Sunfire True Subwoofer, under power, with a probe slip, 20 years ago, I don't probe live boards without DIP clips, and insulated wires. A probe slip on even a low level driver can be a disaster.

I cut the cable tie and RTV securing the output inductor and tried shifting it about 1/2 inch to allow use of a safer SOIC-8 or SOIC-16 test clip. The inductor barely budged and I wasn't in the mood to un-solder it.

To clear the inductor, I decided to use an 11 pin breakaway socket strip and pins to raise the lower edge of the driver board 1 inch above the main PCB. I can always replace the driver's breakaway pin strip. With the socket/pins, I could also remove the driver board to Ohm or photograph it. Photos of the driver and main board are included.

I'll have to install beefier pins and main board socket strip. The Gold contacts are "iffy", but that's all I had on hand. At the moment, I can't do much more than sloppy soldering.

The SOIC-8 and SOIC-16 clips wouldn't hold. The chip profiles are too shallow. They may not be SOICs. I'll check.

I'd originally WAG'd the small, vertical PCB as the Class D driver because:

a) all the topside main board circuitry between the driver PCB, the input jacks and pots are linear op amp preamps and filters. Please ref the left side of the main board photo.

b) there are two 14 pin CMOS 74HC digital devices on the driver board. Since I can't find a crystal, etc, I'll WAG they're an RC clock, digital interface drivers or both.

c) small power transistors on the main board, nested near the driver PCB, are likely the drivers for the output FETs.

d) the stuff at the opposite (right) end of the main board is the PWM power - not a good place to put a Class D driver or low-level analog circuitry, regardless of whether you've been diligent in pin-guarding or avoiding ground loops.

In the interim, I've Ohm'd about 1/4th of the driver board.

The main board photo is without the driver board plugged in. You can barely read component Reference Designators on the main board. I'll have to re-tweak the camera aperture settings to deepen the depth of focus.

Ron
 

Attachments

  • Driver-Frontside-5-11-22.jpg
    Driver-Frontside-5-11-22.jpg
    716.5 KB · Views: 37
  • Driver Backside 5-11-22.jpg
    Driver Backside 5-11-22.jpg
    401.9 KB · Views: 38
  • Main-PCB-Topside-5-11-22.jpg
    Main-PCB-Topside-5-11-22.jpg
    569.6 KB · Views: 37
Jumpers like you're using can be very time consuming. The best option that I've found so far is the jumpers below. These are breadboard jumper wires. Searching eBay for Arduino breadboard jumper wires should get you good results. Be careful not to buy female-female or male-female jumpers. You want male-male jumpers.

This diagram is close to what you have.
http://www.bcae1.com/temp/SLA1500.pdf

Please don't start blindly changing parts. These amps are fairly easy to troubleshoot.
 

Attachments

  • IMG_9461x2.jpg
    IMG_9461x2.jpg
    364.6 KB · Views: 34
  • DSCN9803bx.jpg
    DSCN9803bx.jpg
    154.5 KB · Views: 31
Thanks immensely. The photo you've shown me looks essentially like the same PCB, with readable IC numbers. Mine are barely readable, no matter how I adjust angles and lighting. I haven't yet tried my cheapo USB electronic microscope. I'll assume the schematic is close. I was trying to determine what device your U701 actually is.

I haven't the ambition to rip into anything. My only desire is to "simply?" tap into the Class D controller's analog input pin(s), to connect an external inverse feedback loop, without the phase shift of all the preceding filters. I'll eventually use them.

The last time I tried the same thing with a bridged, full BW 1.2kW Crunch AB four years ago, I ended up with an almost pure 1 Hertz sine wave generator and damaged the battery in a 1 year old RAV4 - a true DIY kluge producer.

I've been experimenting with all sorts of subwoofer linearizing devices, including classical 2nd Voice coil feedback, Velodyne-like accelerometers, MEMS pressure sensors and even an X-band (10 GHz) Doppler radar module. Sitting roadside, I could pick up cars breaking the 30 mph speed limit, half a block away. Maybe I should try lidar feedback and/or the motion feedback forums.
:cop: thread moved to Car Audio.

This is a unique sort of question. If you can not get the info you need after a reasonable time then report the thread again and we can move it back.
:cop: thread moved to Car Audio.

This is a unique sort of question. If you can not get the info you need after a reasonable time then report the thread again and we can move it back.
I have a lot of what I need, thanks to Perry Babin's photos and schematic.


I too am an LTSpice fan. Been using it since it first came out in 1998. Before that, it was vacuum tube macromodels, using netlist-driven Tatum Labs' ECA in 1987. Log/Antilog amps using 1N914 diodes didn't work very well.

This is my first venture inside a consumer Class D. While I understand a bit of the tech, I'd bought and set aside a pair of those jelly bean 600W, $20 Class D amps.

I'll start on the schematics if I can stay awake. I won't even touch a DMM in my present state.

Thanks Again,

Ron
 
Jumpers like you're using can be very time consuming. The best option that I've found so far is the jumpers below. These are breadboard jumper wires. Searching eBay for Arduino breadboard jumper wires should get you good results. Be careful not to buy female-female or male-female jumpers. You want male-male jumpers.

This diagram is close to what you have.
http://www.bcae1.com/temp/SLA1500.pdf

Please don't start blindly changing parts. These amps are fairly easy to troubleshoot.
 
Hi Perry,

Your schematic shows a TL494, which is still a great jump for me. Thank you.

With my audio thinking cap on, I was looking for a Class D driver chip at that location. When I saw TL494 on the schematic, I looked at the data sheet - a general purpose PWM controller, primarily for power supplies? For audio? Why not? Even an 80 dB subwoofer dynamic range with ~ 1% THD should be just fine.

I did find some TI app notes mentioning TL494 audio, but haven't read them yet.

Except - my IC isn't a TL494 or a 16 pin device. I searched for a 14 pin STMicro PWM controller and found only one device. the inexpensive, low-component-count TSM108. Since all the Reference Designators on my board are different from the schematic and photo you sent, I just have to tread slowly. So far, 3 IC pins check out. Pin 7, the OCS (timing capacitor) pin, doesn't.

More to come.

Ron
 
The 494 is used to produce the drive pulses for the power supply, not for the audio.

The main board should be very much like the diagram. I can see the regulators and the optocouplers. This appears to be a generic clone that's very much like the SLA1500.
What you're saying makes sense. I know, physically, where the optos are. With the schematics in hand, I'll start looking on the main board.

I've also purchased the square pins and headers. They're a lot beefier than my original kluge connectors. An intermittent connection might also be a catastrophe.
 
What you're saying makes sense. I know, physically, where the optos are. With the schematics in hand, I'll start looking on the main board.

I've also purchased the square pins and headers. They're a lot beefier than my original kluge connectors. An intermittent connection might also be a catastrophe.
The audio input end of the main board, near the optos, is all linear circuits, filters, jelly bean TL084s and what looks like FET drivers. The little board may, in fact, be a PWM power controller. On the schematic link you referenced, the TL494 appears to be used as an audio PWM driver. I've found an actual TL494, on the main board, near the SPKR terminals. Now to find out whether it's used as an audio controller or a power controller. It'll still be slow going, since my board's Reference Designator numbers are still different from the schematics.

If it's an audio controller, per the schematic, it should have feedback from one of the speaker terminals. If it is, it may drive the optos at the other end of the board to avoid a GND loop, etc.

It took a sledge hammer to quickly remind me - a general purpose PWM controller can be used as a Buck, Boost, Flyback, Push Pull or even an inverting Cuk power converter. I'd never simulated a Cuk until 2 months ago. The TL494 can also be used as a jelly bean, consumer quality audio PWM driver.

Professor Slobodan Cuk is a major power designer, inventor and former teacher at Cal Tech. Sorry, I can't find the Serbian keyboard characters to properly spell his name.

https://en.wikipedia.org/wiki/Slobodan_Ćuk
 
Last edited:
Hi Perry,

I did a piece-by-piece power path, then a signal path. I virtually understand much of what's happening except for what's on the Tiny Driver Board. There are a number of floating supplies, a consequence of using all NFET output devices, because they generally have lower ON resistance than their PFET brethren.

For speed, every NFET bank is fed by a push-pull PNP/NPN Emitter Follower, referenced to that bank's Sources.

As you've stated, the 494 is indeed the power PWM driver/controller for all the Emitter Followers driving the NFETs of the two center-tapped push-pull transformer primaries, T1 and T2, that share a common toroidal core in my case??

The only GND or Chassis-referenced transformer secondaries are the main +/60V, the raw +/- 12V and PV+/-. Every other secondary is part of an NFET Source-referenced floating supply.

But, where are all the big rectifiers???

Are power FETs being used as synchronous rectifiers for the beefy supplies? Not unusual. Boosts efficiency and eases filtering, but the schematic I have shows diodes.

The Audio Signal Path, working backwards:

The upper output NFETs, with +60V on their Drains, have their own floating +12V, +5V and push-pull NPN/PNP Emitter Follower Gate drive, referenced to their Sources. The Emitter Follower Bases are driven by a digital opto whose supply is that upper bank's +5V. That digital opto's input LED is driven by pin 11 of the Tiny Driver Board.

The lower output NFETS , with -60V on their Sources, have their own floating +12V, +5V and push-pull NPN/PNP Emitter Follower Gate drive. referenced to their Sources. The Emitter Follower Bases are driven by a digital opto whose supply is that lower bank's +5V. That digital opto's input LED is driven by pin 10 of the Tiny Driver Board.

I hope I haven't swapped pins.

The Tiny Driver Board:

Since I don't yet have a Driver schematic, at this point, I can only surmise what it's doing.

It looks like its analog audio input enters header Pin 1, from one of the main board's op amps.

Pins 2 and 8 are supposed to be GNDs. I only read continuity to GND on pin 8.

Pin 3 is MUTE from the main board.

Pin 4 is RAW (minimally filtered), GND-referenced -12V to the driver board

Pin 5 is RAW (minimally filtered), GND-referenced +12V to the driver board

Pin 6 is the output drive, directly to the Emitter Follower feeding the lower output NFET's Gates.

Pin 7 is the output drive, directly to the Emitter Follower feeding the upper output NFET's Gates.

Pin 8 is the GND that doesn't Ohm yet.

Pin 9 is GND-referenced +5V to the driver board.

Pin 10 is drive to the GND-referenced input LED for the opto that feeds the lower output bank's Emitter Follower.

Pin 11 is drive to the GND-referenced input LED for the opto that feeds the upper output bank's Emitter Follower.

What I think I know so far:

1) Other than the missing rectifiers, FEEDBACK signal, Soft Start, Overload Monitoring. etc.............

2) A single analog audio input feeds the Tiny Driver Board and leaves as digital to the two opto's input LEDs.

3) That same driver chip that looks almost like a 14 pin STMicro TSM108, is still the likely Class D analog-to-PWM converter.

On to continued Ohming of the driver board tomorrow

BTW - after the amp sat for a week without power, I drew a small, clicky spark Ohming the driver header. I think it was either the RAW +12V or -12V, still around 3V. Good electrolytics. I'll look for bleeders.
 
Last edited:
The diagram tells you what the pin configuration for the driver board is.

Do you get a good square wave on the output pins (10 and 11) of the driver board if you drive a 50-100Hz signal into the amp?

Aren't 2 of the components in the red squares (below) the rectifiers for the rails.

In #3 above, what is the circuit board designation for the 14 pin IC you're referring to for the TSM108?
 

Attachments

  • Main-PCB-Topside-5-11-22x.jpg
    Main-PCB-Topside-5-11-22x.jpg
    102.4 KB · Views: 8
The diagram tells you what the pin configuration for the driver board is.

Do you get a good square wave on the output pins (10 and 11) of the driver board if you drive a 50-100Hz signal into the amp?

Aren't 2 of the components in the red squares (below) the rectifiers for the rails.

In #3 above, what is the circuit board designation for the 14 pin IC you're referring to for the TSM108?
I've just noticed Baldin said my 14 pin driver IC may be a LM219 Quad Comparator. That may also make sense if you're rolling your own Class D driver.

Thank you, Baldin.

-----------------------------------------------------------------------

Based on my 1995 circuit probe-induced meltdown, I'm still hesitant to probe a live power circuit until I understand what it's doing. Now that I know the driver is, in fact, a driver, I can either defeat the output FET drives, the +/- 60V or better yet, run the socketed driver board on the bench, all by itself.

Wednesday, I wanted to move a light mounted above my workbench. The light was kluged in place by a couple of those large BLK cable ties. When I went to cut one of the ties, There was a big spark, the tips of my newest diagonal cutters melted and the breaker tripped. The BLK cable tie was a live BLK 120VAC cord. Hadn't done that since 1985.

So I'm gun-shy.

I'll test the driver offline.

Remember, I haven't been down this path before. Those devices in RED, may, in fact, be the rectifiers. They're designated "Q" something, but don't have 1k series Gate resistors, like all the other NFETs.

The supposed TSM108 and the 8 pin chip below it have no visible designators. They may be underneath. In fact, none of the 4 ICs on the driver have visible designators. The two upper left ICs are 74HC02 quad NOR Gates.

Speaking of DIY - note my moniker and email address - the Raytheon CK722 was the first DIY transistor. I was a vacuum tube baby at age 8 and switched to the CK722 at age 11. Raytheon didn't tell us CK722s were re-packaged noisy hearing aid rejects.

---------------------------------------------------------------

Perry - if you're still in Louisiana, my buddies and I used to picnic on the banks of Lake Pontchartrain with two girls from Bay St Louis, Mississippi.

Baldin - my next door neighbor, Lisbeth, is from Denmark.

Those conversations belong on email, not here.

My Thanks to both of you,

Ron

[email protected]
 
Last edited: