Sherwood Car Amp

[jheatac]

Hello.

I am new here. I was interested in electronics as a kid but never really developed the hobby. I have worked around electricity all of my life.  I have decided now that I am going to pick it up and try to understand it.

Resistors, capacitors, diodes, I understand. I'm struggling getting transistors straight in my head to where I can look at a schematic in my head and simulate what I think should be happening.

My first project is to fix a car audio amp that my brother reversed polarity to 20 years ago. I tinkered with it a little back then, but never got anywhere with it.

Sherwood GA-1072BP
I have found a service manual/schematic that is listed for this unit, but is slightly different.

The differences I have found:
R558 - 10k in the unit, not the 22k listed
R568 - 5.1k in the unit, not the 100k listed
D520 - Does not exist. This line was fed from the AC line just beyond transformer T501. Instead, on this unit, it comes from the rectified line feeding the amp.
Between pin 2 and pin 13/14 of IC501 is what I believe are two thermistors in series. One mounted on each side of the case heat sinks, wired in series at about 170 ohm at room temp.

What I have discovered.

Highlighted components on the attached partial schematic have been tested
I put 14.4 volts into the unit as it lists on the schematic.
I measure 4.16v at pin 2 of the 494C instead of 4.3v listed.
I measure 3.75v at pin 1 of the 494C instead of 4.3v listed.
Output is at about 57v, instead of the listed 66v listed
I do have the 5v Vref at pin 13 of the 494C
I do have 10.6v at pin 12 of the 494C (10.5v listed)
Diode D514 was blown in half when I started and has been replaced with listed diode
Frequency of the oscillator is about 35khz
Pulse width of the two channels are the same.
The values of the highlighted components have been verified (probably some addition before I started highlighting. The pink highlighter was removed from circuit to check.)


I have a few questions:

1) Obviously, if someone can give me some direction, I would be excited to hear.
2) How does feedback work on 494C dual error amplifier work if they are tied together?
3) While powered on, if the error amps are working, should the inputs be the same? (Pin 1 = Pin 2 and Pin 15 = Pin 16) -
4) Am I barking up the wrong tree? as in, is the power supply close enough that the problem is probably in the amplifier section?

Thanks for any input

[rcowboy78]

I don't quite understand the nature of the problem. It sounds like its on, do you just not have any audio out?

[Rerouter]

If D514 was blown, then its likely that Q511 is shorted Base to Collector. If so, i would also be suss on Q510, these are the main output stage drivers.

I would have to a much greater extent have expected D512 to have blown to protect the device, as it appears its sole purpose is as a reverse protection diode intended to pop the fuse. which is suss to begin with, having a 3A diode try and pop a 30A fuse. (a 30A fuse may take 120A to pop in under 1 second)

Aswell i would assume C516 and C519 to have been damaged.

The remote wire has diodes protecting it against reverses, so i would say anything relating to that should be ok.

Pull out a multimeter, and measure between all 3 pins of the transistors on the board, and make sure none of the diodes are open or short. i suspect you will find 1 open diode, and 2 shorted transistors.

[xwarp]

Can you post a complete image of the amplifier's schematic?
When you refer to the "output" being at 57 volts, what "output" are you referring to?

[CJay]

What we are looking at here is the power supply from a car amplifier, there's no audio amp in the attached schematic.

To the OP, have you got the rest of the schematic and can you post it?

[jheatac]

I don't quite understand the nature of the problem. It sounds like its on, do you just not have any audio out?

Yes, the ultimate problem is no audio out. Reversing input polarity was the cause.

In looking at the schematic, it says that I should be getting +33 and -33 out of the power supply circuit and I am not, so I started tackling the supply circuit first before looking at the amplifier.

[jheatac]

Can you post a complete image of the amplifier's schematic?
When you refer to the "output" being at 57 volts, what "output" are you referring to?

I have attached the service manual. This unit is the 1002. At some point they published a manual that lists the GA-1072BP and 1002 both, but I do not have it. With only a few exceptions, it seems to match.

I am checking output on the secondary side of the transformer (AC) and just after the rectifier before entering the amplifier (DC)

[jheatac]

What we are looking at here is the power supply from a car amplifier, there's no audio amp in the attached schematic.

To the OP, have you got the rest of the schematic and can you post it?

Yes, this was just the power supply circuit. The full schematic is attached to the post above

[jheatac]

If D514 was blown, then its likely that Q511 is shorted Base to Collector. If so, i would also be suss on Q510, these are the main output stage drivers.

I would have to a much greater extent have expected D512 to have blown to protect the device, as it appears its sole purpose is as a reverse protection diode intended to pop the fuse. which is suss to begin with, having a 3A diode try and pop a 30A fuse. (a 30A fuse may take 120A to pop in under 1 second)

Aswell i would assume C516 and C519 to have been damaged.

The remote wire has diodes protecting it against reverses, so i would say anything relating to that should be ok.

Pull out a multimeter, and measure between all 3 pins of the transistors on the board, and make sure none of the diodes are open or short. i suspect you will find 1 open diode, and 2 shorted transistors.

Thank you for the direction. I am nearly certain that I pulled either Q511 or Q510 and tested it, but it was before I started highlighting.

I know I have not pulled those caps. I will check on it and let you know what I find. Thank you

[jheatac]

Can you post a complete image of the amplifier's schematic?
When you refer to the "output" being at 57 volts, what "output" are you referring to?

I believe you are correct. I think D512 was what was in half. D514 has been checked. I know it was a IN5402 that was replaced.

[WaveyDipole]

Yes it makes sense for D512 to have blown itself apart first. Easy to tell which one it is as the 1N5402 would physically be much bigger and have thicker leads than the 1N4002. Without the rest of the circuit it is difficult to see what might be causing the slightly low voltage on pin 1 of the 494C, although it is encouraging that other voltages around that IC seem OK, the osicllator is working and you are getting power output. Possible causes based on what we can be seen of the circuit might be a resistor gone high or a leaky C520.

How did you measure that AC voltage across the output transformer? If you are not using a multimeter that is capable of  measuring TrueRMS at 35kHz (the oscillator frequency) then the reading may be misleading. Since you say there is no audio output, a fault in the output circuitry might be dragging the voltage down as well.

[jheatac]

Update:

I pulled and checked C518 & C519 test 957uf 970uf... A little low but within 5%

I pulled Q510 and Q511. Neither will work with my meter on the diode setting.

I put them in the little transistor checker and Q511 shows up as bad/no part and Q510 tests out of spec, I believe.

Q510 shows as:
hFE=321 spec I believe is 320 max
Vbe=1.81v spec I believe is 1.2v max
Ic=4.7ma spec I believe is 1a max

The part is a 2SD600F (also marked with 9M, which I'd guess as a date code?)
The spec sheet I found was for a 2SD600K

Can someone give me a common replacement? I can order some from eBay, but I'd rather place an order with an electronics site.

[rcowboy78]

Before you place an order for those transistors, check all of your output transistors and make sure there not shorted too.

[jheatac]

Yes it makes sense for D512 to have blown itself apart first. Easy to tell which one it is as the 1N5402 would physically be much bigger and have thicker leads than the 1N4002. Without the rest of the circuit it is difficult to see what might be causing the slightly low voltage on pin 1 of the 494C, although it is encouraging that other voltages around that IC seem OK, the osicllator is working and you are getting power output. Possible causes based on what we can be seen of the circuit might be a resistor gone high or a leaky C520.

How did you measure that AC voltage across the output transformer? If you are not using a multimeter that is capable of  measuring TrueRMS at 35kHz (the oscillator frequency) then the reading may be misleading. Since you say there is no audio output, a fault in the output circuitry might be dragging the voltage down as well.

Voltages were measured on a Fluke 45. I didn't give a thought to the frequency, I have a Fluke 289 or a scope, but at this point the amp is apart waiting on transistors. I've checked the DC after the rectifier and had comparable voltages.

In thinking that maybe the output circuit is dragging down voltage, I've also considered removing the inductors after the rectifier in order to isolate the power supply from the amp. I am powering from a benchtop power supply and my amp draw is only 0.59a @ 14.4vdc and I have no noticeable heat generated by the power transistors, so I didn't think it was loaded.

Also, given that D512 blew apart at the time of reversed polarity, if the fuses were replaced without noticing the reverse polarity that is when the damage would really occur because at that point there is no longer a diode to blow the fuse, so the circuitry takes the abuse?

[WaveyDipole]

Here is some data on the 'F' variant:
http://alltransistors.com/transistor.php?transistor=19778

If you click the cross-reference search you get the following:
http://alltransistors.com/crsearch.php?mat=Si&struct=NPN&pc=8&ucb=100&uce=100&ueb=5&ic=1&tj=150&ft=130&cc=20&hfe=160&caps=TO126

Using the search criteria on Farnell and RS I got the following:
http://uk.farnell.com/nte-electronics/nte373/bipolar-transistor-npn-160v/dp/4197288
http://uk.rs-online.com/web/p/bipolar-transistors/9034305/
The second one is on the list above but hopefully someone can confirm that these choices are OK.

Like rcowboy says, before you place an order, check the other transistors and any capacitors you might want to replace while you are at it. The postage cost mounts up otherwise.

Electrolytics are often up to 20% tolerance although you can get 10% or tighter. Your readings are well within spec, however electrolytycs are easily damaged by having reverse polarity applied to them. They might well be OK but the cost of a good quality (Panasonic/Nichicon/Vishay etc) 1000uF/16v capacitor is not so prohibitive so I would probably replace both of them for peace of mind anyway. If there is any sign of bulging or electrolyte leakage on any other electrolytic capacitor then it should be replaced as well.

The Fluke 45 does TrueRMS up to 100kHz so no problem there. I was thinking more about someone using one of those cheap things from the far east.

[jheatac]

Before you place an order for those transistors, check all of your output transistors and make sure there not shorted too.

The output transistor is a 2SD1213-R

I do not know which one came from which location but here is what I found:

All have a BC & BE diode test of 0.569 to 0.576 mV
None have any registered resistance between BC, BE, CE in either polarity
All have continuity between the back plane and the Collector - I assumed since all did, this was normal, but is it?

Tested as following on transistor checker:

hFE    vBE    Ic
177    550   6.4mA
177    568   6.4mA
177    558   6.4mA
448    549   6.4mA
279    550   6.4mA
298    554   6.4mA

I would guess the last 3 are from the same bank. The spec sheet was listed as a max hFE of 200 on -R

Are these bad?

[rcowboy78]

All have continuity between the back plane and the Collector - I assumed since all did, this was normal, but is it?

Yes, on most TO-220 packages it is. Some are, some aren't. Just depends. However I assume there is some insulator between it and the amps case. All you really had to do is a simple diode test to make sure they weren't shorted.

I think your DC to DC converter was fine if your amp was powering up, which it sounded like it was in your first or second post, because you said you had 57 or some odd volts on the rails.. I guess go ahead and replace those transistors, but your going to have to trace your input signal to find out where it is getting lost at. I was just trying to help save you some on shipping cost.

I'm still slightly confused on the amps current state. Does it power up? What is the voltage on the output transistors at their collectors?

First things to tackle would be DC converter, then preamp, then output stage. I also assume that there are probably a couple of filters to check too.

[jheatac]

It did have insulators between the transistors and the case/sink.

I will wait for parts to show up before testing further. Earlier, I did notice I had 18v between the case and ground connection. The board itself does not ground to the case. It has 6 or 8 screws with ground trace under them, but the ground trace is painted over.

[rcowboy78]

The amps case should be referenced to ground (i.e. 0V). Having 18V there is a little concerning. Tells me something is still shorted.

[jheatac]

The amps case should be referenced to ground (i.e. 0V). Having 18V there is a little concerning. Tells me something is still shorted.

I did - there were no amps when checked with meter.

[jheatac]

I was going to replace the two 1000uf/16v caps.

It seems everything in the size is considerably smaller in physical size. I have a found a 1000uf/50v with the same dimensions and pin spacing. Is there any reason not to use these?

[rcowboy78]

It seems everything in the size is considerably smaller in physical size. I have a found a 1000uf/50v with the same dimensions and pin spacing. Is there any reason not to use these?

Should be fine , you can use them. If the amp was made in the 90's, the cap sizes probably will be different. The dielectric over the years has gotten better allowing for a decrease in size in capacitors 

Just got done working on an old HP power supply from the 70's, and I almost ran into an issue where the newer axial capacitor leads were almost not long enough to cover the body length of the older capacitors. 

[jheatac]

The amps case should be referenced to ground (i.e. 0V). Having 18V there is a little concerning. Tells me something is still shorted.

Just a thought - I pulled a couple of the power transistors a while back and checked them. When I put them back, I cleaned and reapplied thermal grease to both sides of the mica. The thermal paste I have looks like a silver metallic. Would this conduct from the collector to case to cause the voltage?

[rcowboy78]

My suggestion would be to but it back the same way it came apart. If they used the mica with thermal compound, then put it back that way. There are some electrically non-conductive thermal pads that you can purchase. However, the way you did it should be fine because most thermal greases are electrically non-conductive. I would look at the data sheet for the grease that you used. It should tell you.

If the compound that you used has silver in it, I would replace it with an aluminum oxide one.

[floobydust]

Note the audio section is floating from DC GND, up to a diode vdrop D521, D522 and R575 220R, to prevent ground loops with the head section. (these are by bridge rectifier)
These parts might have been damaged too.

[Armadillo]

Just a thought - I pulled a couple of the power transistors a while back and checked them. When I put them back, I cleaned and reapplied thermal grease to both sides of the mica. The thermal paste I have looks like a silver metallic. Would this conduct from the collector to case to cause the voltage?

Aside from the mica, Screws needs plastic sleeve that you have to install. Check continuity between collector and the heatsink, if you are worried.

I would recommend as attached.

first, ensure that you have 10.5V at pin 12 of IC501.

then, use a digital logic probe, probe pin 8 and pin 11 for pulses. Any sign of life?

if yes, then you continue probe the Q510 and Q511 bases.... and continue .... until you find no pulses.

logic probe is not expensive, $10 I think.

Note: set logic probe to cmos, tie positive to Battery+ and negative to ground.

[jheatac]

Just a thought - I pulled a couple of the power transistors a while back and checked them. When I put them back, I cleaned and reapplied thermal grease to both sides of the mica. The thermal paste I have looks like a silver metallic. Would this conduct from the collector to case to cause the voltage?

Aside from the mica, Screws needs plastic sleeve that you have to install. Check continuity between collector and the heatsink, if you are worried.

I would recommend as attached.

first, ensure that you have 10.5V at pin 12 of IC501.

then, use a digital logic probe, probe pin 8 and pin 11 for pulses. Any sign of life?

if yes, then you continue probe the Q510 and Q511 bases.... and continue .... until you find no pulses.

logic probe is not expensive, $10 I think.

Note: set logic probe to cmos, tie positive to Battery+ and negative to ground.

I had a little higher than 10.5v. It was 10.65 or so.

I had a pulse from each. I had voltage out of the supply but not enough.

The wave, if I remember correctly, had a considerable flutter in the middle. It would cross 0v and rebound about 30% then go negative. Not that I know exactly what I'm supposed to see, but it didn't look right.

Sent from my SM-G920V using Tapatalk

[Armadillo]

I had a little higher than 10.5v. It was 10.65 or so.

I had a pulse from each. I had voltage out of the supply but not enough.

The wave, if I remember correctly, had a considerable flutter in the middle. It would cross 0v and rebound about 30% then go negative. Not that I know exactly what I'm supposed to see, but it didn't look right.

Sent from my SM-G920V using Tapatalk

What is the voltage at;

C518 =
Q511 Vbe =
Q510 Vbe =

Q512 Vbe =
Q513 Vbe =
Q514 Vbe =

Q515 Vbe =
Q516 Vbe =
Q517 Vbe =

[jheatac]

I will hopefully have parts to put this together tomorrow. Then I can get you some measurements.

Can someone explain to me how the to error amps in the 494 ICC work with a common output? I'm not very knowledgeable, but I can't seem to grasp what is supposed to happen or what to expect the output to be

Sent from my SM-G920V using Tapatalk

[floobydust]

TL494 has two error-amplifiers, wired together in a "OR" configuration.
"... With both outputs ORed together at the inverting input node of the PWM comparator, the [error] amplifier demanding the minimum pulse out dominates..." {datasheet}

So Error Amp 1 OR Error Amp 2 can lessen or stop PWM.

In the Sherwood, Error Amp 1 is used to regulate the DC-DC converter output (feedback to pin 1) compared to pin 2 as the setpoint voltage.

But pin 2 is used for another thing.
I think Q520/Q519 ON instantly shuts down PWM. So REMOTE ON turns them off to enable the TL494. Kinda confusing. As long as pin 2 can have its ~4.3V setpoint for Vout.

Error Amp 2 doesn't do much, pin 16 GND and it seems to limit max. duty cycle less than 50%.


You should be able to run the TL494 from REMOTE only (leave BATT unconnected), unless Q520 gets in the way.
It is one way to check signals with DC-DC driver/output transistors unpowered. The pre-driver stage Q508/Q509 should still be working.
The TL494 should just go full tilt, max. PWM trying to bring up the DC-DC output voltage.  R556/C521=33.3kHz I would expect.

[jheatac]

TL494 has two error-amplifiers, wired together in a "OR" configuration.
"... With both outputs ORed together at the inverting input node of the PWM comparator, the [error] amplifier demanding the minimum pulse out dominates..." {datasheet}

So Error Amp 1 OR Error Amp 2 can lessen or stop PWM.

In the Sherwood, Error Amp 1 is used to regulate the DC-DC converter output (feedback to pin 1) compared to pin 2 as the setpoint voltage.

But pin 2 is used for another thing.
I think Q520/Q519 ON instantly shuts down PWM. So REMOTE ON turns them off to enable the TL494. Kinda confusing. As long as pin 2 can have its ~4.3V setpoint for Vout.

Thank you for taking the time to explain that. If I read it a couple more times it may sink in.

Error Amp 2 doesn't do much, pin 16 GND and it seems to limit max. duty cycle less than 50%.


You should be able to run the TL494 from REMOTE only (leave BATT unconnected), unless Q520 gets in the way.
It is one way to check signals with DC-DC driver/output transistors unpowered. The pre-driver stage Q508/Q509 should still be working.
The TL494 should just go full tilt, max. PWM trying to bring up the DC-DC output voltage.  R556/C521=33.3kHz I would expect.

[jheatac]

I have another situation now.

I pulled the 6 power transistors out - Q512-Q517 - thinking they were all the same.

The parts list says they are 2SD1213-S (with the exception of 1 2SP1213-S, which I think was a misprint)

I goofed and ordered -R, based off of the one I had in hand when ordering. I realize now, that 3 were -R (Hfe 100-200) and 3 were -S (Hfe 140-280)

My problem now, is I don't know which came from where. I thought I compared one at each side before removal, but now I am questioning that.

Based off of the parts listing, I would think it was 2 of the S and 1 R on one side and 1 S and 2 R's on the other. Can I put all 6 as R's?

At this point I'll replace the Rs with new, and put the Ss back in, however by my testing the Ss were more out of spec.

[WaveyDipole]

I would have expected them to all be identical spec rather than a mixture of R and S types although this is not likely to be much of a problem in practice. I don't think putting in R types in all 6 positions would be a problem and having all of them the same spec would be preferable.

I had a pulse from each. I had voltage out of the supply but not enough.

The wave, if I remember correctly, had a considerable flutter in the middle. It would cross 0v and rebound about 30% then go negative. Not that I know exactly what I'm supposed to see, but it didn't look right.

Sent from my SM-G920V using Tapatalk

So you are observing this with an oscilloscope?
When measured with reference to V+ (pin12) the pulses at pin 8 and pin 11 as well as the base of Q508/Q509 should be identical but in opposite phase. When measured with reference to GND, the pulses at the base of Q510/Q511 should also be identical but in opposite phase.

I was going to ask at which point in the circuit did you observe the wave that had the flutter? Of course, once you replace the transistors if the situation changes then it may not matter. Lets see what happens when that is done first.

[jheatac]

I am getting -35v and +28v at the amplifier section now

I have replaced all 6 of the 2SD1213-R(or S) with 2SD1213R (Q512-Q517)
I have replaced both of the D600F with NTE373 (Q510 & Q511)

I have attached a shot of the scope on pins 8 & 11 of the 494C chip with reference to pin 12.

It looks like it is maxed out to me, but I am not knowledgeable.

The output wave heading to the rectifier looks like hell to me, but again, I am not knowledgeable.

[jheatac]

Here is the waveform at the input of the step-up transformer - one pole referencing the other pole

The other shot is the base of the power transistors referenced to ground.

[floobydust]

The waveforms look right to me for an 80's bipolar push-pull switcher with no snubber and Darlington output trannies.

TL494 pins 8, 11 go low to turn on the transistors so the scope trace shows they are off most of the time= low duty cycle, idling. It seems to be ~2usec PW but base-drive scope trace different suggests ~8usec on time so not sure why the difference. Point is the PSU is not stuck working really hard and things are symmetrical.

Are the speaker protection relays going click?

I'm concerned the rails are not so equal +28V vs -35V.
The +ve rail is usually a bit lower than the -ve rail because the prot. relays are powered off it.

The power amp sections draw current from +ve to -ve rails (no current flow to GND unless SPKR connected) so they should not be able to imbalance the rails.

I would look and see if the four power amp sections are balanced, having near zero volts on their outputs.
Just to see if anything is wonky and loading the rail down to +28V.  Or the rectifier diodes or filter caps might be off.
Make sure D521/D522 are OK and you get 220R from input DC GND to Audio GND. The amplifier side uses this as a reference point.

I think this is a good old school Class AB amplifier, I tried to buy something like it but everything nowadays is Class D with higher distortion. It's worth repairing. Must be pretty big, like Alpine 3544.

[jheatac]

Okay. So I made some progress. Some of it was solving my own stupidity, but still... progress.

I had the amplifier functional last night. Once I figured out the voice coil was blown in my test speaker and got a second speaker, it started working.

I was playing with it a little more and noticed I had about a 3v wave coming from the output with no input. I put a 5W, 4.7ohm resistor as a dummy load (with no input) and it did not change this.
It was identical on two channels and about half as much on the other two channels. (Scope pic attached)

But in the process of poking around something went wrong. I don't think I did anything, but I noticed that the amperage draw went from 0.69 to maxing out the 5amp bench supply)
The pulse from 8 & 11 have a large spike in the middle of the off cycle.

I have to go to work but I will work on it later this evening.