SVT 7 pro no speaker output

[lowell]

Wikipedia has a great page on PWM that I just browsed.

Its odd to me, as there WAS a PWM signal on R131 and R129 when I first started scoping the amp. Guess its possible I screwed something up while poking around.

ok so should there then be a PWM signal at U21 even without signal? I was getting 5v pulsating DC on U21A and U21C outputs last I checked. Can't remember if that was with or without signal. I will check this tomorrow and also check to see if this is also at U21B output.

Finally, how does the 5v from U37 become pulsated?

BTW thanks for all the help!

[lowell]

Ok so yes there is 5v of pulsed DC on all outs of U21A, B, and C. The amp is apparently not muted, and the ~5V on the collector of Q33 tells us that. So... now I'm lost. Guess I may build this filter to see where the AC signal is stopping.

well I just used my signal probe and there is indeed signal reaching the output of U21C pin 2. There is nothing on the gate of Q10 on the other hand.

[MarkusBass]

Lowell, try to go back to this suggestion, please.

This may be not that easy because, as far as I understand, now there is no signal on inputs. But general rule is the same; if you don't have signal on the otput -> check input. So, if there is no signal on U16A, check where the signal is coming from. And in this case it is coming from the amp output. If there is no square signal there, the amp is not working at all. But previously it was working. So what happened? Isn't it muted?

Mark

[lowell]

Not sure if it's muted or not. There is ~5vDC on the collector of Q33. I cannot find anything in the 20955 datasheet that explains what DC voltage I should be seeing on the CSD pin. It talks of the th1 and th2 thresholds but it does not explain what these thresholds are either. I'm confused.

It's looking as though I may need to remove the entire board of this amp. The 20955 ICs have what I believe are heatsinks mounted on them so I cannot access the 20955's pins directly. I think I need to unsolder the heatsinks from the board in order to access these pins. I can then use my signal tracer to confirm as to whether or not signal is reaching IN of the 20955.

[Jazz P Bass]

We ruled out a low level mute command.
Q33 is off, pin 2 is high.
If CSD pin 2 is high, then the amp is not muted.
Unless...
If the IC (high level mute) commands a mute function (because a malfunction is inputing a mute signal) then pin 2 will be low. Briefly.
The 10uf capacitor is a time out until the IC can try to unmute.
The 10uf will slowly let CSD rise back up to 4.7 Vdc.
At this point the IC will try to run, by first checking the conditions for a mute function. If the check mute command fails again then the IC will internally pull the CSD pin low again.
Then the 10uf will delay the retry.
Ad nauseum until the input that caused the mute is removed.
So look at pin 2 with the scope & see if this is what is happening.
Is pin 2 bouncing low then high.

[Jazz P Bass]

No
It's looking as though I may need to remove the entire board of this amp. The 20955 ICs have what I believe are heatsinks mounted on them so I cannot access the 20955's pins directly. I think I need to unsolder the heatsinks from the board in order to access these pins. I can then use my signal tracer to confirm as to whether or not signal is reaching IN of the 20955.

Trust your DVM. Leave the IC's alone for now.
If you can get to the IC pins at all. (use a jumper wire & a sewing needle)
Run a continuity check from an outboard component (IC21-C/3 to U300/2 & IC21-B/5 to U17/2.
While you are in there verify Q33/ collector R39/ 100 ohm is good & that it does connect to Q33 & IC17/2.

[lowell]

Okay I will check that on Monday. Meanwhile I'm studying as much as I can about how Class D amps work. Is the PWM coming from D55? Or before that? If that's it, HOW are those diodes and the 5v from U37 creating the PWM?

[Jazz P Bass]

PMW Signal

Can't help on this one.
I am clueless, being at a decided disadvantage that I do not have the equipment in front of me, nor have I ever had the (dis)pleasure to work on one.
Try this link from Sound. Westhost. Class-D Amplifiers
Here is a snapshot of the PMW Wave.

[lowell]

You know, I had read that page. Very good info.. but it still doesn't explain how the PMW is created. It mentions using a triangle wave generator, but I don't think that the Ampeg uses that topology. Even then, the page doesn't describe the triangle wave generator in great detail.

[Jazz P Bass]

Well that bottom wave form is what the signal should look like.
You can see how the width is changing.
Now I would have to assume that without a preamp signal, the PMW will still be going through the output section.
But at a 50% duty cycle, which should consume very little power.
And the speakers should not see it because of the output filter.

[MarkusBass]

Lowell,

Plugging headphones into this amp causes that the power amp gets muted. I hope that you don't have headphones plugged in :-). This shows how easy you can check whether muting functionality works. Just plug and unplug headphones and measure voltage on R39 or C174 (which is the same as pin #2 of IRS20955). You shouln't worry about U37 and D55 diodes. They just make sure that the input of U21 does not get more that 5V (which would destroy the IC). You may consider that this is a protection for U21. The PWM signal is created in the preceeding IC. If you look at its datasheet, you will notice that this is an op amp with 100MHz bandwidth.
If you get square signal on outputs of U21, I would start checking IRS20955. But I hope that you know that you cannot just desolder heatsink from the IC and turn the amp on. This would be a disaster. I would start checking power supply voltages (there are several power supply voltages required including a reference voltage - read the datasheet). If you don't have access to IC's pins, just measure voltages on parts connected to the IC. Be careful when connecting an oscilloscope to gates of the output transistors - I already warned you not to do it.

Mark

[lowell]

Yes you're right. And no the amp is not muted. The ~5vDC Vc on Q33 changes and reads 0v when I insert a plug into the headphone jack.

Ok so here we go:

U300
CSD I cannot directly access this pin however the Vc on Q33 tells me that it should be ok, assuming there is continuity to this pin.
IN pulsating 5vDC
VREF -50v (again cannot access the pin but this voltage is on the other side of R107
OCSET -50v
DT -55v
VSS 0v
COM -55v
LO -55v
VS -1v
HO -1v
VB 5v
CSH -1v


U17
CSD "same as above"
IN pulsating 5vDC
VREF -50v on opposite side of R220 (couldn't access the pin directly)
OCSET -50v
DT -55v
VSS 0v
COM -55v
LO -55v
VS -1v
HO -1v
VB -18v
CSH -1v

U300 seems to have a couple voltage issues. VDD pin1 is low at 4.2v. Also VB pin 15 is low at 5v. These are low compared to the voltages on U17. It seems more logical to me that U300's voltages are low rather than U17 voltages being low, just from the surrounding components and voltages listed on the schematic. I'm sure what to do with these voltages other than replace U300.

[MarkusBass]

U300 seems to have a couple voltage issues. VDD pin1 is low at 4.2v. Also VB pin 15 is low at 5v. These are low compared to the voltages on U17. It seems more logical to me that U300's voltages are low rather than U17 voltages being low, just from the surrounding components and voltages listed on the schematic. I'm sure what to do with these voltages other than replace U300.

Please note that you haven't specified Vdd voltage for U17 so it's difficult to say whether it's too low or not. Before you start desoldering U300 I would check diodes in the power amp. Have you checked (with magnifying glass) how surrounding components are soldered? I would also look (very carefully) at HI and LO outputs of both U300 and U17. Can you post a photo of how U300 is soldered?

Mark

[Steve Conner]

Those gate drive chips have about 10 different kinds of protection, all of which shut them down. I'd check that all supply rails are present, and none of the power MOSFETs is blown.

Is that Vb really -18V?

[Jazz P Bass]

It sure would be nice to know the Vdd voltage on pin #1 of both IC's.
Internally to the IC is a 10 volt zener diode. (hint).
It was asked previously what the voltage should be on CSD pin.
It is a percentage of Vdd.
Which sets Vtr1 & Vtr2.