Help fix amp - rail voltage gone crazy

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I had an issue with my amp - a Pioneer GM-x922.

Long story short, now it turns on however there is a difference in the rail voltage

The positive is 31V and the negative 45. According to the instructions, it should be in the 30V range so the negative is way out of range, especially taking into consideration that the rail capacitors are 35V rated.

I checked the output voltage of the TL494 Pin 9 and 10 and there is a difference here. Pin 9 is 1.1V while Pin 10 is 0.2V

My understanding is that these two should be equal however I can not figure out why they are not.

Any suggestion?

Thanks
 
If you place the black meter probe on one of the non-bridging speaker terminals and the red probe on the center legs of the output transistors, do you still get different positive and negative rail voltages?

Are you sure you have the pin numbering on the 494 correct (see attached)?
 

Attachments

  • pstl594pinnumbers01.gif
    pstl594pinnumbers01.gif
    9.5 KB · Views: 257
Thanks Perry - I was hoping you was going to answer.

I actually removed the finals as this amp mount some old Sanken that are difficult to find here - I'd like to avoid to fry them.

I double checked the numbering and I did it right. Pin 9 and 10 are the one driving the IRFIZ44N according to the amp service manual.

This amp has the rail capacitors rated 35V so anything in the 50V range is not well tolerated by them and if they short, due to the overvoltage exposure, the Sanken may toast - this was my assumption, at least.

Tomorrow I'll re-solder the Sanken and read the voltage as suggested.

In the meantime, I have measured the voltage at all pins of the 494. Pins 2 and 3 looks high to me - 3.52V
 
If 9 and 10 have significantly different voltage, the IC may be defective, something is loading one of the outputs or some input to it is causing it to malfunction.

You can measure the rail voltage at the center legs of the rectifiers (D961/962) without reinstalling the outputs.

Without seeing the voltage on all of the pins, it's difficult to determine what 2 and 3 should be. Can you post the DC voltage on all of the pins of the 494? Place the black meter probe on chassis ground. Place the red meter probe on the point where you need to measure the voltage.

IC951
Pin 1:
Pin 2:
Pin 3:
Pin 4:
Pin 5:
Pin 6:
Pin 7:
Pin 8:
Pin 9:
Pin 10:
Pin 11:
Pin 12:
Pin 13:
Pin 14:
Pin 15:
Pin 16:
 
If 9 and 10 have significantly different voltage, the IC may be defective, something is loading one of the outputs or some input to it is causing it to malfunction.

You can measure the rail voltage at the center legs of the rectifiers (D961/962) without reinstalling the outputs.

Without seeing the voltage on all of the pins, it's difficult to determine what 2 and 3 should be. Can you post the DC voltage on all of the pins of the 494? Place the black meter probe on chassis ground. Place the red meter probe on the point where you need to measure the voltage.

IC951
Pin 1: 0
Pin 2: 3.5 (after 1 minute it goes up to 4)
Pin 3: 3.5 (after 1 minute it goes up to 4)
Pin 4: 0.28
Pin 5: 1.66
Pin 6: 3.7
Pin 7: 0
Pin 8: 13.2 (the amp is supplied with 13.2 volts)
Pin 9: 1.28 (after 1 minute it goes down to 0.57)
Pin 10: 0.44 (after 1 minute it goes up to 0.57)
Pin 11: 13.2
Pin 12: 13.2
Pin 13: 5
Pin 14: 5
Pin 15: 2.4
Pin 16: 2.4

I originally measured the voltage at the diodes and re-done it today: +34 and -47.5. Measured after 1 minute the amp is ON, they are +33 and -48 while some of the IC951 voltages changes significantly (to me)

I also measured leg of the IRFIZ. Q963/5/7 have 0.7V while Q 964/6/8 have 0.65.

Giving the high negative rail voltage, I removed the rail capacitors. Hope this is not causing any problem to the readings
 
For now, I think the IC may be OK. The voltage could be different because the soft start causes the voltage to vary when starting. If you don't read them at the same time, they may be different.

Which legs had ~0.7v (gate?)?

You need to have rail caps in while testing. Lower value caps rated for 50v would be OK.

Does the voltage adjustment potentiometer for the rail voltage (VR951) have any effect on the rail voltage? If so and you can get the negative rail to -35, reinstall the original rail caps.

You should pull the rectifiers and check them. Confirm that both diodes in each rectifier are OK. With your meter on diode check, you should read ~0.4 in one direction and open in the other direction (from the center leg to either outside leg).

The rail caps (one positive and one negative rail cap) have two terminals connected together (one terminal will be positive and one negative). Do you read 0 ohms from that point to the chassis ground terminal?

And also from the chassis ground terminal to the non-bridging speaker terminals?
 
Soldered the 50V caps - the negative rails goes up to 55V while the positive is still OK.

I moved the VR951 and I can see the positive voltage going up and down (25V to 36V) while the negative has minimal variations (52-55).

The 0.7V were on the leg1.

Diodes are OK and the readings are ok - 0 ohm

Looks like the VR951 has no effect on the negative rail. I'll have a look to the components between it and the IC951
 
I spent few time looking at all traces and did not found any open one.

I fitted two FETs back (one per rail side) and the voltage is still crazy.

I checked another 922 amp and found that the IC951 pin 9&10 voltage is higher - it's in the 4V range vs the 0.5V of the defective amp.

Now, randomly, the amp does not start when I connect the remote. To be precise, it turns on but the leds flash and the rail voltage stays low (10-15V) - when stop flashing, the rails are back to the "normal" +30/-52V.
 
Is that PSU regulated? Does it have output inductors right after the secondary side diodes without electrolytic capacitors first? Does it have two separate inductors or one with two windings on the same core? A couple of pictures would help.

Does the negative rail drop when some load is applied? (Use a power resistor to draw 0.1A to 1A or so...)

Taking a look at transformer waveforms with an oscilloscope would give a lot of information about the problem. Higher than expected output voltage in unregulated SMPS can result from unexpected resonances (shorted turns in the transformer, small ceramic caps shorted, etc...)

Note that some not-so-good push-pull SMPS designs will produce higher than expected output voltage just when the output is not loaded. Operating the SMPS at the wrong frequency can do the same.

Once I got an amplifier with wrong rail voltages that cooked supply electrolytics and snubber resistors. Oscilloscope revealed that it was operating at 300Khz rather than 30Khz. Timing capacitor was open. The SMPS was no longer operated in resonant mode (zero voltage turn on for the primary side MOSFET) resulting in heavy overshoot.
 
An externally hosted image should be here but it was not working when we last tested it.


Here it is the amp as it stands right now.

It's regulated an hads two inductors (left side of the pic) just after the diodes.

As I removed the finals, I can not apply any load.

Unfortunately, I have no access to a scope so it's hard to verify what you mentioned, Eva.

As only the negative rail is wrong, has to something in the negative regulator path.

I have the pdf of the service manual, if this may help to help me.
 
I mean applying a load from the negative rail to ground (or from the positive rail to the negative one), drawing current directly from the supply rails. The voltage will probably drop with some load current.

On that kind of power supply, due to the separate inductors, the negative rail is usually not under control and voltage is very likely to rise if there is not a minimum load current. The bias current of the output stage was probably helping a lot with that (most car amplifiers are class B but I think this one is true AB as older classic ones from Pioneer).

So with output devices removed the negative rail is very likely to be higher than expected, even on a perfectly working amplifier.

If you have changed output devices (particularly if the new ones are of a different type), you should check bias current too.

btw: The image does not expand when I click on it, I think it already appears full size (quite small).
 
Last edited:
OK -Understood. I'll apply the load as you said.

I may have reduced the pic when I uploaded - let me try again.

The output device are still the "old" ones. I just removed them to check if they were still OK and never re-mounted as I was afraid to fry them.

Pioneer claims this is a class-A amp so it has to be an AB with an high bias.

An externally hosted image should be here but it was not working when we last tested it.
 
3.9 ohms was a too low load (over 10A and 500W with 47V), and 470 is a bit on the high side (100mA with 47V) but should have been enough to produce a substantial voltage drop.

The inductor of the negative rail may have shorted turns. Try to swap the output inductors of the PSU.

If you have access to a frequency meter, try to check the operating freq of the SMPS.

Also, could you mail me the service manual? Use my messenger address (it appears by clicking on the icon under "Location:" in my posts).
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.