Ecler PAM 1360E Fault - DC on speaker output

[paul_g_787]

Hi everyone. I have an old Ecler PAM 1360E power amplifier here which has both protect LEDs showing up.

I have found that there is approx +70VDC on both of the speaker outputs (pre-relay), which is what is causing the DC protect to kick in. However I am a bit stumped as to what might be the cause and I am hoping one of you might see what I am missing. I might just be missing something really silly.

I disconnected both power amplifier modules leaving just the main board which contains the power supply and pre-amp and the fault is still present with approx +70V on both L and R channel speaker.

See the main board schematic below. The connectors shown in red are the wires to the power amp modules which are unplugged.

My power rails are stable, the three Linear regulators are outputting a solid +18 +5 and -18V and the two high voltage rails are both sitting spot on at +90 and -90V too. (coloured Green).

When the unit is powered down and I have discharged all the power supply capacitors I get an infinite DC resistance across all of the capacitors from +90/-90 to GND suggesting there is not an immediate short on any of them.

I have noticed that if I power off the mains power and unplug from the AC outlet that the -90V rails stays charged overnight due to the capacitors being charged, however the +90V discharges itself in a few minutes so I think this is confirming there is a short from the +90V rail.

Note: All voltage measurements I mentioned above are in relation to GND.

[paul_g_787]

Also, I have just noticed that once the amp is disconnected from the AC mains power the Protect LEDs stay lit up  at a reduced brightness for quite some time, probably explaining the drain on the +90V capacitors??

[timenutgoblin]

Check the 10V rail. It is located directly under where you have typed +70V DC. Check at the anode of D110 7.5V zener diode. The cathode is connected to the +18V rail.

[paul_g_787]

Check the 10V rail. It is located directly under where you have typed +70V DC. Check at the anode of D110 7.5V zener diode. The cathode is connected to the +18V rail.

Ah good suggestion! Thanks. I hadn't measured this point. Must have overlooked it.
Just checked and it is a solid +10.02V  at the anode of D110, so looks ok here. A steady +18V at the Cathode of D110 also.

So that proves the 10V rail is good.

Location of service manual for those wondering:
https://elektrotanya.com/ecler_pam1360e_pam960e_pam560e_pam360e_amplifiers.pdf/download.html

[cozza]

Is there any component glue on the board which might have become conductive?

[paul_g_787]

Is there any component glue on the board which might have become conductive?

Nothing that obviously stands out in the open that I can see but it is a possibility for sure. The board is a bit dirty but nothing seems to come off other than dust.

[paul_g_787]

So I have been doing some probing and measured all marked test points shown in the schematic.

Please see updated schematic with my readings. (In slightly higher resolution this time).

Using the very informative service manual (linked previously), it appears to be that the DC protection circuit is working as intended, which is nice! 


Now I am looking suspiciously at the clip detect circuit. I am a little confused with the readings on the clip circuit.

I am measuring +90V at the cathodes of D117 and D115 (as I should be), however I am also measuring +90V at both of the anodes too!
And also at all three pins of Q103 and Q104! are at approx +90V.

However checking D117/D115 and Q104/Q103 I cannot see that they are shorted, the junctions all are reading approx 600mV on the diode test mode on my meter and all read infinite in reverse.

I think I am getting close. Any tips what to check next? 

[timenutgoblin]

Try removing - one at a time - zener diode D115 or D117 and check the corresponding protection LED. Removing the zener diode should stop the corresponding transistor Q103 or Q104 from switching on. The corresponding input voltage level to the NAND gate IC106 should drop. However, it looks like both halves of the circuit are causing the +70VDC on both lines, not just one line.

[dietert1]

Connect two small 5 to 10 ohms resistors that you have to the outputs. Then output voltage should be near zero. With the amplifiers out the clipping detectors inject enough current from +90 to confuse measurements. Yes, the 5.6 V zeners should have 5.6 V then and setup the clipping threshold at about 90-5.6-0.6 = 83.8 V nominal. To operate the clipping detectors you can use the low ohms resistors to drive the outputs to +90.

Regards, Dieter

[paul_g_787]

Try removing - one at a time - zener diode D115 or D117 and check the corresponding protection LED. Removing the zener diode should stop the corresponding transistor Q103 or Q104 from switching on. The corresponding input voltage level to the NAND gate IC106 should drop. However, it looks like both halves of the circuit are causing the +70VDC on both lines, not just one line.

OK so I tried your suggestion and removed D115. Now I am getting +7.9VDC at this channel output and 3.8V at the emitter of Q103.
I repeated this with the other channel by removing D117 and I am reading the same voltages on this channel.

Of course +7.9V is still enough to trigger the DC protect circuit and therefore both LEDs are still lighting up.

However D115 and D117 seem to test fine with my multimeter? How could they be passing such a high voltage when in circuit?  Are they faulty and I am just not testing them correctly?

So now the mystery; where is this remaining voltage coming from?

[timenutgoblin]

Can you check the voltage levels at the inputs (pins 1, 2, 5 and 6) of the NAND gate IC106? In your previous schematic, you measured 8.2V on pins 1 and 6. The NAND gates that drive the LEDs should produce a logic high if one of the inputs on each gate is logic low. If both inputs of one gate are logic high then the output should be logic low to switch off the LED.

Also, can you check the voltage levels at the inputs (pins 1, 6, 8 and 13) of the NOR gate IC107? In your previous schematic you measured 8.2V at the inputs (pins 1, 6, 8 and 13) of IC107. 8.2V is the clamping voltage of the zener diodes D109, D111, D112 and D113. I suspect that the fault could be IC107. Is it possible for you to remove it? If so, measure the voltages again.

After checking the voltages, reinstall the zener diodes D115 and D117. Make sure orientation is correct - cathode to +90V.

Regarding the voltage on the speaker output, I think that there will always be a voltage, but it depends whether a speaker or (resistive) load is connected. Without a speaker or (resistive) load connected, you will always measure some voltage. Thinking more about the +70V DC, I think that it is an arbitrary voltage that depends on the load connected to the speaker terminals. The higher the resistance of the load, the higher the voltage measurement. The lower the resistance of the load, the lower the voltage measurement.

[Audiorepair]

As I understand it, in this fault condition, the speakers are disconnected from the power amplifiers and fault sense circuit by the speaker relays, that are not engaged under fault conditions.

Any dummy load for testing puropses would have to be on the fault sense rails, not the speaker terminals.

Also, if the power amplifiers are not connected to this sense circuit, then part of the sense circuit is missing - the power amplifiers outputs.
Whether this makes any difference to the voltages measured perhaps should be tested.

Regarding the voltage on the speaker output, I think that there will always be a voltage, but it depends whether a speaker or (resistive) load is connected. Without a speaker or (resistive) load connected, you will always measure some voltage. Thinking more about the +70V DC, I think that it is an arbitrary voltage that depends on the load connected to the speaker terminals. The higher the resistance of the load, the higher the voltage measurement. The lower the resistance of the load, the lower the voltage measurement.

[timenutgoblin]

As I understand it, in this fault condition, the speakers are disconnected from the power amplifiers and fault sense circuit by the speaker relays, that are not engaged under fault conditions.

Yes, thanks for pointing that out. I was refering to the OP's own terminology regarding the speaker connector (prior to the relay).

I have found that there is approx +70VDC on both of the speaker outputs (pre-relay)

Note: All voltage measurements I mentioned above are in relation to GND.

Apologies for any confusion caused.

[dietert1]

That amp is a pretty serious model, i think with Lab Gruppen involved. There is a service manual at electrotanya including check procedures for locating errors. I think the OP should try to follow that procedure. He will read that each amplifier module has a bias current of 420 mA, which at +/- 90 V generates about 70 W of heat per channel. That's no fun. You can't run that amp without fan and probably not with the cover open - no hobby project. The recommended check procedures require a 1500 VA variac.

Regards, Dieter

[paul_g_787]

I have had another go today and I tested with both modules connected again but with the two Zener Diodes still removed.

The voltage of +7.9V has vanished with the modules connected! hoorah! However the protection is still switched on! Boo Hiss.

As you can see the 4001 (IC107) has switched back to it's normal state with D117 and D115 Removed and there is no longer a DC voltage on the speaker lines.
I have written the voltages in blue as I am not sure if these are correct, however I believe they are as pin11 is now measuring 10V as the schematic indicates.

So are the Zener diodes bad? They test fine on my meter and component tester so I am confused. I will re-install and check voltages again with modules connected.

The second problem still is that IC106 is still switched to the wrong state on pin 10 and 11. This I am confused by as both inputs (pin 8 and 9) are now 10V (high). The output of the NAND gate should be low 0V.


I also noticed that the CH2 heat-sink gets warm but the CH1 Module heatsink does not. Another fault maybe?

[paul_g_787]

I have soldered the two zener diodes back on D115 and D117 and the readings are the same as the last post.

So Next I need to work out the problem with the NAND Gate (pins 8,9,10 of 4011).

[timenutgoblin]

As you can see the 4001 (IC107) has switched back to it's normal state with D117 and D115 Removed and there is no longer a DC voltage on the speaker lines.
I have written the voltages in blue as I am not sure if these are correct, however I believe they are as pin11 is now measuring 10V as the schematic indicates.

Did you measure pin 3 of IC107? The schematic is not marked. Otherwise, it looks like IC107 is working.

So are the Zener diodes bad? They test fine on my meter and component tester so I am confused. I will re-install and check voltages again with modules connected.

The zener diodes can be tested in circuit later.

The second problem still is that IC106 is still switched to the wrong state on pin 10 and 11. This I am confused by as both inputs (pin 8 and 9) are now 10V (high). The output of the NAND gate should be low 0V.

In an earlier schematic, you measured 10V at the output (pin 10) of IC106, but the input was logic low measuring 0V. This time the output is the same, but the input is logic high measuring 10V. Strange! Perhaps it is IC106 that is faulty and not IC107.

I have soldered the two zener diodes back on D115 and D117 and the readings are the same as the last post.

So Next I need to work out the problem with the NAND Gate (pins 8,9,10 of 4011).

With the zener diodes soldered in as before, try desoldering the lead of resistor R145 which is connected to pin 10 of IC106 - do NOT desolder the other lead of the resistor! Connect the desoldered resistor lead to 0V (i.e., GROUND rail) and the relay K101 should click or turn on. This test should energize the relay coil and extinguish both LEDs D101 and D102. If this test works, resolder the resistor R145.


I am running out of ideas now, good luck with it!

[paul_g_787]

Did you measure pin 3 of IC107? The schematic is not marked. Otherwise, it looks like IC107 is working.

Yes I did and I forgot to mark it on the drawing. Pin 3 of IC107 is at 0V now. I have taken all the measurements again to ensure I have made no mistakes and attached an updated diagram.

In an earlier schematic, you measured 10V at the output (pin 10) of IC106, but the input was logic low measuring 0V. This time the output is the same, but the input is logic high measuring 10V. Strange! Perhaps it is IC106 that is faulty and not IC107.

Yes I am starting to suspect that the pin 8,9,10 NAND Gate of IC106 is malfunctioning. It is currently at 1, 1 and the output should be 0, but is instead 1. This in turn is causing the next NAND Gate (used as a NOT gate) to read incorrect at IC106 pin 11.

Looking at my updated diagram do you think this is faulty also from my readings?

With the zener diodes soldered in as before, try desoldering the lead of resistor R145 which is connected to pin 10 of IC106 - do NOT desolder the other lead of the resistor! Connect the desoldered resistor lead to 0V (i.e., GROUND rail) and the relay K101 should click or turn on. This test should energize the relay coil and extinguish both LEDs D101 and D102. If this test works, resolder the resistor R145.
I am running out of ideas now, good luck with it!

Will give this a go!

[paul_g_787]

With the zener diodes soldered in as before, try desoldering the lead of resistor R145 which is connected to pin 10 of IC106 - do NOT desolder the other lead of the resistor! Connect the desoldered resistor lead to 0V (i.e., GROUND rail) and the relay K101 should click or turn on. This test should energize the relay coil and extinguish both LEDs D101 and D102. If this test works, resolder the resistor R145.
I am running out of ideas now, good luck with it!

OK good suggestion! Thanks. This test worked as you theorised. (See attached diagram).
So I think it is probably safe to say I need to replace the 4011B IC. I will order one of these, only 28p!

Will post back when ordered and fitted.

Next, onto the power amp modules!

[timenutgoblin]

With the zener diodes soldered in as before, try desoldering the lead of resistor R145 which is connected to pin 10 of IC106 - do NOT desolder the other lead of the resistor! Connect the desoldered resistor lead to 0V (i.e., GROUND rail) and the relay K101 should click or turn on. This test should energize the relay coil and extinguish both LEDs D101 and D102. If this test works, resolder the resistor R145.
I am running out of ideas now, good luck with it!

OK good suggestion! Thanks. This test worked as you theorised. (See attached diagram).
So I think it is probably safe to say I need to replace the 4011B IC. I will order one of these, only 28p!

Will post back when ordered and fitted.

Next, onto the power amp modules!

Thanks for the update! Good that the test worked.

Make sure you order the correct part number as per the schematic. The part number is printed at the bottom as HEF4011B. This part number is rated up to 18V. There is also CD4011B rated up to 20V. It looks IC106 is a DIP package.

[paul_g_787]

Actually believe it or not, even though it is a 1999 amp it is a surface mount design and the 4011, 4001 etc are SOIC packages. See pictures.

I can get the 4011 from here:
https://cpc.farnell.com/on-semiconductor/mc14011bdr2g/logic-nand-quad-soic-14/dp/SC16329

[timenutgoblin]

Actually believe it or not, even though it is a 1999 amp it is a surface mount design and the 4011, 4001 etc are SOIC packages. See pictures.

I can get the 4011 from here:
https://cpc.farnell.com/on-semiconductor/mc14011bdr2g/logic-nand-quad-soic-14/dp/SC16329

Yes, of course. The symbol at the bottom of the schematic is exactly that - just a symbol for the IC, not its physical package. I did notice from the photo that it has a Philips logo. Philips is owned by NXP now (I think).

https://assets.nexperia.com/documents/data-sheet/HEF4011B.pdf

It's good that you have a photo of the PCB because now you have a reference for the correct orientation for the IC. It looks like there is red-coloured glue or cement on the PCB. I wonder if the IC is glued down? If necessary, extra heat should help with softening the glue for removing the IC. Try touching the tip of a soldering iron to the glue to see how it behaves - be careful of the fumes!.

[paul_g_787]

OK so my 4011 ICs arrived and I have fitted a replacement.

I have installed a replacement IC and now when I turn the amp on the main board is now functioning as normal! Upon switching on the power, both protect LEDs are on for a few seconds, then the relay clicks on and the LEDs go off. 

All the voltages now read as they do on the schematic! So this fault is fixed. Yay!

Thanks for the very helpful advice. I have not had to troubleshoot a logic gate circuit before so this is a good learning experience for me.

So now I need to check the power amp modules. I have observed that the CH2 module seems to get hot but the CH1 does not. Both channels were working prior to the now rectified fault. However it is entirely possible that I have two faults at the same time.

[paul_g_787]

Hi Everyone. Thanks so much helping me troubleshoot this issue with this PAM 1360E amplifier. 

Regarding the CH2 Module getting hot and CH1 Module not, I have potentially discovered something and I am going to make a new thread for this as it is a different issue.

Edit:
New thread is here https://www.eevblog.com/forum/repair/ecler-pam-1360e-module-current-fault/