sashijoseph
Member
ZD1's anode should be -ve and ZD2's cathode +ve,wrt gnd.
could you confirm the orientation of the zeners?
Group Buy: Mauro Penasa MyRef Rev C kit
- Thread starter linuxguru
- Start date
- Status
ZD1's anode should be -ve and ZD2's cathode +ve,wrt gnd.
could you confirm the orientation of the zeners?
Actually I got one channel up and running late last night.
The only thing changed was to replace the TL072 with the spare LM318 I bought. When I got sound for the first time, it was at least making sound on both channels, so I guess the fundamental construction was fine. it just seems to be a case of bad opamp.
Switched on, measured the numbers and everything looked perfect, including a DC offset of just under 3mA. So I fired it up on the speaker and there was absolutely no hum or funny noise. I'm playing only one channel and it's playing very sweet. And the 40 Watts is more than enough to turn up the volume to ear splitting levels.
Just for the record, the measurements after using the spare LM318 are:
1. ZD1 anode to ground (too tough to access the cathode terminal so could not measure it): -12.3V DC. So by interpolation the cathode to ground drop is +12.3 VDC.
2. ZD2 cathode to ground: +12.5 VDC
3. LM 318 V- input: -12.3V
4. LM 318 V+ input: +12.5V
5. LM 3388 V+ to ground: didn't note down exact number but it was 34.x V.
6. DC offset: +2.3 mV
On the other channel, I mainly suspect that I may have blown the rectifier diodes. Will desolder them and check one by one.
The only thing changed was to replace the TL072 with the spare LM318 I bought. When I got sound for the first time, it was at least making sound on both channels, so I guess the fundamental construction was fine. it just seems to be a case of bad opamp.
Switched on, measured the numbers and everything looked perfect, including a DC offset of just under 3mA. So I fired it up on the speaker and there was absolutely no hum or funny noise. I'm playing only one channel and it's playing very sweet. And the 40 Watts is more than enough to turn up the volume to ear splitting levels.
Just for the record, the measurements after using the spare LM318 are:
1. ZD1 anode to ground (too tough to access the cathode terminal so could not measure it): -12.3V DC. So by interpolation the cathode to ground drop is +12.3 VDC.
2. ZD2 cathode to ground: +12.5 VDC
3. LM 318 V- input: -12.3V
4. LM 318 V+ input: +12.5V
5. LM 3388 V+ to ground: didn't note down exact number but it was 34.x V.
6. DC offset: +2.3 mV
On the other channel, I mainly suspect that I may have blown the rectifier diodes. Will desolder them and check one by one.
Some progressive/retrogressive reports on the non-working channel:
Since I wasn't getting either +/-34 volts nor the +/-12 Volts after the rectifier stage, I concentrated on this area of the circuit.
But before desoldering anything, I reflowed all solder points but net result was DC supplies are still absent. I mainly suspected that I might have blown the rectifier diodes so I desoldered all four F08H60S fast Schottky diodes and tested all and found all are fine. Next up, I swapped out both 10000 uF caps with fresh pieces of 6800uF/50V since that's all I had on hand. Still the same results as before, but I could at least measure the voltages on the legs on the capacitors. I got +/-35.x volts, which was an encouraging sign as it meant that the rectifier bridge and smoothing caps are working. Next, I checked both 12V zener diodes by desoldering them and found them to be OK, but I still inserted fresh pieces. Still the same results - there's no +/-12 volt. Next I checked all four 1K resistors and found them to be OK. Next I checked both the 47 uF zener bypass caps by desoldering them and read 43.x and 43.y uF on the two. So bypass caps also fine. Next I removed the two 1N4004 diodes and tested them and found both intact.
So I've removed all power supply components and found all components are normal, but on putting them back into the board I still don't get +/-35V or +/-12V. I'm basically at the ends of my (already limited) wit. Any pointers and advice most welcome.
I bought spare zeners (ZD1 and 2) and 1N4007 (to replace the two 1N4004. Also got a couple of 10K uF caps. I can swap out all.
One weird observation: when I switch off the power, the working channel's relay clicks with a single "tack" sound. This is almost instantaneously after power switch is put off, whereas on the non working channel the relay clicks "tick tack" after a delay of about 3 to 4 seconds after the power switch is put off. The click is two distinct sounds unlike the single click of the working channel. Is it possible that some fault condition is latching the relay shut? Does the LM3886 have some sort of fault detection that forces a shutdown?
Since I wasn't getting either +/-34 volts nor the +/-12 Volts after the rectifier stage, I concentrated on this area of the circuit.
But before desoldering anything, I reflowed all solder points but net result was DC supplies are still absent. I mainly suspected that I might have blown the rectifier diodes so I desoldered all four F08H60S fast Schottky diodes and tested all and found all are fine. Next up, I swapped out both 10000 uF caps with fresh pieces of 6800uF/50V since that's all I had on hand. Still the same results as before, but I could at least measure the voltages on the legs on the capacitors. I got +/-35.x volts, which was an encouraging sign as it meant that the rectifier bridge and smoothing caps are working. Next, I checked both 12V zener diodes by desoldering them and found them to be OK, but I still inserted fresh pieces. Still the same results - there's no +/-12 volt. Next I checked all four 1K resistors and found them to be OK. Next I checked both the 47 uF zener bypass caps by desoldering them and read 43.x and 43.y uF on the two. So bypass caps also fine. Next I removed the two 1N4004 diodes and tested them and found both intact.
So I've removed all power supply components and found all components are normal, but on putting them back into the board I still don't get +/-35V or +/-12V. I'm basically at the ends of my (already limited) wit. Any pointers and advice most welcome.
I bought spare zeners (ZD1 and 2) and 1N4007 (to replace the two 1N4004. Also got a couple of 10K uF caps. I can swap out all.
One weird observation: when I switch off the power, the working channel's relay clicks with a single "tack" sound. This is almost instantaneously after power switch is put off, whereas on the non working channel the relay clicks "tick tack" after a delay of about 3 to 4 seconds after the power switch is put off. The click is two distinct sounds unlike the single click of the working channel. Is it possible that some fault condition is latching the relay shut? Does the LM3886 have some sort of fault detection that forces a shutdown?
sashijoseph
Member
Hope you have replaced the TL072 in the defunct channel with a LM318/TL071.
Try a continuity check with a multimeter to test if the pcb traces are ok.
like between ZD2's cathode and pin7 of the opamp and so on.
Are the rectifier diodes over-heating?
and the big caps have +\- 35v across them,so at which point does this voltage disappear?
Try a continuity check with a multimeter to test if the pcb traces are ok.
like between ZD2's cathode and pin7 of the opamp and so on.
Are the rectifier diodes over-heating?
and the big caps have +\- 35v across them,so at which point does this voltage disappear?
The wrong TL072 is replaced with an LM318 I bought locally. It's not a TI. I forgot the branding. Will look up and tell.
Good point about the rectifier diodes. Honestly I haven't checked. I had only checked how hot the LM3886 and the resistors got. Will check tomorrow.
I'm at a loss as to how the +/-35V appearing at the big caps don't show up across the resistors or the zeners. BTW, I have swapped out both zeners too and they are currently on the board.
Reading some older posts on this thread as well as at diyaudio, I think I'll recheck that I've inserted correct values of resistors and capacitors. Also to be checked if there is any unusual DC voltage across the speaker terminals.
Thanks for the pointers :thumbsup:
I like what I heard with the working channel so at least there's incentive to rectify this faulty channel
Audiodoc
Well-Known Member
I believe that you will have to desoldering the lm3886 as you have already done the trouble shooting of the power supply section and all components are fine.
The relay clicking that I used to get at shutdown is similar to your clicking which usually means the speaker protection circuit is fine. You have a problem at the chip level or compensation of lm318.
The relay clicking that I used to get at shutdown is similar to your clicking which usually means the speaker protection circuit is fine. You have a problem at the chip level or compensation of lm318.
Tomorrow I'm anyway headed to the part of town near the electronics market so I'll pick up one more LM3886, but the T version as it is substantially cheaper than the TF version that I bought. I'll also pick up a few extra LM318s and may be a few TL071s too. I also see a few BC5xx transistors. I think they're in the speaker protection circuitry? Any chance the relay itself would have conked off?
Audiodoc
Well-Known Member
The lm 318 are usually TI other nat semi.
I believe you should check for a compensation problem. Try to reflow the solder of ic base and check for continuity and shorts by placing mm probes on each pin on both sides. If that is fine check with a tlo71. If it works then it is a compensation problem otherwise probably a problem with the 3886.
I believe you should check for a compensation problem. Try to reflow the solder of ic base and check for continuity and shorts by placing mm probes on each pin on both sides. If that is fine check with a tlo71. If it works then it is a compensation problem otherwise probably a problem with the 3886.
Audiodoc
Well-Known Member
No I don't think the relay is the problem. I has similar problems initially and I had done nothing wrong. No assembly error or dry solder. The circuit is picky of the lm317 not so for the tl071 as a tester. But if you are not getting 24 very between pins 4 and 7 I suspect something is wrong with the soldering or there is a short circuit. Was busy finishing my audiosector parallel lm4780 amp for my 4 ohm speakers so late on the forum. In the final phases of assembly.
Audiodoc
Well-Known Member
One more thing that will tell if you a stability problem is to attach the red led on the non working channel. As soon as you switch of the power the relay makes 2 clicks and led blinks for a few seconds
Some updates:
I gutted out the entire power supply section of the non working board. I traced the PCB traces. I followed Russ White's schematic in post#3 here: My new monobloc layout for Mauro Penasa's MyRef Rev. C - diyAudio
I found out that the PS cap nomenclatures are reversed in Linuxguru's board (C3 and C8 in Russ' schematic are C8 and C3 in Linuxguru's board). That threw me off for quite some time as I wasn't getting continuity on the traces where I expected. I found that the traces are fine. Can someone share Linuxguru's schematic for the Rev C please?
Since I had bought spares, I used MUR860 in place of the existing Schottkys. I also reverted to the original zeners ZD1 and 2 as I didn't find anything wrong with them. Desoldered and tested both 1K/6W resistors and found them OK too. On assembling it back, I now get +/-35V as well as +/-12V on either ends of the 1K/6W resistors. I also confirmed that the opamp now has +/-12V supply at the appropriate pins. However, +/-35V is still not reaching the LM3886. Replaced this LM3886 with a fresh one, but there's no improvement in the status - +/-35V does not reach amp chip, there's no relay action when powered on, there are two clicks when switched off (with a delay of about 3-4 seconds), and offset is about -2.4 mV.
The only thing changed is the rectifier diodes, though none of the original four were faulty as per my testing. So it could also be bad solder on any of the diodes or other PS section components.
Should I concentrate troubleshooting efforts around the opamp TLO71 compensation network, or poke around the LM3886 chip? I'm not ready to desolder the LM3886 chip again as I'm certain the traces will not stand up to the punishment for a third time.
And, mysteriously, the working channel has also stopped working. No relay activity when switched on. Symptoms are same as those described above: opamp has DC supply but amp chip does not see DC supply. Offset measured on this channel is around 4.2 mV. Relay does not click at all when switched on or off. The only change I had done is to use a TLO71 in place of LM318 I had used earlier. Reverting to the working LM318 does not help.
I'm getting a bit tired of this.
I gutted out the entire power supply section of the non working board. I traced the PCB traces. I followed Russ White's schematic in post#3 here: My new monobloc layout for Mauro Penasa's MyRef Rev. C - diyAudio
I found out that the PS cap nomenclatures are reversed in Linuxguru's board (C3 and C8 in Russ' schematic are C8 and C3 in Linuxguru's board). That threw me off for quite some time as I wasn't getting continuity on the traces where I expected. I found that the traces are fine. Can someone share Linuxguru's schematic for the Rev C please?
Since I had bought spares, I used MUR860 in place of the existing Schottkys. I also reverted to the original zeners ZD1 and 2 as I didn't find anything wrong with them. Desoldered and tested both 1K/6W resistors and found them OK too. On assembling it back, I now get +/-35V as well as +/-12V on either ends of the 1K/6W resistors. I also confirmed that the opamp now has +/-12V supply at the appropriate pins. However, +/-35V is still not reaching the LM3886. Replaced this LM3886 with a fresh one, but there's no improvement in the status - +/-35V does not reach amp chip, there's no relay action when powered on, there are two clicks when switched off (with a delay of about 3-4 seconds), and offset is about -2.4 mV.
The only thing changed is the rectifier diodes, though none of the original four were faulty as per my testing. So it could also be bad solder on any of the diodes or other PS section components.
Should I concentrate troubleshooting efforts around the opamp TLO71 compensation network, or poke around the LM3886 chip? I'm not ready to desolder the LM3886 chip again as I'm certain the traces will not stand up to the punishment for a third time.
And, mysteriously, the working channel has also stopped working. No relay activity when switched on. Symptoms are same as those described above: opamp has DC supply but amp chip does not see DC supply. Offset measured on this channel is around 4.2 mV. Relay does not click at all when switched on or off. The only change I had done is to use a TLO71 in place of LM318 I had used earlier. Reverting to the working LM318 does not help.
I'm getting a bit tired of this.
He posted it here - (needs eagle to open)
The new "My Ref" Rev C thread - Page 105 - diyAudio
I am puzzled at all the issues.
My original Russ White PCBs based amp from the diyaudio group buy (don't
remember when - perhaps 2009) is totally trouble free.
Audiodoc
Well-Known Member
Try the tl071 instead of lm318
If it works you will have to change the compensation capacitor value to 100pf.
I believe both your boards are functional per se. It is a problem of the compensation leading to unstability
Can you post the values of compensation capacitors that you are using
If it works you will have to change the compensation capacitor value to 100pf.
I believe both your boards are functional per se. It is a problem of the compensation leading to unstability
Can you post the values of compensation capacitors that you are using
Last edited by a moderator:
Yes, I did.
There are three per board - two BC547B and one BC539. The only change observed after replacing all three with new, tested pieces is the offset increased from about -4.2 mV to about -22mV.
I didn't bother swapping out the other channel.
Thanks for the suggestion. Except for resistors and small value capacitors I've changed everything.
I swapped out to TLO71. Still no cigar.
I think the compensation circuit has two caps in parallel. Which one should I change? C34 is 10 pF and C20 is 22 pF. Both these values are way too small for my MM to measure. Mine goes down only to 20 nF, so I could not verify their values by measurement. But their colour and description matches the hand written notes of Linuxguru in the BOM.
Last edited by a moderator:
I have one suggestion. I just saw that PCB layout from above link. It has no problem if you don't install LM3886 chip and test power supply rails. Opamp is required due to zener regulated 12V rails. Then once you tested PS rails, you test if relay is working. From output it goes to small transistors for DC sensing. Relay is just cutting off speaker ground.
sashijoseph
Member
Please check for continuity between pin7 of 3886 and Gnd.
Just noticed R11(1 ohm) which connects GND to pin7.
If this resistor is open,it would explain the mysterious disappearance of +/- 35v on the 3886 pins.
Just noticed R11(1 ohm) which connects GND to pin7.
If this resistor is open,it would explain the mysterious disappearance of +/- 35v on the 3886 pins.
- Status
Purchase the Audiolab 6000A Integrated Amplifier at a special offer price.
