...a very expensive way to replace a piece of wire?Did you know Caddock makes and Mouser sells a 0.ZERO33 ohm MP-930? I did not, but I know now.I don't know why anyone would need that small value,
Isn't that the res for R3 position, Rev A mod? I could be interested, I'd like to swap to that configuration soon or later, but just the Caddocks have way too high postage costs.Did you know Caddock makes and Mouser sells a 0.ZERO33 ohm MP-930? I did not, but I know now.
Peace,
Tom E
I get it now. You ordered 0.033 when you wanted 0.33 Ohms for the Evo mod A.
Now Unixman's comment makes total sense. An expensive wire indeed.
Jac
Dear Tom,
We are joking about your disadvanture.. but feel the pain and me personally feel my own part of it.. l mean if we don't insist you would not have jumped into this misfortune.. sorry.
Hope it will get straight and the result pay back .. hm with all this it rather should!
Ciao, George
We are joking about your disadvanture.. but feel the pain and me personally feel my own part of it.. l mean if we don't insist you would not have jumped into this misfortune.. sorry.
Hope it will get straight and the result pay back .. hm with all this it rather should!
Ciao, George
Well, yes, J K, it is all your fault. I was perfectly happy with my FE's until you started making improvements. You and those damn decimal points that start moving to the wrong place before I even notice.
To put this in perspective, I have probably spent considerably more than a couple thousand $ building many (six?) different versions of MyRef and FE over the past nine years.
I perhaps take second place only to Dario for number of different parts tried in various positions. But I've got two amazing MyRef's and two more amazing FE's, plus a friend with a pair, to show for it all. Learned a lot, too, especially from all you fellows on this forum.
Proper value resistors already on order. What's another $13 (including postage)?
Peace,
Tom E
To put this in perspective, I have probably spent considerably more than a couple thousand $ building many (six?) different versions of MyRef and FE over the past nine years.
I perhaps take second place only to Dario for number of different parts tried in various positions. But I've got two amazing MyRef's and two more amazing FE's, plus a friend with a pair, to show for it all. Learned a lot, too, especially from all you fellows on this forum.Proper value resistors already on order. What's another $13 (including postage)?
Peace,
Tom E
Can someone shed some light on my problem please. I just finished the build. One channel works perfectly fine but the second one worked for maybe 2-3 min and stopped. The diode simply turned off and the sound stopped. I checked the cables assuming the connection problem, powered off and on but the channel was still dead.
I noticed that about 10-15s after power off the diode flashes (for less then a second) and the relay clicks.
After one more power on I heard a pop and I saw a flash somewhere in the middle of the board. The fuse was blown. I found that 5 of 8 rectifier diodes are faulty (1 failed as closed, the rest failed open).
And now the main question, how to troubleshoot the issue? I assume resoldering new rectifier diodes whould be the starting point but how can I power on only the PSU on the PCB to test it (desoldering J1, IC101, R104, R204,T201, T101,R14??) ? Can someone advise on the minimum current rating of the diodes?
I previously used MUR820 which are 8A but is it realy needed? I have some spare 1-3A diodes that I could temporarily use.
I wonder what potential damages could have happened to the rest of the circuit. Maybe someone has simmilar experience.
Thanks
Greg
I noticed that about 10-15s after power off the diode flashes (for less then a second) and the relay clicks.
After one more power on I heard a pop and I saw a flash somewhere in the middle of the board. The fuse was blown. I found that 5 of 8 rectifier diodes are faulty (1 failed as closed, the rest failed open).
And now the main question, how to troubleshoot the issue? I assume resoldering new rectifier diodes whould be the starting point but how can I power on only the PSU on the PCB to test it (desoldering J1, IC101, R104, R204,T201, T101,R14??) ? Can someone advise on the minimum current rating of the diodes?
I previously used MUR820 which are 8A but is it realy needed? I have some spare 1-3A diodes that I could temporarily use.
I wonder what potential damages could have happened to the rest of the circuit. Maybe someone has simmilar experience.
Thanks
Greg
Check caps first
Check your main filter caps. Then the 3886. It takes something serious to blow the diodes. Maybe one of the diodes was oriented incorrectly.
I think the diodes can handle huge surges. The 8 amp rating is continuous.
Check your main filter caps. Then the 3886. It takes something serious to blow the diodes. Maybe one of the diodes was oriented incorrectly.
I think the diodes can handle huge surges. The 8 amp rating is continuous.
Desolder
I do not know your set up. You said one channel worked fine the other had issues and let out some smoke.
Compare the oriention of the main filter caps from channel to channel. Same with rectifier diodes. Verify the rail to rail caps are caps. Look for swollen filter caps. If you blew even one rectifier there is a dead short somewhere.
If the power transformer is connected properly, the filters caps look good, and the other items look okay remove the 3886. Replace the rectifiers and hook up with no power chip. If more smoke is most likely a shorted filter cap.
I always check the filter caps with an analog meter (Simpson 260) to see they charge up. If there is a shorted cap the resistance will not go very high. Usually after checking both polarities the meter change all up and they read high resistance.
I do not know your set up. You said one channel worked fine the other had issues and let out some smoke.
Compare the oriention of the main filter caps from channel to channel. Same with rectifier diodes. Verify the rail to rail caps are caps. Look for swollen filter caps. If you blew even one rectifier there is a dead short somewhere.
If the power transformer is connected properly, the filters caps look good, and the other items look okay remove the 3886. Replace the rectifiers and hook up with no power chip. If more smoke is most likely a shorted filter cap.
I always check the filter caps with an analog meter (Simpson 260) to see they charge up. If there is a shorted cap the resistance will not go very high. Usually after checking both polarities the meter change all up and they read high resistance.
all good suggestions.
I concur with George.
Don't forget the main, original cause is still hidden, so just replacing broken parts is not enough.
I'm a bit thinking that maybe maybe it could also be oscillation of the output stage with consequent heating up and brake.
It would mean all compensation elements should be checked first, that is, all caps and resistors.
But it is strange, have never seen such a violent form of oscillation in practice.
So parts of the power section are more suspect, agree with George
Heat transfer from chip to heatsink was all ok?
heatsink grounded at the HS point?
Stray contact of powersupply leads & tracks to case, heatsink excluded?
Ciao, George
I concur with George.
Don't forget the main, original cause is still hidden, so just replacing broken parts is not enough.
I'm a bit thinking that maybe maybe it could also be oscillation of the output stage with consequent heating up and brake.
It would mean all compensation elements should be checked first, that is, all caps and resistors.
But it is strange, have never seen such a violent form of oscillation in practice.
So parts of the power section are more suspect, agree with George
Heat transfer from chip to heatsink was all ok?
heatsink grounded at the HS point?
Stray contact of powersupply leads & tracks to case, heatsink excluded?
Ciao, George
Last edited:
The set up is configured as two mono block completely separated (separate transformers, fuses, etc) but the heat sink is temporarily shared.
Joseph, what do you mean HS point? My heat sink was not grounded.
I have used thermal paste so the heat transfer should be ok.
I checked visually all components. Everything seems to be fine (orientation, no swollen cap etc).
Should the 8A diodes be safe when 2.0A fuse was installed on the power line?
I must admit that as it was only the testing power on, I kept all PCB, transformers, heat sink outside of the case with no grounding.
Joseph, what do you mean HS point? My heat sink was not grounded.
I have used thermal paste so the heat transfer should be ok.
I checked visually all components. Everything seems to be fine (orientation, no swollen cap etc).
Should the 8A diodes be safe when 2.0A fuse was installed on the power line?
I must admit that as it was only the testing power on, I kept all PCB, transformers, heat sink outside of the case with no grounding.
Greg,
You don't have to 'admit' anything..��
It's all ok, we are not at the court.. small errors can happen that's all..��
But it's better to have a clear look. The ungrounded heatsink could pick up and transfer RF to the high impedance input circuit of the LM3886. So it is always good to think of a good, low RF impedance grounding path for it. The point with the sign 'HS' is a good one for these occasions.
So this could have contributed to some oscillation. But still this all seems to be just too much violent..
Control carefully if the compensation elements are all functional, not disconnected, not short circuited.
Also, at this point you would spare yourself a lot of extra pain, by doing a clean restart..
Change all active components, not only the diodes. An extra chip or two is not that much more cost.
The next trial do with a bulb tester a la' Andrewt..
Ciao, george
For a start, we are talking about the standard configuration, aren't we? No mods, I mean.
You don't have to 'admit' anything..��
It's all ok, we are not at the court.. small errors can happen that's all..��
But it's better to have a clear look. The ungrounded heatsink could pick up and transfer RF to the high impedance input circuit of the LM3886. So it is always good to think of a good, low RF impedance grounding path for it. The point with the sign 'HS' is a good one for these occasions.
So this could have contributed to some oscillation. But still this all seems to be just too much violent..
Control carefully if the compensation elements are all functional, not disconnected, not short circuited.
Also, at this point you would spare yourself a lot of extra pain, by doing a clean restart..
Change all active components, not only the diodes. An extra chip or two is not that much more cost.
The next trial do with a bulb tester a la' Andrewt..
Ciao, george
For a start, we are talking about the standard configuration, aren't we? No mods, I mean.
Another, very unlikely possibility. There are two kinds of LM3886, TA and TF, if memory serves. One of them, TF I think, has insulation over the tab and back of the chip. It uses heat sink grease to improve heat transfer but is electrically isolated from the heatsink. The TA has a metal tab and could short the chip to the heatsink. Is there any chance that the bad channel has a metal tab LM3886 and didn't have additional electrical isolation from the heatsink? Most of us use the TF, but I know I have to be careful because I have some of the TA's hanging around.
Regarding your 2A fuse, we tend to use a slo-blow fuse because of the rapid current flow to charge the filter caps on start up. It is possible to not blow your slo-blow fuse and have enough current to blow your 8 A diodes.
The fact that the diodes blew but the filter caps appear OK may suggest that the short is before the caps. That's a good reason to check that the filter caps are OK. I might be tempted to remove the filter caps so I could check them "off board" and it would allow me to check the diode circuit for shorts independent of the rest of the pcb.
Running without case ground shouldn't cause the problem. I've been running mine on a piece of plywood for 5 years with out problem.
One other suggestion, once you find the source of the problem. I recently blew up some components on another project because I didn't use a "light bulb tester" on power up. If you can, I recommend using a light bulb tester for your next power up. It can save components and give you a visual indication of OK or fault.
Edit: I see that George just beat me to the light bulb tester suggestion. Good job. But seriously, learn from my stupidity. I had a light bulb tester sitting 2 meters away and didn't use it.
Last edited:
Try this
If you have an analog VOM check the positive rail. Then the negative. If the meter is connected properly you will see the resistance go low and then go to a high value.
It can be done with a digital meter but you do not see the needle swing. You will need to either pull the bad rectifiers or replace.
This should tell you if the issue is power supply and which rail (+/-). Then report back.
If you have an analog VOM check the positive rail. Then the negative. If the meter is connected properly you will see the resistance go low and then go to a high value.
It can be done with a digital meter but you do not see the needle swing. You will need to either pull the bad rectifiers or replace.
This should tell you if the issue is power supply and which rail (+/-). Then report back.
I've never done this, so a couple of questions for clarity.
This test is unpowered by mains, correct? Or capacitor alone.
You are using the VOM's battery in the resistance measurement to put power into the rails.
By rails, can I assume that you are attaching the VOM after the diodes?
No analog meter here. I played around with a digital bench meter and a 25,000 uF cap. I saw resistance less than 100 Ohms when I first attached the leads, climbing to 2+ MegOhms. When I discharged and switched polarity, it rose quickly then flattened out about 10k, Is this roughly what to expect from this test?
Jac
Jac,
Yes, yes, and yes. The actual resistance will depend on the meter and circuit. One tip, for the positive rail connect the red lead to the + terminal and the black to ground. For the negative rail connect the red lead to ground and the black lead to -.
If the issue is a cap this should tell you. Also if a diode. If doubt if a bad 3886 will blow rectifiersbefore the chip explodes. The rectifiers are more rugged than the 3886.