I don't know why all this design talk. Confuses the OP.
This is an existing design with a unit that had been working before.
Trying to help him w/b the thing to do.
the driver DC bias current min is approx 2*(Vbe + 0.02)/100R - I therm ~ 13mA
the max of course depends on Iload and output betas
This is an existing design with a unit that had been working before.
Trying to help him w/b the thing to do.
the driver DC bias current min is approx 2*(Vbe + 0.02)/100R - I therm ~ 13mA
the max of course depends on Iload and output betas
I just thought of the fact that the fuses are 6A, so it is not possible to do the manual test I suggested at full voltage with an 8 ohm load. You could do it at about 45V on the rails with 8 ohm, or the full value with a 16 ohm load.
Just wanted to warn you.
The other obvious concern is fake parts so be careful who you purchase the parts from and inspect them.
Just wanted to warn you.
The other obvious concern is fake parts so be careful who you purchase the parts from and inspect them.
I have often thought that drivers were the reason for output stages failing and I just wanted to mention a worst case calculation. The drivers in this design will be badly stressed when the output is shorted and the fuse is the only protection. The devices see a high Vce so they are low on the SOA curve and fuses take a relatively long time to blow in terms of the thermal time constants so we do not get much if any of the dynamic benefit on the SOA curve.
On the other hand I have probably blown fuses more than a few times with these amps and never a problem.
On the other hand I have probably blown fuses more than a few times with these amps and never a problem.
Fred, do you think you might have bumped the bridge switch when it blew? I think you said you were changing speaker wires or something like that.
C3 Voltage Rating
You asked about this I think. I have seen an amp that had, Q2 short B-C and put the full rail voltage on what would be C3. The cap was a low voltage part and exploded, foil and paper spilled out the end, lol! That is probably why they raised the voltage rating. Put C3 in with the negative to ground, even better use a non-polar type.
You asked about this I think. I have seen an amp that had, Q2 short B-C and put the full rail voltage on what would be C3. The cap was a low voltage part and exploded, foil and paper spilled out the end, lol! That is probably why they raised the voltage rating. Put C3 in with the negative to ground, even better use a non-polar type.
Hi Pete,
The original orientation for C3 is correct. Base current through the resistor to common will pull that point slightly negative W.R.T. common ground. No matter, since electrolytic capacitors should be able to withstand a couple hundred millivolts reverse bias for a long time. Just a fine point.
Those DC blocking capacitors may vent if the output is stuck high or low. The average voltage is no longer zero (or close) and will either reverse bias the part, or exceed it's breakdown rating. Why do they tend to use 6.3 VDC parts? I refuse to use any normal electrolytic rated less than 16 volts for reliability reasons. Not unless I am forced to use the original type due to size issues, or contractual issues (warranty work).
Unless the replacement capacitor is a foil type, bipolar caps are not as good as polarized capacitors. You're better off using polarized unless the application forces the use of bipolar types. The normal signal voltage from base to common will be very small, unless the amp has left the linear zone. Things can get ugly then. I think that Douglas Self suggested placing a silicon diode across the polarized capacitor, both to save it from terrible abuse, but possibly to clamp high amplitude peaks. A pair back to back would be used in that case. I'm working on memory from reading his book, so check me on this. BTW, "Self - Audio Power Amplifier Design Handbook" is an excellent text to read. It explains a fair amount, and with other books on the same subject, provides a strong basis for understanding how these things actually work. Highly recommended reading if anyone has an interest. Walt Jung, Bob Cordell and a few other people have a wealth of information on their web sites. Most allow free downloading of their own work. The mark of an expert in the field, and someone who is helping people coming up the ranks in learning.
Hi infinia,
Yes, you are right. Thank you for clarifying the failure modes.
Hi Fred,
To properly test those diff pair transistors, you should use the jig I sent you the information on. Another hint, look at the base voltages w.r.t. common. They should be within mV of each other. By reading the voltage to common, you can figure out how much current is flowing into the bases, knowing the tail current when in a balanced state you can then figure out how much beta the transistors have. These parts should be in thermal contact with each other. Heat shrink tubing, or silicon, will help keep them at a similar temperature and slow down changes due to heat sources in the area. That's one of those fine detail things I do. It improves the stability of the amplifier for DC offset drift and even how much THD is not canceled.
-Chris
The original orientation for C3 is correct. Base current through the resistor to common will pull that point slightly negative W.R.T. common ground. No matter, since electrolytic capacitors should be able to withstand a couple hundred millivolts reverse bias for a long time. Just a fine point.
Those DC blocking capacitors may vent if the output is stuck high or low. The average voltage is no longer zero (or close) and will either reverse bias the part, or exceed it's breakdown rating. Why do they tend to use 6.3 VDC parts? I refuse to use any normal electrolytic rated less than 16 volts for reliability reasons. Not unless I am forced to use the original type due to size issues, or contractual issues (warranty work).
Unless the replacement capacitor is a foil type, bipolar caps are not as good as polarized capacitors. You're better off using polarized unless the application forces the use of bipolar types. The normal signal voltage from base to common will be very small, unless the amp has left the linear zone. Things can get ugly then. I think that Douglas Self suggested placing a silicon diode across the polarized capacitor, both to save it from terrible abuse, but possibly to clamp high amplitude peaks. A pair back to back would be used in that case. I'm working on memory from reading his book, so check me on this. BTW, "Self - Audio Power Amplifier Design Handbook" is an excellent text to read. It explains a fair amount, and with other books on the same subject, provides a strong basis for understanding how these things actually work. Highly recommended reading if anyone has an interest. Walt Jung, Bob Cordell and a few other people have a wealth of information on their web sites. Most allow free downloading of their own work. The mark of an expert in the field, and someone who is helping people coming up the ranks in learning.
Hi infinia,
Yes, you are right. Thank you for clarifying the failure modes.
Hi Fred,
To properly test those diff pair transistors, you should use the jig I sent you the information on. Another hint, look at the base voltages w.r.t. common. They should be within mV of each other. By reading the voltage to common, you can figure out how much current is flowing into the bases, knowing the tail current when in a balanced state you can then figure out how much beta the transistors have. These parts should be in thermal contact with each other. Heat shrink tubing, or silicon, will help keep them at a similar temperature and slow down changes due to heat sources in the area. That's one of those fine detail things I do. It improves the stability of the amplifier for DC offset drift and even how much THD is not canceled.
-Chris
Didn't get the message in time about checking EBC voltages before pulling Q1 and Q2. Oh well. Here's what I found today:
R17, R18, R19 and R20 all check good, dead-on in fact.
Q1 no leakage, 85hfe
Q2 no leakage, 53hfe
C4: 10pf (replaced with new cap tested at 22pf)
C3: can't test (meter only goes to 20pf), but I replaced it with new 100v electrolytic on the grounds that it was a 25 year old electrolytic.)
With Q1 and Q2 out I was able to go to full power without incident. With no signal in, power full +/- 84.5vdc on the rails yielded -84.4vdc at the speaker outs.
At full power, no signal in, there was zero voltage across R24-R31. As in exactly zero, less than one millivolt. (And yes, my meter is working).
I had the scope hooked up to the output and it showed a flat, steady trace that dropped as the negative DC voltage increased.
I'm thinking I should replace Q1 and Q2 with new devices that have in spec gain and then measure EBC voltages. What do you guys think?
Your clipping is on the positive transition, so I think while you have it setup for this test I would do it with the resistor to send the output positive. Now, if it is the output stage we are expecting this to fail so take it slow and cautious. You've got ~84V on the rails? The service manual lists 81V there and I would not take it over 75, unless you want to, and have tested up to 75 with a lot of confidence.
Yes, I would replace Q1 and Q2 after you do the above test. While the IT-18 measures
no leakage, it is doing the test with a 1.5V battery supply so it is not at full voltage. They
could still be bad.
There seems to be no mention of the pot on the driver transistor in the service manual, anyone have the adjustment procedure?
C4 should be 39 pF, it wouldn't hurt to put another 22 in parallel if we believe that the 39 value is correct.
Hi Fred,
The entire point of a differential pair is that they are matched for beta and held to the same temperature. Note that most parameters of a transistor shift with temperature. That's why my jig uses a pair of transistors together with a current source, the temperature shift is compensated for if they are the same beta.
To begin, you should buy about 20 transistors at the same time. You should be able to get a couple tight matches out of that. If you can buy the parts "on tape", do it. They will be closer than in bulk with small amounts.
Hi Pete,
The bias adjustment procedure is arduous to say the least. This will take an hour or so. I have detailed the procedure here in another thread, and I really don't want to type it out again. Can you do a search? The overall method is the same with all these amps. The bias current may be a little different between units.
-Chris
The entire point of a differential pair is that they are matched for beta and held to the same temperature. Note that most parameters of a transistor shift with temperature. That's why my jig uses a pair of transistors together with a current source, the temperature shift is compensated for if they are the same beta.
To begin, you should buy about 20 transistors at the same time. You should be able to get a couple tight matches out of that. If you can buy the parts "on tape", do it. They will be closer than in bulk with small amounts.
Hi Pete,
The bias adjustment procedure is arduous to say the least. This will take an hour or so. I have detailed the procedure here in another thread, and I really don't want to type it out again. Can you do a search? The overall method is the same with all these amps. The bias current may be a little different between units.
-Chris
I put a preset in the tail of the long tailed pair to trim the output dc offset.
On my amps it settles down pretty quickly and doesnt budge over temperature.
On my amps it settles down pretty quickly and doesnt budge over temperature.
This was all I could find, but it is in a MKII thread so I don't know if it is the same.
http://www.diyaudio.com/forums/soli...55-mk2-factory-output-trannys.html#post489691
I have the factory procedure right in front of me, but there are two pots, one for bias and there is no mention of how to set the other in the manual. I'd think that it is for minimum distortion, not sure.
http://www.diyaudio.com/forums/soli...55-mk2-factory-output-trannys.html#post489691
I have the factory procedure right in front of me, but there are two pots, one for bias and there is no mention of how to set the other in the manual. I'd think that it is for minimum distortion, not sure.
Fred,
The schematic that you have in this thread has R3 as 33.2K which doesn't make sense. My service manual has it as 33.2 ohms.
I would not worry too much about matching the diff pair right now, just get it working. If the output offset voltage is low then you are fine.
The schematic that you have in this thread has R3 as 33.2K which doesn't make sense. My service manual has it as 33.2 ohms.
I would not worry too much about matching the diff pair right now, just get it working. If the output offset voltage is low then you are fine.
Hi Pete,
Yes, the resistor is 33.2 ohms, has to be. I also know from experience that it is. I agree also that Fred should worry about getting it working before sweating the finer points.
I have to ask you what you are looking at. I'm not able to remember a pair of controls off the top of my head. Seeing it may jog my memory. In no way do I doubt you either, please don't get that message from my post.
-Chris 🙂
Yes, the resistor is 33.2 ohms, has to be. I also know from experience that it is. I agree also that Fred should worry about getting it working before sweating the finer points.
I have to ask you what you are looking at. I'm not able to remember a pair of controls off the top of my head. Seeing it may jog my memory. In no way do I doubt you either, please don't get that message from my post.
-Chris 🙂
No problem Chris, there is the expected bias pot. Then there is a 1K pot from EB on the upper driver transistor.