I wonder in such diamond configuration all parts want to be somewhat matched, Q12 to Q13 and Q11 to Q8 so as the diodes D12 to D13 and D11 to D14. Have you been able to find reasonable matches ?
Also, I am not sure if 5% resistor tolerance is ok, I would rather see 1% used preferably with good temp. coefficient.
Can you set P2 such that Q9 conducts the most, e.g. minimum bias.
Can you trim the amp output to zero with P1 ? If you can, this should equalize the voltages across the circuit to symmetry, then with P2 you should be able to increase the voltage spread on basis Q12 to basis Q13, also even with symmetry.
What are the resistance values for R21, R30 ? Are these 1/4 W ?
With respect to ground, what voltage do you get on anode of D16 and cathode of D15 ?
Also, I am not sure if 5% resistor tolerance is ok, I would rather see 1% used preferably with good temp. coefficient.
Can you set P2 such that Q9 conducts the most, e.g. minimum bias.
Can you trim the amp output to zero with P1 ? If you can, this should equalize the voltages across the circuit to symmetry, then with P2 you should be able to increase the voltage spread on basis Q12 to basis Q13, also even with symmetry.
What are the resistance values for R21, R30 ? Are these 1/4 W ?
With respect to ground, what voltage do you get on anode of D16 and cathode of D15 ?
So right now I seem to have equal voltages feeding Q10/Q11/Q13 as based on comparing to the positive section. Test points D and C are equal (albeit +/-) meaning everything feeding into Q10/Q7 are the same. The change occurs at Q11 collector as it is around -0.75V and should be -1.2. The positive side is +1.14V. When I bypass the bias section I get around +0.4V meaning I have lost -0.4V somewhere. Since the change is at the collector of Q11 I have been testing and changing everything around this point.
D10/D14/D13 are all new. C14 is new and I confirmed proper orientation. Q13 is new and appears to be functioning as it is dropping voltage appropriately (it is just being fed too low a voltage from Q11). The only other components in connected to this point are C13/C17/C11/C19/C20. I have a few 0.1 polyester caps I can use but some of the smaller value ones I will have to dig around or order.
The only other thing I can think of is the output section which is just the three mosfets and three resistors. Is it possible I am back feeding something though Q13 to cause the voltage issues?
If that is the case. Can I remove Q12/Q13, eliminating the output section and safely test the circuit? Now that I think about it I could just remove the wire at J9 and see what that does.
D10/D14/D13 are all new. C14 is new and I confirmed proper orientation. Q13 is new and appears to be functioning as it is dropping voltage appropriately (it is just being fed too low a voltage from Q11). The only other components in connected to this point are C13/C17/C11/C19/C20. I have a few 0.1 polyester caps I can use but some of the smaller value ones I will have to dig around or order.
The only other thing I can think of is the output section which is just the three mosfets and three resistors. Is it possible I am back feeding something though Q13 to cause the voltage issues?
If that is the case. Can I remove Q12/Q13, eliminating the output section and safely test the circuit? Now that I think about it I could just remove the wire at J9 and see what that does.
I did not match Q8/Q11/Q12/Q13 as I was under the impression that the more upstream transistors were more critical of matching and given their price I wasn't about to buy a bunch to match.
Currently I have the bias section bypassed. Prior, when I adjusted the bias too high the relay would kick in even though the bias reading across the resistor was barely over 150mA. Also when I adjusted the bias and read the voltage off Q8 and Q11 I could see the voltage change over Q8 (positive side) but wouldn't change the voltage over Q11. This would just stay at the -0.7ish volts.
Offset can be adjusted to zero and is fairly steady around that value.
R21/R30 are both 82 ohm and this coincides with the schematic. They are 1/4 watt per the schematic.
For D16/D15 I can't remember what the exact value was but I am definitely sure they were equal. These were replaced as I had exact replacements available.
Yes, you could temporarily disconnect the OPS on point 4 and 9. You could then also temporarily install two equal dummy resistors connecting 4 to 6 and 9 to 6 so the driver transistors would mimic a weak OPS. If things then balance out as expected, the problem then likely is found in the lower half of your OPS. Perhaps on of the power Fets is dead or half dead.
You could do quick measure if the gates float or not.
You could do quick measure if the gates float or not.
I did pull the output gets and measure them with a DMM. I used this method:
https://m.youtube.com/watch?v=gloikp9t2dA
Is this what you mean by checking if the gates float? If so the all passed. I'll remove the output stage from the equation tomorrow to try and further narrown this down.
Is it possible one of those caps I mentioned is the culprit. I don't have a way of testing them but I don't want to change all these parts just because and most likely for no reason.
https://m.youtube.com/watch?v=gloikp9t2dA
Is this what you mean by checking if the gates float? If so the all passed. I'll remove the output stage from the equation tomorrow to try and further narrown this down.
Is it possible one of those caps I mentioned is the culprit. I don't have a way of testing them but I don't want to change all these parts just because and most likely for no reason.
Yes, simple test of the Fet suffices for now.
I am not sure if a cap could throw off the balance just slightly like that, usually caps fail badly shorted or wide open and remain so.
It is possible that Q8 vs. Q11 and/or Q12 vs. Q13 have to much difference in current gain as they are of opposite types. Generally these want to be reasonable well matched in current gain.
If you have extra spares, and if you have a DMM with hfE test, see what difference you get when testing the opposite types.
I am not sure if a cap could throw off the balance just slightly like that, usually caps fail badly shorted or wide open and remain so.
It is possible that Q8 vs. Q11 and/or Q12 vs. Q13 have to much difference in current gain as they are of opposite types. Generally these want to be reasonable well matched in current gain.
If you have extra spares, and if you have a DMM with hfE test, see what difference you get when testing the opposite types.
In the December 2006 issue of Audio Express there's an article about the DH-500 and it's failure mechanisms. The amp in the article also had the base leg of Q13 missing, D13 and D14 were shorted, and R9 was bad. Some of the output FETs had failed OPEN!!!!! Anyway see if you find the article if you can not maybe I'll try to scan it and send it to you. PM me with your email address.
Craig
Craig
Go figure. I'll see what I can dig up. Hopefully I'll get a chance this weekend to remove the output section and see what happens. I have also heard that some of the 47K resistors can be suspect even if they measure OK. I might try and replace those.
I do have a lead on some original outputs at decent prices so worst case scenario I won't have to spend an area and a leg to get working mosfets.
So I went ahead and desoldered the wires to pin 4 and 9. Brought the unit up on the variac with the dim bulb in place. Stayed a bit brighter than before and the relay never clicked on. Did some measurements and everything was all over the place. Pin 4 and 9 were both negative voltage but was downtrending quite fast. The Tops of Q11 and Q8 (suppose to be -/+ 1.2V, respectively) were also both negative and downtrending (~ -2V). I also measured pins 5, 6 , 7, and 8; and all were around -13V except for 7 which was still zero. I was going to let it settle down but I realized the Q8 was getting very hot so I stopped.
Not really sure if this is helping me narrow down the problem. Would installing the resistors across 4-6 and 6-9 help stabilize? Is there any specific resistors I should use? Should I have removed the rail voltage supply to the output mosfets also?
My next step at this point will be remove the rail voltage to the outputs completely and see what that does.
If any of the above makes sense I would appreciate some insight. Thanks
Not really sure if this is helping me narrow down the problem. Would installing the resistors across 4-6 and 6-9 help stabilize? Is there any specific resistors I should use? Should I have removed the rail voltage supply to the output mosfets also?
My next step at this point will be remove the rail voltage to the outputs completely and see what that does.
If any of the above makes sense I would appreciate some insight. Thanks
Yes, as without there is no feedback and the amp floats towards one rail side.
Just two equal resistors, lets say 470 Ohm to 1k should suffice.
Yes, you should also disconnect the Fets from the rails to be safe.
Now with the two resistors temporarily installed the amp should balance out to zero(ish) and you should measure pretty close the equalized voltages in this circuit. If this is the case investigate the Fets and the gate resistors.
If not there is still a problem in the circuit elsewhere.
Now we are getting somewhere!!
I removed voltage to output stage and placed 1K resistors across points 4-6 and 6-9. I now have -50mV at the shorted C6 cap. The relay kicks on and I could cautiously play some very low music through the one channel. DC offset is essentially zero. I don't have any voltage at points 4 or 9 but I am thinking that is to be expected given no outputs.
Looks like my problem is at the output stage. I checked the resistors and they are within spec. So this most likely points to a bad mosfet. I do not know of any way I can further test these outputs. I was thinking I could remove all of them and install one pair at a time to see at least which pair is the culprit. Could I narrow it down in regards to the positive or negative side given I had an overall positive voltage at C6 initially?
That will be the plan unless someone can give me a better way of testing.
Thanks again for the help. There is definitely light at the end of the tunnel!
I removed voltage to output stage and placed 1K resistors across points 4-6 and 6-9. I now have -50mV at the shorted C6 cap. The relay kicks on and I could cautiously play some very low music through the one channel. DC offset is essentially zero. I don't have any voltage at points 4 or 9 but I am thinking that is to be expected given no outputs.
Looks like my problem is at the output stage. I checked the resistors and they are within spec. So this most likely points to a bad mosfet. I do not know of any way I can further test these outputs. I was thinking I could remove all of them and install one pair at a time to see at least which pair is the culprit. Could I narrow it down in regards to the positive or negative side given I had an overall positive voltage at C6 initially?
That will be the plan unless someone can give me a better way of testing.
Thanks again for the help. There is definitely light at the end of the tunnel!
Last edited:
Driving hard into low impedance loads burns out the gates of the outputs in Hafler amplifiers.
At that point you need a complete rebuild, outputs, drivers, zeners, etc.
The outputs will test OK from Drain to Source, but have a low impedance on Gate to Source.
http://www.exicon.info/PDFs/ecf10n20.pdf
http://www.exicon.info/PDFs/ecf10p20.pdf
At that point you need a complete rebuild, outputs, drivers, zeners, etc.
The outputs will test OK from Drain to Source, but have a low impedance on Gate to Source.
http://www.exicon.info/PDFs/ecf10n20.pdf
http://www.exicon.info/PDFs/ecf10p20.pdf
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.