Removed the two offending standoffs, grounded the heat sink via a wire to the still dis-assembled chassis and it biases fine!
Am now re-assembling the erector set and will test again.
I will post some pictures of the offending standoffs when I disassemble the other board. The standoffs I am using come very close to the output trace. Clearly, too close. Don't have insulators, so I will assemble without the two standoffs.
Now trying to contain my excitement and not make another stoopid mistake!
Am now re-assembling the erector set and will test again.
I will post some pictures of the offending standoffs when I disassemble the other board. The standoffs I am using come very close to the output trace. Clearly, too close. Don't have insulators, so I will assemble without the two standoffs.
Now trying to contain my excitement and not make another stoopid mistake!
Thanks generg, but the 
, you can't expect much from an infrequent Pub dummy 
Thanks again, Itsmee, for picking that up. The board is in the chassis and, so far, working fine.
I do have another problem though with the second channel. When it was built it required a 6.8K R9 to get the bias over 100mV. I unmounted the board to swap the resistors and after reassembly, it promptly blew both fuses (2.5 A slo blows). No smoke. Unmounted the board and found a very small speck of metal stuck to the bottom of one of the n fets. Reassembled it and it biased fine (on the ungrounded heat sink) though never run for more than a minute or two. This was a few days ago.
Today, I removed the two standoffs and powered it up on the grounded but detached heat sink for a minute or so and it seemed fine at a 150 mV bias. I confidently attached it to the chassis, powered it up and no bias (measured about +/- 3v on the rails, P1 and P2 unresponsive).
After picking myself up off the floor, I removed the heat sink/board (I’m getting good at this
) from the chassis and, now, whether the sink is grounded or not, no bias.
OK, so something is wrong with the board. Physically, it looks fine. Is it most likely the MOSFETS, shunt regulator? Any way to tell short of removing them? Not sure how best to proceed.
I do have another problem though with the second channel. When it was built it required a 6.8K R9 to get the bias over 100mV. I unmounted the board to swap the resistors and after reassembly, it promptly blew both fuses (2.5 A slo blows). No smoke. Unmounted the board and found a very small speck of metal stuck to the bottom of one of the n fets. Reassembled it and it biased fine (on the ungrounded heat sink) though never run for more than a minute or two. This was a few days ago.
Today, I removed the two standoffs and powered it up on the grounded but detached heat sink for a minute or so and it seemed fine at a 150 mV bias. I confidently attached it to the chassis, powered it up and no bias (measured about +/- 3v on the rails, P1 and P2 unresponsive).
After picking myself up off the floor, I removed the heat sink/board (I’m getting good at this
) from the chassis and, now, whether the sink is grounded or not, no bias.OK, so something is wrong with the board. Physically, it looks fine. Is it most likely the MOSFETS, shunt regulator? Any way to tell short of removing them? Not sure how best to proceed.
Sounds like you may have punctured some of the FET insulators; disconnect the amp from the PSU and check for continuity between each of the supply rails and heatsink, also check rail to rail continuity; note, the anodization on the heatsinks is an insulator, you will need to pierce it with the tip of your probe.
If all of the above measure open circuit, check for continuity between the output and heatsink; if thats OK, check for shorted (dead) FETS - continuity between drain and source.
The above should find most faults, the TL431 is probably OK, don't expect any bias with +/-3V rails!
If all of the above measure open circuit, check for continuity between the output and heatsink; if thats OK, check for shorted (dead) FETS - continuity between drain and source.
The above should find most faults, the TL431 is probably OK, don't expect any bias with +/-3V rails!
No continuity present in any of the above.
Sorry for my previous post. It was written at the end of a long and frustrating day.
OK, time to connect it back up to the PSU, not forgetting the heatsink to chassis connection.
Temporary fit a shorting links across D1 & D2 and the signal input; also use a bulb tester in line with the supply to prevent any further damage, if the bulb stays on permanently at any stage, there is a problem.
Power up and measure the +/- supply voltages, are they OK? If they are, power down and remove the shorting links from D1 & D2, leave the input shorted.
Turn the bias down - P1 at minimum value, power up and measure the supply voltages again, are they OK? If they are, power down and remove the bulb tester.
Power up and monitor the supply voltages; then measure the voltage across the TL431, slowly increase the bias on P1 untill you read 5.5 to 6V, the FETs won't be (shouldn't be) conducting yet.
Move the probes to measure the bias across one of the source resistors (R19 to R21) it doesn't matter witch one; very slowly adjust P1 untill you read something on your meter (0.01V) then stop, measure the voltage across all the source resistors (R16 - R21), some FETs may be conducting more than others and some may not be conducting at all with such a low bias, what your looking for is a significantly high value (current hogging).
Measure the supply rails again, then make another small adjustment to P1, and remeasure all the source resistor voltages; wait 10 minutes then repeat the mantra until you have achieved 1A bias.
Once you are satisfied the amp is working correctly, fit it back into the chassis, then carry on with the bias set up as recommended with the lid on the chassis.
Temporary fit a shorting links across D1 & D2 and the signal input; also use a bulb tester in line with the supply to prevent any further damage, if the bulb stays on permanently at any stage, there is a problem.
Power up and measure the +/- supply voltages, are they OK? If they are, power down and remove the shorting links from D1 & D2, leave the input shorted.
Turn the bias down - P1 at minimum value, power up and measure the supply voltages again, are they OK? If they are, power down and remove the bulb tester.
Power up and monitor the supply voltages; then measure the voltage across the TL431, slowly increase the bias on P1 untill you read 5.5 to 6V, the FETs won't be (shouldn't be) conducting yet.
Move the probes to measure the bias across one of the source resistors (R19 to R21) it doesn't matter witch one; very slowly adjust P1 untill you read something on your meter (0.01V) then stop, measure the voltage across all the source resistors (R16 - R21), some FETs may be conducting more than others and some may not be conducting at all with such a low bias, what your looking for is a significantly high value (current hogging).
Measure the supply rails again, then make another small adjustment to P1, and remeasure all the source resistor voltages; wait 10 minutes then repeat the mantra until you have achieved 1A bias.
Once you are satisfied the amp is working correctly, fit it back into the chassis, then carry on with the bias set up as recommended with the lid on the chassis.
Hi RonD2,
I noticed a mistake in my earlier post
P1 should be at maximum value for minimum bias.
I noticed a mistake in my earlier post
P1 should be at maximum value for minimum bias.
Hi RonD2,
You will need to measure across P1, it looks like the easiest place would be the end of R7 furthest away from the FET; and one end of R9, it doesn't matter which end for the moment, set P1 for maximum resistance.
Using the same end of R7 and the end of R9 that measures ~9K75 rather than 5K, are the two points that you will need when measuring the voltage across the TL431.
You will need to measure across P1, it looks like the easiest place would be the end of R7 furthest away from the FET; and one end of R9, it doesn't matter which end for the moment, set P1 for maximum resistance.
Using the same end of R7 and the end of R9 that measures ~9K75 rather than 5K, are the two points that you will need when measuring the voltage across the TL431.
To turn the bias down, I disconnected the board from the PS (I don’t have bleeders). I maximized the resistance between R7 (far side from FETS) and R9 (near side to FETS and the higher of the two sides). It settled at 21.5K with P1 clicking and measured 14.5K to the other side of R9.
I reattached the PS and powered up while monitoring the voltage between R7 (far side) and R9 (near side). It was immediately over 20V and I quickly powered down.
Note that I am using a 6.8K resistor for R9 as that was required to get the bias over 100mV on initial testing after which it biased fine (on the ungrounded heatsink).
Does this hone in on the problem?
Hi RonD2
It looks like your earlier suspicion was correct, a duff TL431, with P1 set to maximum and a value of 6K8 for R9, the voltage should have been just over 7V; see my calculator.
For some reason I can't upload the excel spread sheet.
It looks like your earlier suspicion was correct, a duff TL431, with P1 set to maximum and a value of 6K8 for R9, the voltage should have been just over 7V; see my calculator.
For some reason I can't upload the excel spread sheet.
Attachments
Last edited:
I replaced the TL431 but before powering up thought I’d poke around a bit.
With the PS unconnected, I get the same resistances as I did prior to the TL431 swap, i.e. 21.5K between R7 (far side from FETS) and R9 (near side to FETS). If I turn P1 completely in the other direction, I get 16.7K between the same R7 and R9 measuring points.
My power supply has no bleeders and has been off/disconnected for almost 24 hours. It still measures +/- 5.5V (is this normal?).
I connect the board to the PS to see what happens with just the +/- 5.5 volts.
The voltage across R7 and R9 when P1 set at 16.7K resistance: 2.25V
The voltage across R7 and R9 when P1 set at 21.5K resistance: 2.52V
If the voltage difference is significant, it is the opposite of what I expected and suggests to me that I had P1 wide open when set at the higher resistance.
Does the 2+ volt TL431 reading with the +/- 5.5v rails seem too high?
I am reluctant to power this up (at the lower resistance P1 setting) as my instincts have been so consistently wrong.
Also, if this saga is not appropriate for this thread, please let me know!
With the PS unconnected, I get the same resistances as I did prior to the TL431 swap, i.e. 21.5K between R7 (far side from FETS) and R9 (near side to FETS). If I turn P1 completely in the other direction, I get 16.7K between the same R7 and R9 measuring points.
My power supply has no bleeders and has been off/disconnected for almost 24 hours. It still measures +/- 5.5V (is this normal?).
I connect the board to the PS to see what happens with just the +/- 5.5 volts.
The voltage across R7 and R9 when P1 set at 16.7K resistance: 2.25V
The voltage across R7 and R9 when P1 set at 21.5K resistance: 2.52V
If the voltage difference is significant, it is the opposite of what I expected and suggests to me that I had P1 wide open when set at the higher resistance.
Does the 2+ volt TL431 reading with the +/- 5.5v rails seem too high?
I am reluctant to power this up (at the lower resistance P1 setting) as my instincts have been so consistently wrong.
Also, if this saga is not appropriate for this thread, please let me know!
Ron,
Of course this is the appropriate thread!
I don't know if the tl431 is useful at 5V, it probably is, and ot seems to be trying to make the bias voltage as best it can, although not having enough total rail voltage to make the Vgs bias voltage it is coming up short.
Add some bleeders to your PSU. Any 3W resistors you have in your box, from about 2K-10K will be fine.
Of course this is the appropriate thread!
I don't know if the tl431 is useful at 5V, it probably is, and ot seems to be trying to make the bias voltage as best it can, although not having enough total rail voltage to make the Vgs bias voltage it is coming up short.
Add some bleeders to your PSU. Any 3W resistors you have in your box, from about 2K-10K will be fine.
Hi RonD2
With the given resistance values, you should be getting:-
voltage across the TL431 should be 5.06V when P1 set to maximum
voltage across the TL431 should be 5.8V when P1 set to minimum
Double check the value of R9, it doesn't look to be 6k8.
With the given resistance values, you should be getting:-
voltage across the TL431 should be 5.06V when P1 set to maximum
voltage across the TL431 should be 5.8V when P1 set to minimum
Double check the value of R9, it doesn't look to be 6k8.
A while ago I was struggling with fluctuating DC at the output on one of the channel. After a long hiatus, for some reason the problem solved itself, I have no idea why.
Rail voltage is ±23.6V. Bias is set to ~200mV across the source resistor. DC at output is ~0V. Looks perfect. But it sounded soft and distorted on both channels.
I have no idea what to do now. I'd highly appreciate if you guys can point me in the right direction.
Rail voltage is ±23.6V. Bias is set to ~200mV across the source resistor. DC at output is ~0V. Looks perfect. But it sounded soft and distorted on both channels.
I have no idea what to do now. I'd highly appreciate if you guys can point me in the right direction.