I think there are two most likely possibilities:
1) The amp was oscillating in the system.
2) The amp was inserted in the system without proper
burn-in and adjustment, and when it drifted up to 150 mV
offset, this was corrected with an increase in bias of the
opposite side, and then it drifted even higher until the
amp broke.
1) The amp was oscillating in the system.
2) The amp was inserted in the system without proper
burn-in and adjustment, and when it drifted up to 150 mV
offset, this was corrected with an increase in bias of the
opposite side, and then it drifted even higher until the
amp broke.
yes. I am assuming you have these resistors mounted on the source leads of the mosfets or on a separate output pcb.
I think you could also probably use the one resistor on the PCB and run separate wires for each mosfet. Maybe Nelson can confirm the better approach.
I would also consider, not using the current limiting circuit as another option. So if you don't use it pull it all out - I have no idea how the current limiting circuit might react with the amp circuit if active components are left on the pcb that are not properly connected as intended (could cause oscillation etc).
I think you could also probably use the one resistor on the PCB and run separate wires for each mosfet. Maybe Nelson can confirm the better approach.
I would also consider, not using the current limiting circuit as another option. So if you don't use it pull it all out - I have no idea how the current limiting circuit might react with the amp circuit if active components are left on the pcb that are not properly connected as intended (could cause oscillation etc).
Last edited:
check your mosFETs and reduce back to the standard F5 to check your circuit is OK.
Run it for a while, repeatedly re-checking the output bias and the output offset.
When you recognise the pattern that it follows from cold through warm up to hot, then add the extra pair. But first you must select a matched pair for the Ns & the Ps
Match Vgs @ your intended Id (Ibias)
Run it for a while, repeatedly re-checking the output bias and the output offset.
When you recognise the pattern that it follows from cold through warm up to hot, then add the extra pair. But first you must select a matched pair for the Ns & the Ps
Match Vgs @ your intended Id (Ibias)
Yes, one heatsink. I was running 1.7A (.800V)per device before - rather hot.
When I dialed it down to .525V it gave 1.12A per device which is a little hotter in total. When it was running the sinks were certainly hot but I could keep my hand on them for a few seconds.
It is placed in a well ventilated spot.
When I dialed it down to .525V it gave 1.12A per device which is a little hotter in total. When it was running the sinks were certainly hot but I could keep my hand on them for a few seconds.
It is placed in a well ventilated spot.
Well I added the 1K resistors to complete the protection circuit and it has been idling steady for 2 hours now. .525v and <4mV DC at the outputs - nominal drift.
I will let it run for a couple days and check periodically. I thought when I lost the CHR70s I lost at least one 2sk170 but that is not the case...at least I'll have some spares now.
I would prefer to keep the protection circuit in place for now until I have run it with speakers for a while...
The sinks are hot but I can touch them for about 8 seconds before it starts hurting.
My rail voltage is a confirmed 22.2V so yes I am dissipating 99.1 watts per channel...I may very well dial the bias down to as close to 1W per fet...
I will let it run for a couple days and check periodically. I thought when I lost the CHR70s I lost at least one 2sk170 but that is not the case...at least I'll have some spares now.
I would prefer to keep the protection circuit in place for now until I have run it with speakers for a while...
The sinks are hot but I can touch them for about 8 seconds before it starts hurting.
My rail voltage is a confirmed 22.2V so yes I am dissipating 99.1 watts per channel...I may very well dial the bias down to as close to 1W per fet...
Provided you have the necessary heat sinking anything under 40W per fet should be fine. 20W to 30W per device will be more than safe.
You probably don't want to exceed 100W total dissipation per channel with those heatsinks.
If you have a thermocouple you could also measure the temperature of the devices compared to the heatsink temp.
You probably don't want to exceed 100W total dissipation per channel with those heatsinks.
If you have a thermocouple you could also measure the temperature of the devices compared to the heatsink temp.
there's the problem.
One or more of the mosFETs is hotter than the remainder. Overstress the FETs by fitting too small a heatsink and then running one of them at a higher bias current and your amplifier tells you it does not like what you ask it to do.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.