Thank you Marcus for the link. Mark Johnson, in his own special way, hinted at this behavior.
I think the bulb tester is useful only to prevent fire/catastrophic failures IF there is a dead short... like voltage rails shorted to ground (if a MOSFET case is not insulated properly, i.e. sitting at the ground potential). Another scenario where the bulb tester would be useful is if there's a wiring mistake in the power supply section and the mains (IEC socket..?) fuse is not chosen correctly.
If there's no smoke AND the rails are showing equal voltage after a while (2-3min), ditch the bulb tester and connect the amp straight to the mains supply... to be able to correctly set the DC operation points, e.g. bias and offset. Do not attempt to set the DC operation points with the bulb tester in use.
It would be nice to see what happens when the bulb tester is in use with a perfectly good working amp, properly set... Maybe someone can shoot a short video..?
If there's no smoke AND the rails are showing equal voltage after a while (2-3min), ditch the bulb tester and connect the amp straight to the mains supply... to be able to correctly set the DC operation points, e.g. bias and offset. Do not attempt to set the DC operation points with the bulb tester in use.
It would be nice to see what happens when the bulb tester is in use with a perfectly good working amp, properly set... Maybe someone can shoot a short video..?
Last edited:
For others who have lab supplies with current limit, this is how I test.
Set the lab supplies to the correct voltage.
You can probably start with +/-20V without any difference.
Make sure power MOSFETs are properly mounted on adequate heatsink.
Set the current to 1.5x intended bias, in this case 1.8A.
At power up, the lab supplies will only hit current limit for a few seconds when the caps get charged up.
After that it should settle to the intended bias, i.e. ~1.2A.
If it is not the case within say 15 seconds, you should switch off and check.
The good thing about current limit is that when it is triggered, the voltage will drop to a very low level.
So sort of damage limitation.
But any active device that has seen such current limits, I normally do not retain for long-term use.
It may or may not have suffered permanent damage which might not be immediate obvious.
Which is also why you should first test build with cheap, unmatched IRFP MOSFETs and get it working first.
Patrick
Set the lab supplies to the correct voltage.
You can probably start with +/-20V without any difference.
Make sure power MOSFETs are properly mounted on adequate heatsink.
Set the current to 1.5x intended bias, in this case 1.8A.
At power up, the lab supplies will only hit current limit for a few seconds when the caps get charged up.
After that it should settle to the intended bias, i.e. ~1.2A.
If it is not the case within say 15 seconds, you should switch off and check.
The good thing about current limit is that when it is triggered, the voltage will drop to a very low level.
So sort of damage limitation.
But any active device that has seen such current limits, I normally do not retain for long-term use.
It may or may not have suffered permanent damage which might not be immediate obvious.
Which is also why you should first test build with cheap, unmatched IRFP MOSFETs and get it working first.
Patrick
I nominate member @Extreme_Boky to shoot a short video of a dim bulb tester connected to a perfectly working amp. Thread readers who are immensely inquisitive would have their feverish curiosity satisfied, while those who don't care at all can simply choose not to view. It's a win win scenario.
Just found out that Rochester no longer has any Harris IRFP9240's left.
Maybe now only ones left at Pass Labs ?
What a pity.
https://www.diyaudio.com/community/threads/burning-amp-ba-3b-balanced.201281/post-7044310
Patrick
Maybe now only ones left at Pass Labs ?
What a pity.
https://www.diyaudio.com/community/threads/burning-amp-ba-3b-balanced.201281/post-7044310
Patrick
No Harris IRFP9240s, but Harris IRFP9141s can be had. Not from Rochester, though.