Wow, looks like I got really lucky here: don't ask me why I have 18R metal film in stock, but they were exactly what I needed 
! With a proper PSU hooked up (not the final one that will end up with this build though) and armed with a 100R potentiometer, I have determined 237.5R and 238.6R for the bias resistors. Went with 220+18 in series for each channel, which gives me 99mA for the Left channel and 93mA for the Right one. Close enough, it will stay like that!
Output offset is impeccable, too: +0.6mV Left and -0.6mV Right, according to my Fluke 8840A. Nice!
It won't be easy to run some power through the amp and hook up the computer for distortion measurements right now, so that will probably have to wait until everything is nicely fitted together inside the proper case. Should I become too impatient and do it earlier though, I'll post the results as soon as I have acquired them
! With a proper PSU hooked up (not the final one that will end up with this build though) and armed with a 100R potentiometer, I have determined 237.5R and 238.6R for the bias resistors. Went with 220+18 in series for each channel, which gives me 99mA for the Left channel and 93mA for the Right one. Close enough, it will stay like that!Output offset is impeccable, too: +0.6mV Left and -0.6mV Right, according to my Fluke 8840A. Nice!
It won't be easy to run some power through the amp and hook up the computer for distortion measurements right now, so that will probably have to wait until everything is nicely fitted together inside the proper case. Should I become too impatient and do it earlier though, I'll post the results as soon as I have acquired them
Looks like the Koa Speer BPR58 is the one to go for. Best alternative from ones linked to by Anand looks to be the Vishay/Draloric resistors Mouser p/n 594-AC050002207JAC00
Thanks guys
Geoff
Thanks guys
Geoff
My existing BJT power amp (Self's Load-Invariant) has associated speaker protection for power-on delay, power-down disconnect, overcurrent, dc offset and overtemperature and uses relays.
As I set out on my first outing with a MOSFET amp, and will also take the opportunity to go SSR for the speaker protection, what elements of protection do I need to include? Start-up, shutdown, overcurrent and dc offset are obvious ones. Something along the lines of Bonsai's speaker protection board seems a reasonable direction to go but do I still need over-temp protection?
As I set out on my first outing with a MOSFET amp, and will also take the opportunity to go SSR for the speaker protection, what elements of protection do I need to include? Start-up, shutdown, overcurrent and dc offset are obvious ones. Something along the lines of Bonsai's speaker protection board seems a reasonable direction to go but do I still need over-temp protection?
The switch on delay is the essential one and it needs to be reliable and foolproof under all conditions.
I used Doug Self's basic design from his Precision Preamp of 1996 as that gives a 99.9999% (it has never once been caught out) reliable delay and instant drop out on power off or even a 'long' mains glitch although we don't get those normally in the UK. Disconnect the AC for a couple of cycles and it resets and you get the full delay again. The time constants were altered to give around 6 seconds delay. You can not catch it out even with rapid mains on/off switching and generally totally abusing it.
DC offset protection feeds into the above on mine and again I used Doug's designs from (I think) the original Blameless Class B articles. Any standard offset sensing is much like another.
These solid state replacements I made were a direct physical replacement for the relays originally used:
https://www.diyaudio.com/community/threads/output-relays.191449/post-2659578
Bonsai's design should be fine I imagine.
I used Doug Self's basic design from his Precision Preamp of 1996 as that gives a 99.9999% (it has never once been caught out) reliable delay and instant drop out on power off or even a 'long' mains glitch although we don't get those normally in the UK. Disconnect the AC for a couple of cycles and it resets and you get the full delay again. The time constants were altered to give around 6 seconds delay. You can not catch it out even with rapid mains on/off switching and generally totally abusing it.
DC offset protection feeds into the above on mine and again I used Doug's designs from (I think) the original Blameless Class B articles. Any standard offset sensing is much like another.
These solid state replacements I made were a direct physical replacement for the relays originally used:
https://www.diyaudio.com/community/threads/output-relays.191449/post-2659578
Bonsai's design should be fine I imagine.
Hi Mooly. Thanks for the feedback.
My current amp /PSU / protection setup is all Doug Self's designs he sold via the Signal Transfer Company. I may well replace the existing relays on the protection board with solid state equivalent similar to those you did.
For the "Mooly" amp, I'll be doing the Bonsai circuit.
My current amp /PSU / protection setup is all Doug Self's designs he sold via the Signal Transfer Company. I may well replace the existing relays on the protection board with solid state equivalent similar to those you did.
For the "Mooly" amp, I'll be doing the Bonsai circuit.
How are people setting the bias on the prasi dual output board?
The instructions for biasing in post #121 were to measure 22mV across one of R23,R24, R43, R44.
On the prasi board there is no easy way to measure across one of these resistors.
Could I measure 44mV from T8 drain to T9 drain?
Also I have some pretty hefty heatsinks for this project. 1.6kg each.
Is there any reasons not to run a higher bias for a wider class A envelope?
Say 500mV?
The instructions for biasing in post #121 were to measure 22mV across one of R23,R24, R43, R44.
On the prasi board there is no easy way to measure across one of these resistors.
Could I measure 44mV from T8 drain to T9 drain?
Also I have some pretty hefty heatsinks for this project. 1.6kg each.
Is there any reasons not to run a higher bias for a wider class A envelope?
Say 500mV?
100ma per pair was the recognised sweet spot for the Hitachi/Renesas lateral FET's and is the point where the tempco was effectively zero. If you run two parallel pairs you need 100ma per pair. Measurement can be between the two drains of a pair and as there is 0.44 ohm total resistance doing that you would need 44 millivolts.
You can run whatever bias you want as long as the device is kept within its safe limits.
You can run whatever bias you want as long as the device is kept within its safe limits.
Thanks, I thought as much.
I have been trying to get my feeble brain around the parallel pair arrangement.
Am I correct that the if I measure 44mv through 1 pair, that the current through the other pair should be the same?
Thinking out loud. I could of course check the voltage across the second pair...
The 44mv is measured across the two series resistors which are effectively in series and that is the same as measuring from drain to drain on one pair.
The other pair should end up fairly close, how close depends on how well matched the FET's are. If they are from the same production run they should be close. I would aim to have the lowest reading pair correct and the other pair would then be automatically slightly higher.
The other pair should end up fairly close, how close depends on how well matched the FET's are. If they are from the same production run they should be close. I would aim to have the lowest reading pair correct and the other pair would then be automatically slightly higher.