Juma,
better to build amps and be obnoxious, than not be obnoxious and build nothing.
It's complex, but I'll take your word for it - world needs obnoxious amps...

NP once wrote a paper about loudspeaker cables, outlining some of the things that could degrade signals going through them. I know this is not a cable thread and the issue is not directly that.
I read a paper yesterday from a Dutch radio amateur who stated that loudspeaker cables make up a good dipole antenna that could inject up to 2Vs of RF EMC backwards into the amp. It seems reasonable from what I know about the topic. How does a F5 cope with that - does it matter or does it not matter soundwise ?
Radio amateurs and Test & Mesurement people tend to look differently at things.
2Vs seems to be quite much compared to the output signal at low listening levels.
The remedy is supposed to be 2 different ferrit toroids placed close to the amp, where the loudspeaker cable has to be wound 5 - 7 times through both toroids on top of each other. That should kill the RF signals before they reach the amp. The toroids are made of different ferrit material, to attack different frequency bands.
Do we have a good idea here, or is it woodoo ?
I read a paper yesterday from a Dutch radio amateur who stated that loudspeaker cables make up a good dipole antenna that could inject up to 2Vs of RF EMC backwards into the amp. It seems reasonable from what I know about the topic. How does a F5 cope with that - does it matter or does it not matter soundwise ?
Radio amateurs and Test & Mesurement people tend to look differently at things.
2Vs seems to be quite much compared to the output signal at low listening levels.
The remedy is supposed to be 2 different ferrit toroids placed close to the amp, where the loudspeaker cable has to be wound 5 - 7 times through both toroids on top of each other. That should kill the RF signals before they reach the amp. The toroids are made of different ferrit material, to attack different frequency bands.
Do we have a good idea here, or is it woodoo ?
After the mistake, they settled down at 0.6V and were working fine.
During the mistake they were running at nearly 3V. (6A all be it briefly)
I'm guessing that even though the heatsink was still cold after such a short period, the junction temperature might have rocketed.
Everything is quite flat and cosy on the heatsink. I'll be giving everything a thorough check when the new parts arrive.
During the mistake they were running at nearly 3V. (6A all be it briefly)
I'm guessing that even though the heatsink was still cold after such a short period, the junction temperature might have rocketed.
Everything is quite flat and cosy on the heatsink. I'll be giving everything a thorough check when the new parts arrive.
Last edited:
"The remedy is supposed to be 2 different ferrit toroids placed close to the amp, where the loudspeaker cable has to be wound 5 - 7 times through both toroids on top of each other. That should kill the RF signals before they reach the amp. The toroids are made of different ferrit material, to attack different frequency bands."
Would that not put an inductive load in both directions ?? how would that alter band width of music going to speaker and change the reactive loading of the outputs???
Regards and good day to you all,
Elwood
Would that not put an inductive load in both directions ?? how would that alter band width of music going to speaker and change the reactive loading of the outputs???
Regards and good day to you all,
Elwood
Last edited:
I'm no engineer, but this doesn't sound right. Speaker cables might pick up 2v RF if you were next door to a radio broadcast antenna and the load on the wires at both ends was extremely high. Yes, they'd make a decent antenna if connected to a radio's RF input; I've often made crappy folded dipoles out of cheap twin-lead speaker wires.
But real speaker wires usually have a load in the general range of 8 ohms at one end and a decent damping factor makes the other ends at the amp act like an extremely low impedance. That would mean your environment has so much RF that you could make a completely passive 50 watt AM crystal detector radio as loud as most stereos. You could light your house by splitting your lamp's line cord instead of plugging it in. The damping factor at higher frequencies is high. Maybe a really wild guess could be a damping factor of 100 and that the load at the amp end acts like .08 ohms. 2 volts into that would be 25 amps and 50 watts. That would mean cheap amps with 24 gauge speaker wires would be at risk of having their speaker wires burst into flames without any audio signal.
But real speaker wires usually have a load in the general range of 8 ohms at one end and a decent damping factor makes the other ends at the amp act like an extremely low impedance. That would mean your environment has so much RF that you could make a completely passive 50 watt AM crystal detector radio as loud as most stereos. You could light your house by splitting your lamp's line cord instead of plugging it in. The damping factor at higher frequencies is high. Maybe a really wild guess could be a damping factor of 100 and that the load at the amp end acts like .08 ohms. 2 volts into that would be 25 amps and 50 watts. That would mean cheap amps with 24 gauge speaker wires would be at risk of having their speaker wires burst into flames without any audio signal.
Last edited:
Perhaps, Katie or Dad, you might consider adding a dc protection unit to your shopping list - the cost is less than 1/10 of your replacement drivers ...
very sad to read this
think it's the first time we hear about fried speakers
this amp was originally designed to work with the most fragile fullrange speakers
could the reason be fake transisters ?
think it's the first time we hear about fried speakers
this amp was originally designed to work with the most fragile fullrange speakers
could the reason be fake transisters ?
I hope they weren't fake, I bought them as a kit from H.
AFAICT the damage seems to be limited to Q3,Q4,R11,R12,R14 and R18.
AFAICT the damage seems to be limited to Q3,Q4,R11,R12,R14 and R18.
I can relate to one of the MOS-FETs failing and popping the other and the two Source Resistors.
I can almost see why R14 Gate Stopper might fail.
I can't see how R18 would have overheated - it hasn't failed it's just got too hot.
R11 also hasn't failed but it too looks as though its got very warm.
I can almost see why R14 Gate Stopper might fail.
I can't see how R18 would have overheated - it hasn't failed it's just got too hot.
R11 also hasn't failed but it too looks as though its got very warm.
Perhaps, Katie or Dad, you might consider adding a dc protection unit to your shopping list - the cost is less than 1/10 of your replacement drivers ...
Don't worry, I am going to do so.
It is probably not related, but was there a mains fuse and did it hold at the breakdown? I didn't see it in any of your comments.
My guess would be that R12 failed first. It was R11 and R12 that were cooked initially after seriously over biasing the amp ? It would have briefly been exposed to 15W, some 5 x its rating.
3V squared, divided by 0R47, makes a little over 19W.
6A times 21V (24-3) makes 126W dissipation per output device.
Even if the heatsink would be infinitely large, think what 126 J/s will do if Rjc is close to a C/W.
If you look at the SOA curve of e.g. an IRFP240, you can estimate from the lines for 10µs-100µs-1ms-10ms, where the line for 100ms will be.
19W times 0.1s is 1.9J.
The MOSFET is dead in 100ms, but the resistor still alive.
6A times 21V (24-3) makes 126W dissipation per output device.
Even if the heatsink would be infinitely large, think what 126 J/s will do if Rjc is close to a C/W.
If you look at the SOA curve of e.g. an IRFP240, you can estimate from the lines for 10µs-100µs-1ms-10ms, where the line for 100ms will be.
19W times 0.1s is 1.9J.
The MOSFET is dead in 100ms, but the resistor still alive.
All this surmising.
Just use a Mains Bulb Tester for EVERY power up of any modified project.
Just use a Mains Bulb Tester for EVERY power up of any modified project.
Wow; you are good with the above explanation.3V squared, divided by 0R47, makes a little over 19W.
6A times 21V (24-3) makes 126W dissipation per output device.
Even if the heatsink would be infinitely large, think what 126 J/s will do if Rjc is close to a C/W.
If you look at the SOA curve of e.g. an IRFP240, you can estimate from the lines for 10µs-100µs-1ms-10ms, where the line for 100ms will be.
19W times 0.1s is 1.9J.
The MOSFET is dead in 100ms, but the resistor still alive.
Well, I've repaired the components that I found goosed. I'm waiting for the epoxy to go off setting the thermistors against the MOS-FETs.
I'm always dubious about switching on a repaired amplifier, I might sell the magic smoke to Cinderella, she uses plenty of it at this time of year.
I'm always dubious about switching on a repaired amplifier, I might sell the magic smoke to Cinderella, she uses plenty of it at this time of year.
All this surmising.
Just use a Mains Bulb Tester for EVERY power up of any modified project.
Good Luck!!
- Home
- Amplifiers
- Pass Labs
- F5 power amplifier