F5 Turbo Builders Thread

Jump to Latest
#3,761
SRMcGee said:
Folks:

One of the F5T V2 amps I built for a friend has found its way back onto my workbench. He reported "nice flames and some smoke" when the amp failed last weekend. I opened it up and found (by visual examination only) that one of the MOSFETs on a P channel output board had failed. The dark marks on the photo are mostly smudges, not shadows.

When I delivered the amp to my friend, the offset on that channel was 1.0 mv after an hour's burn-in, the bias was set at .359 A (P side, .349 on the N side) and the rail voltage was 44.9 VDC. We used a 5U "Deluxe" chassis purchased from the Store. He had the amp sitting out in the open; there was no constricted air flow around it.
Click to expand...

Yikes! 😱

Ok now that got me scared, so now two people flamed their builds!

I'll have to reconsider my bias now.
My F5TV3 build is a 4 case one with 4x 10"x10" heatsinks per channel, and although biased at 0.39V the heatsinks are barely 20C above ambient, its seems a little close the the 0.4V and 25C above ambient FAB that NP talks about where thermal runaway may happen.....
 
#3,762
buzzforb said:
Thermal runway. MOSFET got hot and started drawing more current than the others and then it began a vicious cycle that lasted seconds and ended in flames. F5t is pushing 5u chassis, IMO.
Click to expand...

Buzz:

So I should be able to get away with simply replacing the two MOSFETs on that board and reducing the bias? If so, does aiming for the .3 amp range sound reasonable?

Regards,
Scott
 
#3,765
dazed2 said:
Yikes! 😱

Ok now that got me scared, so now two people flamed their builds!

I'll have to reconsider my bias now.
My F5TV3 build is a 4 case one with 4x 10"x10" heatsinks per channel, and although biased at 0.39V the heatsinks are barely 20C above ambient, its seems a little close the the 0.4V and 25C above ambient FAB that NP talks about where thermal runaway may happen.....
Click to expand...

I should have done what I was doing with the power off.... so no, not to worry about mine. Unless you plan to be an idiot like me.
 
Last edited:
#3,767
Thanks SRMcGee. This was probably the stupidest thing I have *ever* done. I mean why not turn it off first?

Regarding the existence of the diodes..... if you have, say, 200 watts RMS Class A into 8 ohms ( ie peaks to 400 ) then what's the point of using the diodes at all?
 
#3,768
The diodes act like smaller value source resistors as the current increases, (so the power isn't burnt up as heat in the resistors) letting you swing closer to the voltage rails and get more power to the load.

The diodes effectively let you have more power in a smaller amp.
 
#3,769
Advice for correct biasing for F5 v2 Turbo

Hi at all!

So..I have F5 v2 Turbo amp with jims audio pcb and 5U 400mm deep hi fi 2000 cabinet.


For two time with 0.6mV voltage accross Rs my left ch is dead.


I have r17-r24 with 0.91R so i have 0.455rs which is the best mV value for biasing in my case?

Thanks in advance for advices
 
#3,771
Tonimxp,

Assuming you mean .6V (not mV) across your output Source resistors, that's probably the reason you are blowing your outputs. Reread the article - you are a little too fearless and driving your diodes into conduction at idle, reducing the thermal stability. Limit your bias to .3V (or .4 if you are fearless). The source resistor value doesn't matter, the voltage limit is to prevent the diodes from conducting too early.

Note in the article NP says you can increase the net value of the Source resistors to 0R68 (2 1R36 in parallel) to increase bias stability, which sounds like your problem.

And yes, follow Jim's thermal guidelines even if you don't reach 0.4V across the Source resistors.

Good luck.
 
#3,773
6L6 said:
The diodes act like smaller value source resistors as the current increases, (so the power isn't burnt up as heat in the resistors) letting you swing closer to the voltage rails and get more power to the load.

The diodes effectively let you have more power in a smaller amp.
Click to expand...

Thanks 6L6.

Would you use them if you could bias to 200 watts RMS without them?
 
#3,776
SRMcGee said:
Folks:


When I delivered the amp to my friend, the offset on that channel was 1.0 mv after an hour's burn-in, the bias was set at .359 A (P side, .349 on the N side) and the rail voltage was 44.9 VDC. We used a 5U "Deluxe" chassis purchased from the Store. He had the amp sitting out in the open; there was no constricted air flow around it.


If I understand your numbers correctly you were running only around 15.75W(.35A x 45V) per device and yet the device failed?
Could it be that the Mosfet was improperly mounted?

Nash
Click to expand...
 
#3,777
nashbap said:
SRMcGee said:
Folks:


When I delivered the amp to my friend, the offset on that channel was 1.0 mv after an hour's burn-in, the bias was set at .359 A (P side, .349 on the N side) and the rail voltage was 44.9 VDC. We used a 5U "Deluxe" chassis purchased from the Store. He had the amp sitting out in the open; there was no constricted air flow around it.


If I understand your numbers correctly you were running only around 15.75W(.35A x 45V) per device and yet the device failed?
Could it be that the Mosfet was improperly mounted?

Nash
Click to expand...

Nash:

I'd be pretty surprised if that were the case. I'm using the Store boards and chassis, so the MOSFETs and diodes are evenly spread. The surfaces were all clean, I used Keratherm insulators and built spreader bars (one for the P side and another for the N side) to keep the MOSFETs tight against the heatsink.

Something else triggered this. My friend told me, in a rare embarrassing confession, that he and his wife had been dancing to show tunes when the amp blew. Not surprisingly, my theory is the amp immolated itself on purpose. But if you have any other theories, I'm certainly interested.

Regards,
Scott
Click to expand...
 
#3,780
SRMcGee said:
Folks:


When I delivered the amp to my friend, the offset on that channel was 1.0 mv after an hour's burn-in, the bias was set at .359 A (P side, .349 on the N side) and the rail voltage was 44.9 VDC. We used a 5U "Deluxe" chassis purchased from the Store. He had the amp sitting out in the open; there was no constricted air flow around it.

Scott
Click to expand...


Ok I re-read this again. I'm guessing its 0.359V bias not A?
A Stereo channel biased that high In a 5U chassis is going to be hot. I would hazard to guess the heatsinks are right on the edge of 25C above ambient?

If I read you correctly, you delivered the amp only after 1 hour of burn in?
In my experience bias will drift up after the first 24hrs of it being on, guaranteed. I have found that in fact it will drift up at least 0.03 - 0.04V within that time, and offset will sometimes drift up 10-20mV!
It needs to be rebiased within the first 24 hours.
Then after 1 week it will drift again another 0.01-0.02V.

So with you already biasing 0.36V after 1 hour of initial turn on. I would almost guarantee that after a week or two the amp may have drifted up to a dangerous 0.4-0.41V @ something above 25C above ambient or about +50C which as NP says, is close to the diodes conducting all the time, and causing thermal runaway.
 
Last edited:
You must log in or register to reply here.
Top Bottom