This response is from purely looking at spec sheets and amplifier topologies - I haven't built my hybrid amp yet - it would seem that higher bias (up to a point) tends to boost the effective small signal transconductance of each device, increasing (and reducing the effective dynamic nonlinearity) of the output stage's damping factor particularly in a push-pull topology.
Nelson says that it is the total bias and not the bias of each individual mosfet that matters. These may well be true for Vertical Mosfets running in Class A with a relatively high bias current!
One question in this regard is, what if one were using Lateral Mosfets of say, the Hitachi type; these devices require an optimum bias of 100mA. Hence, when parallelling these devices, would one multiply the number of output devices with 100mA or would TOTAL BIAS be set without taking into account the actual current flowing through each mosfet? In the latter case, would these devices still exhibit their negative TC characteristics?
Would Nelson and/or someone else like to throw some light on the above question?
One question in this regard is, what if one were using Lateral Mosfets of say, the Hitachi type; these devices require an optimum bias of 100mA. Hence, when parallelling these devices, would one multiply the number of output devices with 100mA or would TOTAL BIAS be set without taking into account the actual current flowing through each mosfet? In the latter case, would these devices still exhibit their negative TC characteristics?
Would Nelson and/or someone else like to throw some light on the above question?
To be clear, total bias is the dominant effect, but it is still not
exactly the same for different numbers of parallel devices.
No reason to think the lateral devices would be different,
in fact given their lower temp coefficient, I would expect them
to parallel more nicely, assuming matching, of course.
exactly the same for different numbers of parallel devices.
No reason to think the lateral devices would be different,
in fact given their lower temp coefficient, I would expect them
to parallel more nicely, assuming matching, of course.
Brian Donaldson said:All of the rail voltage is dropped across each fet. you use more fets on larger alephs to decrease current/fet so you don't exceed max wattage per fet. Aleph 2 is .5A at 45V per fet= 22.5Watt. Smaller Alephs have smaller voltage and maybe slightly more current over fewer fets while still not exceeding 25 watts/fet.
I had the bias cranked a little too much and had a fet and source resistor melt down.I think the faital blow was about 4 months ago when I was putting them through the ringers and enjoying too much bias. and they sounded great. I let the sink temp reach 175 deg before I knew it. It seemed to survive, but I bet that weakened the chips bond to the tab and let one of the current source (pos rail) fets short and the active set tried to fight it. I'm hell on fets. 80 volts at 25 amps and 160000 ufs can make a lot of magic smoke very quickly. The 8 source resistors looked like little halogen lights on a dimmer.
25 Amps?!
Tell us more
My tansformers are 1000va 50v x2 @10A. But they are totally inaudible up to that point. But they were humming like mad and I'm guessing that the rails (rail to rail) sagged from 110 v to around 80 and likely over 25A. I think one fet on the pos rail shorted source to drain and the active fets all fought like hell to hold them back. All 8 source resistors on the active side glowed dull orange and the one on the current source side was bright white with fire shooting out it's ****! There were little black filliments floating around the room like an acetylene torch makes with no oxygen. I think that was the magic smoke that made the amp sound so good.
I'm glad it wasn't hooked to my speakers at the time. Future Active will get a few hundred dollars from me soon. Anyone else need some matched irfp244s. you need a lot to match sets of 8 and I think I'll replace all of mine to save replacing speakers when the other one goes. Since it saw he same abuse, I think failure is inevitable.
I'm glad it wasn't hooked to my speakers at the time. Future Active will get a few hundred dollars from me soon. Anyone else need some matched irfp244s. you need a lot to match sets of 8 and I think I'll replace all of mine to save replacing speakers when the other one goes. Since it saw he same abuse, I think failure is inevitable.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.