I can see many people using IRFP240 and other similar fets in simple amps. Reading Nelson Pass article "Zen Amplifier Revisited" i could see that IRFP040 gives half the distortion compared to IRFP240
This is from that article. http://www.passdiy.com/pdf/retofzen.pdf
I bought a couple IRF3706 that should give even lower distortion.
They can only take 20 volts so i will have to cascode them with a higher voltage type on top. That should lower the distortion even more?!
Why does anybody use IRFP240 in a circuit with only 20 volts across them?
This is from that article. http://www.passdiy.com/pdf/retofzen.pdf
I bought a couple IRF3706 that should give even lower distortion.
They can only take 20 volts so i will have to cascode them with a higher voltage type on top. That should lower the distortion even more?!
Why does anybody use IRFP240 in a circuit with only 20 volts across them?
Last edited:
Im sorry if my question is a bit unclear. I find myself having a hard time writing exactly what i want to in english. I dont find the right sentenses and words.
Anyway. Thanks for the answer. I will try to explain.
There is the IRFP3703. A 230 watt, 30 volt TO247 package - device, and ofcourse IRFP040, IRFP054 and some others. I never see people using them, even though the amp never uses more than 40 volts rails.
I hope a TO220 device can handle 2 amperes and 5 volts. That should be what it will see in a cascode, in something like the Zen v9 but with a low voltage Hexfet instead of a J-Fet. (and ofcourse some bias-circuitry).
Any one of Nelson Pass different Zen/Aleph/First-Watt-amps. This is in the Pass Labs forum....
The question is more a general one about the use of high voltage fets in low voltage applications, when Papa himself has written about lower distortion from low voltage devices. It seems this point is not considered ( or forgotten) when trying to squeeze maximum performance out of simple circuits. If Nelson Pass could reduce distortion to 1/2 just by using IRFP040 instead of IRFP240, everyone should want to use IRFP040?!?! But i never see them used by Pass or anyone else....????
is there some other reason that i dont know about? It doesnt make sense?!
Anyway. Thanks for the answer. I will try to explain.
There is the IRFP3703. A 230 watt, 30 volt TO247 package - device, and ofcourse IRFP040, IRFP054 and some others. I never see people using them, even though the amp never uses more than 40 volts rails.
I hope a TO220 device can handle 2 amperes and 5 volts. That should be what it will see in a cascode, in something like the Zen v9 but with a low voltage Hexfet instead of a J-Fet. (and ofcourse some bias-circuitry).
Any one of Nelson Pass different Zen/Aleph/First-Watt-amps. This is in the Pass Labs forum....
The question is more a general one about the use of high voltage fets in low voltage applications, when Papa himself has written about lower distortion from low voltage devices. It seems this point is not considered ( or forgotten) when trying to squeeze maximum performance out of simple circuits. If Nelson Pass could reduce distortion to 1/2 just by using IRFP040 instead of IRFP240, everyone should want to use IRFP040?!?! But i never see them used by Pass or anyone else....????
is there some other reason that i dont know about? It doesnt make sense?!
Why not use the IRFP140?
That does seem like a good question. The IRFP040 does not seem to be available anymore, although I see a IRFP044 available from Digi-Key. The IRFP044 is a 55 Volt device and its use would be marginal at best in an amplifier with +/-25 Volt rails. The IRFP140 is a 100 Volt device and does seem (to me, anyway) like it would be a good candidate for use in various amplifiers which use the IRFP240.
That does seem like a good question. The IRFP040 does not seem to be available anymore, although I see a IRFP044 available from Digi-Key. The IRFP044 is a 55 Volt device and its use would be marginal at best in an amplifier with +/-25 Volt rails. The IRFP140 is a 100 Volt device and does seem (to me, anyway) like it would be a good candidate for use in various amplifiers which use the IRFP240.
Yes ofcourse, but according to Pass own words you get a more linear respons from a lower voltage part, due to more conductans and more possible feedback, local or global. The class A is device-independent.
Yes, i have read something along these lines somewere....
That shouldnt be a problem since the steadystate voltage across the device will be 20 - 25 volts. A mosfet rated 55 volts will take 55 volts without a problem. Probably it could take 60 volts too, if not to hot or running to much current at the same time.
I will buy some IRFP054 to my next project.
Johannes.
It isn't the steady state voltage that is the problem, but the peak voltage. If you play at full volume, and the amp has a +/- 25V supply you will have at least 45V across the FETs and perhaps as much as 48V - let's call it 50 V-peak. If the line voltage is 10% high, then you could be at 55V-peak.
If you are comfortable with those safety margins, then everything is fine. It's DIY and mostly for ourselves, so skimping on the device safety margins can be part of the fun. After all, who hasn't considered building a SOZ amp with way undersized heatsinks just to see how it sounds, knowing it won't last very long.
Jeremy
Yes, but at full output Vds will swing from near 0 to near 50V. That does not provide much of a safety factor, especially when varying lines voltages are considered.
FWIW, the IRFP054 has considerably higher gate charge and input capacitance (2900 pF) compared with the IRFP044 (1500 pF) so I would expect about half the bandwith from an amp built with them. IRFP140 and IRFP240 have similar input capacitance to the 044 part, in the range of 1500 pF.
Cheers,
Thats true. The inputcapacitance is maybe a little high on those devices. The IRFP044 is not available from ELFA -- Elektronikdistributör i norra Europa in sweden were i buy my stuff. Thats why i i want to buy IRFP054 instead.
The high input capacitance shouldnt matter when the device is in a cascode???
These are original to the Zen v4 as published by NP...which I have on mine.
The Zen V4 as published has a single 42 Volt regulated rail. According to the original article:
An unregulated supply is a different beast, as the rail values are at the mercy of the power supply company. How much do the mains vary? I don't know, but if you believe it is +/- 10% then you can expect to be on the hairy edge at some point with a nominal +/- 25 Volt rail that is at +/-27.5 Volts.
Call me paranoid, but I prefer a higher safety margin than a couple of percent.
YMMV
It's not really the voltage rating that effects distortion. Instead it's the transconductance. Devices with higher transconductance require less change in Vgs to provide the required change in Id. In general, devices with higher current ratings have higher transconductance and lower voltage ratings.
It still matters, but not as much. But operating the device in a cascode is a different discussion, no? I've been operating under the assumption this was a general discussion about Zen/Aleph configurations.
Anyway, the IRFP044 or IRFP054 seem like a reasonable selections if you will be using them at the bottom of a cascode. I think I see where you are going with this now
.Pass uses cascode in Zen v8 and v9. Cascode can be applied to any version of the Zenamp.
I want lowest possible distortion from a single stage amp.
A IRFP3703 is a 30 volt 210 ampere device. Cascoded it should provide a lot of gain and there by a lot of feedback. I think a lot of feedback is good as long as it is around only one stage. Nelson Pass writes
this in the same article. I think that holds true in one stage amps.
Part of the key here is what impedance can your preamp drive
comfortably?
The larger parts can be biased up higher, they have higher
transconductance and so they can be made lower distortion,
but the capacitance of the part comes in an creates distortion
at the higher frequencies.
If your source impedance can be made low, you can really
improve the performance for a single stage amp.
comfortably?
The larger parts can be biased up higher, they have higher
transconductance and so they can be made lower distortion,
but the capacitance of the part comes in an creates distortion
at the higher frequencies.
If your source impedance can be made low, you can really
improve the performance for a single stage amp.
It has always puzzled me when we discuss FETs with high transconductance, like the latest power jfets etc. The Yfs figure in the dazta sheet has no real meaning unless they are specified at the same Id. I have yet to come across a FET which gives a transconductance > 4S at 1A bias, irrespective of size, power capability, capacitance, ...., with the Japanese UHC-MOSFETs being the only exception.
Perhaps Nelson can correct me with some actual measurements of Devices ??
Patrick
Perhaps Nelson can correct me with some actual measurements of Devices ??
Patrick
The Zen V4 as published has a single 42 Volt regulated rail. According to the original article:
An unregulated supply is a different beast, as the rail values are at the mercy of the power supply company. How much do the mains vary? I don't know, but if you believe it is +/- 10% then you can expect to be on the hairy edge at some point with a nominal +/- 25 Volt rail that is at +/-27.5 Volts.
Call me paranoid, but I prefer a higher safety margin than a couple of percent.
YMMV
Well there you go...usually NP's diy offerings the output transistors are biased at 24-25V. My regulator is tuned for 50V and with enough heatsinking, the 044 is quite happy with 24V and 2A which is what I have for 5.5 years now.
...
Call me paranoid, but I prefer a higher safety margin than a couple of percent.
YMMV
Btw, I have some left-over IRFP240s, 2 quads matched at 12V/1A. It's been a while but I'll retest them at 24V/1A then offer them for sale...if they pass.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.