Hi,
Some basic questions:
- Resistors in the BOM are 1/2 watt unless otherwise specified?
- What are for audio purpose the most important specs from a power FET.
- Do you have to double the input capacitance when applying a second driver and compare this with applying a FEt with more power. So applying two IRFP240 or one IRFP1405?
- Where can I find this CLC-filter schematic.
Thanks,
Jan
Some basic questions:
- Resistors in the BOM are 1/2 watt unless otherwise specified?
- What are for audio purpose the most important specs from a power FET.
- Do you have to double the input capacitance when applying a second driver and compare this with applying a FEt with more power. So applying two IRFP240 or one IRFP1405?
- Where can I find this CLC-filter schematic.
Thanks,
Jan
Concerning Power FETs in general:
Every manufacturer gives us unbelieveable power ratings and current specs, especially for the modern low-voltage types (have a look at IRF2804 for example). This is alle very pretty, but absolutely useless in RL. Those specs are extrapolated values for chip temparatures of 25°C or so. You won't be able to keep the die on that level, no matter what you try.
Also, those current ratings of +150 amperes are BS, because the case pins just can't stand it. At least IR has a very small note in their datasheets covering this issue. As far as I remember, they allow max. 50 amperes on a TO-220 pin.
So, all those modern high power types don't really help us because we can't use them at their theoretical limits. When chosing a suitable FET, I would care for the following points:
- low capacitance figures. This usually applies to the higer voltage types (100-200V).
- a case that is capable of transferring some power to the heat sink. This is TO-247/TO-3P these days, as the classical TO-3 is more or less dead.
- As you won't use the FET at more then 50 Watt power loss and 10 Amperes of current, the related specs are not that important.
Of course there are some more points: linearity, distortion figures, and so on. I would stick to the types that seem to do a reliable job: IRFP240 (wich should do better then IRFP140), IRFP044 for example.
Every manufacturer gives us unbelieveable power ratings and current specs, especially for the modern low-voltage types (have a look at IRF2804 for example). This is alle very pretty, but absolutely useless in RL. Those specs are extrapolated values for chip temparatures of 25°C or so. You won't be able to keep the die on that level, no matter what you try.
Also, those current ratings of +150 amperes are BS, because the case pins just can't stand it. At least IR has a very small note in their datasheets covering this issue. As far as I remember, they allow max. 50 amperes on a TO-220 pin.
So, all those modern high power types don't really help us because we can't use them at their theoretical limits. When chosing a suitable FET, I would care for the following points:
- low capacitance figures. This usually applies to the higer voltage types (100-200V).
- a case that is capable of transferring some power to the heat sink. This is TO-247/TO-3P these days, as the classical TO-3 is more or less dead.
- As you won't use the FET at more then 50 Watt power loss and 10 Amperes of current, the related specs are not that important.
Of course there are some more points: linearity, distortion figures, and so on. I would stick to the types that seem to do a reliable job: IRFP240 (wich should do better then IRFP140), IRFP044 for example.
--Half-watt resistors will be just fine. You can even use 1/4W if you want. Signal levels are really quite small until you get to the output. Be sure to use larger resistors (I used 3W) for the output current sensing array.
--The important numbers for a MOSFET are the voltage, current, and watts of heat dissipation that it will take. It's not a bad idea to stay below half of the published specifications. If you want to take it a level deeper, you can start looking at the Gate capacitance and such. In general, less is better, as it takes more current to drive a higher capacitance. The gate capacitances will add together for devices in parallel.
--A CLC filter is so simple that you don't neeed a schematic: Cap from rail to ground. Inductor in series, heading for the next cap. Cap to ground. Then off to the circuit. Some books refer to it as a PI filter because when you look at it on paper it's shaped like the Greek letter.
--TO-220s work fine for outputs. I used IRF644s for my Aleph 2s. Yes, the TO-220 has less heat dissipation capability than some of the other packages, but as long as you as you watch your wattage, it'll do the job.
--The IRFP140 won't require any changes.
Grey
--The important numbers for a MOSFET are the voltage, current, and watts of heat dissipation that it will take. It's not a bad idea to stay below half of the published specifications. If you want to take it a level deeper, you can start looking at the Gate capacitance and such. In general, less is better, as it takes more current to drive a higher capacitance. The gate capacitances will add together for devices in parallel.
--A CLC filter is so simple that you don't neeed a schematic: Cap from rail to ground. Inductor in series, heading for the next cap. Cap to ground. Then off to the circuit. Some books refer to it as a PI filter because when you look at it on paper it's shaped like the Greek letter.
--TO-220s work fine for outputs. I used IRF644s for my Aleph 2s. Yes, the TO-220 has less heat dissipation capability than some of the other packages, but as long as you as you watch your wattage, it'll do the job.
--The IRFP140 won't require any changes.
Grey
Well, engineers also said that 4558 was a good audio opamp.
Actually i think a coupple of parameters worth checking before choosing fet: Drain-Source resistance during on-state, especially if You plan to use the amp with higher Q speakers.
Most important, amount of blue smoke contained in the device.
This is easily checked by, for instance, creating a short in a driver transistor and thereby letting the entire supply reserve rush through one of the output fets. The amount of blue smoke it then lets out is directly proportional to how good it could have sounded.
Seriously : The new TO 247 and other plastic siblings of TO3 are good but need to be bolted very tight to the heatsink.
Just one tiny m3 screw wont do, use washer or metalplate to distribute fastening torque evenly across the transistor body and always use heat transfer compound even if You use silicon isolators.
Actually i think a coupple of parameters worth checking before choosing fet: Drain-Source resistance during on-state, especially if You plan to use the amp with higher Q speakers.
Most important, amount of blue smoke contained in the device.
This is easily checked by, for instance, creating a short in a driver transistor and thereby letting the entire supply reserve rush through one of the output fets. The amount of blue smoke it then lets out is directly proportional to how good it could have sounded.
Seriously : The new TO 247 and other plastic siblings of TO3 are good but need to be bolted very tight to the heatsink.
Just one tiny m3 screw wont do, use washer or metalplate to distribute fastening torque evenly across the transistor body and always use heat transfer compound even if You use silicon isolators.
Actually i think music and audio etc is too important a matter too be dead serious about.HBarske said:I hope this belongs to the joke part of your post...
But seeing that IRF sais their hexfets are not suited to audio (actuyally there is an audio page on the IRF site but concerned with class D) meanwhile there are mosfets labeld as audio devices, for instance magnatec BUZ900 series and several Hitachi and so on, the BUZ900 devices have RDS on ratings in the region 1-1,5 ohms, where a comparable IRF device weighs in at around 20-100 milliohms. You cant stare at just one parameter and sacrifice, for instance, capacitance for RDS figures but the third generation Hexfet devices from IRF offers a way to find a decent
path in between. The 240 and 260 families are good examples while APT devices show capactince figures up in nF ranges if You want high current and low RDS, but. I´m the first to admit that i´m wrong most of the time.. so tell me. (Been married to the same wife for 20 years so somewhere i did something right)
(She is in Germany today, bringing home some of Your nice beer)
RDS is only an issue if you use a MosFet as a switch, means: saturated. Give it (at least) 10 Volts between gate an source and you will benefit from the on-resistance.
But we are building amplifiers, and we have to do anything but never saturate an output device, because we want it to use on the linear part of the curve.
Once again: I don't think you will ever benefit from the absolute current ratings of such a device, because it's a) pure theory and b) not needed in an ampifier. The output devices are as good as always limited by their power losses (and by voltage issues), but never by current.
And, you are absolutely right, those not-so-sophisticated devices like gen 3 IRF offer much lower capacitance then those modern current monsters. So go with it and you will be fine.
But we are building amplifiers, and we have to do anything but never saturate an output device, because we want it to use on the linear part of the curve.
Once again: I don't think you will ever benefit from the absolute current ratings of such a device, because it's a) pure theory and b) not needed in an ampifier. The output devices are as good as always limited by their power losses (and by voltage issues), but never by current.
And, you are absolutely right, those not-so-sophisticated devices like gen 3 IRF offer much lower capacitance then those modern current monsters. So go with it and you will be fine.
MOSFET equivalents
After reading the above posts, will the following be good substitutes for IRFP044 (in Aleph-X application):
Basic comparisan:
.................: Vdss Id Pmax Cin Cout
IRFP044......: 60 57 180 2500 1200pF
MTP52N06V.: 60 52 188 2660 810
NTP45N06L..: 60 45 125 1700 480
http://www.onsemi.com/site/products/summary/0,4450,MTP52N06V,00.html
http://www.onsemi.com/site/products/summary/0,4450,NTP45N06L,00.html
I noticed they are in TO-220 packages, but I always mount these with a flat "spreader bar" across all FET's. Get very good mounting pressure and thermal conduction this way.
After reading the above posts, will the following be good substitutes for IRFP044 (in Aleph-X application):
Basic comparisan:
.................: Vdss Id Pmax Cin Cout
IRFP044......: 60 57 180 2500 1200pF
MTP52N06V.: 60 52 188 2660 810
NTP45N06L..: 60 45 125 1700 480
http://www.onsemi.com/site/products/summary/0,4450,MTP52N06V,00.html
http://www.onsemi.com/site/products/summary/0,4450,NTP45N06L,00.html
I noticed they are in TO-220 packages, but I always mount these with a flat "spreader bar" across all FET's. Get very good mounting pressure and thermal conduction this way.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.