IRFP240 IRFP9240 design

[Alan0354]

Hey Steve,

I notice you have quite a few double post. I notice there is a problem with this forum that they keep asking you whether you want to leave this page. If you say yes, then it will go to a more advanced posting mode with your post in it. But actually your post has been posted by that time. I found when I get to this point, instead of click posting again, I just go to the top of the page and click the "Theory and Design". You'll find your post is actually in already. I have to do that all the time to avoid having to go manually delete the second identical post.

[lowell]

JM,
Didn't get a response from you on this. What if my meter is measuing gate voltage to ground before the gate resistor and not directly on the gate lead? Is this not a reliable method?

[Alan0354]

JM,
Didn't get a response from you on this. What if my meter is measuing gate voltage to ground before the gate resistor and not directly on the gate lead? Is this not a reliable method?

It's ok for DC measurement. It is not like tubes that current conduction if you go beyond 0V. MOSFET is nothing like that, it is high impedance all the way, there is absolutely no voltage drop across the resistor......that is until you put so much voltage that pass the Vgs max and blow the gate.

[lowell]

I think that contradicts what JM is saying. So which is it yall?

[Alan0354]

Are we talking about the circuit shown in the first post of this thread? It's a sack of sh!t. Here's the original 1970s Hitachi circuit that it was based on:
A Paul Kemble web page - Hitachi Fet designs.

Note that: 1) There's a trimpot for adjusting the bias and 2) It specifies lateral MOSFETs for the reasons discussed in this thread.

For more information on audio use of VMOS devices read this: Using HEXFETs in High Fidelity Audio

See, your schematics have current limiting resistor for individual FET. You get back the low output impedance by negative feedback like an opamp circuit..........It is an opamp in every respect, just an opamp build with discrete components.

Just because a few guitar amp use the circuit don't mean it's a good design. A lot of people just blind copy because they don't really understand.

But the two schematics you posted still have DC offset problem and a Cap is needed to blocked DC from getting to the speaker. It might be only 100mV, but for 8ohm, you still have quite a few mA. The alternative is to use a dual transistor at the front end differential pair. The two transistor are fabricated on the same die and the offset voltage is within 3mV or so. Also, the temp will be closely track also.

I like the second circuit better, the one "Using HEXFET in HF audio". It use only one differential amp stage, then using BJT pre-driver to drive the MOSFET. One problem of the MOSFET is the input capacitance. They can easily goes beyond 1000pF for a reasonable size FET. Bigger ones goes beyond 3000pF. If it is not because of the input capacitance, MOSFET is a perfect device.

It is always the driving requirement that really limit the use of MOSFET. In fact, LDMOS is popular in RF power amp because people use the input capacitance as part of the matching network. Matching can be done by a pi type network with shunt cap first, then through an inductor and then a shunt cap at the end. The capacitance of the LDMOS serve as the end shunt capacitor.

[Alan0354]

I think that contradicts what JM is saying. So which is it yall?

I might be new to tubes, I worked with MOSFET for years. Don't trust me, read the data sheet, it's all there. They are not incomplete like the tube data sheets. Usually they give the limitation of +/-20V or so. Below that, the gate is open circuit at DC. Leakage current is in sub pA range. At high temp, leakage increase to say nA. But for what you working on, it's open circuit.

The structure of MOSFET is that the gate is a capacitor plate insulated from the channel by a thin layer of oxide. It is totally insulated. They are not like JFET that you can actually turn the gate junction on. It is too involve to explain the way it work, for enhance mode N-MOS the +ve voltage at the gate draw electrons to the surface of the channel below and cause an inversion layer beyond the threshold voltage, this make the channel conductive. It is truly like a capacitor, there is no conductive path from gate to source and drain until you punch through the gate oxide......then you burn the FET. This is part of the famous MOS Cap!!!!

[Steve Conner]

All of this is true provided the MOSFET and gate resistor aren't burnt out. If you see any voltage across the gate resistor, something is wrong. However, if the gate has got disconnected, your meter or scope could change the voltage on it.

I gave up on MOSFET output stages. This runs off BJTs. scopeblog » The “Selfless” Amplifier

[Alan0354]

I meant if everything is in working order. If you are trouble shooting and something burned, all bets are off. You go into trouble shoot mode and measure every point, not taking anything for granted. You can have open gate resistor also as current can be high driving MOSFET in AC because of the input capacitance.

One thing I even forgot about the original circuit. The input capacitance of the big MOSFET is in nF range, you drive with the current source ( the second stage diff amp with no buffer), you might run into slew rate problem. The bottom schematic you post use BJT common collector as pre drive to conquer the problem. Also, it has idle adjustment also.