Now what's my role here? Observer, I should answer to myself!
But I'm glad that no one got hurt in this marvellous discussion, I
learnt a little of differences between lateral and vertical mosfets, however what version should I build, the first one that everybody
trashed, with lateral mosfets, taking the risk of still blowing it up,
or the last one Mike says it has the usual mistake in the bias circuit
with the IRF?????????????????????????????????????????????????
It seems to me that at some point you guys get lost in the arguments
more or less "flamed" and forgot about this poor guy who doesn't
know didle about Mosfets and wants to rectify his opinion about the sonic qualities of amplifiers built with them....
Me too I thought the Slone amplifier built with those laterals was
the ultimatum, but hallas, NO, it sounds mediocre, with no life at all....
But I'm glad that no one got hurt in this marvellous discussion, I
learnt a little of differences between lateral and vertical mosfets, however what version should I build, the first one that everybody
trashed, with lateral mosfets, taking the risk of still blowing it up,
or the last one Mike says it has the usual mistake in the bias circuit
with the IRF?????????????????????????????????????????????????
It seems to me that at some point you guys get lost in the arguments
more or less "flamed" and forgot about this poor guy who doesn't
know didle about Mosfets and wants to rectify his opinion about the sonic qualities of amplifiers built with them....
Me too I thought the Slone amplifier built with those laterals was
the ultimatum, but hallas, NO, it sounds mediocre, with no life at all....
jmateus:
a few quick points:
1. one can make a great sounding amplifier with either vertical/hexfet mosfets or lateral mosfets
2. one can also make a lousy sounding amplifier with either vertical/hexfet mosfets or lateral mosfets
One of the secrets to have results closer to 1. and further from 2. is to APPLY THE DEVICES CORRECTLY. That's what a lot of this thread was discussing.
3. the choice of output devices is not the sole factor determining performance of the entire amplifier
4. the lateral devices have characteristics that enable their successful use in simpler circuitry ... but ...
5. the lateral devices are a lot more expensive - no competition really; few manufacturers produce them
although there are some who think otherwise, i do not think you can definatively say one device is "categorically" better than the other - each has strengths and liabilities, depening on the application, your requirements and design approach.
i have never completed a power amp that uses vertical/hexfets, although i did buy a SKA - i haven't finished it yet, so i cannot commment on their robustness.
however, i have been building lateral mosfet based amplifiers since getting my first lateral mosfet samples from hitachi in the late 1970s. i've been using stock that i purchased over 20 years ago. In all my crazy experiments, i have only destroyed a few devices by "accidentally" applying 65 volts of gate to source voltage. even then, it took a couple of hours for the devices to fail hard enough for me to notice.
just my thoughts, for what it's worth ...
mlloyd1
a few quick points:
1. one can make a great sounding amplifier with either vertical/hexfet mosfets or lateral mosfets
2. one can also make a lousy sounding amplifier with either vertical/hexfet mosfets or lateral mosfets
One of the secrets to have results closer to 1. and further from 2. is to APPLY THE DEVICES CORRECTLY. That's what a lot of this thread was discussing.
3. the choice of output devices is not the sole factor determining performance of the entire amplifier
4. the lateral devices have characteristics that enable their successful use in simpler circuitry ... but ...
5. the lateral devices are a lot more expensive - no competition really; few manufacturers produce them
although there are some who think otherwise, i do not think you can definatively say one device is "categorically" better than the other - each has strengths and liabilities, depening on the application, your requirements and design approach.
i have never completed a power amp that uses vertical/hexfets, although i did buy a SKA - i haven't finished it yet, so i cannot commment on their robustness.
however, i have been building lateral mosfet based amplifiers since getting my first lateral mosfet samples from hitachi in the late 1970s. i've been using stock that i purchased over 20 years ago. In all my crazy experiments, i have only destroyed a few devices by "accidentally" applying 65 volts of gate to source voltage. even then, it took a couple of hours for the devices to fail hard enough for me to notice.
just my thoughts, for what it's worth ...
mlloyd1
Hi Al,
Thank you. I was just about to request the exact same thing.
Anatoliy, no one recommended just dropping in an IR device for a BJT. I know from experience that the lateral Mosfet has a negative temperature co efficient that become dominant in the 100 mA range. I also know, that for all intents and purposes, an IR hexfet does not have this effect. That is until you have reached currents that have overheated your parts.
That aside, you still have to deal with gate charge issues. That is why I suppose you are using a source follower to drive them.
-Chris
Thank you. I was just about to request the exact same thing.
Anatoliy, no one recommended just dropping in an IR device for a BJT. I know from experience that the lateral Mosfet has a negative temperature co efficient that become dominant in the 100 mA range. I also know, that for all intents and purposes, an IR hexfet does not have this effect. That is until you have reached currents that have overheated your parts.
That aside, you still have to deal with gate charge issues. That is why I suppose you are using a source follower to drive them.
-Chris
Chris, all FETs have such effects, but on different currents. For IR devices I want to use (cheap as a dirt) it is about 7A of current, too high for such genuine simple solution.
No problem to thermocompensate them on less currents with some effort, but I had a different idea of FET+BJT pairs, and asked for an advise on the forum hoping somebody have some tasty solutions, but nobody helped; I am too lazy to experiment and short in time, so I gave up the idea for now and use BJT complementary followers driven by class A FET driver in the current project.
Speaking of schematics, I am not ready yet, but I shared already a lot of ideas. It is safe to share ideas with people who can and want to understand, but I don't want to end anymore with my solutions patented by people who are dumb to invent so I can't use them anymore for good.
No problem to thermocompensate them on less currents with some effort, but I had a different idea of FET+BJT pairs, and asked for an advise on the forum hoping somebody have some tasty solutions, but nobody helped; I am too lazy to experiment and short in time, so I gave up the idea for now and use BJT complementary followers driven by class A FET driver in the current project.
Speaking of schematics, I am not ready yet, but I shared already a lot of ideas. It is safe to share ideas with people who can and want to understand, but I don't want to end anymore with my solutions patented by people who are dumb to invent so I can't use them anymore for good.
anatech said:
"People consider her whimsical only because do not know her character." (C)
Aikido hint (not Aikido amp that has nothing to do with the real Aikido): don't fight against gate charge, use it for good.
Chris, you did not want to show me Mike's schematics (I suppose, patent safe), why do you expect me to share mine?
Anatoliy,
There is a common phrase in scientific and engineering circles, "Extraordinary claims require extraordinary evidence". You admit yourself that your claims go against perceived wisdom, yet you offer no evidence to back them up. Who would you believe if you were in our shoes? Several well respected members who have helped many people and published many good projects here, or a relatively new member who shows no evidence?
There is a common phrase in scientific and engineering circles, "Extraordinary claims require extraordinary evidence". You admit yourself that your claims go against perceived wisdom, yet you offer no evidence to back them up. Who would you believe if you were in our shoes? Several well respected members who have helped many people and published many good projects here, or a relatively new member who shows no evidence?
pinkmouse said:
No, my claims are not against perceived wisdom, they are against well established superstitious beliefs. Once I got a flaim about "partial feedback", later we discussed the topic and found that nothing is extraordibary in the claim that parallel feedback by current in the 2'nt cascade may cause increase of distortions;
the second time we discussed the topic of CCS and found that there is nothing extraordinary in increasing differential resistance by removing 2 extra components;
now you claim that IR FETs can't be used without a huge idle current, but it is your turn now to proove your claim. I am out.
And, I never ask you to believe anybody. I always provoce to check, compare, and use own mind to conclude. As I always did being a teacher.
mlloyd1 said:you CANNOT (well, I suppose you could , but you SHOULD NOT, if you want to be successful
the biasing arrangements is one area where the lateral fets are simpler/easier to use. Set your circuit to provide Vgs bias on the output devices to set quiescent drain current to apprx 100mA per device (somewhere around 0.5V Vgs). since this is the zero tempco point, you can use a simple variable resistor to set the current. this is not perfect, but it works and is simple.
there is more but let's keep it simple ...
slone's amps use lateral mosfets ...
mlloyd1
The circuit of luixssj3 used 240/9240 devices but I think the problem was the vbe multiplier which in my posted schematic was already been corrected.
Here it is again: http://img55.imageshack.us/my.php?image=misurevx3.gif
any other points to consider? thanks for any comments.
Thanks,
JojoD
Wavebourn said:
Since it's written about the middle of page 3 of this thread, FOR THE RECORD this is what the question was:
And your reply:
I've taken the liberty of bolding out the relevant parts, just so that there is no misunderstanding. I don't see how it could have been misunderstood on my part, no matter how much you are now trying to say it was about high bias currents. Not that I am nitpicking, but it simply WAS NOT the question.
If you have misunderstood the question yourself and intended to say something about the bias current, you should have corrected yourself, and not suggested courses at college for me - thanks for the sentiments, but I already had them. I have also used both lateral and vertical MOSFETs by the hundreds in all sorts of topologies, and this is nothing compared to the experience other members that have contributed to the thread have in this matter.
BTW yes you can use low bias currents. You can also use smaller HEXFETs in a MOS-BJT follower arrangement - I've done it myself.
@JojoD818, IRFP240 and 9240 are not electrical complements, not in a sense usable to an amp designer. They are very often used as such though, due to reasons beyond my understanding, which probably have to do with knee-jerk reflexes.
IRFP240 and 9240 are, however, process geometry complements - in essence the same die has N and P regions reversed. But, because the carriers move at different speeds for N and P semiconductors, the P and N channel parts end up having different characteristics. In order to have similar characteristics of the N and P parts, you end up with quite different construction for each one of them, which in the HEXFET family, means you use different maximum Vds for the N and P part. A few suggestions: IRFP240 and 9140 for supply rails up to 50V, IRFP340 and 9240 for rails up to 100V. Note that all parts are without the 'N' at the end - the latter are quite different, with signifficant improvement to gate charge, but do not form very good pairs in either combination, unfortunately, because the improvement is much better in the N parts compared to the P parts.
JojoD818 said:
Hello JojoD818
( looks like your topic caused some HOT feelings, not only HOT MOSFET
)That schematic might work well,
but if you have checked out what Nelson Pass actually said and recommended
he suggested you used a TO-220 HEXFET for temperature control / sensing.
MJE340 or BD139 might work,
but to thermally compensate for HEXFET we should maybe use HEXFET
and not bipolar transistors.
At least this was what The One and Only, Mr Pass
suggested in my linked topic. it was the very point of what he said!
http://www.diyaudio.com/forums/showthread.php?postid=1025310#post1025310
**************************************************
As my reply, which tried to give some useful info, to you, JojoD818 and jmateus,
sorta got lost in the storm of all posts that followed
I will quote only a little bit here:
lineup said:
JojoD818 said:
Jojod;
a first I'dd add couple of Zeners with fast small Shottky diodes in series from output to bases of MJE350 and MJE 340 to limit output current. Second, if you power your driver from higher voltage (say, +10V per rail) it will be more effective, and needed for the following below additional suggestions.
Second, I'd split the resistor between emitters, and used couple of trimmers of much less value so idle current through your emitter followers would be about half an ampere, but FETs would see a voltage needed for small idle current about 10 MA each. Center point between trimmers I would connect to the output through a resistor about 47 Ohm. The driver would have own feedback loop for a gain about 10 times of the gain of the overall amp. I did such thing in 1977 calling it "A+B nonlinear approximation", but I did not know yet that this solution was patented by Acoustical Manufacturing Company calling it "Current Damping" that was a wierd name to hide obvious solution in the cloud, we are happy nobody patented "Hitemp Gluing", so the industry did not have to invent something to avoid paying huge money for soldering, and to explain the fact that "Soldering causes bad effect on a sound quality".
Later I tried this A+B approximation with FETs and got good results. I used op-amp with 2 transistor voltage amplifier (asymmetrically distorting one) loaded by a CCS and IR FETs, actually as the result I had an inverting amp with 3 different feedbacks to the same input where input current was applied to from one more inverting amp.
However, on higher end of the spectrum class A driver sees less load impedance because of input capacitances of FETs, but it is not a big deal comparing with a 47 Ohm resistor you would have already.
Probably, some additional RC networks may be needed for a stabilization.
Hi Wavebourn,
1. I signed a non-disclosure agreement (and I always keep my word).
2. I can not scan the information, plus the enormous amount of scanning that would be needed.
I am therefore bound by my word at the very least. To clarify to anyone else, I was asked if I could provide manuals or schematics for Michael Elliot's Counterpoint products. You also implied that I didn't wish (personally) to comply with your request.
I have no idea why you would bring this up since it had no bearing at all on what was being discussed. I do think that was a little underhanded. You failed to mention any of the reasons you were given, therefore you distorted the truth by omitting the facts known to you.
Al requested you post the diagrams and I agreed that they would help understand what you were getting at. So you weren't really denying a request from me. I will have to assume that you are blowing smoke, based on my own experience with these devices.
The other issue I have is that every other member who has commented has some experience. Others are true experts. I will consider djk and ilimzn as experts. You declare them wrong but refuse to offer up any proof. Not a wise move my friend.
-Chris
Well, I hate to have this dragged up, but I gave you two reasons.
1. I signed a non-disclosure agreement (and I always keep my word).
2. I can not scan the information, plus the enormous amount of scanning that would be needed.
I am therefore bound by my word at the very least. To clarify to anyone else, I was asked if I could provide manuals or schematics for Michael Elliot's Counterpoint products. You also implied that I didn't wish (personally) to comply with your request.
I have no idea why you would bring this up since it had no bearing at all on what was being discussed. I do think that was a little underhanded. You failed to mention any of the reasons you were given, therefore you distorted the truth by omitting the facts known to you.
Al requested you post the diagrams and I agreed that they would help understand what you were getting at. So you weren't really denying a request from me. I will have to assume that you are blowing smoke, based on my own experience with these devices.
The other issue I have is that every other member who has commented has some experience. Others are true experts. I will consider djk and ilimzn as experts. You declare them wrong but refuse to offer up any proof. Not a wise move my friend.
-Chris
anatech said:
Do you still need schematics, or this explanation is enough?
http://www.diyaudio.com/forums/showthread.php?postid=1025971#post1025971
Some more explanations.
1) Q. Why no run-away?
A. Because of very small iddle current.
2) Q. Why approximation?
A. Because transistors are nonlinear by definition. Even working in AB class they arfe nonlinear. In case of approximation we have a control over nonlinearities.
3) Q. Why input capacitances do not play a big role?
A. Because they are small against a feed-forward resistor from A-class driver.
4.) A. Why to limit gain of stages by multiple feedbacks?
Q. To get control over transitive processes when the next stage is overdriven.
More questions?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.