What's so great about MOSFETS over BJTs and why are they always so advertised? The layman consumer has been convinced that MOSFETS raise the bar to a whole new quality level.
I used to think MOSFETs were gods gift to amplifiers. Huge input impedance, no gate current -> easy to design. Then I found out about parasitic gate capacitance and random oscilations... And a Vt of 4 volts actually kind of sucks.
Also, the early voltage of a MOSFET is typically less than the VA of the BJT, so the BJTs are more constant in their F.Active region than the MOSFETS in saturation which leads to less distortion.
I've now been convinced that, at best, FETS are good for the input stage but BJTs are far better for VAS and especially the output stage. Lower distortion, and less feedback needed to linearize, so more gain.
--
danny
I used to think MOSFETs were gods gift to amplifiers. Huge input impedance, no gate current -> easy to design. Then I found out about parasitic gate capacitance and random oscilations... And a Vt of 4 volts actually kind of sucks.
Also, the early voltage of a MOSFET is typically less than the VA of the BJT, so the BJTs are more constant in their F.Active region than the MOSFETS in saturation which leads to less distortion.
I've now been convinced that, at best, FETS are good for the input stage but BJTs are far better for VAS and especially the output stage. Lower distortion, and less feedback needed to linearize, so more gain.
--
danny
It's just Ford vs. Chevy, Nissan vs. Toyota, Acura vs. Lexus, Mecedes vs. BMW. They are different but with either one you can get where you need to go.
The debate is just so entertaining, though.
The debate is just so entertaining, though.
Dave said:
isn't that called igbt?
🙂
I think it is pretty much moot to claim superiority based on the type of output devices. MOSFETs have their +s and -s, just as the BJTs.
I happen to like MOSFETs more because I have used them for a long time and have lots of them in my drawer. I would not based on purchases on MOSFETs or BJTs, because doing so makes zero sense.
It's a bit more complicated. At first, power mosfets looked like audio nirvana. They had high input impedance, a more linear G(m), and were free from the dreaded 2'nd breakdown that made transistors fail. When we actually tried them, however, we found big problems. First, the Hitachi 'lateral' power mosfets, while rugged and high voltage, had LOW G(m) and low peak output current.
The American 'vertical' fets had lots of G(m), fairly high voltage and current, BUT they were prone to breakdown. You can't use them at anywhere near their rated voltages, except as switches. And over the years, bipolar transistors got faster and more rugged.
In the last 5 years or so, power mosfets have actually gotten worse. Many parts that are really superior, have been discontinued. Still, I will use selected mosfets, before I use bipolars, where I can, because they tend to give lower amounts of higher order distortion in my designs.
The American 'vertical' fets had lots of G(m), fairly high voltage and current, BUT they were prone to breakdown. You can't use them at anywhere near their rated voltages, except as switches. And over the years, bipolar transistors got faster and more rugged.
In the last 5 years or so, power mosfets have actually gotten worse. Many parts that are really superior, have been discontinued. Still, I will use selected mosfets, before I use bipolars, where I can, because they tend to give lower amounts of higher order distortion in my designs.
One big reason that they are worse:
They don't have to be good to make an H-bridge. All N-types, full on, or full off......how good do they need to be?
As long as the gate capacitance and Rs is predictable, what else do they need?
Jocko
They don't have to be good to make an H-bridge. All N-types, full on, or full off......how good do they need to be?
As long as the gate capacitance and Rs is predictable, what else do they need?
Jocko
Sounds like these breakdowns may have occured when the MOSFETs were operated out of their SOAs. Wrt SOA range, the Toshiba 2SK1530 and 2SJ201 seem to have much more robust pulse SOAs in the 1-100msec range than, say the IRFP240 & IRFP9240 even if their absolute peak current ratings aren't quite as high. On top of that, their Vgs thresholds are significantly lower than average for devices of their type and their g(m)'s appear competitive.
Quote from a mail I received from Randy Slone related to bipolar vs mosfets (2SK1058/2SJ162); I had a 'peak current capability' question relative to his OPTI-MOS kit.
* On an "overall" basis, L-MOSFETs can provide much greater peak output currents than bipolar devices. You see, a typical bipolar power transistor that may be rated for an Ic (max) of 15-amps will probably have a secondary-breakdown point (depending on where the SOA falls for the operating conditions) of about 2-amps, or less.
Thus, the overload protection circuitry must restrict output currents very tightly within certain operational regions, and such restrictions can result in distortion
from the protection circuitry under extreme impedance variations. This is the reason that many "esoteric" audiophiles have abandoned inherent protection circuitry, which is "always" a bad idea (if you desire any type of reliability from the amplifier). In contrast, MOSFETs do not have a secondary-breakdown mechanism, and this is the primary reason for their much
greater reliability. Hitachi's tests on these devices show that they can conduct over 700-amps for 8-milliseconds with no damage.
Unlike bipolar devices, which can be destroyed in microseconds by exceending maximum current ratings, the 7-amp current rating on the specified L-MOSFET devices is an "average" parameter (such as in a simple resistive device). Thus, MOSFET devices are capable of much, much higher peak current outputs than comparable bipolar devices.
However, the Rd(on) parameter of L-MOSFETs is a little higher than the "on" resistance of a saturated bipolar power transistor, so the overall efficiency of L-MOSFET output stages is a little
less than comparable bipolar devices. This equates to the need for slightly higher rail voltages than needed for bipolar devices (for equivalent power outputs), depending on the number of parallel legs in the output stage design. I hope this helps. 🙂
* On an "overall" basis, L-MOSFETs can provide much greater peak output currents than bipolar devices. You see, a typical bipolar power transistor that may be rated for an Ic (max) of 15-amps will probably have a secondary-breakdown point (depending on where the SOA falls for the operating conditions) of about 2-amps, or less.
Thus, the overload protection circuitry must restrict output currents very tightly within certain operational regions, and such restrictions can result in distortion
from the protection circuitry under extreme impedance variations. This is the reason that many "esoteric" audiophiles have abandoned inherent protection circuitry, which is "always" a bad idea (if you desire any type of reliability from the amplifier). In contrast, MOSFETs do not have a secondary-breakdown mechanism, and this is the primary reason for their much
greater reliability. Hitachi's tests on these devices show that they can conduct over 700-amps for 8-milliseconds with no damage.
Unlike bipolar devices, which can be destroyed in microseconds by exceending maximum current ratings, the 7-amp current rating on the specified L-MOSFET devices is an "average" parameter (such as in a simple resistive device). Thus, MOSFET devices are capable of much, much higher peak current outputs than comparable bipolar devices.
However, the Rd(on) parameter of L-MOSFETs is a little higher than the "on" resistance of a saturated bipolar power transistor, so the overall efficiency of L-MOSFET output stages is a little
less than comparable bipolar devices. This equates to the need for slightly higher rail voltages than needed for bipolar devices (for equivalent power outputs), depending on the number of parallel legs in the output stage design. I hope this helps. 🙂
I would agree with sloan on this. I have used V-mosfet since the 1980s and have always loved their ruggedness (?). I have not managed to blow a single one of them. each time, the fuse is always the one to first go away on mosfet-based amps.
From that point of view, I think DIYing mosfets has a big advantage.
From that point of view, I think DIYing mosfets has a big advantage.
"Hitachi's tests on these devices show that they can conduct over 700-amps for 8-milliseconds with no damage. "
Sloan is blowing smoke again.
He quotes numbers for V mos when he uses L mos in his designs.
At room temperature the 2SK1058/2SJ162 he uses have an Rds of about 1 ohm, about 1.7 ohm over temperature.
They put out about 1A per volt of gate drive, and the gate will be permanently damaged above 14V.
It is not possible to get 700A out of one of these things under any real operating condition.
Sloan is blowing smoke again.
He quotes numbers for V mos when he uses L mos in his designs.
At room temperature the 2SK1058/2SJ162 he uses have an Rds of about 1 ohm, about 1.7 ohm over temperature.
They put out about 1A per volt of gate drive, and the gate will be permanently damaged above 14V.
It is not possible to get 700A out of one of these things under any real operating condition.
@djk: hmm.. I just ordered the kit so I'm already fried 
My Infinity Reference Standard II speakers behave naughty with 1.5ohms impedance at 40Hz and @6000Hz.... will see what comes out.
@others: 'related' thread:
http://www.diyaudio.com/forums/showthread.php?threadid=1392
My Infinity Reference Standard II speakers behave naughty with 1.5ohms impedance at 40Hz and @6000Hz.... will see what comes out.@others: 'related' thread:
http://www.diyaudio.com/forums/showthread.php?threadid=1392
Azira,
I thought the idea that FETs were inherently better sounding in audio power amp applications than BJTs died in the early 80s.
I thought the idea that FETs were inherently better sounding in audio power amp applications than BJTs died in the early 80s.
azira said:
True if you want to run lower bias currents. If you are running
high bias currents, the Mosfets start getting better. If you want
an amp that measures well and runs cool and doesn't have too
much feedback, BJT's are a good choice. If you want to make
a simple Class A amp that sounds good, Mosfets are a good
choice.
For some reason, unknown to me, I have never liked the result
of any of my efforts to drive Bipolar with Mosfet. Similarly,
I haven't had much enthusiasm for BJT's driving Mosfet outputs.
😎
Re: Re: MOSFETS vs BJTs
Is the high feedback factor important?
Will a SE class A mosfet amp with just a little feedback not be a good amp?
Nelson Pass said:
Is the high feedback factor important?
Will a SE class A mosfet amp with just a little feedback not be a good amp?
Maybe here is a good thread to get a little help. I built a discreet mosfet poweramp with 2SK1058 and 2SJ162. While I testing with sinus or something else (triangle or square) I have no problem, everything working well with +-15V...+-70V power. But when I connected some music to this circuit, the output DC shift to -Vs for a very short moment. Therefore when I want to use to listen music, the DC protection always off the speakers. I saw the result on scope, this is not a problem of protection circuit (which is UPC1237). Somebody can help me why wrong this circuit with music? The DC offset voltage very low (around 0.1-0.3mV) when no input. When something loud in music (for example drum) the offset voltage jumping to the -Vs as I said. Both channels do same. I already changed R49 and R50 to 200 Ohm trimmer pot, but nothing happend. 
I think the schematics can help to send me some advice:
http://rentahost.net/stratofan/pdf/Schematic_poweramp.PDF
The board:
I think the schematics can help to send me some advice:
http://rentahost.net/stratofan/pdf/Schematic_poweramp.PDF
The board:
An externally hosted image should be here but it was not working when we last tested it.
Re: Re: Re: MOSFETS vs BJTs
shows up at the end of the month.
Feedback works perfectly in perfect circuits. The less you need
feedback, the better feedback works.
Feedback is like a credit card - very convenient, but the billBricolo said:
shows up at the end of the month.
Feedback works perfectly in perfect circuits. The less you need
feedback, the better feedback works.
Re: Re: Re: Re: MOSFETS vs BJTs
that reminds me of bankers: the less you need money, the more they want to lend money to you. The moment you need money, they are nowhere to be found.
Nelson Pass said:
that reminds me of bankers: the less you need money, the more they want to lend money to you. The moment you need money, they are nowhere to be found.
lucpes said:@djk: hmm.. I just ordered the kit so I'm already fried
Sorry for replying to myself in this old thread, but I have something to correct. I built the kits from Randy Slone and I must say that the sound is astonishing. Good quality parts, idiot proof instructions (this being my first ever project in DIY), best assistance one can desire. You can't go wrong with these! Unless you try to
them using a hammer.- Status
- Not open for further replies.