Hi Bob,
This is something I'm messing about with at present.
Aside from the added complexity is does appear to have a couple of obvious advantages. The first being that you can run a lower driver bias and use smaller faster devices especially if the drivers are cascoded. The second obvious advantage is that the PCB traces between the drivers and the MOSFET gates can be very short.
Cheers
Paul
I do agree with this , the first watt is the most important in preserving the sound
That's right!
Output rail is 80VDC.
And this's layout of Telos 1000 OPS.
The PCB traces between the drivers and the MOSFET gates are very short.
MOSFETs have ferrite beads in G and D pins.
Attachments
Hi Bob
The idea is to increase the slew rate and frequency response of the amplifier using a separate driver for each MOSFET. Have a look at this application note from Texas Instruments at page 19:
http://www.ti.com/lit/an/snaa045a/snaa045a.pdf
This is true, but it is not necessarily true that BJTs outperform MOSFETs at low power. Some BJT designs are terrible at low levels, partly due to a low bias current where distortion is optimized (especially so for Self's CFP output stages). MOSFETs like higher idle bias current, so the class A region is larger and the crossover distortion is softer. Yes, it is true that you pay a little more in heat generated in the quiescent state, but in my opinion it is worth it. The numbers are not bad anyway, and are certainly well below the amount of dissipation the heat sinks must be designed for for higher power levels. MOSFETs are happy with 150 -200 mA per pair.
Cheers,
Bob
Has anyone tried this setup ...?
I'm a fan of the big Accuphase, good sonics. Big price ...
Maybe, because some of the early amps like my first Sony V-Fet integrated had a veiled midrange and no bass, these concepts stuck and gave the reviewers something to talk about. I heard no such thing in my B&K (but it had no dynamics) or my modified DH-120. Pretty sweet, but not the match for my HCA's. Neither were my DH220's.
The lack of "balls" in the B&K was also a common complaint. Well if you omit the drivers or have too small a transformer, that's what you get. Erno knew better.
Now if a US reseller would just pick up Exicon and Alfet so I did not have to make a car payment for shipping, I would be building the lateral version of my reasonable amp, not the BJT. Does anyone have any subjective comments about Exicon vs Alfet or even lateral vs HEXFET? What do you all think about using the smaller Alfet as a driver for HEXFETs instead of BJT's as the old mid-size laterals are no longer in production?
My question referred to the first part of your comment.
A proper bias figure for MOSFETs is at least 100mA, goes for lateral and verticals, preferably 125mA (or more).
A decent PP power amp has an output stage with at least 2 pairs of output devices, which means 1W or more in class A.
For BJT's, optimum class AB is biased to a Vre of 26mV.
A 2-pair output stage with 0R22 emitter resistors would also come in reach of that 1st class A watt.
Which makes this comment rather meaningless in the real world :
''In all respects, except for Class A, BJT output devices simply outperform MOSFETs in all low level distortion respects''
Which ones do not ?
1-pair BJT output stages, or those with emitter resistor values of 0R33.
A really good BJT output stage will have 0R1 emitter resistors.
Optimum class AB biasing at 26mV would imply a quiescent current figure of 1/4A per output device.
So for large output stages, which serious power amps have, optimum biased BJT power devices are a drawback.
(to repeat myself : a pair of 300W Marantz power amps cost some $15k in the US. They have balanced BJT output stages with 0R1 emitter resistors, which are biased at 5mV. Which means a total idle dissipation for the output stage of 30W, 1/3d watt in class A, and imho sound exactly as such)
Last edited:
Whatever they used. They advertised V-FET. I thought they had the early Hitachi's vertical MOSFETS in them, but that was in the 70's so I really don't know. I bet the reviewers didn't either.
I found I got away with 10mA per mosfet on my irsp240/9240 amplifier.
I applied a sine wave signal to input and monitored output on a scope.
I turned bias right down.
I turned up bias very slowly until cross over distortion went and that was at 10mA across the mosfet source resistor which was 0r22.
I thought I had done something wrong after reading lots of other people using a much larger bias current. Then someone mentioned that Peavey use similar currents in their amplifier's.
My amplifiers sounded fine with a small bias current. I could clearly tell when there was crossover distortion as you can clearly hear it with little or no bias.
I applied a sine wave signal to input and monitored output on a scope.
I turned bias right down.
I turned up bias very slowly until cross over distortion went and that was at 10mA across the mosfet source resistor which was 0r22.
I thought I had done something wrong after reading lots of other people using a much larger bias current. Then someone mentioned that Peavey use similar currents in their amplifier's.
My amplifiers sounded fine with a small bias current. I could clearly tell when there was crossover distortion as you can clearly hear it with little or no bias.
My question referred to the first part of your comment.
A proper bias figure for MOSFETs is at least 100mA, goes for lateral and verticals, preferably 125mA (or more).
A decent PP power amp has an output stage with at least 2 pairs of output devices, which means 1W or more in class A.
For BJT's, optimum class AB is biased to a Vre of 26mV.
A 2-pair output stage with 0R22 emitter resistors would also come in reach of that 1st class A watt.
Which makes this comment rather meaningless in the real world :
''In all respects, except for Class A, BJT output devices simply outperform MOSFETs in all low level distortion respects''
Which ones do not ?
1-pair BJT output stages, or those with emitter resistor values of 0R33.
A really good BJT output stage will have 0R1 emitter resistors.
Optimum class AB biasing at 26mV would imply a quiescent current figure of 1/4A per output device.
So for large output stages, which serious power amps have, optimum biased BJT power devices are a drawback.
(to repeat myself : a pair of 300W Marantz power amps cost some $15k in the US. They have balanced BJT output stages with 0R1 emitter resistors, which are biased at 5mV. Which means a total idle dissipation for the output stage of 30W, 1/3d watt in class A, and imho sound exactly as such)
15K Marantz.... really now.
A really good BJT output stage will have 0R1 emitter resistors.
Optimum class AB biasing at 26mV would imply a quiescent current figure of 1/4A per output device.
So for large output stages, which serious power amps have, optimum biased BJT power devices are a drawback.
Using RE as low as 0.1 ohm in each emitter is thermally dangerous, especially when beta variations with temperature can interact with small DC drops across base stopper resistors. 0.1 RE may be safer with ThermalTraks, but I've never tried it.
Cheers,
Bob
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.