If I remember correct ...IRF9640/640 as driver , with gate protection zener/diodes...
and standard MJLxxxx as OP. Multiple pair OP can be used.
Simulates the same as a renesas IGBT.
The REAL igbt (renesas) is more optimized with a much lower Vsat ..
protection diodes would still be needed.
Renesas Electronics 7th-Generation 650V and 1250V IGBT Series sets a new technology benchmark enabling high efficiency solutions | Renesas Electronics
This device can take 10 m/sec. short circuit conditions , MJLxxxx in the
"fake IGBT" would be toast at this point.
I would try this for a LF amplifier application .... standard EF2/3 are
much better in the HF arena.
OS
Why do you consider such high capacitance mosfet in an IGBT as "driver"? If discrete IGBT, Lazy Cat has one example (SSA High Performance BIGBT).
Neat idea. I've just being doing some TINA simulations using a BUZ901 and an IRF540. Looks as if using the vertical MOSFET (in class-C) in parallel with the lateral is a very workable concept.
You guys are great! Sharing all your ideas and best efforts. Great stuff! OStripper, looks like a new OPS is in the making to add to your CFA Face-off. 
We all benefit from everyone's interaction and cross-pollination of ideas and experiences. I have to do some thinking now.... meanwhile, got to go to an old geezers birthday (80th) now. Be back soon. Hope I dont miss something.
Thx-RNMarsh
We all benefit from everyone's interaction and cross-pollination of ideas and experiences. I have to do some thinking now.... meanwhile, got to go to an old geezers birthday (80th) now. Be back soon. Hope I dont miss something. Thx-RNMarsh
RNMarsh (4365)
You guys are great! Sharing all your ideas and best efforts. Great stuff!
As for OPS that I liked the idea here
http://www.diyaudio.com/forums/solid-state/222816-new-lineup-idea-power-follower-output-stage.html
This may be a good alternative to OPS with ECH
best regards
Petr
You guys are great! Sharing all your ideas and best efforts. Great stuff!
As for OPS that I liked the idea here
http://www.diyaudio.com/forums/solid-state/222816-new-lineup-idea-power-follower-output-stage.html
This may be a good alternative to OPS with ECH
best regards
Petr
Here is an example of one structure field effect transistors in the OPS
http://s019.radikal.ru/i635/1402/3b/7606002bc8e2.png
best regards
Petr
http://s019.radikal.ru/i635/1402/3b/7606002bc8e2.png
best regards
Petr
Thanks
And it might even qualify as a CFB design. Did you ever decide whether feedback applied to an emitter of the input stage was CFB or not ?If you did I can post the design.
Attachments
Is this your opinion only or exists there detailed technical details? Maybe you are able to provide us some links or references?
This amplifier easily delivers > 220W 8R and > 440 4R using a 2EF bjt ops and has a very good low distortion performance.
BR, Toni
The EF2 for powers of up to 100 W is not a rule- just a guideline. For a 200 W EF2 amplifier, typically you would see 4 or 5 output devices and a driver stage. Into a worst case load of 3 Ohms, you would see a significant increase in distortion, and into 4 ohms, more than most designers would consider acceptable.
You've used many OP devices in parallel, and you've paralleled your VAS devices. This as you probably know means that for the most part, the OP devices operate where the hFE vs IC is high, helping to keep the distortion low. With the doubled up VAS devices, each device operates over a narrower IC range, and narrower hFE range also helping with distortion (same of course applies to the output stage).
The techniques you have used thus also achieve low distortion. The other option here would be an EF3, where with you design I would expect a further 10 to 20 dB reduction in distortion i.e 1-2 ppm range. However, chasing numbers this low is wasted effort in my view.
However, your amplifier looks like a very nice design - congratulations!
You've used many OP devices in parallel, and you've paralleled your VAS devices. This as you probably know means that for the most part, the OP devices operate where the hFE vs IC is high, helping to keep the distortion low. With the doubled up VAS devices, each device operates over a narrower IC range, and narrower hFE range also helping with distortion (same of course applies to the output stage).
The techniques you have used thus also achieve low distortion. The other option here would be an EF3, where with you design I would expect a further 10 to 20 dB reduction in distortion i.e 1-2 ppm range. However, chasing numbers this low is wasted effort in my view.
However, your amplifier looks like a very nice design - congratulations!
Last edited:
Hi Toni,
Nice work and results in that design. 2EF designs can provide good performance, but are much more vulnerable to beta droop and ft droop. Even in somewhat lower-power designs, I generally believe that a Triple is a more cost-effective way to get high performance under a wider variety of conditions, such as high current into low-impedance loads. In lower-power amplifier designs, where SOA considerations might allow as little as one output pair, it is more cost-effective to add the pre-driver to make a Triple than to add more output devices just to avoid problems with beta droop and ft droop.
So there is no hard and fast rule, just a range of preferences. Self has historically preferred 2EF, while I have usually preferred Triples. Both can yield good performance. Self achieved his so-called load invariance by paralleling output devices, I achieved it by going straight to a Triple.
Cheers,
Bob
I simulate my triples at 2R . Even as I never expect this , I'll get "anal" and
go for 1 or lower. Just a slightly larger "suckout" cap at the driver stage.
On a different tangent , anyone ever tried a class A OP on a CFA ?
Here is one (below) that is quite promising. H2 only and total THD20 about
as low as my standard EF3 (w -gnfb).
OS
go for 1 or lower. Just a slightly larger "suckout" cap at the driver stage.
On a different tangent , anyone ever tried a class A OP on a CFA ?
Here is one (below) that is quite promising. H2 only and total THD20 about
as low as my standard EF3 (w -gnfb).
OS
Attachments
Ostripper, I also sim at 2 ohms - I don't think it's anal, just good engineering
Your approach above looks very interesting - I'll need to have a play with that.
For another class A CFA, you can look here Ovation sx-Amplifier: A 15 W class A Amplifier
15 W but great for classical and jazz and it drives my 89 dB speakers. The bias arrangement may be of interest.
Interesting you ask the question because I am just looking at this for a high power class A using output stage a la NP F5 with sx/nx front end sans VAS stage.
Your approach above looks very interesting - I'll need to have a play with that.
For another class A CFA, you can look here Ovation sx-Amplifier: A 15 W class A Amplifier
15 W but great for classical and jazz and it drives my 89 dB speakers. The bias arrangement may be of interest.
Interesting you ask the question because I am just looking at this for a high power class A using output stage a la NP F5 with sx/nx front end sans VAS stage.
Last edited:
Ostripper, I also sim at 2 ohms - I don't think it's anal, just good engineering
Your approach above looks very interesting - I'll need to have a play with that.
For another class A CFA, you can look here Ovation sx-Amplifier: A 15 W class A Amplifier
15 W but great for classical and jazz and it drives my 89 dB speakers. The bias arrangement may be of interest.
Interesting you ask the question because I am just looking at this for a high power class A using output stage a la NP F5 with sx/nx front end sans VAS stage.
With the above values , 4 pairs would be good. I have this running 80V p-p,
not quite the "power hog" as the genesis class A OPS.
Nice thing is that ALL the THD residual is H2.
PS - this "alien" OP negates itself thermally ... the MJE's offset the NJW's nearly perfectly.
OS
Attachments
Last edited:
Thanks! Think it's worth a try using the current input stage/doubled VAS driving a 3EF ...
Thank you and thanks for clarification! Simulation of 3EF shows indeed an increase in THD performance e.g. using 2.5R compared to 8R load hence better load invariance. But square wave simulation also showed more signs of instability / oscillations...
Maybe in real world designs we lose performance getting a 3EF ops stable compared to a 2EF ops? (Of course using the same input/VAS stage)
BR, Toni
This one does not like FET's , I tried it ...
PS - the "genesis stage" does VERY well with IRFPxxx !!
OS
no oscillations ??
triple (below 1) puts out wave (below 2) into 1u/2r load.
I can't get this triple to oscillate - whatever I do !
20 different input stages - the OPS NEVER misbehaves ....
OS
