F5 power amplifier

[jackinnj]

Last question, what is your supply voltage?

It's +/- 24VDC

The DC offset is tweaked down to about 10mV which is nice, but this takes a bit of patience.

The DUT was the one without the thermal compensation, although that shouldn't make a difference.

 

[Nelson Pass]

I asked because the design was optimize for 24V, and as you
go down, any peaking is exacerbated.


Your vocabulary word for the day-

Exacerbate: to make more harsh or severe.

usage: "This solid state amplifier exacerbates the peak in my tweeter"

 

[Zen Mod]

...........
Exacerbate: to make more harsh or severe.

usage: "This solid state amplifier exacerbates the peak in my tweeter"

cure :

just exactoknife that exacerbating tweeter .......

 

[jackinnj]

Exacerbate: to make more harsh or severe.

usage: "This solid state amplifier exacerbates the peak in my tweeter"

I'll use that with my wife this evening.

 

[GRollins]

Just a passing observation...but if all you've got is a tweeter, shouldn't you be trying one of those spam e-mail pills that claims to give you a long-throw woofer?

Grey

P.S.: I've always been partial to the word sesquipedalian, myself.

 

[chinsettawong]

Your vocabulary word for the day-

Exacerbate: to make more harsh or severe.

usage: "This solid state amplifier exacerbates the peak in my tweeter"

Now, we are learning not only to make amps here.

 

[fab]

PROFET Question

Much of the difference is due to the high degeneration of their
input JFETs, the lower transconductance of Laterals and their relatively
high output gain. This means a lot less feedback, as seen by the
output impedance and distortion specs.

For Common Source topologies, damping factor = feedback.

Since I do not have complementary vertical mosfet in hands but lateral mosfet, I have tried the "Profet" version and I can not make it work on the bench when loading with 8 ohms. When I connect the 8 ohms load, the signal amplitude drops to about 70%. The open loop gain is way too low. I even used lower degeneration resistors for the jfet and less close loop gain than the original profet. I used only one output pair as the Profet. And the THD is quite high. It seems that laterals 2SJ49/2SK134 (or 2SJ162/2sk1058) have about 10 times lower transconductance than IRF244 for example. Has someone successfully built and heard an F5 or Profer with lateral outputs? I know that jackinnj has described a lateral F5 adaptation in one post somewhere... he had 3 output laterals but even that is a lot less transconductance than 2 output verticals...

Thanks

 

[fab]

Profet like schematics

Here is the schematics for the previous post.

M1 and M4 are laterals.
Bias current is about 0.6A

J1 and J2 are SK170/J74

Thanks in advance for your comments

 

[jam]

Fab,

.................Nice! My thoughts exactly.........byt do you need the compensation capacitor (C21)?

Jam

 

[Nelson Pass]

Re: PROFET Question

Since I do not have complementary vertical mosfet in hands but lateral mosfet, I have tried the "Profet" version and I can not make it work on the bench when loading with 8 ohms. When I connect the 8 ohms load, the signal amplitude drops to about 70%. The open loop gain is way too low. I even used lower degeneration resistors for the jfet and less close loop gain than the original profet. I used only one output pair as the Profet. And the THD is quite high. It seems that laterals 2SJ49/2SK134 (or 2SJ162/2sk1058) have about 10 times lower transconductance than IRF244 for example.

That appears to be consistent with "Chuck" Hansen's review of the
Profet in AX magazine, in which he measured a 6.8 ohm output
impedance. That doesn't mean that the amp sounds bad, but it
tends to limit its application.

 

[EUVL]

> I asked because the design was optimize for 24V, and as you go down, any peaking is exacerbated.

Would you be kind enough to elaborate the mechanism that causes the exacerbation at lower supply voltages ?

I suppose the capacitance of the MOSFET increases ......


Patrick

 

[Skorpio]

Any specs for the NTC resistors?

There are several types......also in 4k7 range

Also details about mounting? Direct on housing with glue, in area around, direct on heatsink, etc....?

 

[TimS]

Has anybody tried the 2SK1530's and 2SJ201's in the F5?

 

[Nelson Pass]

Would you be kind enough to elaborate the mechanism that causes the exacerbation at lower supply voltages ?

I suppose the capacitance of the MOSFET increases ......

That would account for at least part of it. I'll look into it further when
I have time.

 

[EUVL]

If it were the capacitances of the MOSFETs, then perhaps one need to use IRFP240 for the N-FETs ?? If I remember correctly, the FQA19N20C has somewhat different values to the IRFP240.

> I'll look into it further when I have time.

Thanks in advance.


Patrick

 

[plep]

FQA12P20 ?

Dear Nelson,

I’m unclear.
Do you mean that we have better to use FQA12P20 instead of IRFP9240 in general? I mean in the F3 and F4 also?

How about replacing the IRF240 by FQA19P20?

Best regards.
Philippe (who, like many other here, is waiting for your “buffer “ paper because, today, he uses a complementary fet buffer to drive an *old* power amp that has an input impedance of 1.2 KOhm and 32 db of gain)

 

[Nelson Pass]

Re: FQA12P20 ?

Do you mean that we have better to use FQA12P20 instead of IRFP9240 in general? I mean in the F3 and F4 also?

How about replacing the IRF240 by FQA19P20?

waiting for your “buffer “ paper

You can use the FQA12P20 as a general replacment for the
IRFP240, and it will usually work as well or better.

In unbalanced Common-Source applications, it makes a difference
using the IR made parts versus the alternatives (equivalent P
channel parts made by some else).

In Common-Drain or Common-Gate unbalanced applications it
makes less of a difference, but it is still an improvement to use
the non IR made parts.

In balanced applications, such as the PL X amplifiers, we find
virtually no difference, and we use them interchangeably,
although the Fairchild parts are cheaper.

There is no reason to worry about the N channel IR parts, such
as the IRFP240.

And I'm still working on the buffer article. I expect to be done in
a week, but I have to decide where to publish, as AX has a 2
month lead time.

 

[plep]

Thank you Nelson.

If you allow me a last question for now...
Is this new buffer cell the basis of the future crossover you already talked about?

Best regards
Philippe.

 

[Nelson Pass]

Is this new buffer cell the basis of the future crossover you already talked about?

Yes, although Part 1 deals only with the buffers and some tweaks
to get the lowest distortion or the desired ratio of 2nd to 3rd
harmonic.

Part 2 will cover some applications.

 

[fab]

Fab,

.................Nice! My thoughts exactly.........byt do you need the compensation capacitor (C21)?

Jam

thanks Jam

C21 made no difference on my proto board but I put it there only by precaution. The value could be increased though since the frequency cutoff it creates is so high that is had probably no effect.