Pass JFET Power Amplifier

Status
Not open for further replies.
Jump to Latest
#204
Thanks for your kind confirmation.

As you have explained, I have thought that the upper FET, as a Common Base structure, is there just to provide the almost constant voltage across the lower FET and to dissipate the most heat, and it has unity current gain. Therefore, I have believed that the internal capacitance values of the upper FET can be ignored or none.

If I¡¯m wrong, please correct me.

Regards
jH
 
#205
Im going to try IRF3706 as Q1, with a IRFP250 as Q2 a soon as I get some time of.

I think that you want a realy high transconductance device as Q2 to minimize the voltage fluctuations over Q1, but I have no idea how much difference it does.
Some of it will be swapped by the modulation the circuit provides.
This might be just purely theoretical, but its fun expirementing.
 
#209
Tyimo said:
What about the PSU requirements?

300VA 35V, 4x 10.000 uF 63V per channel is enough?

Tyimo
Click to expand...
Add a 3 mH aircore inductor between the banks of capacitors for much better filtering. A normal aircore loudspeaker coil is fine. Preferably at least 1,2 mm wire to keep dc resistans down.
Or better yet, build a Capacitansmultiplier.
 
#214
Re: Device-selection? Mr Pass???

Circlomanen said:
What would be the absolutely best transistor as Q1 (not counting the Lovoltech j-fets) that you can think of ???
Click to expand...

I don't want to think that there is a "best" simply because this
is an area where you consider the tradeoffs in getting the
sound you want. A lateral Mosfet or a small vertical Mosfet will
have better response at the highest frequencies, and I think
will sound different than a big vertical Mosfet, which will shine
at lower frequencies.

Having said that, what you have is a fine opportunity to try all
sorts of different devices in that spot and see which ones you
like - it doesn't require anything by way of changing the rest of
the circuit.

😎
 
#216
Mr. Pass, thank you for delving into the finer points of what makes your favorite gain devices your favorites. You would have to be expecting more questions??? Could you please entertain us more regarding what makes the devices we want to use sound like we would like them to sound??? For instance You elude to only small verticle devices??? Are the larger verticle types covering excessive die area and creating larger capacitance??? therfore only good in the low end??? Is there some balancing going on with these C's, Coss and Crss Parameters that are also important??? Is the transfer characteristic curve being convex or concave an indication of parameters at work in your designs??? Thanks again😀 😀 😀
 
#217
The transfer curve is always important. It defines the gross deviations from linear behavior that you can expect from the device. The various capacitances within the device (from Gate to Source and Drain, and from Source to Drain) all factor in as frequency-dependent filters that modify the response on the fly.
Is there a perfect device? No. Is there an optimal one? Maybe. It depends on how many parameters you are willing to lock in. The flexibility required of commercial product leaves the designer balancing on a knife edge. DIY folks can pick and choose the capacitance, transconductance, and other factors in order to fine tune their circuit for use with a super tweeter or a subwoofer.
The peril arises in expecting an absolute answer to a relative question.

Grey
 
#218
flg said:
You elude to only small verticle devices??? Are the larger verticle types covering excessive die area and creating larger capacitance??? therfore only good in the low end??? Is there some balancing going on with these C's, Coss and Crss Parameters that are also important??? Is the transfer characteristic curve being convex or concave an indication of parameters at work in your designs???
Click to expand...

There is always a balance, and there is always a trade-off, but
you can pick the characteristics you like, and this being a
subjective area, you get to decide what you like.

The concave transfer curve (specifically the current vs voltage
across the device for a given input value) is interesting in that
it potentially allows for some cancellation against the current
vs input at a given voltage across the device. We generally
look to triodes for this sort of thing, but the power JFETs
also have a useful region with this curve.

😎
 
#219
Well, I feel reguvinated. Thank you Mr. Rollins and of coarse Mr. Pass. I guess you wish to let me evaluate the details of the answer to my question... But, I have a few problems acoomplishing that. I have a natural (w/age, concerts, guns, motorcycles, drag racing etc.) filter in 1 ear similar to a 6 pole box in the trunk. The other ear has a response that makes me wonder why I hear enough to enjoy music at all. But I still have ears, and a scope, and I just like this audio reproduction stuff...
All that aside. I notice most of the mosfet transfer curves, from various manfacturers, have what I would call a concave trend on the Vgs vs. Id curve. However, when evaluating the IRF parts Mr. Pass uses so often, you can see an opposite trend. What I would call a convex curve. I guess it could also be said that these devices have decresing gain with higher Id or Vgs or both??? They also have a threshold (knee) below the curve totally??? There are many devices out there called mosfets that have a knee that dosen't flatten out until the device is burning up (very high Id's). I beleive the ideal would be to get above the threshold into the linear portion of this curve without excessive Pd. But, is there a correlation of the Vgs vs Id characteristic and the subjective or objective performance.
Mr. Pass, you commented on the JFET's Vds vs. Id, I realize I'm in a Zen8 thread but, I guess the only way to acheive that type of curve is with JFETs or triodes??? I also notice there is no Tranconductance spec in that datasheet. I'm gussing these things have high Gfs based on the Id going strait up in that transfer characteristic curve??? I have a couple "variations" ready to be built that I think you would like...Thanks again...
BTW, Mr. Pass, do I see a mistake on the Blues JFET input Zen4. If I read that right, isn't that jfet juntion melted by now??? Is there 10V on the gate and +3 more on the source??? forward biasing the "J" part of that FET. I need a couple volts more negative on my source??? But then I'm not real fluent with all these reverse upside down depletion things. It seems like these datasheets avoid using + and - on the Bv's too. What is that? Part of the source and Drain interchangability???😀 😀 😀
 
#220
flg said:
I notice most of the mosfet transfer curves, from various manfacturers, have what I would call a concave trend on the Vgs vs. Id curve. However, when evaluating the IRF parts Mr. Pass uses so often, you can see an opposite trend. What I would call a convex curve. I guess it could also be said that these devices have decresing gain with higher Id or Vgs or both?

Mr. Pass, you commented on the JFET's Vds vs. Id, I realize I'm in a Zen8 thread but, I guess the only way to acheive that type of curve is with JFETs or triodes??? I also notice there is no Tranconductance spec in that datasheet. I'm gussing these things have high Gfs based on the Id going strait up in that transfer characteristic curve?

BTW, Mr. Pass, do I see a mistake on the Blues JFET input Zen4. If I read that right, isn't that jfet juntion melted by now??? Is there 10V on the gate and +3 more on the source???
Click to expand...

When we talk of convex and concave curves, it's easy to get
confused. The concave curve is one that goes up with an
exponential shape and looks a bit like 1/2 of a parabola. The
convex looks more logarithmic - it continues to go up, but at a
slower rate.

Triodes, Pentodes, Mosfets, and Jfets are what we would think
of a concave when considering transconductance versus current.
For a given voltage across the device (Drain to Source, for
example), the current through the device goes up exponentially,
up to some limit. Bipolars do not tend to do this.

When considering transconductance versus voltage across the
device, they all go up with more voltage, but triodes go up
exponentially. Pentodes, bipolars, Mosfets, and Jfets tend to
level off above a certain voltage level (thus the convex shape)
but the power JFETs have a region at lower voltages where the
curve has an exponential (concave triode shape), and then at
higher voltages they start looking like the others.

We are measuring two things here, and when we talk about
having "concave" shapes to these two different curves, we are
discussing the potential for some distortion cancellation in how
we place the working load line of the gain device.

And yes, the voltage called out should have been 10V not 13V,
and there was a subsequent post in that thread which corrected
it. The junction would not have melted in any case - you would
simply not see that voltage.

😎
 
Status
Not open for further replies.
Top Bottom