F5 power amplifier

[Nelson Pass]

That is a possibility. To check, you need to measure
the transconductance as a function of frequency. I did
write an article on how to do that, as I recall.

[tinitus]

After having thought that laterals, like renesas wouldnt work, I thought I better read a bit, which put a small smile on

Regarding bias pots
Would someone please tell how to orient them
Original F5 schmatic shows the two bias pots mounted opposite each other
Others dont, but mount both the same way

[Nelson Pass]

I don't believe it matters.

[alazira]

Originally Posted by juma

There is a very good reason for 2W resistors in that position.
R1 and R2 are part of the feedback network and they see as much as one sixth (16,67%) of amp's output voltage which is 20V peak (max. value).
20V / 6 = 3.33V
Power dissipated on each of those resistors = (3.33 * 3.33) / 10 = 1.11 W (peak).
1/4W is OK only if you don't listen to your amp too loud.

Juma & Jackinnj, thanks for the reply.

As the discussion concerning LTSpice shows, theoretical and actual experiences can differ. Many here have commented that the commercial F5 using 25V rated caps seem to be pushing it. I would guess more people than just Peter Daniel have 1/4 watt in the R1&R2 positions without any reports of failure that I can recall seeing anywhere on this or the F5 group buy threads. Is this 1.11W max peak reached in class A mode or when pushing into class AB? Since its a peak and unlikely sustained, perhaps the resistors can manage it for a short period.

In any case, if its just a failed resistor, I'm ok with taking my chances (an expensive speaker driver would be a different story ). Many here already build the amp without circuit protection or push the bias way past standard F5 spec. DIY's tend to be creatures who like to take some risks to push the performance envelope

[alazira]

Originally Posted by jackinnj

1)
Note the paralleled devices:



2) There are TO-220 devices which are pretty similar in Fairchild's lineup -- the FQP12P20 and FQP19N20C -- they will handle a bit less current, have slightly higher gm. The IRF/Vishay IRF240 devices will handle 20A continuous vs 21.8A for the Fairchild FQA19N20C and 19 for the FQP19N20C.

Hi Jackinnj,

Perhaps it's the picture, but those traces look silver. How did you manage that?

Garrett

[enochRoot]

the traces are tinned probably

[juma]

Originally Posted by alazira

.... theoretical and actual experiences can differ.

No, they can't. Interpretations do.

Originally Posted by alazira

...more people than just Peter Daniel have 1/4 watt in the R1&R2 positions without any reports of failure...

One more time, power dissipated on that resistors is directly proportional to volume level that you use to listen to your amp. Without signal they dissipate from 0.25 mW to 0.4 mW (depending on JFET's Idss).
If you don't turn the volume knob past 8 o'clock you can use 1/16W resistors.

[jackinnj]

Originally Posted by alazira

Hi Jackinnj,
Perhaps it's the picture, but those traces look silver. How did you manage that?
Garrett

In the boards I burn myself I use a product called "Liquid Tin" from MG Chemicals. The ones above I sent out.

[generg]

I like the sound of the F5 without current limiter more, may be because my Thiel SCS 4 has only 4 Ohm.
But my current is going up and up, slowly but sure. Will be there a stop? Or is it impossible to get a balance?

Oh I forgot I put away the thermistors too!

[Russellc]

Originally Posted by Peter Daniel

I bought 16pcs some 15 years ago at Active Surplus at $9.95/pc. Those were the good times

I had them sorted out for much better looking chassis, but these days I prefer simplicity and open concept when possible.

Peter, is this aluminum the case (not the heat sink) is made from? Very nice.
I would like to build my F-5 similar to this example.

Russellc