jFET input ,full symmetrical class A circuit

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
This is a circuit I found at this sight for headphone amps:


I really like this circuit and I need some help from someone to modify it into a discrete op amp.
I was thinking of using it in an I/V and filter circuit of a DAC.

The input and driver parts I think will work perfect. The output stage is a high current one for driving headphones.

Using only one pair of bjts instead of 4 is a good idea?
Should the emitter resisotors be smaller like 4R7 or 10 Ohm ?

Could I just avoid the servo?
I don´t have so much space for it.

Also it has no overall negative feedback. Can you just apply it from output to negative input like an opamp ?


  • gilmore3_1.gif
    12.6 KB · Views: 2,069
The one and only
Joined 2001
Paid Member
All the things you want are do-able as you speculate
until you get to the op amp part. The gain of the circuit
is set at a low figure, and you will want to increase the
open loop gain so as to have some leftover for feedback.

To get this, you can take the 200 ohm resistors off the
diff pairs, and consider less resistance off the Emitters
of the 2SA1015 and 2SC1815, and maybe also adjusting
the values off the Drains of the input diff pairs, using pots
to get proper bias and offset values.
Thanks a lot Nelson. I will give it a try.
Will your active crossover be using something like this? I would guess probably something simpler. :)

Jocko , what would be the best for a stage like this?
You know a lot on this subject. What is the dream I/V and filter stage ?

I will be using PCM1704 btw
Glad to see you are not still using TDA1541s [joke].

Assuming you are using 8X oversampling, you can get away with just a capacitor across whatever load resistor is in your circuit for a filter. Yes, higher order filters will get rid of more stuff, but you will find that there is not all that much to get rid of.

As to what type of circuit, this is one place I prefer BJTs in the input stage. Try a common-base scheme, with the emitter at ground reference. There are several ways to go from that point, but a simple resistor in the collector path works. I would suggest something biased at 5 mA, or higher, as a starting point. That should get you going.

What is this information based on ?

Simpler design I would guess. You are right about the filter.
Not much to get rid of anyway above 20KHz.

I don´t know what exactly you mean but I can search.
So a common base circuit with one BJT is enough.
Where should the filter be , at the output of this bjt?
I understand a capacitor in series with the output load.
After that would a buffer stage be a good idea.
A buffer like with 2 jFETs or 2 BJTs.
Yeah I have seen it. IT´s very good but not quite my taste.
I am looking at the Nelson Pass solution that works with a balanced DAC with positive and negative input.

Its also a very simple circuit something like what Jocko is talking about.

Jocko what do you think about it?
It is based on having done this many times, over many years.

You can put a capacitor across the load resistor. A JFET of BJT follower after it would be a good idea. (Actually.....you will need a few more parts to bias the whole thing up......and as they liked to say in college...."the solution is left to the reader as homework....".)

Borbley's version has too much stuff, and he mucks it up with feedback, as I recall.

I have some ideas on making a differential output circuit using only one DAC. At >$33 each for a '1704, using a total of 4 seems too expensive. However, it is still in the idea stage, as I have more pressing problems.

Like no job.....and thanks to crooks like Bernie Ebbers.....no retirement fund anymore....., so it may be some time yet before I get back to it.

Kevin Gilmore's amp is outstanding just the way it is. Reducing the number of output BJTs will increase the output impedance and is not a good idea. The DC servo is an excellent, noninvasive method of protecting the output from DC at the input. Since it integrates the output signal to produce a bias voltage to correct DC offset, there is no reason to use a discrete circuit in place of the opamp.
I thought this circuit wouldn´t work for me because I thought it was balanced and a one circuit.

After you saying about a capacitor across the load at the outptut of the DAC I rechecked Nelson´s DAC and saw that it was 2 independant circuits, one for every phase.

I think the filter is in the input, I am still not sure how it works and how the volume works. It also has caps in the output which I dont know if they are supposed to be there. If the Circuit has DC at the output than I will need to use those.

I like it because it is simple.
I am really sorry I made a mistake here.
You can´t see it at first but the two parts are connected through the volume control.
Strange volume control.
I wouldn´t know how to make something like this for unbalanced signal.
The input IS a common gate mosfets configuration, but I think the filter is at 20 KHz only for CD use.
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.