EF86-EL84: too much gain!

[pftrvlr]

Hi,

I spent the last week modifying a 12ax7-el84 headphone amp. The major change is replacing the 12ax7 with ef86 and getting ride of the nfb. The project went fine and the amp is now working fine. There are still a few things I want to improve. The first one is I need to reduce the gain. It is way too sensitve to vol control and the normal listening level is at 9 o'clock.

I want to keep the idle current (ef86 @ 1.8ma, el84@35ma) as it is. Any suggestion as to reduce the overall gain? The schematic is below:

[planet10]

My current amps are similar, but the EL84s are wired in pentode, and there is plate to plate feedback (search for RH84 here and on the net). more than twice the power and at least similar quality sonics.

dave

[dsavitsk]

Originally posted by pftrvlr
getting ride of the nfb.

The lack of a cathode bypass caps implies some local nfb. Whether this is bad or not is up to you, but it is there (and it reduces gain.)

Anyhow, you could swap the output tubes for some 12B4's, or you could shunt the volume pot by putting a resistor (about equal to the pot resistance) in series with the pot's input. Or you could add a resistor connecting C8 with R20 to build a voltage divider (which is similar to the shunted pot, just in a different place). Lower R20 to 500K and use a 500K for the series resistor. Or, you could just drop the driver altogether -- not sure why you would need it for a headphone amp.

[engels]

Originally posted by dsavitsk



Lower R20 to 500K

[pftrvlr]

Thanks guys.

When I said I got rod of the NFB, I mean the global NFB. I add the local NFB intentionally to reduce the gain of each stage.

I will do some PSPICE sim of the RH84 to see how it work out.

[GordonW]

One other idea- try running the EF86 in triode mode. This will drop the gain significantly.

Of course, this will also increase the Miller capacitance of the input. However, with low-gain settings on the EF86, this shouldn't be much of a problem, if any...

Regards,
Gordon.

[planet10]

Originally posted by GordonW
One other idea- try running the EF86 in triode mode. This will drop the gain significantly.

According to his schema he already is...

dave

[dsavitsk]

In the end, it is odd to use a high gain tube, and then complain that you have too much gain. There are plenty of other choices out there that will solve the problem.

[Wavebourn]

Congratulations;
replacing the tube you got more gain so GNFB may be deeper now that means less distortions and better sound if to compare against "slight" GNFB that is enough to make a specter of distortions wider, but is not enough to bring level of distortions down. A negative feedback in tube amps is such sharp thing so you have only 2 options: use it as deep as possible, or do not use at all. To prevent oscillations you may need an additional minimal possible capacitor from anode of the 2'nd toob to cathode of the 1'st.