6EM7 Headphone Amp

[Paul Barker]

Is one of those bare rails a bus rather than star ground? Just wandering. Not that it would cause the em7 to run away. But if the darlings are on the same coupling caps, those are not implicated. I always thought the 2meg grid resistance for the power section made no sense. In use as audio rather than TV I’d suggest 500k max which is same as 2a3 limit. Did you use grid stoppers?

Don’t worry though take a break from the 6em7 for now. But if you go back to it; give it an easy life no more current than 50mA no more anode voltage than 150. First time I used them as audio they sounded forsed nothing like the beautiful sound I ended up with keeping to the above limits. Stephie back then explained why you have to keep it to those above limits but I can’t recall the technical term for the syndrome. Val was talking to me about 6em7 circuits he had found which achieved more power out but I saw that they were pushing it too far, so things learned 25 years ago no longer in the modern designs of the 6em7….

Certainly for a headphone amp easy to keep it in its happy place.


I’ll search for a link to me mensioning this problem many years ago soon after I learned it from Stephie in the email group we were both in with Jono and Gary et al.

[Paul Barker]

I cant find a post early enough to when I mentioned what Stephy had told me. But my view confirmed here. I had said the 150v 50mA operating point . It was an era when the bottlehead led bias and the battery bias were flavour of the era. I wouldnt change back to either any longer. I prefer resistors these days.

https://ibb.co/zV1mfV9f

[Cressy Snr]

Cheers Paul,
I always use bus bar earthing. Star earthing sometimes makes more noise than busbars, probably because I don't do it properly, but there you go. In the future I'll lower my voltages. There's no reason for high voltages in a headphone amp. At least I've learned something.
With the 1626s I've halved the voltages by putting the secondaries in parallel, so I have 0-120V into a solid state bridge package. Noise is lower.

[Cressy Snr]

Schematics:

1626HPA_v2.0.png (37.29 KiB) Viewed 16382 times

1626PSU_V2.0.png (51.18 KiB) Viewed 16382 times

[Cressy Snr]

Applied some feedback to the headphone amp:

1626HPA_v2.1.png (36.8 KiB) Viewed 14192 times

It was too loud at only 9 o'clock on the volume control. Now I have a bit more range and the background is virtually silent, even with AC heaters.

[Paul Barker]

Schematics:
1626HPA_v2.0.png1626PSU_V2.0.png

You got the operating point right.

[Cressy Snr]

Cheers Paul,
Strictly speaking, I need to replace the cathode resistor on the ECC82 with a higher value in order to maintain the same bias point I had before applying the feedback.

[Nick]

Cheers Paul,
Strictly speaking, I need to replace the cathode resistor on the ECC82 with a higher value in order to maintain the same bias point I had before applying the feedback.

A cap in the feedback loop would avoid that for DC, and would also remove the small DC offset across the output transformer. But it may need to be a biggish value (20uf maybe for 7Hz) to avoid a gain peak at LF as feedback reduces.

[Cressy Snr]

A cap in the feedback loop would avoid that for DC, and would also remove the small DC offset across the output transformer. But it may need to be a biggish value (20uf maybe for 7Hz) to avoid a gain peak at LF as feedback reduces.

Copied Gary Devon's feedback config for the WAD HD83, from the HFW DIY supplement - Dec 1998. Slightly different component values due to me having no feedback winding, but it works well giving great sound.

1626HPA_v2.2.png (40.43 KiB) Viewed 13511 times

Screenshot 2025-07-24 at 15.38.30.png (183.11 KiB) Viewed 13501 times

[Nick]

Yep, but the feedback winding allows him to do that without worrying about DC on the headphone output.

[Cressy Snr]

Oh bugger!
I’m gonna fry my headphones if I don’t get a fecking grip!
Too thick for this hobby

[Nick]

Just go back to your previous circuit, with the 100uf in the feedback link from the output transformer.

[Cressy Snr]

I checked the offset on the OPT secondary; it was 1.9mV. Whether that’s bad for headphones, I’m not sure, but it’d probably be better without.

[Nick]

Probably won't do much if any harm. But the cap would let you go back to the original DC conditions for the input valve.

[Cressy Snr]

OK thanks Nick.