Output transformers and frequency response

[dsavitsk]

Hi,

I have a headphone amp that I have been working on for a while. It is up and running, and in its current version is a 6N6p on each side (with both sides in parallel) transformer coupled to a pair of headphones.

The headphones have a 32R impedance. The transformers are 5K:8. I have a 33R resistor in parallel to the headphones, so 10K o nthe plates.

Anyhow, the question is that the bass response is a bit weak. Sound otherwise is great, and on recordings with plenty of bass the bass is nice, but if the recording is bass shy at all, then the amp is all midrange.

So, is there a better way to set this up? No resistor, a bigger resistor, or would I be better off with transformers with a true 32 ohm tap? Is it the tube choice -- though 6922s sounded about the same?

Also, for now the tube's cathodes are resistor biased and are not bypassed in case that matters. Basic schematic is attached

[Geek]

Hi,

Are you using a SE (gapped core) transformer?

If not, the core is saturating.

If so, the primary inductance may not be sufficient for the plate Z, or Z ratio.

[AndreasS]

The plate resistance of the tube is increased by degeneration; try bypassing the cathode resistor.

Andreas

[Sherman]

Originally posted by dsavitsk
Hi,

...Anyhow, the question is that the bass response is a bit weak. Sound otherwise is great...

... though[t] 6922s sounded about the same?

Don't know if it is related to your issue but I have a KT88 SE amp (monoblocks actually) that use a 6N1P as the input/driver. When I changed that for a 6922 the amps definitely became bass-shy. Listening tests with other people confirmed it.

If you have a 6N1P laying around try plugging it in and see if it changes things. (Be careful though, heater current on the 6922 is 300mA and the 6N1P is 600mA. Make sure your heater secondary can supply the needed current.)

[dsavitsk]

Originally posted by Geek
Hi,

Are you using a SE (gapped core) transformer?

If not, the core is saturating.

If so, the primary inductance may not be sufficient for the plate Z, or Z ratio.

SE Trafo, and I am running tham at about 20mA while the transformers can take up to 65mA or so.

Originally posted by AndreasS
The plate resistance of the tube is increased by degeneration; try bypassing the cathode resistor.

I figured that since the plate resistance of these tubes is low enough, especially in parallel, that not bypassing would not be a huge deal since the plate load is 10K. I did try LED biasing, and while the bass is a bit better, the midrange is not as good -- I am not sure I like this LED bias thing so much, sounds a little flat and hard to me. I'll try some capacitor bypasses tonight, but I was hoping to keep them out of the signal path.

Originally posted by Sherman
If you have a 6N1P laying around try plugging it in and see if it changes things. (Be careful though, heater current on the 6922 is 300mA and the 6N1P is 600mA. Make sure your heater secondary can supply the needed current.)

No problem on the heaters, but no real change either.

[Dave Cigna]

According to theory, when you 'ratio' a transformer it shifts the frequency response. For example, if you were to take a transformer designed for 5k:8 and instead used it as 20k:32 then the frequency response would be shifted up. Both the low end and high end cutoff would go up; you would lose low frequency response in exchange for high frequency response.

I understand that with the parallel resistor you are not really doing that, so you shouldn't be losing bass this way... according to the theory. I do believe in theory, but I also believe in the tremendous amount of unknowns we're dealing with.

Anyway, the lack of bass could be due to a lot of things, but my thought is that it would be pretty easy for you to experiment with different parallel resistors just to see what happens.

We could also talk about taking the transformers apart and experimenting with the gap. It's not as difficult or scary as you might think. If the OT's were designed to handle more DC idle current than you're putting through them then a smaller gap might bring out more bass. Finding the balance that you like is largely a matter of experimentation.

-- Dave

[dsavitsk]

The lack of cathode bypasses does seem to be a big part of it. I added 470uF BG caps bypassed by 0.1uF Vit Q's and the bass is back. However, I don't think the midrange is nearly as good. It is just a little sloppy-er sounding. I think I'll try ultrapath bypasses next to see how that does.

As for regapping the transformers, they are potted in epoxy, so I am guessing that it is not going to work. However, I am going to order some electra-prints to upgrade a bit so perhaps I'll talk to Jack about optimum gapping.

-d

[Zibi]

http://www.aikenamps.com/CommonCathode.htm
http://www.turneraudio.com.au/education+diy.html

[Zibi]

33R resistor in parallel to the headphones = 50% output power loss. I think, that Ra=10k is no proper for two parallel connected triodes, rather 3.5k or 2.5k .
http://www.shinjo.info/frank/sheets/113/6/6N6P.pdf

[AndreasS]

Originally posted by Zibi
33R resistor in parallel to the headphones = 50% output power loss. I think, that Ra=10k is no proper for two parallel connected triodes, rather 3.5k or 2.5k .
...

The output power isn't important for a headphone amp; but a higher load decreases the distorsions (btw. the transformer is 5 kohm to 8 ohm).

Andreas