Which DHT for Line stage?

[andyjevans]

My current line stage is 1626 in differential pair with a CCS under it. I like this quite a bit but am keen to go one step higher in terms of sound. I plan to order some DHTs but since I need two matched pairs I want to get a good idea of what to order first. I'd like to keep the mu and plate impedence (Ra) lowish in case I go PP with an output transformer.
Seems like 31, 46 (in triode) and 71A are suitable
12A and 26 and 843 (IDHT) are close
01A and 30 is just possible (1LE3 is getting too high mu/Ra)
I'd like help choosing from you guys that have heard these or used them. By all means suggest any others (though not stupidly expensive - I need a few)
I'd also like some idea of which are low on hum and which require special measures so I know what to expect. Bear in mind it's a balanced line stage, if that helps with the hum at all.
Thanks!!! Andy

[Cassiel]

This guy knows something about DHT's:

http://www.vt52.com/

[Rocky]

71A in a linestage does require some current from the stage before (mr.miller), you will have to check the driving component is upto the task. Also, gain will be very low, but that can be a good thing as well. I am actually building an amp these days with the 71A as output tube.

12A is very close to the 26 but has conciderably higher Rp than the 71A. It has however a good advantage vs. the 26 what Rp regards. About 2/3. Also 12A can take a little more current, and has slightly higher mu than the 26, and much friendlier filaments 5V/0.25A vs 1.5V/1A. The 26 is known for notorious hum and microphonics, but once you avoid those issues, many people claim it's the very best sounding tube ever made (a differential pair will probably minimize the hum issue, but remember it's difficult to work with microphonic tubes when gain in the following stages are high).

I would think the 01A and 30 has way too lowish current to be useful in a line-stage, they also has high Rp.. The 31 looks very nice to me, and it is specially made to be operated by batteries again, if the miller doesn't take your HF away... you will have to make sure your source can deliver the current you need to drive your tube choice.

[kevinkr]

I like and use the 26 in my transformer coupled pre-amplifier, microphonics and hum are fairly easily dealt with.
See my article in the positive feedback online magazine archive or go to my site at www.kta-hifi.net for a link to that article.

IMHO going with a single triode in SE with transformer coupling or some variation of parafeed seems to be the way to go. Simplicity really has its charms and the 26 or 01 are both so linear that the benefits of differential operation are doubtful in my mind anyway.

Kevin
www.kta-hifi.net

Edited for goofy line spacing

[dhaen]

71A in a linestage does require some current from the stage before (mr.miller)

Since you are using balanced toplogy, you can cancel Miller by cross-coupling a few pf's (the calculated Miller value) from the other side's anodes. 2 caps and Mr Miller is dead.

[Rocky]

Originally posted by dhaen

Since you are using balanced toplogy, you can cancel Miller by cross-coupling a few pf's (the calculated Miller value) from the other side's anodes. 2 caps and Mr Miller is dead.

Very clever.

Thinking a little on the matter, if I were to build a balanced DHT linestage, I'd be curious to try the 31. But then again, I'm not going to build a balanced DHT linestage. I might try it as output tube in single ended though

[andyjevans]

Since you are using balanced toplogy, you can cancel Miller by cross-coupling a few pf's (the calculated Miller value) from the other side's anodes. 2 caps and Mr Miller is dead>

Could you explain a little for my benefit - I haven't tried this. What value caps, and where exactly between the anodes? I ask because this may be happening with my present 1626 line stage, i.e. loss of treble. the input is directly from the DAC which has a differential output of about 1.5v no doubt. It also has about 2v DC on the outputs, not that this probably matters. would we expect some reaction with Miller here? Very helpful! Let me know. Andy

[dhaen]

Clever but not new. It's used widely in RF where miller C is a significant factor, which is where I learned it.
A quick Google shows that WE used it in their '92 in the 30's. Lynn Olson has a copy on his page here (image linked too) with some discussion on how he couldn't get it to work in his transformer coupled circuit due to ringing. That does indeed seem a tough task. But IMO making it work in a stage with an RC coupled input should be a piece of psss as long as the output Z is significantly lower than the input Z. But hey, that's what line stages do, isn't it?
The value of the caps should be equal to the miller capacitance for that stage element: Cm=Cga(V+1).

[andyjevans]

thanks for the link to the WE design. The input to my stage, however, isn't RC coupled. It comes straight out of the DAC as a differential signal with 2v DC on it, which gets absorbed in the bias of the stage. The coupling cap and stepped attenuator comes after the tube stage. Does this make it difficult?

I've been looking at some of the options for DHTs in this line stage again. I haven't even considered the Sovtek 2a3, but I don't see why not, since it's a currently available tube and not too expensive. Maybe this is another option.
71A seems a good choice, but prices have really rocketed
46 seems another good choice, not cheap though reasonable and not as bad as 10Y which is a great option. I need two matched pairs, and I can see my wallet suffering here.
12a is more expensive than 71a and I'm not sure if it would be a better choice - maybe has the edge on sound
843 has no hum problems being IDHT - not much data on users of this except it's a good tube. Not sure how good.
31 by all accounts is a lovely tube, but a nightmare for microphonics
That leaves the 26 - still reasonably priced though we're now up at 7k Rp - marginal for PP.
27 and 37 are higher Rp
01A has higher Rp again and is quite expensive, so we're drifting out of the Rp range where we could use a PP output transformer.
1LE3 is cheap but we're really cranking up the Rp here

The 71a is tempting on all accounts - maybe why prices have risen, but also this is needed for old radio sets.
46 and 12a would be good choices
843 is interesting - need more data
31 and 26 would be a real challenge hum and noise-wise, though cheaper than the others.
No obvious free lunch here. One would have to have thought of all this a couple of years ago when 71a could be had for under $10. No longer the case, sadly.
So back to investigating the 2a3 as a line stage - anyone tried this? Andy

[dhaen]

thanks for the link to the WE design. The input to my stage, however, isn't RC coupled. It comes straight out of the DAC as a differential signal with 2v DC on it, which gets absorbed in the bias of the stage. The coupling cap and stepped attenuator comes after the tube stage. Does this make it difficult?

No, I don't think so. The complication for Lynn (but not for WE) was the input transformer ringing.
Whatever is "hanging" on the grids will be influenced equally by the signals from miller c and -miller c, which should cancel.