Dealing with Mr. Miller (phono)

[mashaffer]

The problem of miller capacitance with high mu triode input for phono amps is well known. One solution of course is to use a lower mu tube such as 12AY7, 6N1P or such. Of course this reduces overall gain and possibly increases the overall noise of the circuit.

What I would like to kick around is the possibility of buffering the input in such a way as to isolate the cartridge from the Cm of the first gain stage. Would putting a CF or SF in front of the first gain stage do more harm than good?

Silly idea?

[SY]

In a word: noise.

A phono stage input is the ideal place to use a cascode, and I'm not the only one who thinks that a FET does very well in that spot.

[Geek]

Cascode ECC88 does as spectacular as any tube can in that position. For better performance, yeah, youll need a jfet or the like on the bottom.

Cheers!

[mashaffer]

Indeed noise was my main concern. So a medium mu triode in traditional circuit would be better than trying to buffer. The mention of the cascode reminded me of this article

http://www.hawthorneaudio.com/photos...evisited~0.pdf

Which examines cascodes at some length. I would be interested in your opinions of the last circuit in his write up which he refers to as the ultra linear cascode. That one looks like a really good candidate.

I have some high gm tubes on hand but most are rf tubes which have a reputation for microphonics. Too bad the AX7 has such poor gm as I have lots of those. Maybe I need to use the RF tubes for the CF in my crossover so that I can free up the 6N1Ps for phono duty.

What would you think of 6N1P cascode input stage feeding a 12AX7? Another possibility (if my 6SN7s turn out to be good) is the russian tubes as a cascode input feeding the SN7. Since the cascode would give very high gain is it possible that the SN7 would have enough gain to finish the job?

Had hoped to knock up something with a pair of AX7s as a starting point but since I am using a Shure M92 I am inclined to think that it is not worth the trouble. I also worry that a CC 6N1P followed by 12AX7 would be a bit short on gain. I had hoped to use an existing chassis with two 9 pin sockets but if I am going to have to cut more holes any way the a three stage is not out of the question either.

So here is a concrete question. You have on hand the following.

- A tranny that can give about 300V with a doubler supply.
- 2 6N1P
- 2 or 3 12AX7s
- 2 6AU6
- 3 6CB6
- 3 6EM7
- A chassis with two 9 pin sockets installed but a little bit of room for some more (output trannies and can cap removed).

How would you proceed? I had even considered using the 6EM7 for the last two stages. Plenty of output drive.

[SY]

I commend Eddie for his energy and enthusiasm, though I take some issue with his model. I would still use a more conventional cascode, paying attention to power supply issues. For the theory, you're better off reading Morgan Jones. For the practicalities of cascodes, Allen Wright's little book can be very useful.

[kenpeter]

Another method that comes to mind is Neutralization.

Take an ordinary differential stage, and cross couple
two small caps from the plates back to opposing grids.
If your chosen caps equal miller, bad mojo cancels out.
In theory... I've never tried it.

[Eli Duttman]

Guys,

There's a MUCH simpler method for dealing with Miller capacitance at the preamp's I/P. Use the 6GK5 in the 1st gain block! Talk about a near perfect fit for a situation. The 6GK5 is high mu, low noise, and low Miller capacitance. Real world stage gain, with resistive loading, is greater than that obtained with a 'X7 section. The I_B the 6GK5 operates at is sufficient to ensure that LED bias is reliable. Bye, bye, problematic cathode resistor bypass capacitor.

[dsavitsk]

Originally posted by Eli Duttman
The 6GK5 is high mu, low noise, and low Miller capacitance.

I am using this tube at your suggestion for the first stage of my phono amp. It may just be the examples I have on hand, but I needed to build little aluminum foil shields around each tube to get rid of hum and noise.

No comments yet on overall sound as my filter seems to be off a little -- treble is totally rolled off -- and I am not sure what the cause of that is yet.

FWIW, I used a CCS load at 12mA. Bias is via a 22R resistor -- I figured the value was low enough to not increase rp too much, and high enough for enough bias. An LED would be too high I think. Even biased to a quarter volt, the plate is at nearly 110V which is a bit higher than the datasheet would suggest.

[Eli Duttman]

IMO, shielding phono stage tubes should be a matter of routine.

[kmaier]

Another idea that some audio buddies and I kicked around many years ago (long before the internet) was to use a medium mu tube, like a 6072 or 12AY7 configured as a typical common cathode amplifier. Then splitting the RIAA EQ so the first gain stage drives a passive 75-microsecond de-emphasis to the following stage. This would help to lower the noise. The rest of the design could be implemented using either a passive or active circuitry to handle the 318- and 3180-microsecond turnover points. Unfortunately none of us ever got around to finishing the design much less building one. Still, another idea to toss in.

Regards, KM