Hi all,
I have just converted my 12au7 preamp from SRPP to Mu-follower topology and love it. Besides the new sound I like, also higher gain as well.
Now I have noticed one slight problem, L/R gain imbalance. Although I used two 12au7, they are not same brand. Therefore, in order to achieve best balance, I used one tube as the top part and another tube as the bottom part. The two tubes also have different gains as well.
I swapped the top and bottom tubes and I got a lower overall gain but a balanced L/R output. I know the bottom tube determines the gain and the top is basically a cathode follower. Therefore, I concluded one tube is imbalanced inside the same envelope - one triode gives higher gain then the other inside the same envelope!
Is this common ? I am relatively new to tubes but these 12au7s are old and the prints are coming off and hard to read. But they do sound good to me (ok, I know you guys would say get 6sn7 but that's all I have to play with for now).
Thanks.
I have just converted my 12au7 preamp from SRPP to Mu-follower topology and love it. Besides the new sound I like, also higher gain as well.
Now I have noticed one slight problem, L/R gain imbalance. Although I used two 12au7, they are not same brand. Therefore, in order to achieve best balance, I used one tube as the top part and another tube as the bottom part. The two tubes also have different gains as well.
I swapped the top and bottom tubes and I got a lower overall gain but a balanced L/R output. I know the bottom tube determines the gain and the top is basically a cathode follower. Therefore, I concluded one tube is imbalanced inside the same envelope - one triode gives higher gain then the other inside the same envelope!
Is this common ? I am relatively new to tubes but these 12au7s are old and the prints are coming off and hard to read. But they do sound good to me (ok, I know you guys would say get 6sn7 but that's all I have to play with for now).
Thanks.
Tubes are mechanical creatures, assembled by human hands, even at the best of days they will easily have +-30% differences from tube to tube ( or from
triode to triode in the same envelope as an ecc82)
Designers of the past know this and made circuits that did not depend ( as much) and usually tolerated randomly selected tubes.
( hint : feedback)
When proper design did not help tubes had to be selected, some did this
in-house and usually marked the selected tubes in some way ( "red tip tube"),
The rest of the tubes were sold as new, which they were. Some of these
are still on the market today and sold as NOS.
You have two roads here : alter your circuit to accept tubes with wide
tolerances or select tubes ( or pay someone to select, and you have to be clear
what grounds the selections should be done).
triode to triode in the same envelope as an ecc82)
Designers of the past know this and made circuits that did not depend ( as much) and usually tolerated randomly selected tubes.
( hint : feedback)
When proper design did not help tubes had to be selected, some did this
in-house and usually marked the selected tubes in some way ( "red tip tube"),
The rest of the tubes were sold as new, which they were. Some of these
are still on the market today and sold as NOS.
You have two roads here : alter your circuit to accept tubes with wide
tolerances or select tubes ( or pay someone to select, and you have to be clear
what grounds the selections should be done).
Last edited by a moderator:
Thanks for insight. I knew there are certainly differences triode to triode in the same envelope but did not expect such a big difference in the tube I'm using.
I usually measure and match the resistors I use between L/R channels so I will get the best possible match. I don't think there's anything in the schematic I want to alter, especially its a very standard Mu-follower circuit.
I think I'm down to these two - 1) Find more 12au7 and swap them till I find the one I like in terms of sound and balance, or 2) add a balance pot (not what I prefer but can be a quick fix till I find the right tube).
I usually measure and match the resistors I use between L/R channels so I will get the best possible match. I don't think there's anything in the schematic I want to alter, especially its a very standard Mu-follower circuit.
I think I'm down to these two - 1) Find more 12au7 and swap them till I find the one I like in terms of sound and balance, or 2) add a balance pot (not what I prefer but can be a quick fix till I find the right tube).
The gain in an SRPP (and even more so in a mu follower) is set largely by mu, not gm. Mu tends to vary less among samples of a particular valve than gm. However, two different designs of, say, 12AU7 could have somewhat different mu.
I have seen rumours that RFT twin triodes are often particularly well matched. You can find these labelled as RFT, but also under other NOS brands such as 'Mazda foreign' (often with blue print on the glass). Note sure how many of these made it across the Atlantic, though.
I have seen rumours that RFT twin triodes are often particularly well matched. You can find these labelled as RFT, but also under other NOS brands such as 'Mazda foreign' (often with blue print on the glass). Note sure how many of these made it across the Atlantic, though.
Some vendor select -Gm- "matched" tube pairs near/within 1%.
For example:
2 NOS RCA Conn 12AU7A Clear Top Audio Tubes U.S.A. 1960's | eBay
If Gm is almost equal between two halves, the inner structure also very similar, so tube halves output impedance (Rp) is similar.
Mu=Gm*Rp
Almost equal Gm probably means almost equal Mu values.
/ BTW RCA clear top is one of the most "musical" 12AU7 tube. /
For example:
2 NOS RCA Conn 12AU7A Clear Top Audio Tubes U.S.A. 1960's | eBay
If Gm is almost equal between two halves, the inner structure also very similar, so tube halves output impedance (Rp) is similar.
Mu=Gm*Rp
Almost equal Gm probably means almost equal Mu values.
/ BTW RCA clear top is one of the most "musical" 12AU7 tube. /
Pics of my tubes....
Thanks for all the info. Here's a pic of all my 12AU7.
The left two are being used in the mu-follower. The one on the far left is the top tube, second from the left is the lower tube (the suspect that gives imbalance but higher gain). I could hardly read what's print on it. I am not even able to tell what brand it is.
The two on the right side are used on another preamp I built - two stage direct coupling with second stage as a cathode follower. My idea was to minimize usage of coupling capacitor. The result was so-so, clean sound as a preamp but not too much like a tube amp. I suspect this may be the Sylvania tubes ? (not too good).
Is current production also a good option such as JJ tubes ?
Thanks for all the info. Here's a pic of all my 12AU7.
The left two are being used in the mu-follower. The one on the far left is the top tube, second from the left is the lower tube (the suspect that gives imbalance but higher gain). I could hardly read what's print on it. I am not even able to tell what brand it is.
The two on the right side are used on another preamp I built - two stage direct coupling with second stage as a cathode follower. My idea was to minimize usage of coupling capacitor. The result was so-so, clean sound as a preamp but not too much like a tube amp. I suspect this may be the Sylvania tubes ? (not too good).
Is current production also a good option such as JJ tubes ?
Attachments
I would not assume that. Different things set mu and gm, so you can certainly have similar mu with very different gm. I would not be too surprised to see similar gm and different mu. Mu is set largely by geometry, while gm comes from geometry and other things such as cathode emissivity.euro21 said:
clean is what you aim for in that design. but if you want less of that, use resistors other than metal films. And no metalized polypropylene caps of any brand. tube changes are typically a 5% change. But if you aim for the circuit to have a certain character, it doesn't matter what tube gets rolled in.
That is why nos tubes are not a big deal to me.
JJ tubes are ok, never had issues with them.
Last edited:
Here is an example of a datasheet which calls out initial minimum and maximum values for GM, Mu, and more.
http://www.sphere.bc.ca/test/tubes16/ge5670-2c51.pdf
Look at sheet 3.
GM 4500 to 6500
Mu 26 to 44
Plate current 5.9 to 10.9mA
http://www.sphere.bc.ca/test/tubes16/ge5670-2c51.pdf
Look at sheet 3.
GM 4500 to 6500
Mu 26 to 44
Plate current 5.9 to 10.9mA
Also note that this is a "special quality tube" where parameter spread is lower then in "run-of-the-mill" types.
Another possibility, without other info, is that the higher gain circuit has exposed greater imbalance in the volume control. They're always somewhat imbalanced, and worse at "lower" settings. A good test would be to *write down* which channel is stronger, then swap signal paths left to right (through the active stage), then re-check.
All good fortune,
Chris
All good fortune,
Chris
That's the same thing I have been thinking also since I posted. I have two volume pots. One in the preamp and one in the power amp. With the higher gain mu-follower, I have to dial down on both to around 9am position. Anything position higher than that will produce a volume that makes the wife flip the main circuit breaker.
I will do some testing and post result.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.