|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
79.49.212.251
Sometimes a new asylum post makes you wonder about a novel 'what if'.
A recent series of comments on IMD has triggered my interest in 'what if I replaced (mutatis mutandis) one EL84 with two 6SN7 tubes in the same circuit?'.
We are talking about a class A 4-5W amplifier.
In sim world I found out that this all 6SN7 circuit performs much better than any EL84 circuit I have ever simulated.
How much better?
The decision is up to you.
You can trade in much lower IMD in exchange for a higher gain.
Schematic, output signal graph, output currents and the spectra (between 9 and 10 seconds, between 10 and 11 seconds, between 11 and 12 seconds speak by themselves.
Omnes feriunt, ultima necat.
Follow Ups:
What are you going to use to measure IMD in your prototype amp?
Have you set up an IMD test circuit to valve-swap and order the performance of the valves you have?
I have only tested a batch of 6SN7 for HD levels and gain balance, and that certainly showed quite a variation between samples.
I plan to use a set up made of some resistors, plus my PC (with FFT/IMD calculation software) and a couple of audio tones.
Omnes feriunt, ultima necat.
Are you going to test each sample tube separately, or tube roll the tubes when the amp is up and running?
Do you have a way of changing your circuit simulation triode model to make it compare well with your real measured performance?
I only plan to insert one at the time (and eventually swap) the 8 output 6H8C tubes between channels to get a good current balance of each 4 tubes sets before bias adjustment.
Omnes feriunt, ultima necat.
Work in progress.
Star grounding scheme, short cables to the loudness switch and sensitive loudness circuitry housed in a shielded box will hopefully reduce noise.
Plywood and aluminium foil are much easier to work than metal.
Omnes feriunt, ultima necat.
So far, so good.
Omnes feriunt, ultima necat.
Start of final integration.
Omnes feriunt, ultima necat.
What is the primary impedance of the output transformer that you are using?
2X110 V, 2X6V, greater than 30 VA toroid.
One of the 6V outputs feeds the headphone with a 33 Ohm in series.
Omnes feriunt, ultima necat.
So far so good.
Omnes feriunt, ultima necat.
DT667 tried the 6SN7 driving parallel SE 6SN7.
Not sure if this idea will be revisited or not.
DT667
Why would you put yourself in the truly painful position of paying that kind of $$$ for "good-enough" 6SN7 tubes (there aren't many IME).
Tungsols OK but I prefer CV181-T.
seems like a pair of CV181-T runs North of $200. I suppose that's not too bad in the world of hyper-priced 6SN7's, but having tried things like Sylvania metal base 6SN7W, Tung Sol roundplates, National Union grey glass, it's difficult not to keep thinking about what the proposed 6sn7 amplifier would sound like running THOSE tubes...
I am so tempted to build it with the tone/loudness control depicted in the 'Universal tone control for DIY' of a previous post, just breadboarded.
The loudness section can use either 12AX7 or 12AT7 (with marginal gain loss).
Omnes feriunt, ultima necat.
It's pretty easy to add a 12V/6V slide switch to a design, and I have dozens of NOS 12SN7s, including chrome dome Sylvanias that offer truly superb performance.
--------------------------
Buy Chinese. Bury freedom.
.
Have Fun and Enjoy the Music
"Still Working the Problem"
My system no longer uses any 6sn7s. I guess it's time to start selling?
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
I've got new production Tung Sol tubes in my preamp, and it sounds as good as with the NOS RCA's they replaced, which were $$$. Didn't last as long, but then again were less than half the price.
I have many russian 6SN7 tubes (still available in quantity in Europe).
Those coming from Ucraine are easily found on ebay.
Omnes feriunt, ultima necat.
Wow... that is an interesting data sheet.
The quirk that jumped out at me is with regard to the Ri, Gm and µ graph. From the numbers below and using the formula Ri=µ/Gm I get an Ri of 7166Ω but when you look at the plots of these values vs current things don't add up.
First the Ri label only goes to 3KΩ and if you take the value @ 10ma you get 750Ω. My first thought was that they simply dropped a zero from the scale and the Ri is really 7500Ω which may be close enough for government work, but not what I would expect for government military work. The published values for S and µ do have the ± designation for some wiggle room so my next thought was this is not a plot for a bogey tube. The next problem arose when I applied the Ri=µ/S formula to the 10ma values and I get 7500≠24.6/2.8 which has me scratching my head. The shape of the Ri, S & µ plots follow the pattern of the 6SN7 but the numbers sure don't add up. It almost looks as if the Ri axis is mislabeled and the two of the values were from 1/2 of a non-bogey tube and the third from the other half.
dave
Those tubes were top secret technology. The USSR published nonsensical data sheets to confuse Western scientists.
--------------------------
Buy Chinese. Bury freedom.
Are you comparing a pentode to parallel triodes? That wouldn't be much of a contest. The difference should be less significant using a 6V6 in triode mode.
--------------------------
Buy Chinese. Bury freedom.
It was driving Magneplanar SMGs. Sounded great. It's a tempting build, and most of us will have 6SN7s hanging around.
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: