Sovtek 6B4G Evaluation
Chris Brady
base and 6.3v filament) I have been anxious to try out
the new single plate Sovtek 6B4G's. I got a pair of
the Sovteks from Welborn Labs this week and found some
interesting results.
For the much anticipated listening session I had three
pairs of 6B4G's on hand to compare, Tung-Sol's, Sylvania
and of course the new Sovtek's. Everyone that listened
clearly preferred the sound of the Sovtek's. I had planned
to switch tubes from time to time during the initial
listening session. However, the Sovtek's sounded sooo good
(relative to the Tung-Sol's that I normally listen to)
that the other tubes never made it into the amps. The
Sovtek's had a more detailed, life-like presentation. Also
the bass was a little stronger and better controlled. I
am not sure if it is the single plate construction or some
other factor that makes the difference but I sure liked
what I was hearing.
However, there is a strange, unexplained twist. After the
initial listening session I did a few bench measurements.
To my surprise the Sovtek's generate piles of harmonic
distortion. Thinking there must be something wrong with
my test setup I plugged the Tung-Sol's back in a reran the
tests. Here are the results for 1khz HD at different power
levels:
Tung-Sol Sovtek
0.5w .11% 1.1%
1.0w .33% 1.8%
2.0w .58% 2.6%
3.0w .90% 3.7%
Yikes, the numbers for the Sovtek's are lousy! Three to
ten times as much distortion as the Tung-Sol's?? At this
point I took a close look at the operating points for the
Sovteks. I am running SE with cathode bias, a B+ of 340v
and a 1k cathode resistor. This results in 51ma of plate
current. The operating points were almost identical for
the Sovtek's and the Tung-Sol's.
The distortion results cast doubt on the validity of
the previous listening session. Possibly tainted by high
expectations? So we did another listening session and this
time listened to all three tubes types. The results were
clear and unanimous, the Sovteks still sounded the best.
What gives? I never have thought that HD measurements
correlated that well with sound. But 3+ percent?? Is it
euphonic distortion that we all seem to like? I don't
think so. I know what 2 order distortion sounds like and
I certainly would have guessed that the Sovtek's had lower
distortion. I am missing something here?
To sum up, if you want an amp that measures well then NOS
6B4G's are the way to go. However, if you want an amp that
sounds good you may want to try the Sovtek's.
Matt Warren
Matt
Chris Brady says...
Chris Brady
plate resistor, biased at about 7.5ma via a 510 ohm cathode resistor.
The 6SN7 is cap coupled to the 6B4Gs. The gain is on the low side
but just enough for my setup.
I was also wondering about the distortion spectrum. Unfortunately
my distortion analyzer only reports total distortion. I was thinking
about trying some PC based spectrum analyzer software that uses a
sound card to do some measurements. However, the quality of the
soundcard is probably insufficient to accurate results. Anyone
out there tried a PC based spectrum analyzer?
Chris
Mr. James C. Mcshane wrote:
>
> One possibiltiy that strikes me is distortion cancellation when the
> tubes are installed in your amp. I'd be interested to know how you were
> driving the Sovteks. I'd also be interested in the distortion spectrum
> generated by the Sovtek vs the NOS tubes.
>
> Thanks for the nice work.
>
> Jim McShane
> MLJ...@prodigy.com
Chris Brady
the same as a 2A3 except for the base and filament voltage. But
the going price for NOS 6B4G's is only about $35 each. I got my
Sovteks from Welborn Labs for even less, $50 a pair.
The Sovtek build quality is only fair. The glass to base
joint is lousy. The inside components look rugged but not
as elegant as the NOS tubes. As for reliability, time will tell.
But given the great sound and low price I could live with less than
stellar reliability.
Chris
>
> Considering I can usually find NOS 2A3s for 75.-100 USD per pair if I keep a
> sharp eye open, how is the build on these tubes?
> Mark S.
Mr. James C. Mcshane
TROOBKA
>I am driving the 6B4Gs with a single section of a 6SN7 with a 28k
>plate resistor, biased at about 7.5ma via a 510 ohm cathode resistor.
>The 6SN7 is cap coupled to the 6B4Gs. The gain is on the low side
>but just enough for my setup.
>
>I was also wondering about the distortion spectrum. Unfortunately
>my distortion analyzer only reports total distortion. I was thinking
>about trying some PC based spectrum analyzer software that uses a
>sound card to do some measurements. However, the quality of the
>soundcard is probably insufficient to accurate results. Anyone
>out there tried a PC based spectrum analyzer?
>
>Chris
>
>Mr. James C. Mcshane wrote:
>>
Bob Fitzgerald
Chris Brady wrote in message <35F1F584...@cray.com>...
>I was also wondering about the distortion spectrum. Unfortunately
>my distortion analyzer only reports total distortion. I was thinking
>about trying some PC based spectrum analyzer software that uses a
>sound card to do some measurements. However, the quality of the
>soundcard is probably insufficient to accurate results. Anyone
>out there tried a PC based spectrum analyzer?
>
>Chris
Chris - I was wondering as well when I saw your first post. If you are
interested, I may be able to help you get something going to measure
distortion - note, for the levels you're talking about, 1% or so, a sound
card is a precision swiss instrument!
Where it falls down is at .01% type stuff.
Bob
Chris Brady
I loaded up a PC based spectrum analyzer and have some distortion
spectrum measurements for the Sovteks.
I tried using "Spectra Plus" but their 30 free license code
apparently is faulty. I ended up using "HpW Works Lite" and
it works quite well. I have a cheapo 16bit sound card and
got meaningful results down to about .02%. Much better than
expected.
Back to the original topic, here are the 1khz distortion spectrum
results for the Sovteks and Tung-Sols:
Order Sovtek Tung-Sol Ratio
-------------------------------------
2nd 1.716 0.532 3.281
3rd 0.480 0.177 2.712
4th 0.104 0.035 2.971
5th 0.012 0.008 1.500
6th 0.008 0.007 1.143
It's clear that the distortion for the Sovteks is always much higher
than the Tung-Sols. There could be something interesting going on
with high order (6th +) that we can't see due to noise, but it seems
unlikely that it would be audible.
I sure don't see anything in these results that would explain why the
Sovteks sound better. The mystery continues....
Chris
iga
In Russia this tube ( 6C4C ) is out of production for many years, aprox.
from 1985. I have Svetlana made 6C4C and they looks and work fine.
Bad glass to base connection was typical for Novosibirsk plant ( small
diamond logo ), but AFAIK they never made this tube.
Please, describe your tubes with more details - base height, number of
pins, any stamp or mark on glass etc...
Best regards,
Igor.
Chris Brady escribió en mensaje <35F216B4...@hobbes.denver.sgi.com>...
>One of the things I like about 6B4Gs is their price. They are
>the same as a 2A3 except for the base and filament voltage. But
>the going price for NOS 6B4G's is only about $35 each. I got my
>Sovteks from Welborn Labs for even less, $50 a pair.
>
>The Sovtek build quality is only fair. The glass to base
>joint is lousy. The inside components look rugged but not
>as elegant as the NOS tubes. As for reliability, time will tell.
>But given the great sound and low price I could live with less than
>stellar reliability.
>
>Chris
>>
Chris Brady
The base material looks like black phonemic and is
about 2cm high. The glass has a prominent red logo that
looks like a star/bird and says "SOVTEK Made in Russia".
The inside of the bottom 2cm of glass is silvered from
the getters.
I hope this helps,
Chris
Eric Barbour
> wrote:
> > Tung-Sol Sovtek
> >0.5w .11% 1.1%
> >1.0w .33% 1.8%
> >2.0w .58% 2.6%
> >3.0w .90% 3.7%
These figures correspond to my own results. The Sovteks
have 4 to 8 times the distortion of NOS 6B4Gs.
Not a favorable indication of construction quality......
Acrosound
Eric Barbour wrote:
>These figures correspond to my own results. The Sovteks
>have 4 to 8 times the distortion of NOS 6B4Gs.
>Not a favorable indication of construction quality......
Hey Eric. Hello. Any idea of why the distortion might be so high? Are the
electrical parameters (mu, AC plate resistance, and etc) the same as the NOS
units or is this a case of creative numbering, i.e., picking something close
and putting a well known tube type number on it?
Would be nice to have a relatively small triode that used a 6.3 volt heater and
sat in an octal socket. The 6B4G tube really looks like a great tube on paper.
Has Svetlana looked at this tube type?
BTW... your 300B tube is quite, quite nice. Don't yet have many hours on them
but this really is a great sounding tube. Good job.
Mike LaFevre