|
Home
/ FAQ
/ News Classifieds / Events |
Audio Asylum Thread Printer |
Get a view of an entire thread on one page |
71.32.60.237
In Reply to: RE: NOS tubes posted by cityjim on May 24, 2009 at 23:15:14
The post made me go what?
Maybe I better go check the Eimac 8877s I bought in the 80s and 90s to make sure they are real mil surplus?
"He (R.M. Nixon) was a foul caricature of himself, a man with no soul, no inner convictions, with the integrity of a hyena, and the style of a poison toad." H. S. Thompson
I inherently distrust threads started by people who are not registered asylum users, unless the thread is in the form of a "please help me with X". Any thread started by a non member that is of this nature usually has some sort of agenda. I'm not saying that this is the case with this thread, but more often than not, it is.
That said, some tubes obviously are NOS - still in the original box with absolutely no signs of use. They can be found, though I agree with popular tube types they are getting rarer, and I also agree there are a lot more people selling NOS tubes (on EBay) that likely are not.
Rob"Let there be songs, to fill the air"
I purchased some tubes on Ebay that were advertised as NOS. I tested them on my Amplitrex tube tester and found they at 60-70% of what they should be to be new.
I love the Amplitrex: With it, the BS ends here.
Donald North
The problem with ANY tube tester is the fact that tube characteristics vary considerably from tube to tube - even the same brands/types/etc.Tubes follow the typical "bell curve" distribution model. While many tubes are near the center point and close to published spec, a good number fall at the edges of the "bell".
Finding tubes at 60 or 70% (% of what - I'm not sure what the poster meant) means only that they are, well, at 60 or 70%! They may be new, unused tubes that simply fall at the low end of the tolerance range. No tube tester, Amplitrex, Hickok, AVO, or ??, can confirm if they are new or not - only how they test.
Edits: 05/31/09
Having just tested about 500 known NOS 6SN7 series using a HICKOK 1575B (Mil version of the 539B/C), some interesting results came to light -
A - Transconductance generally ranged from about 2200 uMhos to
3000 uMhos, with the center about 2500-2600 uMhos-
data minimum being 1700 uMhos per roll chart.
B - Among the 6SN7WGT/WGTA types, the older tubes ( 1950-1965)
were consistently higher than the late 1970-1980's MIL stuff.
C - Test data quite dependent on the AC voltmeter setting, as
line voltage changes caused a shift in Gm and Ip readings.
This is probably linked to changes in filament voltage
and plate voltage during test (triodes sensitive to plate V)
D - Even when the tube sections Gm matched, plate current varied
by as much as 1 ma(generally closer to 0.4 Ma or less). Typical
plate currents were in the range of 5+ mA to about 8 mA at
the indicated 4.5 V grid bias point.
The point being that just a Gm match is close, but not perfect.
E - Jim McS is correct in that a tester can only do so much.
A few obviously used tubes were also tested, with many at
the low end but with some testing at the high end even with
significant visual indications of use present.
TUBOS
For audio use, I like a tube tester that measure small signal tubes in db gain. It puts things into perspective.
One fellow wanted to buy some TFKN ECC88s. I made a nice matched pair of what I believe are/were mil surplus that I bought from Billington's a few years ago. They measured in the normal range for the hundreds of NOS ECC88s I've tested.
He called me complaining that they only measured 70% on his tester. When I asked how he converted his test results into dbs there was silence. I discovered that he defined checks as NOS as 100% on his tester. That's a good way to buy champion tubes, but it gives a lot of false negatives. His view was 30% of the tube's life was gone.
Since it was a small sale, and I knew that if I wanted to sell the tubes for more I could (he bought some 12AX7s and E88CCS also), I simply had him send them back for a refund.
I still have those tubes. TFKN <> E88CC and ECC88s, which I use in my gear; they are worth more, in my hand, than buying some in the future.
When I went to college and studied Marketing Research a little, I read a couple of papers that proved a Sigmoid response to risk under certainty. The curve was more pronounced under uncertainty.
Getting HIGH prices for tubes that can be measured, even when the measurement is not related to how the tube is used, is a good example of how people are willing to spend a lot more to avoid the possibility of buying an under performing tube.
After a while its easy to spot NOS tubes that aren't from the past but are recent production masquerading with an old name. I've still got a few Orange Amprex World Logo JAN E88CCs that are NOS, but they don't have gold pins, they have flying saucer getters and solid Russian type glass.
Best from Hot Tucson
Bob
"He (R.M. Nixon) was a foul caricature of himself, a man with no soul, no inner convictions, with the integrity of a hyena, and the style of a poison toad." H. S. Thompson
.
Rob"Let there be songs, to fill the air"
Thanks for sharing that info!
Just a couple comments if I may:
1. Gm variance from 2200 to 3000 uMho is a difference of nearly 40%. That's typical, and is about the same as I find in other small signal NOS. It's also what I find in new production.
2. 5 ma to 8 ma is a difference of 60%.
I point these out because often I see comments about how new production tubes don't have the "tight specs" of NOS tubes. Well, 40% variance and 60% variance hardly qualify as "tight".
I also appreciate you sharing this:
"A few obviously used tubes were also tested, with many at the low end but with some testing at the high end even with significant visual indications of use present."
SO true! Tester results - any tester - cannot be used as the sole criteria to judge a tube's "newness".
Finally, you wrote:
"...as line voltage changes caused a shift in Gm and Ip readings. This is probably linked to changes in filament voltage and plate voltage during test (triodes sensitive to plate V)"
Yes! A very cogent comment, thanks for sharing it.
Great post, very useful data - thanks again!
JIM McS -
A bit more data re: your post -
From a database of the first 400 US NOS 6SN7 tubes, the following
added info -
A - 19 sections tested below 2200 gm, with an average plate current
for the lower testing section of 5.33 mA (4.4 mA min @ 2175;
6.1 mA max (Gm not noted)
B - 10 sections tested > 3000 Gm, with an average Ip of 8.3 mA,
with one section testing at 4800 Gm / 9.8 Ma Ip
Some of the test variance would probably fall within the
2% reject rate factories allowed distributors in the 1960's.
Among a group of older Japanese (13) and USSR (5) tubes (guess
late 70's to mid 80's production), test results were about the
same as the USA tubes.
However, the section to section Gm and Ip were closer. Japanese
tubes tested mostly 2400-2600 at about 6.6 mA. USSR ran closer to
2750 at about 6.8 mA
How current production tubes would compare I leave to your
experience, as I have none to sample.
The USM-118 (cardmatic) has regulated supplies and a wider range
of test voltages, as does the Triplett 3444 series, the New London
901A and the over-large RCA Lab unit, to name a few.
A BOGEY (design center) tube should test at 75% on the USM-118
(does NOT mean that the tube is at 75 % of the NEW value - I know
know this, but I add the comment for others who may read this
post).
Test data is more reliable using these units - I just happened to
use the Hickok unit for this test series, as Ip is easier to sense.
When matching section to section, a TEK 570 and a computer are
very handy. Alternately, a stairstep generator and a scope or
"sample and hold voltmeter" system would work.
Alternately, a properly configured DIFF AMP circuit with the
proper instrumentation should work well to sense balance and
perhaps tube noise as well.
TUBOS
Post a Followup:
Top of Page ]
[ Contact Us ]
[ Support/Wish List ]
[ Copyright © 2009, Audio Asylum, all rights reserved ]
[ Go Green ]
[ Take5 ]