Re: SACRILEGE! NOS Bull*** Debunked...

In article <43650514.C4905832@xxxxxxxxxxxxx>,
Lord Valve <detritus@xxxxxxxxxxxxx> wrote:

> I didn't *run* them at 32mA.
>
> All the tubes in my inventory are assigned a plate current rating
> as part of the matching process. I picked tubes with a 32mA
> plate current rating so that they'd all be similar (at least, in that
> respect) to each other regardless of brand. It saved a lot of
> bias control tweaking, too. 32mA PC is a spec that nearly
> all batches (of whatever brand) will produce at my test/matching
> parameters. The two exceptions noted in the original article were
> the Shuguang/TAD shortbottles and the JJs, both of which tend to
> read very low at my particular set of test/matching voltages.
> (Hence the necessity of lowering the bias range-setting resistor
> when testing the JJs.) At a plate voltage of 440.5, the current
> setting for the tests was 42mA, which results in a static dissipation
> of 18.501 watts. I've found that the best compromise between tone
> and tube longevity (in most Fenders) is right around 19 watts static.
> I run DT's amps at 19.5 - he loves the tone, and failures are few
> and far between.

So the 32mA is a pre-rating on the tube using your own screening rig. I
shoulda known it wouldn't have been 32mA in use. That's barely class AB
once you get anywhere near saturation.

(I've grown to SERIOUSLY distrust the numbers New Sensor uses in their
"platinum matching"...I'm lucky if "pre-matched" tubes are within 20%
once I put them in a real amp.. I'm sure your matching is more
relevant).

>
> As to those who ask ridiculous questions like "why didn't you
> run them at 550 volts with 30 watts static?", I can only reply
> that it's for the same reason that I don't test light bulbs with a
> sledgehammer. It'd be stupid. Why not make it 1000 volts
> and 50W static? Because no-one uses them that way.
> I'm not about to burn up $1000+ (retail) worth of my inventory
> to prove what I already know.

<nitpick>
But 700V @ 25mA (17.5W) *is* a real world situation. Music Men.
However they use a V(g2) of only 350. Again, field testing is prolly
the way to see reliability factors.
</nitpick>
>
> > And like Robert, I'm also curious about sensitivity. i.e. what was the
> > input voltage which sent the amp to clip. (ok, it's probably more a
> > function of the preamp, but these are interesting data nevertheless).
>
> I wasn't interested in that parameter, only output power.
>
> That's a "feel" thing; I like customer feedback for stuff like
> that, since I don't play the guitar (much).

Hard to quantify "feel" in a lab test. Sort of like being able to
predict what a speaker will sound like based on the T/S parameters. I
know a guy who claims he can do that, but I have my doubts. ;-)
>
> > How 'bout the New Jersey Tung-Sol 5881
> > http://www.vintagetubeservices.com/page16.html
> > vs Mike Matthews' Neo- Tung-Sol?
>
> I only have a pair of the "original" Tung-Sol 5881s, not a quad.

I'm SHOCKED! Just SHOCKED! ;-)

Just for "laughs", you prolly should have included the wafer-base
5881/WGC. I know it's a crummy tube, but some amps HAVE to run it (I'm
thinking specifically of the Marshall JTM30). I don't think it's a
re-based WXT, but I could be wrong.

>
> Mark Baier (Victoria Amplifier Company) and I concur:
> you *cannot* tell them apart by ear. Customer reports
> (so far) are universally favorable. This one's a winner.
>
> > And finally, give these tubes a 24 hour burn in under signal and THEN
> > retest, measuring the pin 5 voltage before and after. Is there much
> > bias drift?
>
> Everything I sell is burned in for 24-48 hours.

>
> It's silly to match tubes that aren't burned in first.

And once they pass 36h and get branded, is there much drift at 100h?
Obviously you can't do this to every quad you sell...at some time you
have to say "good enough", but in the field, hopefully people will be
getting more than 100h out of their toobs. ;-)

Along with tone, reliability ***IS*** my #1 concern. Perhaps even more
so than tone given the settings many of my customers use as well as
their pocketbooks.

Really, I'm not trying to be pugnacious here, just curious.

>
> > Of course the long term test is the hardest to do...you do have a shop
> > to run, but to me, it's probably the most important test. RELIABILITY
> > IN THE FIELD matters more than a 5% difference in power. I **HATE**
> > callbacks!!!!!
>
> Since I sell thousands of tubes per year, I get *plenty* of feedback on
> reliability.
>
> Know what croaks the least? The brown-base Shuguang 6L6GC-M.
> The most? The Ei 12AX7, followed closely by the Groove Tubes
> 12AX7-M. [...snip...] I *quit* carrying
> any of the Ei stuff (besides the EL84) for the same reason - an 80%
> reject rate (mostly for extreme microphonics) is simply not acceptable,

80% for EI 12AX7. I knew they were prone to shot noise, but that's
pretty extreme! I'm phasing them out for the same reason, but the few I
still have I'll stick in as a p/i if the customer wants to save a couple
of bux. They seem to be OK in a cathodyne....not enough data for a
long-tail.

>BTW, my first large batch of the new
> Reflektor/Tung-Sol 5881s (200 pieces) had a ZERO reject rate - *all*
> of them (with the exception of one that committed suicide by rolling
> off the bench) ...

To quote Dr John re: New Orleans: "She didn't die, she was murdered."
>
>
> > Thanks for the test results. You obviously put a lot of work into it.
> >
> > Patiently waiting for this thread to degenerate into a flame toss....;-)
>
> I'm waiting for the wankers like Mulay and TamPax to start squealing that I
> made it all up..

Of course you didn't make it up. It's a carefully done and
well-documented test. However, I am trying to make some real-world
sense of it. There's a 21% difference between the most and least
powerful performers (Matthews-Svet vs Matthews-5881WXT) That's only
-1.1dB. Can anybody really hear this, especially considering the SPL
that 70-90W represents?

Not to discount your hard work, but I think we all agree that tone and
reliability are really the parameters most important to the end user.
Tone is hard to quantify (unless you do a full freq response curve and
can somehow relate the curve to "feel"); and unless you're willing to
devote 6 months and multiple test jigs in a lab test, reliabilty is best
determined by user reports.

Still, it's interesting that the 5881 WXT comes in last. About what I'd
expect based on my own "feel" ;-)

--Mike

--------------------------------------------------------------------
Mike Schway | [Picture your favorite quote here]
mschway@xxxxxxx |
--------------------------------------------------------------------
.

Quantcast