Go Back   Home > Forums > >
Home Forums Rules Articles diyAudio Store Blogs Gallery Wiki Register Donations FAQ Calendar Search Today's Posts Mark Forums Read

Tubes / Valves All about our sweet vacuum tubes :) Threads about Musical Instrument Amps of all kinds should be in the Instruments & Amps forum

Matching 100 tubes ECC803S
Matching 100 tubes ECC803S
Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 27th October 2016, 01:26 PM   #11
petertub is offline petertub  Sweden
diyAudio Member
 
Join Date: Jul 2015
Location: gothenburg,sweden
Quote:
Originally Posted by Ketje View Post
The datasheet from Telefunken gives limits for a good tube.
So does the sheet for the russian (almost equivalent 83) 6N2P.
Mona
tfk 803S is a special quality tube, selected at source. Is not made any more.
6N2P is NOT ecc83.
__________________
My home is at www.tubular-well.se
  Reply With Quote
Old 27th October 2016, 01:58 PM   #12
karsten21 is offline karsten21  Germany
diyAudio Member
 
Join Date: Jan 2010
Location: Germany
Whoww, thanks a lot for all your excellent input!

The manufactor is JJ-electonics and they say only itīs "like" ECC83 but with long plates and "long life" version.

Aahh, you mean LINE 94 but itīs tube #92 Yes, this one an #4 and #20 have quite different systems...

I get a math like this to calculate a "sum" of Ia and mu:
1/4*(2*ABS(Ia1 - Ia2)/(Ia1 +Ia2) + 2*ABS(mu1 - mu2)/(mu1 + mu2) + [other parameters]...)
Not understand this but it was given from a math-expert to me..

regards
Karsten
__________________
/// Never forget your towel \\\
You who are doomed, enter and abandon yourself to despair!
  Reply With Quote
Old 27th October 2016, 02:45 PM   #13
Tubologic is offline Tubologic  Belgium
diyAudio Member
 
Tubologic's Avatar
 
Join Date: Dec 2004
Location: Brussels
I'm wondering what's the purpose of this ?

Original Special Quality tubes like the ECC803S (ECC802S,EF806S,E80CC, etc...) were build to tighter electrical and mechanical tolerances and pre-aged (stabilized), they were rarely used in consumer audio equipment but intended for critical applications like metrology,medical,aerospace, industrial equipment, etc... where reliability was paramount and price a secondary factor. They were much more expensive and usually didn't needed any matching or selection, being already pre-tested/pre-aged/selected tubes.

The problem is that currently manufactured ECC803S are nothing more than rebranded common ECC83/12AX7s with a more luxurious packaging (and higher price) but don't have any of the refinements of the real Special Quality tubes made in the good old days. Also, they don't go through any pre-aging/stabilizing process. Then, the question is: after you have carefully matched your tubes by statistical analysis based on actual measurements HOW LONG will they remain matched once you'll put them into service. A couple of hours ? A few days ? And which parameter is the most important in your circuit ? I've never seen an audio equipment so far who requires critical close matching of the small signal tubes, or it must be a very poorly designed circuit.
  Reply With Quote
Old 28th October 2016, 04:05 AM   #14
PRR is offline PRR  United States
diyAudio Member
 
PRR's Avatar
 
Join Date: Jun 2003
Location: Maine USA
> I simply add all values ( Ia,mu,Ra,Gm) of one system. Next I divide the sum of both systems. If the result is 1 both system matches exactly.

No, because a side could have high Ia and low Mu and give the same score as the other side. Imagine a tube half 12AX7 and half 12AU7.

And as Ra and Mu are larger numbers than Ik and Gm they count for more.

But back up. You have tested at "fixed" bias. Do we ever run preamp tubes in fix bias? No! And one reason is that they vary so much! But if we use self bias (significant cathode *and* plate resistors), the hot units down-bias and the cool units up-bias.

Take tube #26:
Ia 0.99,1.31 (1.32 ratio)
Gm 1.45,1.61 (1.11 ratio)

Significant difference in Ia at same bias *voltages*. About a 1.32 ratio. But put 1.5K under and 100K above it, the difference will be around half. Or perhaps square-root. Now more like 1.15 ratio of Ia. Gm will fall with Ia, and roughly as root of Ia. I predict these two units match in-circuit within 4%.

In small-signal work it probably makes much more sense to "match" audio gain in a typical circuit. For 12AX7 types, 1.5k and 100K with a 220K audio load. Feed 0.1V in, and look for 5V out. Anything outside 4.5V (dud) or 5.5V (unlikely) goes to the guitarists' sale.

With no cathode bypass, more like 2.5V out and much lower variations. 1.5K spoils-down Gm, and levers-up Rp, the same on all tubes, reducing the differences.

If you need better than 5%-10% gain accuracy, Harold Black nailed it: make way too much gain and use NFB to make final gain a function of two stable passive parts.
  Reply With Quote
Old 28th October 2016, 08:33 AM   #15
karsten21 is offline karsten21  Germany
diyAudio Member
 
Join Date: Jan 2010
Location: Germany
Dear PRR,
Thanks a lot for your explanation! You are totaly right; the best thing is to test the tubes in circuit, for sure. Instead to measure static values it makes sense do select the tubes within a small test circuit which has the same design as the target.
This is the result of this nice and very educational thread for me. And Iīm very thankful for all your time and help!

My first aim was to get dedicated and reliable measurments to be able to sort out and pair tubes. Tubes which are to far away from given data by manufactor should be send back. And to do this I have to provide hard facts; I think JJ-electronics will not accept a return shipment if I say "They donīt fit my personal idea of tube quality".

So over all, I will build my own "In circuit" tester. Tubes wich are not fit my requirements, must then be tested again with uTracer to retrieve "hard facts".

The other way is to buy pre selected tubes from manufactor. But... They donīt tell me the tolerance about thier selection. All information I get ( until now ) is that they select by Ia.

By the way why modern ECC803S. For my realy subjective and personal opine this tube works excellent. And Iīm not a beliver of that everything was better,higher,nicer... concerning tubes in old days. Espacial NOS tubes which sometimes hard to get and high price for a not know condition... Sorry, this is totaly Off-Topic....

So: Thank you all for your great input and I have to say: Its every time a big pleasure to be here: So much friendly and wise guys!!!

Best regards
Karsten
__________________
/// Never forget your towel \\\
You who are doomed, enter and abandon yourself to despair!
  Reply With Quote
Old 28th October 2016, 09:59 AM   #16
DF96 is offline DF96  England
diyAudio Member
 
Join Date: May 2007
The original Tfk ECC803S used a frame-grid construction. They also had a conventional ECC83. Then Tesla did a frame-grid ECC83 and a conventional ECC803S (just to avoid confusion). The modern JJ ECC803S has a frame-grid-like anode structure but might not actually have a frame grid - the JJ blurb seems to be deliberately vague on this.

Quote:
Originally Posted by PRR
Significant difference in Ia at same bias *voltages*. About a 1.32 ratio. But put 1.5K under and 100K above it, the difference will be around half. Or perhaps square-root. Now more like 1.15 ratio of Ia. Gm will fall with Ia, and roughly as root of Ia. I predict these two units match in-circuit within 4%.
Yes, this is the advantage of traditional resistor bias and loading - it reduces variation, between samples and due to ageing. Those using CCS (or fixed voltage) bias or loading have to accept wider variation in whatever they are not fixing.
  Reply With Quote
Old 28th October 2016, 10:41 AM   #17
SY is offline SY  United States
On Hiatus
 
SY's Avatar
 
Join Date: Oct 2002
Location: Chicagoland
Matching 100 tubes ECC803S
A couple of follow on comments: because mu is derived from rp and gm, it's not an independent variable. Second, you might want a weighting factor since order of magnitude of rp and mu is 100, order of magnitude of Ia and gm is 1- if you take any row there and set Ia to zero, you'll see that your system derivation factor barely moves.
__________________
"You tell me whar a man gits his corn pone, en I'll tell you what his 'pinions is."
  Reply With Quote
Old 28th October 2016, 07:14 PM   #18
PRR is offline PRR  United States
diyAudio Member
 
PRR's Avatar
 
Join Date: Jun 2003
Location: Maine USA
> because mu is derived from rp and gm, it's not an independent variable.

A different, maybe insightful, way to see it:

Mu is a *geometrical* parameter. Mu may be calculated if you know the internal dimensions exactly. Grid wire diameter and spacing, grid-cathode-plate spacings.

That's not Absolute Truth in real tubes. Mu always drops at low current, because the main high-Mu part of the tube cuts-off, leaving sneak-paths of lower Mu around the ends of the non-infinite grid. Mu often appears to fall at high current, perhaps due to grid "filling with electrons", but also because many tube testers run large grid signals and run into grid current which may reduce effective drive and certainly diverts cathode current. Many older tubes have very "flat" Mu curves. Many of the hotter tubes from the modern era (1950s) are goosed-up at high current and show more Mu slope. (Variable-Mu tubes make the sneak-paths a major part of the useful operating range.)

Gm is how well the cathode emits and how well the grid winding controls it. The absolute limit is proportional to current the same as a BJT. No real tube comes close because the finest grid wires are far too fat (we need crystal-size grid structures, which is why a BJT gets close to the Gm/Ie limit). Gm is set by cathode activity and grid dimensions.

The Rp is the "independent" variable, being essentially Mu/Gm. If you know Mu and Gm there is no need to measure or note Rp. Just derive it when you do circuit computations.
  Reply With Quote
Old 28th October 2016, 07:36 PM   #19
DF96 is offline DF96  England
diyAudio Member
 
Join Date: May 2007
Quote:
Originally Posted by PRR
The absolute limit is proportional to current the same as a BJT. No real tube comes close because the finest grid wires are far too fat (we need crystal-size grid structures, which is why a BJT gets close to the Gm/Ie limit).
No. In an ideal fixed-mu triode valve gm is proportional to the cube root of current. Only in an ideal variable-mu (remote cutoff) valve is gm proportional to current, like a BJT.

Quote:
The Rp is the "independent" variable, being essentially Mu/Gm.
Not really. Given any two (whether from measurement, theory or simulation), you can calculate the third.
  Reply With Quote
Old 28th October 2016, 07:56 PM   #20
petertub is offline petertub  Sweden
diyAudio Member
 
Join Date: Jul 2015
Location: gothenburg,sweden
I would say that any testing of tubes ( to compare with documented values) should be done
at the same working points as the document.
And all (?) ecc83/12ax7 vendors specify typical characteristics at Va=250V and Vg=-2
( it also specifies Va=100 and Vg-1 but that is less typical in applications)
Measuring at any other points will make it difficult to relate to ECC83
( well comparing anode curves would also do but that is more complex then a go-nogo
evaluation)
__________________
My home is at www.tubular-well.se
  Reply With Quote

Reply


Matching 100 tubes ECC803SHide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump

Similar Threads
Thread Thread Starter Forum Replies Last Post
Help Matching RCA 5692 Tubes radiobarnhci Tubes / Valves 5 9th January 2016 11:21 PM
Help matching and crossing FR at 100-150hz lorienblack Full Range 23 22nd January 2014 08:50 PM
Matching tubes for PP h00hbt Tubes / Valves 3 19th May 2011 03:13 PM
Matching tubes pot in old Fenders slideman82 Instruments and Amps 2 27th October 2009 03:06 PM
WTB: (3) JJ ECC803S [12ax7] Tubes mrdon Swap Meet 0 29th November 2007 11:18 AM


New To Site? Need Help?

All times are GMT. The time now is 07:36 AM.


Search Engine Optimisation provided by DragonByte SEO (Pro) - vBulletin Mods & Addons Copyright © 2018 DragonByte Technologies Ltd.
Resources saved on this page: MySQL 14.29%
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2018 DragonByte Technologies Ltd.
Copyright ©1999-2018 diyAudio
Wiki