KMD GV60D hi/lo switch

[pdf64]

Or, even when (as with JJ) they publish credible info for their products, what rating system they're using.

...who knows how closely to the original specs the current manufacturers are keeping, and how much variation they allow?

I've never been able to identify how much variation was acceptable back in the day? eg a tolerance to type nominal for mu etc, or the plate characteristics.

[mhuss]

I've never been able to identify how much variation was acceptable back in the day?

Yes, sadly, this was likely the kind of information passed on from senior engineers to junior ones, and never captured in print.

[pdf64]

Yes, sadly, this was likely the kind of information passed on from senior engineers to junior ones, and never captured in print.

Was there was a standard point at which tube tester displays moved out of the green (or whatever means was used to show performance below the acceptable range for the type was indicated)?

[mhuss]

All the tube testers of my acquaintance (other than the fancy expensive ones like the Hickok 539x series) were really about as useful and accurate as a battery tester. They'd detect obvious faults, but "goodness/badness" was really kind of a ballpark estimate. "OK, it passes current!" They'd run at relatively low anode voltages as well.

Good article here: http://www.alltubetesters.com/articles/tester_guide.htm

If you're DIY handy (as most are here), Ronald Dekker's DIY curve tracer (uTracer) is a nice test/evaluation tool.
https://www.dos4ever.com/uTracer3/uTracer3_pag0.html

[pdf64]

The section starting 'Characteristic Spreads' (p66 of the pdf) of Tomer gives some useful typical tolerances https://frank.pocnet.net/other/docs/...cuum_Tubes.pdf

[mhuss]

Good reading. So even in the Golden Era, transconductance (and therefore plate resistance) were expected to vary +/- 40%!! No wonder all classic designs have lots of local NFB...

[Helmholtz]

The section starting 'Characteristic Spreads' (p66 of the pdf) of Tomer gives some useful typical tolerances https://frank.pocnet.net/other/docs/...cuum_Tubes.pdf

I don't think the spread of performance characteristics allows for conclusions on safe operating limits. Also tolerances are typically bidirectional (+/-) and statistical, so dont mean a headroom for the individual specimen.

[mhuss]

Tomer differentiates between hard limits (max V, max P) and 'fuzzy' ones (transconductance). Design Maximums are really that, not a ballpark number.

[pdf64]

I don't think the spread of performance characteristics allows for conclusions on safe operating limits.

RCA AN-174 shows that both the design centre and design max rating systems include a derating allowance to accommodate tube variation, ie so any tube should be able to handle its type's design max plate dissipation.

[Helmholtz]

RCA AN-174 shows that both the design centre and design max rating systems include a derating allowance to accommodate tube variation, ie so any tube should be able to handle its type's design max plate dissipation.

Not sure what you mean with derating as derating typically means lowering the load/dissipation, which of course is always allowed.

... so any tube should be able to handle its type's design max plate dissipation.

Yes, but acc. to the chart only if relevant circuit components have zero tolerance and all supply voltages incl. heaters are stabilized. In other words design max dissipation must never be exceeded. But typically supply voltages are subject to mains voltage variations, so the designers needs to take into account worst case conditions.

[pdf64]

Not sure what you mean with derating as derating typically means lowering the load/dissipation, which of course is always allowed.

By 'derating', I referred to the process described on that snippet from AN-174, ie the tube manufacturer identifies the absolute max rating typical of the type, and then factors in tube variations (the rating that a (lower) 'limit' tube of that type can withstand) to identify the design max rating. That will necessarily be lower that the type's absolute max rating,
Absolute max ratings are rarely published, an example being the MO KT66 https://tubedata.altanatubes.com.br/...086/k/KT66.pdf whose plate limit has an absolute max of 30W and a design max of 25W.

Yes, but acc. to the chart only if relevant circuit components have zero tolerance and all supply voltages incl. heaters are stabilized. In other words design max dissipation must never be exceeded. But typically supply voltages are subject to mains voltage variations, so the designers needs to take into account worst case conditions.

I agree, my view is that design centre ratings are much better suited to our application. AN-174 shows that design max ratings aren't intended to be used directly, rather that a further 'derating' (ie a coefficient of <1) be applied to the design max value. When biasing an amp, it's rather inconvenient to attempt to derive a suitable value for that coefficient, particularly as there's relevant but unknown data, eg transformer winding tolerances. It's much simpler to use a design centre rating, which as per above, has allowances for typical variations already included.

Based on the published info for 6V6GTA and 6L6GB, a design centre limit looks to have a coefficient of about 0.86 of its design max. If applied to 6L6GC, its design centre plate dissipation limit would be about 26W.
Note that on the 6L6GC fixed bias AB1 'operation characteristics' chart on p9 of https://tubedata.altanatubes.com.br/...93/6/6L6GC.pdf the plate dissipation curve doesn't rise above 26.5W.

6V6GTA design centre plate dissipation limit 12W https://tubedata.altanatubes.com.br/...7/6/6V6GTA.pdf
6V6GTA design max plate dissipation limit 14W https://tubedata.altanatubes.com.br/...9/6/6V6GTA.pdf

6L6GB design centre plate dissipation limit 19W https://tubedata.altanatubes.com.br/...49/6/6L6GB.pdf
6L6GB design max plate dissipation limit 22W https://tubedata.altanatubes.com.br/...27/6/6L6GB.pdf

[Helmholtz]

Note that on the 6L6GC fixed bias AB1 'operation characteristics' chart on p9 of https://tubedata.altanatubes.com.br/...93/6/6L6GC.pdf the plate dissipation curve doesn't rise above 26.5W.

Probably because of lower than max screen and plate supply voltages. So it's not a limiting chart - rather it refers to a typical design example (described earlier in the datasheet).

[pdf64]

Probably because of lower than max screen and plate supply voltages. So it's not a limiting chart - rather it refers to a typical design example (described earlier in the datasheet).

My interpretation of that example of typical operation is that it demonstrates the 30W design max plate limit is intended to be derated. The operating conditions may have been selected to ensure that a ‘design centre’ equivalent plate limit is adhered to, with plenty of a ‘margin of safety’ being left between the actual plate dissipation and the type’s design max limit.
That ‘margin of safety’ perhaps being the allowance for component and supply variation derating coefficient, ie the difference between a design max and a design centre limit.

[Helmholtz]

My interpretation of that example of typical operation is that it demonstrates the 30W design max plate limit is intended to be derated. The operating conditions may have been selected to ensure that a ‘design centre’ equivalent plate limit is adhered to, with plenty of a ‘margin of safety’ being left between the actual plate dissipation and the type’s design max limit.
That ‘margin of safety’ perhaps being the allowance for component and supply variation derating coefficient, ie the difference between a design max and a design centre limit.

Maybe, but could as well be just a good conservative design example. In the old days tube manufacturers certainly cared about rejects.
I am with you that a good design should always incorporate some derating. In my company we had strict derating guidelines.

Interesting to note that with a lower plate load (e.g. Fender typically used a Zaa of around 4.5k) plate dissipation would considerably increase and show a hump at medium power.

[vintagekiki]

https://groups.google.com/forum/#!topic/rec.audio.tubes/SM3iWw8ExCY
_{Some tube history about 6L6}

http://web.archive.org/web/20050317085639/http://schema.web1000.com/tech.data/6l6.txt
_{6L6 FOREVER By Eric Barbour}

1)_{A story from the workshop}

As a radio amateur 30 or more years ago I came to across various types of output tubes, mostly used or written off inventory.
On one occasion a quartet 5881 smaller bottle as 6V6 appeared ^1. I installing this tubes in a Boogie 60W, at some musician who plays in pubs every day. The first pair those tubes changed after 4 years of use.
Those tubes definitely moved all the boundaries of my thinking.

Then older colleagues said me that in their time the cathode and filament were made with a thicker layer of tungsten, the anode was coated with graphite, the pages anode is spot welded, so the fifties years tubes be resistant to red-hot, burn out ...
From them I learned that tubes bias set at a cathode current of mostly 25 mA, which I still do today.

I changed a few tube tester. Everyone mostly tests tubes in cold mode (250V), and when this tube installed at 500V rock 'n' roll begins
Back 10 years I replaced the tube testers with a 1 ohm resistor in the cathode ^2., and the tubes matched in the operating mode (voltage 400 - 550V), the selection of tubes is quite sharp and the amplifiers do not return ...
1.https://i.ebayimg.com/images/g/PL0AAOSw5BNeluBq/s-l400.jpg
2. https://www.tubeampdoctor.com/en/biasmaster-testadapter-pair-for-tubes

2)
http://www.audiotubes.com/6l6.htm
https://www.amplifiedparts.com/tech-articles/6l6gc-comparison-current-made-tubes
https://www.tdpri.com/threads/nos-tung-sol-5881-versus-jan-tung-sol-6l6wgb-differences.757632/