diyAudio (
-   Tubes / Valves (
-   -   Screen grid zeners; which? (

Johan Potgieter 7th March 2008 02:48 AM

Well yes, obviously. But whoever used a large (unbypassed) resistor as a screen dropper in a triode-connected pentode? The odd 100 ohm normally used as a series stopper for screens (not even always there) will cause no more than 1 - 2V drop. Hope nobody considers that of any consequence.

But then we have the recommended all of 1K in series with EL34 screens in their UL operation. And the manufacturer shows that the distoriton is lower with that in place than without it.
So it would seem mainly (only?) a matter of screen dissipation, which will not be exceeded in the first place with G2 tied to anode, if anybody will read their tube specs.

But thanks Anatoliy.

Klimon 7th March 2008 04:19 PM

Johan, excuse me if I've proven short on temper; I must have misunderstood your hermetical first sentence and filled in ambiguity in a manner you did not intended. Understandably it tempts for a technical analysis that's above me but what I do perceive is that it would be an almost unimaginable coincidence that two experienced designers independently drawed a very similar circuit that technically doesn't make any sense AND that this circuit has positive properties (to my experience that is) that are independently perceived by one of the designers and myself in about the same manner? Let's say that I've tasted the pudding and liked it alot while you've read the recipe and said it can no way taste good, no wonder I'm scrutinizing your knowledge of pudding.



Wavebourn 7th March 2008 04:45 PM

Speaking of puddings, it is a well known fact that pentodes powered with screen grid voltages lowered show better linearity. Also, low dynamic resistance of that voltage source reduces variations of gm, mu, and rp. I've mentioned already 6P15P used in video amplifiers with low strict G2 source, and GU-50 used in linear RF amplifiers, also a screen grid voltage was low and stable. The same was used in TVs for more linear deflection of rays in order to get geometrically linear pictures.

I suspect that triode strapped pentodes may exhibit similar behavior, so a CCS from a negative bias source may be connected to a screen grid with a Zener to anode. An experiment is needed.

If to shunt a Zener by a large enough capacitor a negative supply may not be needed since when a voltage on a G2 goes below zero it stops conducting.

rdf 7th March 2008 05:34 PM

An analogous scenario occurs when a pentode cathode follower's screen bypass cap is returned to ground instead of cathode. To cathode is 'the right way' but every EL84 sim I run indicates to ground results in less distortion, both into high impedances and when driving grid current. I vaguely recall the same happens in real glass, take that with a grain of salt.
Currently on the bench is a 6UL8 configured as triode gain/ pentode CF into a GU-50 SE. If time permits I'll try both topologies and post results.

Wavebourn 7th March 2008 05:42 PM


Originally posted by rdf

Currently on the bench is a 6UL8 configured as triode gain/ pentode CF into a GU-50 SE. If time permits I'll try both topologies and post results.

Thanks in advance!

Edit: However, when a G2 is bootstrapped to cathode it should show more gain, but less linearity. Kind of positive feedback, but non-linear one.

Johan Potgieter 10th March 2008 01:03 AM

Klimon, your post #32

No harm done, and sorry about the misunderstanding from my part. To put it better, I did not mean that the arrangement will not work (I would have said that instead of advising you in the beginning about zeners' temperature co-efficient, thus implying: go ahead.) I merely questioned the advantages of such an arrangement (with open mind), as the articles by Grimwood and Perkins in no way inspire confidence. [They make a number of statements that are simply not true (my professional experience goes back 50 years)].

Perhaps my bedside manner was not too good, but without intending to be arrogant those of us in the business do feel that it is somewhat of a duty to point such matters out; as was done to our advantage by others before us and now gone. The internet conveniently contains a wealth of information, but unfortunately also an unholy bundle of misinformation. I would stress, if you honestly felt that the experiment improved matters, fine!

And I fear you are right; I do like pudding but cannot read recipes too well :o and have never tried making dessert myself!

AKSA 10th March 2008 01:32 AM

It does make good sense to me to ensure that AC imposed upon both plate and G2 is identical, but that DC voltage is slightly lower. Indeed, standard resistive strapping achieves this, so I see nothing wrong, particularly with the TL431 approach.

Johan, the guy has a point. You once insulted me gratuitously, and didn't even reply to my email. Here at least you are prepared to apologise. I think your moral crusade at some of the audio misinformation has its smartarse side, and I think you should build some of these things and listen to them before condemning because they fall outside your experience.

Perhaps you are getting old?



rdf 10th March 2008 01:52 AM

The mammoths sleep safe tonight. :) At this point more gain and higher distortion with a less pleasing distortion profile in pseudo-pentode mode vs. returning the screen bypass to ground. That's a first cut, more experimentation required, specifically a regulated/floating screen supply.

All times are GMT. The time now is 04:16 PM.

Search Engine Optimisation provided by DragonByte SEO (Pro) - vBulletin Mods & Addons Copyright © 2018 DragonByte Technologies Ltd.
Resources saved on this page: MySQL 17.65%
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2018 DragonByte Technologies Ltd.
Copyright ©1999-2018 diyAudio