Johan Potgieter said:
Indeed! For example, when some guru suggests better regulation by dropping some voltage on a gas discharge tube in series with the load.
😎
For fun I ran a Spice analysis of an ideal version of Pearl's config against a true triode connection. An EL84 was driven directly by an ideal AC voltage source. Ideal DC voltage sources provided cathode bias and plate-screen bias. Distortion figures were compared between the plate-screen voltage set to an arbitrary low number (0.001 VDC) and a 31.5 VDC approximation of Pearl's circuit. The plate load was an 8kohm resistor, input set to 2 VAC 0-p. Pearl's circuit required a grid-cathode voltage of 6.42 VDC against the triode's 8 VDC to maintain the same 41.5ma IP and 250-251 VDC Vpk. Results below, conclusions none. =D
Attachments
Johan Potgieter said:Seems you and I missed something in our technical upbringing, Anatoliy.
Conclusion: old f**ts must die like mammoths.
Interesting discussion - I'm planning a small SE triode amp using 6LR8 compactrons with the power pentode section triode connected. I was thinking of using the simple coward's approach with the screen nailed to the plate through a resistor, but I may try the Pearl approach to determine if I can hear a difference - I'd forgotten about it until now.
Come on Johan, don't give me that kind of denigrating armchair-smartass bogus.
Promises to be interesting, it's not much work and at the very least you'll learn something.
Simon
Promises to be interesting, it's not much work and at the very least you'll learn something.
Simon
Klimon said:
😕 😕
Simon, I seem to have offended you, and I apologise for that.
But would you kindly perhaps just quote where I was a smart-*** and what bogus you are referring to, so that I am at least not at a disadvantage? I think that is a fair request?
[All I can think of is my remark about Eddie, and that was criticism of his epistle, not you. For the record, there are several matters in that article that are plain nonsense, but I am not going to set myself up as a critic per se. That is not what this thread is about. You would have noticed that others agreed with me - are they also s-a's?]
Back to the topic, and perhaps at my own expense (kindly pardon my resistance to go through the whole lengthy Grimwood and Eddie documents again):
The question crossed my mind before, but was unmentioned. Apart from a fixation with screen dissipation, what are the other claimed advantages again of this enhanced triode thing - not higher output?
One gains by allowing the anode to go higher up the voltage scale on positive signal peaks, but loses at the bottom because the screen has a low voltage limit before onset of 2nd harmonic distortion, above which the anode must now stop by the value of the built-in anode-screen voltage drop, instead of being able to approach its lower limit as in normal triode style.
The question crossed my mind before, but was unmentioned. Apart from a fixation with screen dissipation, what are the other claimed advantages again of this enhanced triode thing - not higher output?
One gains by allowing the anode to go higher up the voltage scale on positive signal peaks, but loses at the bottom because the screen has a low voltage limit before onset of 2nd harmonic distortion, above which the anode must now stop by the value of the built-in anode-screen voltage drop, instead of being able to approach its lower limit as in normal triode style.
Johan;
it seems to me they claim an advantage of AC-coupled screen grid to anode against an approach with a resistor to drop excess of a voltage to meet specs of max G2 dissipation. Looks like guys invented a bicycle and found that the 2'nd grid is a grid, i.e. when biased positively in respect to cathode it draw a current, also it is a control grid as any one grid, so current variations cause control, as the result in case of high AC resistance transfer function will be changed.
it seems to me they claim an advantage of AC-coupled screen grid to anode against an approach with a resistor to drop excess of a voltage to meet specs of max G2 dissipation. Looks like guys invented a bicycle and found that the 2'nd grid is a grid, i.e. when biased positively in respect to cathode it draw a current, also it is a control grid as any one grid, so current variations cause control, as the result in case of high AC resistance transfer function will be changed.
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.
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.
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.
Regards,
Simon
Regards,
Simon
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.
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.
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.
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.
rdf said:
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.
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 😱 and have never tried making dessert myself!
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 😱 and have never tried making dessert myself!
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?
Cheers,
Hugh
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?
Cheers,
Hugh
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.
- Status
- Not open for further replies.