At a fixed voltage, raising the impedance drops power but lowers distortion. Raise the voltage back up you gain efficiency, but the maximum idle bias you can run is lower (dissipation limited). Run the bias down too low, and distortion is worse. That EL 34 application with an 11k load and 800! Volts is how you get the efficiency high enough to get closer to 100 watts. It doesn’t sound as good as 450 volts with a 4 or 5k load.
It is also far easier to make a transformer with a full 20-20k response with a primary impedance of 4K vs. 10k, at a given power level. It’s just hard to make things physically have high impedances. More inductance takes more turns. More turns add more capacitance. So both ends are fighting you. If you really want to start making custom transformers you’ll have a lot to learn, but worth learning for sure if you have the equipment! I make my own custom transformers, but it’s limited to 60 Hz power transformers only. That’s because the materials I can reasonably obtain are not really suitable for high quality output transformers. And I do wind with a hand crank so I’m not doing any 5000 turns.
It is also far easier to make a transformer with a full 20-20k response with a primary impedance of 4K vs. 10k, at a given power level. It’s just hard to make things physically have high impedances. More inductance takes more turns. More turns add more capacitance. So both ends are fighting you. If you really want to start making custom transformers you’ll have a lot to learn, but worth learning for sure if you have the equipment! I make my own custom transformers, but it’s limited to 60 Hz power transformers only. That’s because the materials I can reasonably obtain are not really suitable for high quality output transformers. And I do wind with a hand crank so I’m not doing any 5000 turns.
WRONG.
The 807 has strictly aligned grids and very poor gain in the screen grid, hence its unsuitability like the 6V6 for UL running. You will note the low currents in the screen circuitm compared with say a typical OP pentode
ON a beam tetrode the screen IS THE ANODE.
The real anode is only a virtual electrode.
This why they are pretty non linear devices prone high order 3rd and 5th harmonic THD
The KT88 is a totally different valve with more than 2x the gain and much higher current capabilities because of the huge cathode. The screen can also be run at 550V, because its built for it
It has been noted many times the 43% tap on the 6L6, 807 and 6V6 is no good, but appears to have at least some effect at 20%.
Nothing is hard and fast, and the 43% idea is the stuff of myths and legends.
This level of NFB is entirely dependent on the gain, strength of the screen, and the inter electrode gain.
The 807 was made with a short electron path so it could operate reliably at high powers at over 60mhz, which is why its prone to serious oscillation.
Only the GEC TT21/22 achieved a good compromise because again they could isolate the G1 and G2 from the very high voltages and HF above chassis from below the chassis.
Apparently the idea of making the TT21 was an accident anyhow..
The KT88 as a result was never intended as a TX valve-more useful as a AF modulator.
The 807 with good UX bases could stand a lot more punishment than Octal as did the 5B254M with it's high quality locatal and top cap anode.
Last edited:
OK, i understand, it certainly looks like there are alot of tradeoffs by the sounds of things.
The guy who used to wind the transformers started his business with little to no knowledge, yet was able to pick it all up and his work was widely regarded as high quality.
Im sure I will be able to figure things out with enough time.
I think its safe to say that going for 6.6K is the safest bet for a first go anyway.
One thing ive learned is how you have to interleave the primary and secondary windings on an OT.
This explains why alot of amp builders always like to use the full winding of the secondary, as having a tapped secondary with a 4K load and you are only using one half of those interleaved windings.
I was told one way around it is to have 2 seperate 8 ohm secondarys, and use in parallel or series to suit the load, but then you are limited to the impedance of your speaker of either 4 or 16 ohms.
Interesting, I never knew that the screen acted as the anode on a beam tetrode.
I thought they worked in a similar way to a pentode, except the design used a beam forming plate to deflect the electrons instead of using another grid, which was just a means of getting around Philips' patent.
Ive been reading a few guides on transformer winding, and they say just go with 50% screen taps as you will never hear any difference and makes winding simpler.
It’s not sarcastic at all. When the plate voltage is pulled down below the screen, it is the screen which attracts and accelerates the electrons. That makes it an anode. The screen collects quite a few of them if the plate gets pulled low enough. It’s the difference in surface area that keeps screen current from getting ridiculous. A lot of electrons just zip on by and miss what attracted them in the first place. Sort of like the coyote zipping past when the roadrunner suddenly stops. And then running into a brick wall (causing heat).
Thank you. BUT only mostly true.
The difference and of course the resultant vast improvement in gain and power delivery of the beam tetrode compared with the triode are simply that the B-T has a screen grid which acts as the anode.
The anode therefore becomes largely insensitive to variations in voltage over its entire range...you can easily check that for yourself...you can keep the same current delivery at Va=100V and at 250V even if the screen is at 325V...
The quiescent current of a pentode or B-T is set almost entirely by the working voltage of the SCREEN grid, which is of course why it's always been recommended by GEC and others to STABILISE screen supply voltage to within 2% to minimise distortion and keep power stability under high loads (the corollary of limiting distortion of course).
Of course most people can't be bothered with screen grid stabilisation (too much hard work?) so they go for masses of sound destroying NFB, as well as UL which introduces compression into the entire process - something which Williamson rightly understood and to which Keroes etc never referred.
UL mode because of the serious variations in screen voltage and resultant gain changes in the valve curves becomes a very nice but bad dynamic range compressor.....again making sound sound lifeless and dead....but hey who cares, most music is compressed to heck to only 20dB dynamic range anyhow.
I can't understand why people can't understand that in a triode the anode is a REAL anode, while the Anode in a B-T is only virtual.
The descriptions of the various electrodes are well described here and here in detail
As for running EL34 at 800V, Good luck.
Mullard knew perfectly well that was the way to blow them up, which is why they gave trying to compete with GEC's KT88.
NB:-
The interesting thing about the EL37 (the EL34 proto), and the EL84 are the high gain, achieved by having a g1 very close to the cathode, a screen grid which is not perfectly aligned, and a g3 which cannot by it's very nature be aligned at all... The big downside,- with high g2 voltage the screen current goes largely out of control, the g3 doesn't inhibit secondary emission v well and with large anode swings the screen starts to glow white hot.
The EL37 was better than the much cheaper mass market EL34 because of the huge glass bulb and more generous anode and cathode dims.
Once this starts to happen the time taken for the G2 to start to get serious heat related distortion and collapse is measured in seconds....hence why guitar amps running EL34s at high power are monumentally unreliable compared with the fully aligned and largely bombproof 6L6GC....
Idem EL84, with only specialised more expensive rugged versions survivable at high g2 voltages. In fact matched quads of 7189/6P14P-EV are better than EL34 in most ways, but few people seem to realise that.
Last edited:
Hi,
The EL34 is made to run with 800V. There are some amps which were made for commercial service and PA Euipment with high +b. So Philips, Telefunken and Siemens made good reliable Amps. Dynacord has its "Eminent 2" , the "Gigant" and some others which run EL34 on max. +b . I had 2 of the Siemens 6S ela 2796, 250W rms from 6 EL34. They sounds great. I sold them to finance my homebrew projects.
Hiwatt made a 400W slave amp with 8 EL34 fired with 750V.
An EL34 which didn´t withstand the high +b is not a real EL34. Same with the EL156 also made for 800V. Cheap Chinese replicas pops at 500V...
I got still over 100 EL34 between my stuff , all east german (RFT) production. No problems with them to put them in old existing equipment with 800V +b if you have an eye on the screen voltage. It should not exceed 350V at 800V plate feed.
73
Wolfgang
The EL34 is made to run with 800V. There are some amps which were made for commercial service and PA Euipment with high +b. So Philips, Telefunken and Siemens made good reliable Amps. Dynacord has its "Eminent 2" , the "Gigant" and some others which run EL34 on max. +b . I had 2 of the Siemens 6S ela 2796, 250W rms from 6 EL34. They sounds great. I sold them to finance my homebrew projects.
Hiwatt made a 400W slave amp with 8 EL34 fired with 750V.
An EL34 which didn´t withstand the high +b is not a real EL34. Same with the EL156 also made for 800V. Cheap Chinese replicas pops at 500V...
I got still over 100 EL34 between my stuff , all east german (RFT) production. No problems with them to put them in old existing equipment with 800V +b if you have an eye on the screen voltage. It should not exceed 350V at 800V plate feed.
73
Wolfgang
Actually, I think you’ll find it was conceived by Alan Blumlein in the thirties, I believe. Hafler and Keroes definitely exploited the concept commercially.
I stand corrected, perhaps I should have said "popularised".
Speaking of inappropriate word choice, with regard to statements that the 6L6/807 etc is "unsuitable" for ultralinear, perhaps it should be "less suitable". With 6L6 you lose quite a bit of power output with ultralinear compared to tetrode operation, which you don't see with EL34 and similar more "modern" tubes. It's still a valid design choice, with pro's and con's like any other. I don't personally have any amps that run ultralinear. I've got hybrid triode, pentode (regulated screen supply), and I'm currently working on "schade" type feedback to a differential cascode.
Speaking of inappropriate word choice, with regard to statements that the 6L6/807 etc is "unsuitable" for ultralinear, perhaps it should be "less suitable". With 6L6 you lose quite a bit of power output with ultralinear compared to tetrode operation, which you don't see with EL34 and similar more "modern" tubes. It's still a valid design choice, with pro's and con's like any other. I don't personally have any amps that run ultralinear. I've got hybrid triode, pentode (regulated screen supply), and I'm currently working on "schade" type feedback to a differential cascode.
What exactly do you mean by "lose power"? Of course adding any form of feedback reduces gain. So you need to increase the input signal to get the same power output. That's not losing power.
If it helps, this is what I ended up turning a standard 5-20 build into. My output transformers are 6.6K sowters with 43% taps that I specified with tapped secondaries which bit me in the backside when it came to stabilising the feedback. I was unaware at the time of ordering them that tapped secondaries had the issues mentioned earlier in the thread.
I just tried dropping 6L6 into the spice sim I have for the amp and they appear to work just fine. With a lower supply voltage, they should be fine. My 410-0-410 transformer ends up feeding nearly 460V B+ so pushes an 807 beyond it's dissipation limit in my circuit. With 420V B+, idle dissipation is under 20W. Looking at it has peaked my interest in possibly getting a set of 6L6GC's to try.
I just tried dropping 6L6 into the spice sim I have for the amp and they appear to work just fine. With a lower supply voltage, they should be fine. My 410-0-410 transformer ends up feeding nearly 460V B+ so pushes an 807 beyond it's dissipation limit in my circuit. With 420V B+, idle dissipation is under 20W. Looking at it has peaked my interest in possibly getting a set of 6L6GC's to try.
Attachments
Yes and Reeves at hiwatt like everyone else (eg. Vortexion) junked the EL34 as desperately unreliable because Reeves like my Dad worked at Mullard in Mitcham. Reeves fitted KT88 and became famous because of them.
The Mullard works PA system was the perfect example of just how desperately bad the EL34 is and what NOT TO DO!
The system had 16 PPP EL34 running at 800V 24/7. Not a single day went by without them having to pull yet another melted EL34 out of it.
This persisted for ages until some helpful "sod" said 'nuffs enuff and bunged a dozen KT88 in there whereupon the carnage ceased....
The works director didn't find out for a week or so, then was incensed to find his suicide PA system had been made reliable with their direct competitor in Hammersmith....and insisted on their immediate replacement which started the tedious nonsense all over again.
Basic thing, if you insist on running EL34 at 800V you have a suicide wish, and that's official from Mullard, so it's also why they stopped promoting that rubbish.
Last edited:
yep of course he did
I have seen it with my own eyes
Drive a lambda EL34 amp with a square wave and watch as the screen grids light up like a light bulb, - white hot in seconds.
The secondary emissions of that design are uncontrollably bad, and rely simply on the transient/push pull amp duty cycle to stop it spiralling into catastrophic failure.
Once the electrode structure distorts a little bit it's going to die.
(same on KT120 btw, which is a badly overbloated Russian made pile of junk).
If you look at the aligned structure of the 807 and 6L6 it's simple to understand how they can soak up masses of punishment without any worry.
The same square wave test on 807 or KT8C simply results in a red anode and if measuring the screen current, appears absolutely fine even at a massive Va of 800-900V.
I can run that test for 30 mins much as the tubelab tortures TV scan tubes, and they keep on going without blowing up.
That's why they used the 807 as a TV line scan tube. It works and doesn't melt.
In fact there is no convincing reason ever to use 807 any more.
NOS octal 6BG6 are cheap as chips in the USD5 range, same as the soviet ones.
Ever seen EL34 as TV scan tube?
Of course not it would last about 5 mins,-
so they made PL38/ PL500/PL504 etc stuff which are altogether different beasts like the 6AV5 in the USA which is known to survive 40W on the anode.
'nuff said here, Reeves was right 400W (more actually really 300W) from 6 x KT88 is quite OK especially when paired with Partridge's wonderful transformers.
I have seen it with my own eyes
Drive a lambda EL34 amp with a square wave and watch as the screen grids light up like a light bulb, - white hot in seconds.
The secondary emissions of that design are uncontrollably bad, and rely simply on the transient/push pull amp duty cycle to stop it spiralling into catastrophic failure.
Once the electrode structure distorts a little bit it's going to die.
(same on KT120 btw, which is a badly overbloated Russian made pile of junk).
If you look at the aligned structure of the 807 and 6L6 it's simple to understand how they can soak up masses of punishment without any worry.
The same square wave test on 807 or KT8C simply results in a red anode and if measuring the screen current, appears absolutely fine even at a massive Va of 800-900V.
I can run that test for 30 mins much as the tubelab tortures TV scan tubes, and they keep on going without blowing up.
That's why they used the 807 as a TV line scan tube. It works and doesn't melt.
In fact there is no convincing reason ever to use 807 any more.
NOS octal 6BG6 are cheap as chips in the USD5 range, same as the soviet ones.
Ever seen EL34 as TV scan tube?
Of course not it would last about 5 mins,-
so they made PL38/ PL500/PL504 etc stuff which are altogether different beasts like the 6AV5 in the USA which is known to survive 40W on the anode.
'nuff said here, Reeves was right 400W (more actually really 300W) from 6 x KT88 is quite OK especially when paired with Partridge's wonderful transformers.
Last edited:
It's cheaper and a lot easier to drive PPP 7189/6P14P-EV than EL34.
The combined gain is higher, and they cost very little.
The Russian one is very good.
The EAR Yoshino guy did just that. (but I don't like his stuff)
Parallel pairs take 1.5A heater (same as EL34), runs in the 300-400V range, makes a combined 28W pentode, and has much lower distortion, particularly IMD.
I get 50W totally clean out. of a very similar set up running Va at 460V.
The sound with minimal NFB is phenomenal.
In effect it's a 100W linear amp (you can parallel the 2 x 50W channels) stuffed into a tiny space, and doesn't use UL at all.
I use it every day and shows no sign of blowing up no matter how hard you try.
The 250VA PSU transformer was specially wound for me with 6 different secondaries, by Toroidy Poland, and cost 120EURO delivered by courier.
There are seperate PSU rails for AF amp/g2, and the main anode HT, which makes a huge difference, (a trick learned from good 1940s amps and SW TX.)
We did a shoot out with some JBLs and Vand speakers that against a very good pair of 60W 807 amps.
The 807 sounder really coarse and harsh compared with the beautiful little micro pentode amp which was 1/3 the size and 1/4 the weight.
We compared it with a good solid state amp, and it sounded more musical less clinical than that, esp biwired.
In fact it's so light it goes in my checked airline baggage.
The combined gain is higher, and they cost very little.
The Russian one is very good.
The EAR Yoshino guy did just that. (but I don't like his stuff)
Parallel pairs take 1.5A heater (same as EL34), runs in the 300-400V range, makes a combined 28W pentode, and has much lower distortion, particularly IMD.
I get 50W totally clean out. of a very similar set up running Va at 460V.
The sound with minimal NFB is phenomenal.
In effect it's a 100W linear amp (you can parallel the 2 x 50W channels) stuffed into a tiny space, and doesn't use UL at all.
I use it every day and shows no sign of blowing up no matter how hard you try.
The 250VA PSU transformer was specially wound for me with 6 different secondaries, by Toroidy Poland, and cost 120EURO delivered by courier.
There are seperate PSU rails for AF amp/g2, and the main anode HT, which makes a huge difference, (a trick learned from good 1940s amps and SW TX.)
We did a shoot out with some JBLs and Vand speakers that against a very good pair of 60W 807 amps.
The 807 sounder really coarse and harsh compared with the beautiful little micro pentode amp which was 1/3 the size and 1/4 the weight.
We compared it with a good solid state amp, and it sounded more musical less clinical than that, esp biwired.
In fact it's so light it goes in my checked airline baggage.
Last edited:
lol where this thread is going!
Ive seen a few hybrid triode/ tetrode/pentode amps out there with a switch.
I was looking at the possibility of adding a switch to change between tetrode mode and UL mode.
I have not seen anyone do this in a hi-fi setting, but I know there are guitar amps out there that do this.
Ive seen a few hybrid triode/ tetrode/pentode amps out there with a switch.
I was looking at the possibility of adding a switch to change between tetrode mode and UL mode.
I have not seen anyone do this in a hi-fi setting, but I know there are guitar amps out there that do this.
There are countless threads on Tubes / Valves (Hi Fi amplifiers) with schematics that have switches for UL - Pentode, UL - Beam Power, and UL - Beam Tetrode settings.
And Tubes / Valves has countless schematics with switches for UL / Beam Power / Beam Tetrode,
to Triode wired: Pentodes, Beam Power, and Beam Tetrode tubes.
I do not look at the Instruments & Amps threads (Guitar amps), so I have no idea about how many of those use such switches.
A real good rule for such switches is to be sure the amp is off, and cold (cold filaments, and discharged B+) before changing the position of that kind of switch.
Sometimes the switches are only rated for 250V, and used with 500V B+.
Insulation voltage rating is one thing, but the contacts that are opening may arc due to the very fast transient caused by the less than perfect leakage reactance of the output transformer.
(No output transformer has zero leakage reactance).
Your Mileage May Vary
And Tubes / Valves has countless schematics with switches for UL / Beam Power / Beam Tetrode,
to Triode wired: Pentodes, Beam Power, and Beam Tetrode tubes.
I do not look at the Instruments & Amps threads (Guitar amps), so I have no idea about how many of those use such switches.
A real good rule for such switches is to be sure the amp is off, and cold (cold filaments, and discharged B+) before changing the position of that kind of switch.
Sometimes the switches are only rated for 250V, and used with 500V B+.
Insulation voltage rating is one thing, but the contacts that are opening may arc due to the very fast transient caused by the less than perfect leakage reactance of the output transformer.
(No output transformer has zero leakage reactance).
Your Mileage May Vary
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.