The power valves of a power amp must be EL34.
The amp will be Hi-Fi class and the valves will always operate within manufacturer's specs, no tricks.
The topology will be push-pull. Ultralinear or not will be decided at a later stage.
The output power must be least 100 Watts and preferably higher.
How to proceed from here?
* I thought of paralleling 4 x EL34 per signal side, ie 4 x EL34 for the positive half-wave and 4 x EL34 for the -ve half-wave, 8 valves in total per amp.
* My concern is that I haven't seen this arrangement of 4+4 valves somewhere and the only relevant design and build I've done is the trivial 2+2.
* Are there any problems, like stability which is my main concern, in paralleling so many valves?
or some other thing that should be taken into account regarding 4 valves operating in parallel?
The amp will be Hi-Fi class and the valves will always operate within manufacturer's specs, no tricks.
The topology will be push-pull. Ultralinear or not will be decided at a later stage.
The output power must be least 100 Watts and preferably higher.
How to proceed from here?
* I thought of paralleling 4 x EL34 per signal side, ie 4 x EL34 for the positive half-wave and 4 x EL34 for the -ve half-wave, 8 valves in total per amp.
* My concern is that I haven't seen this arrangement of 4+4 valves somewhere and the only relevant design and build I've done is the trivial 2+2.
* Are there any problems, like stability which is my main concern, in paralleling so many valves?
or some other thing that should be taken into account regarding 4 valves operating in parallel?
More than stability, think in PSU. Each tube draws about 100mA, so 8 of them will be 800mA. If stereo, 1600. So a big traffo will be needed. And powerful filter caps. Please, specify the input sensibility, the EL34's need about 20V pp to full drive, and if you need NFB you will need several times the gain to make proper NFB schema.
Thanks for the reply.
PSU isn't a problem.
The valves will operate in class B push-pull, max current is 400mA, quiescent current needs to be determined but will be in the region of ~5-10mA.
The input sensitivity will be the typical of power amps, around 1V rms, similarly the input impedance will be ~22-47kΩ.
Of course there will be more valves in the signal path to properly drive the EL34 push-pull.
Anything subtle to pay attention to?
I haven't seen any schematics with so many valves in parallel, I suspect there will be some difficulty at some point.
PSU isn't a problem.
The valves will operate in class B push-pull, max current is 400mA, quiescent current needs to be determined but will be in the region of ~5-10mA.
The input sensitivity will be the typical of power amps, around 1V rms, similarly the input impedance will be ~22-47kΩ.
Of course there will be more valves in the signal path to properly drive the EL34 push-pull.
Anything subtle to pay attention to?
I haven't seen any schematics with so many valves in parallel, I suspect there will be some difficulty at some point.
For class B operation you will need a good PSU with good regulation (regulated prefereably), and low impedance source: a triode transformer coupled or better, from a cathode follower or still better, a WCF. Gain, I suppose that 1000 times is sufficient, but 5000 or 10K is better. A pentode or cascode input plus some intermediate stages will do the job I don't know about commercial trades, but google for an interesting amplifier project by J Ross McDonald for his 807's triode wired. DC coupled to the final stage and auto biasing.
Is this for listening or to AM modulation?
Is this for listening or to AM modulation?
In attach a photo of an old project made by me years ago.
I found a good results.
The main thing id the use of a strong stage to drive the grid of the all EL34.
I used a 6SN7 in this one but a 6H30 or 6N6 is also fine .
The filter cacaitor in this ace are four of 470 uf at 500 Vdc, a good reserve of energy.
Walter
I found a good results.
The main thing id the use of a strong stage to drive the grid of the all EL34.
I used a 6SN7 in this one but a 6H30 or 6N6 is also fine .
The filter cacaitor in this ace are four of 470 uf at 500 Vdc, a good reserve of energy.
Walter
Attachments
I found when I built something similar that I needed to use 27R plate stoppers to quench oscillations.
Any reason the tubes MUST be EL34?
Any reason the tubes MUST be EL34?
Wow, here is a lot of prereq's before even starting design. Why ?
100 w can be done with 2 EL34 ( on the limit, 2+2 will be convenient)
But why limit to EL34 ? 6550/KT88 is a more powerful tube) 100 w is
achievable in pentod connection with 2 tubes/ channel, 120w with 4 tubes.
As for the number of tubes, paralleling tubes needs matched tubes. The more
tubes in parallel the more risks of some tube failing ( taking others with it
in the death).
Have a look at tubes4hifo M125 where 125 w by the help of 4 6550. An older is dynaco Mk6 .
Last edited:
You can get 100W with pair of EL34, if you go for Ub around 790V. Old Tesla amps use this way, works fine, but you need very good EL34 with serious sockets.
kodabmx, only EL34 are available.
I plan to use grid stoppers everywhere, the plate stoppers will be in parallel with some inductance, if needed.
petertub, the M-125 monoblock is a 2+2 KT88 valves, like many others.
EL34 are available in quantity.
Also I would object using regulator valves (eg 6L6s and derivatives) in a (aspiring) Hi-Fi power amp. The only good thing with 6L6s is the 150V filament to cathode spec.
Please be more specific about tube matching: when 2 or more tubes are "matched", when Ip is within 1%, or 5%, more? I think cathode resistors will smooth any discrpancies.
baudouin0, you guessed right!
zdenoeddie, longevity matters, especially for a (probably) commmercial prototype, everything has to be within manufacturer's max ratings, ie 250V and 100mA and 25W max plate diss.
scottpetersen, please define 'too hard'.
I plan to use grid stoppers everywhere, the plate stoppers will be in parallel with some inductance, if needed.
petertub, the M-125 monoblock is a 2+2 KT88 valves, like many others.
EL34 are available in quantity.
Also I would object using regulator valves (eg 6L6s and derivatives) in a (aspiring) Hi-Fi power amp. The only good thing with 6L6s is the 150V filament to cathode spec.
Please be more specific about tube matching: when 2 or more tubes are "matched", when Ip is within 1%, or 5%, more? I think cathode resistors will smooth any discrpancies.
baudouin0, you guessed right!
zdenoeddie, longevity matters, especially for a (probably) commmercial prototype, everything has to be within manufacturer's max ratings, ie 250V and 100mA and 25W max plate diss.
scottpetersen, please define 'too hard'.
Each pair of tubes are producing less than 40W I have seen the amp, and the outputs were still original after 30 years, and not cooked.
Just add a fifth pair and go Class A. A 600R a-a load to speaker will do.
There is no reason to do a B amp with that many tubes and that power goal.
cheers,
Douglas
There is no reason to do a B amp with that many tubes and that power goal.
cheers,
Douglas
I built two monoblocks, 6 EL34 PP 120w Class AB1 using Hammond 1650TA OPT's I think. Mains power tfmrs are 650VA toroids, self wound, & over wound, so you'll need 1 x 1000VA or two 500VA at least. HT designed for 1A at 450v loaded (each amp runs at about 650mA at full whack) 500v unloaded. Filter is 1100u ish, no L. Biased at 35mA, fixed bias. Sensitivity 1.5v, though I could have made this smaller. See - 120w PP UL monoblocks
Some good ideas here - 100w-monobloc3-2014
Problems I remember running into - oscillation, so PT has some good solutions for this. Took me a while to design and build a decent driver/LTP, agian see PT's ideas. A protection circuit is a must, see PT's circuit for this. Layout is important, get it wrong, the valves lift off. Also heat is a problem, the chassis' get hot as does the mains tfmrs
Good luck, Andy.
Some good ideas here - 100w-monobloc3-2014
Problems I remember running into - oscillation, so PT has some good solutions for this. Took me a while to design and build a decent driver/LTP, agian see PT's ideas. A protection circuit is a must, see PT's circuit for this. Layout is important, get it wrong, the valves lift off. Also heat is a problem, the chassis' get hot as does the mains tfmrs
Good luck, Andy.
Last edited: