Designing an EL34 Class A Stereo Amp

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

Kid Squid

Well-known member
Joined
Jun 4, 2004
Messages
1,011
Location
Port Toilet, South Wales
Hiya Chaps,

Just been given 4 Mullard EL34's by a friend. Just tested them, and they are all good (AVO Mk3).

I've always wanted a nice Valve hifi amp, and I think now is the time to have a go at building one !
Thinking of running two EL34's per side in parallel class A.

anyone here done the same?, looking for some tips/ help in making this little boyo get off the ground.

what sort of O/P transformers, input valves - ecc88 / ef86 etc, will i need a I/P transformer ? this sort of thing.

also, I want to use some real good components - price not an issue !!

I'll have a go at a chassis tonight.

Steve :thumb:
 
If your planning a PP output stage, here is a schematic for the Knight Kit 60 Watt stereo basic amp:

http://www.triodeel.com/images/kb85-2.gif

This is an ultralinear design with an ec86 as the input preamp and a 12ax7 in a nice differential phase splitter.

If you are thinking single ended, then the Loftin-White circuit is my favorite:

http://www.ispra.net/audio/DCSingleEndedAmplifier/index.php

Of course, in this circuit, you will operate the EL34 triode connected and you will be getting about 12 Watts from the pair. Your output transformer will need to accommodate DC so you will need a gapped core. Audio Note, Hashimoto, Hammond, Magnequest and some others make very good single ended outputs.
 
for outputs if money isn't an ishue go for luhndall's ;)
maybe edcor is also good. i didn't test them yet.

if you have 2 pairs you could go for parallel single ended
and the in triode.
and keep it simple!
when you parallel the tubes your impedance of the tube halves and if you use triode mode id gets even lower.
ofcourse the power becomes less.

tdsl says for one tube in SE that you can get like 6w (@ 8%thd) so 4-5w rms would be more likely correct ;)
now thats for one tube so when you use a pair you will get like 8-10 w rms pure class A triode. and really that can be enough.

tdsl says for one tube the impedance has to be like 3k and the B+ is 375v
the current running trough the tube is 70mA and it is cathode bias with 370ohm so 25.9v is lost there that gives a dissipation of 24.437watts. we have to know that the second grid dissipates some part of this! so don't be scared that the anode dissipation is only 25watts
in triode it is a little more and i looked in a datasheet and they say that dissipation without signal is 27.5watts so it will hold.

for 2 tubes you will have

B+375v
140mA current (70mA per tube).
1k5 primary impedance. (you will need a gapped output transformer for a current of 140mA! that isn't that cheap tough)
output power is like 8-10w rms

you could also use the tube in penthode. but you have to lower your B+ and i think you wouldnt benefit that much powerwise. it will probably get you 15w or something.

as input tube it depends on how much gain but the 5670 or 2c51 i have heard in hifi amps sounds pretty good. you could also use the 6sn7 (or 12sn7 beceause hifi freaks have lifted the price for 6sn7's).
for triode you need a swing of 18,9v. take 19v for instance.
so a simple good preamp tube will do. maybe lookout for miller.

then if you want it balanced for studio use or something i recommend a transformer. 1-1:1 or something. after that you could use a pot or attenuator before the grid of the first tube.

then for supply you will need like 375v
use a CLC filter! the use of an CLC filter works great with SE amplifiers.
so we have got like 375vdc on the plates after that filter.
let see that will be a transformer of around 270v and i think it's quite standard. of course that is if you use solid state bridge rectifier but i think you should but use some good diodes. i use avalanches diodes last time i think. they don't have much switching noise. and don't use a big capacitor after the bridge! use like 47µF or maybe 100µF beceause in total you will be needing 140mA per channel and thats enough so 100µF won't hurt. after that use an L of 10-20H yes that big. and it has to be minimum 140mA but probably 160mA is better to have some reserve.
now as a capacitor after the L you can use a bigger one. the B+ will come in slowly beceause of the big C afther the L and the smaller one just after the bridge doesnt make high current spikes so those spikes we'll be small and will not make much trouble. the C i mostly use after the L is 470µF
then after that als make sure u use 100nF c's over the elco's in that way the spikes that are left are surely out of the way. i alway build a supply that way and it works great.

then after that you can use an RC filter where you can calculate the R by nowing how much B+ you need on the preamp tube and how much current will be drawn. as a C for that filter use something like 47µF.

so thats the supply. i would uce AC heaters simple and just use a 100ohm wirewound 5w pot to adjust if you have hum. works perfect.

hmmm gotta go know but i will update with a schmatic if you want. and it will be readable this time :)
check out the site from luhndall maybe ;)
i think the LL1627 or LL1623 can be an output.
the 140mA current is a lot the 1627 cant take it but the primary is only 1k2 so a little low maybe but in triode it should work.
the 1623 has the problem that it can only hold 120mA so the bass responce will be less. i would go for the LL1627.

for an L you can use a LL1638 or if you find something similar.
you should use 2! one for each channel.
if i had to choose i would use a seperate supply for each channel and maybe even a seperate mains transformer but thats maybe over the top.

speaking of mains transformer. thats like 280mA dc current only fot the output or more or less you will need 400mA @ 270v
for heaters you have 4 tubes that draw 1,5A or 6A! @ 6.3v
so roughly thinking it will be a mains transformer around 200w :)
look maybe at hammond or so they have more choice then luhndall for that.

succes, and try to keep it simple and use good components.


btw, you will get like 20w of power in total and you will be using like 150w of power :p
class A you where saying ;)

thats like 13% of usefull power :p
 
Good Lord, Kid, don't build an amp, sell those on ebay!

Whats the going price for a pair of Mullards?

$650?

so you have about $1200 after taxes, enough for a years supply. :razz:

Please don't say metal base in less you want me to puke!
 
Top Stuff Chaps !!

Got chassis sorted on nightshift last night :grin:



if you have 2 pairs you could go for parallel single ended
This is what I was thinking about !


and the in triode.
oh yes indeed :grin:

and keep it simple!
Please do !


for 2 tubes you will have

B+375v
140mA current (70mA per tube).
1k5 primary impedance. (you will need a gapped output transformer for a current of 140mA! that isn't that cheap tough)
output power is like 8-10w rms

sounds real good to me :grin:
had a look on the sowter website for a suitable output transformer, found this - SE15-s

2x EL34 /KT88 SE,

Pri Z - 1.5K,
DC mA - 180 - yeah bwoy !,
P - 30W,
Plate Res - 1500 ohms

price £136.37 ......................

With regard to the HT section, i was thinking of using a valve rectifier, over a solid state rectifier, GZ34 ?? what do you think ?

regarding the choke - I'd love to have a go at winding my own ( can get my hands on a small winder, but i havent got a clue where to start.....yet!!)

I've got four 6sn7's so I 'd love to use these.
for triode you need a swing of 18,9v. take 19v for instance.
is this the voltage, that will be driving the EL34's ?
what am I looking for with regard to the miller capacitances ?....or avoid !



then if you want it balanced for studio use or something i recommend a transformer. 1-1:1 or something. after that you could use a pot or attenuator before the grid of the first tube.
Didn't think about this, but the more I think about it - the more I like the sound of it :grin: , maybe have Balanced and Unbalanced I/P's.

i would uce AC heaters simple and just use a 100ohm wirewound 5w pot to adjust if you have hum. works perfect

Agreed.

btw, you will get like 20w of power in total and you will be using like 150w of power :p
class A you where saying ;)

thats like 13% of usefull power :p

Why Not ?? :grin: :guinness: :thumb:

hmmm gotta go know but i will update with a schmatic if you want

That would make my day LazyOne !!


kubi Posted: Tue Jan 02, 2007 6:25 pm Post subject:

--------------------------------------------------------------------------------

30W Push Pull amplifier designed by Claus Byrith, Royal Academy of Music in Aarhus, Denmark: http://www.lundahl.se/claus_b.html

Even if you don't want to build it, the PDF is worth to read.
Thanks for the Link Kubi :guinness: :thumb: :guinness:



solder_city Posted: Tue Jan 02, 2007 6:26 pm Post subject:

--------------------------------------------------------------------------------

hey, youre british, why not build a Williamson?

its still one of the best valve hi-fi amps ever designed.
clicky
been thinking SE paralell, Just because I wanted something a little different ! :wink:

Wavebourn Posted: Tue Jan 02, 2007 7:50 pm Post subject:

--------------------------------------------------------------------------------

Kid Squid;
Altec Lansing 1569A is very good prototype for your amp. ;)
just looking over the drawings as we speak - a little bit too much power, also Class AB, I want some Pure Class A mojo :grin:

CJ Posted: Wed Jan 03, 2007 1:23 am Post subject:

--------------------------------------------------------------------------------

Good Lord, Kid, don't build an amp, sell those on ebay!

Whats the going price for a pair of Mullards?

$650?
I F**kin wish, Bro :thumb:

Please don't say metal base in less you want me to puke!

You're safe Old Bean - ........just lowly old Black plates.... :thumb:


Steve :thumb:
 
[quote author="Kid Squid"]

Wavebourn Posted: Tue Jan 02, 2007 7:50 pm Post subject:

--------------------------------------------------------------------------------

Kid Squid;
Altec Lansing 1569A is very good prototype for your amp. ;)
just looking over the drawings as we speak - a little bit too much power, also Class AB, I want some Pure Class A mojo :grin:
[/quote]

Every kid knows that for pure class A you need less of grid bias voltage. :wink:

...also you can strap 2'nd grids to anodes through 1 KOhm resistors for quazi-triode mode.
 
Wavebourn Posted: Wed Jan 03, 2007 11:56 pm Post subject:

--------------------------------------------------------------------------------

Kid Squid wrote:


Quote:
Wavebourn Posted: Tue Jan 02, 2007 7:50 pm Post subject:

--------------------------------------------------------------------------------

Kid Squid;
Altec Lansing 1569A is very good prototype for your amp. ;)


just looking over the drawings as we speak - a little bit too much power, also Class AB, I want some Pure Class A mojo



Every kid knows that for pure class A you need less of grid bias voltage.
:oops:
Sorry if I sounded a bit off, no offence meant
, But I'm still quite new , when it comes to designing things, thought I'd jump in at the deep end a little bit :thumb:


...also you can strap 2'nd grids to anodes through 1 KOhm resistors for quazi-triode mode.

Is a pentode / triode switch possible ?

Steve :thumb:
 
you could do pushpull in class A as well thats true but if you look at the spectrum even if THD is the same you see that a pushpull output cancells out even harmonics. this is good but the odd harmonics are still present and beceause the even harmonics are not there that much it will sound worse.
there is proof of this published in many papers.
but then again a pushpull amp can really sound good as well. and the output is cheaper thats true.
that doenst mean that your se has to have much thd! you can make an se amp that sounds good and measures good as wel.
use a little feedback (not to much!) but then again if you have a good output transformer it probably is quite good. you can do some feedback at the preamp tube and you can use those 6sn7 i think.
the 19v swing is indeed what the el34 needs.
the preamp has to swing it and you have a capacitance so the preamp tube has to drive it also. you got 2 tubes so the capacitance doubles but i think a 6sn7 will do just fine ;)
in fact one 6sn7 can drive the el34's stereo.
one halve per output section.
if you really want to go overkill the parallel the 6sn7's so that you need one tube per channel.
the 6sn7 has a µ of like 20 so it will do.
the rp=7700 and if you parallel that you will get like 3850ohm thats quite overkill! so it will swing without anny problem but really don't forget gridstoppers at the grid of the el34's ore the thing can oscillate as hell. maybe in triode it can go and indeed for triode put 1k to the 2nd grid coming from te plate.
in my opinion a single 6sn7 will do fine ;)

put 2 blocks together with good components and voila :)

if you want something more special there is something experimental i r have read in an elektor article audio special. it was using only one transformer at the output of a se stereo amp.

the used what io think was a 3 legged core and used the outer legs per channel. in that whay the dc current can't saturate anymore just like in a normal pushpull amp and you only have 1 output transformer. they did some tests and it worked perfectly well.
maybe a tryout :)

at this time i didn't have the time to make up a schematic yet beceause to less time but i will try. now i have to make a parallel single ended el84 amplifier. if there are people in belgium i would say come and have a listen if you want ;)

btw i'm just saying my experiences and things i have read i'm only 20 years old so i didn't grow up with tubes... so maybe there are some little faults in my texts ;)
 
I agree with LazyOne, symmetrical push-pull means a compromise between power consumption economy and a sound quality. My best sounding amps are single ended. Hmmm.... No, they are push-pull, but pull voltage - push current... Or push current-pull voltage? Never mnd, I did not invent yet a name to call them by names, but they definitely work in class A, though tubes drive output transistors. They measure excellend, and sound gorgeous. Symmetrical all-tube push-pull amps, as I mentioned before about Altec Lansing, sound less gorgeous, both in A and AB, but anyway even in AB they sound much-much better than best transistor AB amps of the same output power made with complementary voltage followers.

If you want a switch pentode-triode, you may add one switching grids (to anodes through 1K 5W resistors, or to power supply as shown on the schematics).
 
i've put something up now
it's basic and as simple it could get.
the only thing you need to do then is:
-dont forget the gridstoppers to be as close as possible to the grid (the 220ohm and the 2k2 resistors on the schematic ;) ).
-use a good power supply (need to draw that when i have the time for it ;) )
-use good components. dont use cheap elco's or C's it really makes a difference. and use a good output transformer. at the other point don't overdo it i mean hifi people are nuts and pay nuts prices ;) just good components is good enough.
-wire it good and with some logic it can make a difference. i mean leave some space. somethings get warm also. especially the kathode resistors for the bias (around 1.8w per el34).

altough it is just a first try. the anode resistor for the 6sn7 is just something i think will do. maybe needs some tweaking. also for the 4k7 in series wit the supply.

supply has to be 150mA (and 160mA is bette rfor reserver ;) ) and you need 375 at that current.
thats for one channel!
so probably you need 300mA of total DC current out of your power transformer thats gonna be like 400-450 mA AC i think. thats gonna be a big one ;).
for the supply i suggest like 100µF after the rectification and then 2 L's of like 10-20H (one for each channel) after that like 470µF per channel would do i think.
and then you should have a 2 channel supply thats good enough to power the 2 channels.

altough i didn't test the schematic so don't come knocking on my head if it goes up in smoke or something. and maybe there can be some faults but i checked and i think it is good like this.
el34se.png
 
There are some schematics and many people who have built both single-ended and push-pull amplifiers here: http://www.audioasylum.com/forums/KandK/bbs.html. Yeah, you can call them "audiophools", but some of them care a great deal about music, like me :grin: .

Whether you build a single-ended or push-pull amp, you are right to choose the best components, but perhaps even more important is to pick the right topology and tubes to execute that topology. One small piece of advice is to avoid medium and high Rp tubes in the driver position. The robustness of the sound will suffer when you use tubes that are not robust. Sorry to mention it, but a good example of the point is the use of the 6SN7GT to drive paralleled output tubes. Try a 6BL7GT instead to lower the drive impedance and get better sound. Another good idea is to use a transformer to couple the driver to the output stage. The transformer provides a better bias path for the output tube grid and allows for gain or attentuation, as needed. In a push-pull amp you can use the interstage transformer to do the "phase splitting" honors, as well, to sonic advantage over an active phase splitter.

Have fun with the project,
 
if he wants he can just parallel the 6sn7 (one 6sn7 per channel).
bu i think it's a bit overkill.
the 6sn7 has an rp of around 7700ohm and i think thats quite low. (for example the 12at7 that is so much used at driving el34's in guitar amp has an rp of 10000 going to 16000 at lower voltages.
so if you double the 6sn7 the rp will be like 3850 and i think it is low enough to drive el34's. but indeed experiment with tubes. the lower rp the "faster" the amp will sound beceause it can drive the capacitor that is between grid and kathode of the el34 better.
now i think of it. if you really want you could use the 6sn7 as an output for a practice amp giving around 1w (pp) or almost 0.5w in se.
have to try that one day :)
phase splitting with an interstage can be good BUT you have to find a good one and they are ahrd to fing. i think luhndall has some of them. but its hard to find a good one soin practice an active splitter does the job better.
good iron is hard to find....
 
Just FYI -- a 6SL7 under reasonable bias conditions is higher impedance than a 6SN7.

Your schematic looks pretty good to me. You might consider connecting the two output tube cathodes and using a single 190-ohm resistor with associated capacitors.

Peace,
Paul
 
the advantage of thuis is that the tubes can be a little mismatched. by using separated kathode resistors the'll bias each at the right level so the mismatch is sort of cancelled.
but indeed you could use a 190ohm resistor. it is less components indeed.
 
Also, I'd increased cost of the amp adding one more resistor from pot's wiper to ground. And an ultra-linear tap, for a bit more economy...
 
Thanks for the input on this chaps, it really means a lot :grin:

sorry I haven't followed things up sooner, but I've been working a pile of long shifts.

Lazyone, is there any chance I could get hold of the drawing of the circuit, I can start to have a go at building it ( got a couple of days off soon :fingers crossed !!)
the link is down.

Lazyone, Wavebourn, Kevin & Paul - nice one fellas :guinness: :sam:
 
will look at it a bit later at the evening. just got home from school doing an exam in power electronics..... scr's,triacs,diacs.... you know they talk like 1000A and 1kV like its nothing.....

i think i have the schematic saved so it will be probably up some time later ;)



ok should be fixed now ;)

i have a pair of beefy 6as7 tubes that are probably going to be used to build a pushpull amp. good for something like 10-15watts for class ab :p. needs some drive tough. it is triode and the rp is only 280ohms! quite special tubes you could say.
 

Latest posts

Back
Top