• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

2A3 amp part 1,2,3 (merged)

help with amp design

hi, i just designed a 2A3 amp, or should i say im still designing. i was wondering if someone could help out with resistor values (i think i got cap values down, but still wouldnt hurt if u guys could check), and im thinkin of usin a 1627SE Hammond OPT. would this work? also, a friend of mine keeps tellin me i need to bias the cathodes/filaments of the 2A3's, but cant tell me why its needed, would appreciate if someone explained why this is needed and maybe explain how its done (pot on the heater supply right?). thx. heres the schem.
 

Attachments

  • my amp se.gif
    my amp se.gif
    5.5 KB · Views: 2,155
Re: help with amp design

mig-ru said:
bias the cathodes/filaments of the 2A3's,

and the AU7s cathodes too...

A tube works because the grid is at a lower potential than the cathode. Your circuit has a ways to go before it will even work, but do keep working on it. The learnin' comes from tryin' :^)

Time to start digging around on the web or get a few elementary texts on tube amp design (ie Beginner's Guide To Tube Audio Design by Bruce Rozenblit OR Valve Amplifiers by Morgan Jones are 2 i find really useful).

a couple web links (and there are many more)

http://www.eecs.umich.edu/~mmccorq/tubes/introduction/introduction.html
http://www.mc-h.demon.co.uk/vtheory/vtheory.htm

and go out and look at other people's tube amp schematics. There are thousands of them (i'd actually bet that for every SS design you can find there are at least 10 tube amp designs).

dace
 
Re: Re: help with amp design

planet10 said:
...The learnin' comes from tryin' :^)...

That can get expensive... more expensive than I can afford...

Anyway, on a side note, I like your design... let me know by posting here what your final is... it is kind of what I am looking for. I wanted a good simple design for a tube amp, was looking for OTL to conserve money, but this is nice... and I always liked the 2A3's...
 
trespasser_guy said:
Anyway, on a side note, I like your design...I wanted a good simple design

The design looks simple because most of it is not shown... i had a nice post going but my browser did a rare blow-up and i lost it (and now i have to go cook supper). Fire up Sherlock (or go to Google or some such if you aren't lucky enuff to be on a Mac) and look for Fi Primer, 2A3 Free Lunch (and from that site check out the Darling to get to Bob Danielak's pages -- he has a nice 2A3 amp there). There are lots more.

I leave you with this map:

Kismet6922-2a3.gif


dave
 
2A3 amp part 2

ok i have been revising this amp, with some help from some friends which are probably annoyed to hell by me now. if u could take a look and comment on whether it will work or not, and what i should improve in either case. im not sure about the resistor and cap values but i think i got them fairly close, if you could double check. thx.
 

Attachments

  • my amp se single 2a3.gif
    my amp se single 2a3.gif
    7.2 KB · Views: 2,068
Re: 2A3 amp part 2

mig-ru said:
ok i have been revising this amp, with some help from some friends which are probably annoyed to hell by me now. if u could take a look and comment on whether it will work or not, and what i should improve in either case. im not sure about the resistor and cap values but i think i got them fairly close, if you could double check. thx.

I'm not good enuff at figuring tube circuits yet to comment on the values, but you seem to have all the necessary bits now. Do you have the power supply figured - in an SE amp it is at least as impotant as the rest of it.

dave
 
12AX7 is comletely incapable of driving 2A3. the final sound may depend more on the driver than the o/p tube. Have a look at low Rp triodes or triode-connected pentodes with either plate choke load or transformer coupled. 2A3 can produce extremely powerful sound with a proper driver. Of course if you want to stay with two stages the options are more limited.

regards

peter
 
You're definitely missing some things. First, you must have a load resistor for the 12AX7. Second, you need power into that load resistor (as was already pointed out.)

I'm not sure why you've got the cap, pot and switch (is it a switch or a jack?) in front of the 12AX7 grid resistor. If you want a volume pot, substitute it for the grid resistor. The input cap is not necessary unless your source has DC offset at it's output (most modern gear doesn't.)

Also, I agree with the opinion that a 12AX7 won't properly drive a 2A3. At a minimum, I'd use a 12AT7.

To get an understanding of how to design simple valve amplifiers, here is a great resource:

http://www.audioxpress.com/resource/audioclass/index.htm

For SET type amps, you should read parts 1 & 2 of the amplifier series, along with the power supply series.
 

Brett

Member
2002-01-07 6:02 pm
mig,

You might also want to take the k out of the 2A3 cathode resistor.

The 12AX7 won't have enough grunt to drive the grid of the 2A3, even if it might swing enough voltage. Try a 6C45pi or an EL84 triode wired, or an EL84 or C3m pentode. You're going to need about 110 Vp-p and a decent amount of current available. It's neccessary that the driver has a lot of headroom, so the whole amp doesn't clip at once. Of course a load and power to the driver will ensure you get any sound out at all.

Use a 1.25V - 0 - 1.25V heater transformer and return the bypassed cathode resistor to the transformer CT. Make sure your trans has an electrostatic shield.

You'll need a very clean power supply, say tube rectified, CLCLC, or better yet, LCLC.or better still one of these with a shunt reg.

Cheers
 
Mig,

agree with Brett, jeff mai and analog_sa.
ECCV83/12AX7 is definitely incapable to drive the 2A3.

To drive the 2A3 at -45V bias, your driver stage has to deliver clean and juicy 84V, better 90V peak2peak.

Let us assume your 0dB input voltage is 2V rms = 5.65 V p2p. So you need a gain of 16, make that 20 or 22, but not more.
The 12AX7 has µ of 100 and gain of about 50 when used with a resistive anode load. Too much and plate resistance of the 12AX7 is 65k, way too high.

You could try out 2 different options:
1)
if you can get a nice 1:1 (5-10k : 5-10k) interstage transformer, use a 5687 or a ECC99 (made by EI only) with a µ of 18 to 20 and a plate resistance of 2k. Juicy enough, minimum count of stages: 2.
2)
also use 5687 or ECC99 and have an active plateload, either a plate choke od a solid state constant current source.

In both cases, stage gain is µ of the tube.
As interstage transformers and plate chokes are expensivish, i would opt for the solid state CCS option.

Brett,
CLCLC: it may sound nice for sure. But technically, it is a mess. Hint: download the PSUD II from the duncanamps homepage. Toy around with this gorgeous toy. Simulate your CLCLC, :devilr: ... try out how long it takes this mess to recover from a load change. The more LC you add, the longer the load step response. Looking :sad: ... you'll see.
I use: LC or maybe cLC, meaning the 1st C after the (tube) recitifer is a very small one, the choke is big and the C after the choke is as big as i can make it with 1st class foil caps.

MHuber, if you read this, thank you for rubbing my nose into that! :)

mig again,
if you decide to go for tube rectification, consider that the tube rectifier has considerable voltage drop, not only 1.2 Volts like a SS Graetz bridge. And the LC filterning will let that voltage drop to about 0.9 times DC of the peak voltage the rectifier (bridge) is handing out.

Another thing, :att'n: 330V is too much for B+ (supply voltage). The 2A3 is specified for 250V of B+. At a quiescent current of 60mA the 750R cathode resistor drops 45V. 45 + 250 is 295 V B+. Give the output transformer another 5 V to drop.
So: if B+ is over 300V, you fry your precious output tube. Not at once, but you do.
 

Brett

Member
2002-01-07 6:02 pm
dice45 said:
Brett,
CLCLC: it may sound nice for sure. But technically, it is a mess. Hint: download the PSUD II from the duncanamps homepage. Toy around with this gorgeous toy. Simulate your CLCLC, :devilr: ... try out how long it takes this mess to recover from a load change. The more LC you add, the longer the load step response. Looking :sad: ... you'll see.
I use: LC or maybe cLC, meaning the 1st C after the (tube) recitifer is a very small one, the choke is big and the C after the choke is as big as i can make it with 1st class foil caps.
Bernhard,
I use PSUD but have never found the effect you have mentioned. I generally use low values in the first cap too (~<1u) and set the time constants to be different enough so as not to be a problem. However I will take your (and Manfred's) advice and do some more sim work. Using diff circuitry, current sourced and active loaded, with decent supply caps it hasn't been an issue for me. I concede it might be in a SET though, even though the SET mafia seem to love this arrangement so much as to make it a commandment.
Another thing, :att'n: 330V is too much for B+ (supply voltage). The 2A3 is specified for 250V of B+. At a quiescent current of 60mA the 750R cathode resistor drops 45V. 45 + 250 is 295 V B+. Give the output transformer another 5 V to drop.
So: if B+ is over 300V, you fry your precious output tube. Not at once, but you do.
Not if you use the Sovtek single plate 2A3/6B4G. They have a much higher plate rating than NOS ones. Something like 30W and 450V.

Cheers
 
Brett,
look in PSUD for the feature simulating load steps. Intimidating. I should not be that paranoid about it (Q is not whether i'm paranoid but whether i'm paranoid enough :) ), my power amp is a differential, a PP one and demands no load steps from the PS. but a SE amp definitely does.

Manfred tried it out, i trust him. He reported the amp with CLCLC sounds worse and IHO step response is the cause. He has a special gadget to trace messes in a naughty PS; I described it once to Jocko (copied it out of a lengthy thread):
###########
One thing i learned from Manfred: hook a headphone amp with differential AC inputs, high commonmode rejection and adjustable gain to B+ and GND of your amp's PS and let the amp under test play music into an appropriate load. And then listen to the music on the B+ ! Provided enough gain, there always is music. If the music is sounding distorted (the PS then facing a complex load and not being able to handle it), the amp is crap and will sound like crap. Dynamics are wrong then, too.
############
I have such a device on my wishlist for long and as soon as i have some time (if diyAudio lets me :) ), i will build one. To verify what Manfred told me: amps with a PS behaving complex instead of resisitive do not sing, do sound distorted in a very unpleasant way. As is with CLCLC.
 
I believe the 250V operating limit was originally set for the early single plate 2A3. I have seen a few (later?) data sheets listing 300V as the limit.

In addition to the high transconductance triodes and pentodes suggested as driver for the 2A3, I recommend trying a simple, plate loaded high mu triode.

On my 300B amps, I found I preferred medium - low transconductance pentodes (6SG7, 6J7) to the high transconductance pentodes (6AC7, 6AG7, C3m.) I found the high transconductance pentodes to have a glare that I couldn't get rid of. I had initially tried some arrangements with two stages of triodes (6SN7, 76) but immediately stopped once I fired up the amp with the 6J7.

With the 45 amp I'm using now, I did not care for any of the pentodes, easily preferring a simple medium - high mu triode (6J5, paralleled 6N7.) Using the 6J5, gain was marginal for my system (I don't use a preamp) but I think the 6N7 is fine with a gain of around 25. If you're using a preamp, either is fine.

Since a 2A3 is sort of in between (actually the gain requirements are similar to the 45) I think the simpler circuits are worth a listen and are certainly easier to work with than the high Gm stuff. I melted down a couple of 6AC7 because of parasitic oscillation.
 
Jeff,
you give good advice :up:
And you are right, 250V is for the original 2A4 single plate. If mig-ru uses Sovtek 2A3 MP, he has almost a 300B as far as rating is concerned, 25W plate dissipation and much higher B+. And they sound very good; Manfred Huber is using them in his amp.
The reason i am into high-G_m tubes is because i like the sound of choke-loaded amp stages. Here atleast medium G_m is needed, otherwise the lower BW limit goes to the dogs. I have no idea where to get plate chokes for 12AX7, noone i know has chokes that big in appropriate quality.
But with 6SN7 or E80CC, one can build fine choke loaded stages already.

With a CCS, this becomes another matter, a CCS can be adapted to suit any insanely high plate resistance :)
 
continuation of 2A3 amp design

hehe, again posting the updated design, if ud notice changed the 12AX7 for a 6SN7. now im wondering if i could use a 12AT7 instead of a 6SN7? if its going to significantly alter the design, i dont wanna mess with it. so here it is, critise and suggest improvement, hehe.
 

Attachments

  • my amp se single 2a3.gif
    my amp se single 2a3.gif
    7 KB · Views: 1,138
you could use a 6CG7 instead of a 6SN7.. they are supposedly the same tube in a different bottle.

AT7 has higher gain.. If you must use any of the 12A_7 series, AU7' could probably be the closest to 6SN7.

I dont know much about the drive required for the 2A3, but are you sure half of a 6SN7 will cut it? seems unlikely..