2A3 Design Issues (Long)

[JPeitzman]

First I would like to start by saying a very warm thank you to all that helped and contributed to my 300B build thread. Unfortunately thanks to school and a financially draining / very active summer the project never really got off the ground. Passives where bought, choices where made, chassis where designed, but the amp never came to be . I quickly realized that 300B amplifiers can be a very costly undertaking, especially with a depleted junk draw and no 300B's around. I did however learn a lot and plan on revisiting this project next summer. I am keeping my fingers crossed for a higher paying internship and more time. Again thanks to all .

I was looking over my stock of tubes and dwindling stock of parts and realized however I do have the parts for another project that I have been itching to built. Really this should have been the project I tackled first over the 300B. I have always lusted for a 2A3 amp and I realized that I have the parts minus some iron to do one. I am planing on keeping these very small and cute, they are going to be bedroom / private listening amplifiers. I have a wealth of driver tubes but unfortunately due to an accident involving 20+ years of bad storage and a "guard cat" I now only have one RCA 2A3 . Luckily I also like the 2A3 idea because a pair of Sovtek 2A3's costs about as much as one 300B on sale... a pair of generic Chinese 2A3's costs about as much as a NOS 6SN7.

So I made the executive decision that this will be simple, small, and 100% open to future upgrades...basically an experimenters amp. I have poured over 2A3 schematics and decided just to do a simple RC coupled 2 tube amplifier. I have taken a lot of inspiration from Bottlehead and decided to use 12AT7's for the inputs. This is an area though where I am open, I have also considered the 6SL7. This will be coupled through a 0.22uF capacitor to the 2A3. The 2A3 is run at it's, I call classic, data sheet operating point: 250v @ 60mA with -45v Bias & 2.5k Load.

As of now the B+ will be 295v - 300v provided by a 600vct Hammond 372DX Transformer. This will be rectified by a RCA 5U4G then through a C-L-C-R-C filter. The L is a 20H Hammond 193C and the R is a 100R wire wound. At a later date if I desire the R can be removed and replaced by a Hammond 5H 158Q Choke without really changing the B+...though PSUD II showed that the C-L-C-L-C was a better PSU compared the ...L-C-R... The C's are in order; 4.7uF, 47uF, 100uF.

I am having an issue though with calculating the 12AT7 operating point. Reading though some information of different sites, forums, and the data sheet, it seems that about 180v - 190v @ 4mA - 5mA is recommended. I am hoping in the future to use a C4S or other CCS but for now just a resistor. By my calculations I determined roughly 30k for Rp and 500R for Rk. Recently though I read a post on the Bottlehead forum that an Rk of 2.5k to 3k is correct for about 4mA, though this is using the C4S at the load. Am I doing my calculations completely wrong?

Couple other real quick questions:
1. Is a grid stopper of 1k or so needed? I have seen it in many designs and most have found it to be a good preventative idea.
2. Is 220k for the grid leak an OK value? I would like to use a grid choke, but that is for another day. I honestly don't know how to calculate this value, please help to enlighten me.
3. Does 100uF sound OK for the Ck on both the 2A3 and 12AT7? I am hoping in the future with a CCS mod and an Ultra-Path (maybe Parafeed) mod these will be avoided, but no such luck for now.

For right now I am planing on using the Hammond 1627SEA output transformer. I was however looking over some sites and realized that for about $10 more total I can get some Electra-Prints. They would be wound 2.5k & 65mA, is this OK or is it pushing it to much, worried about saturation. Should I just stick with Hammond's since they don't seem to sound to bad and they seem to be tough as nails(?) Was also seeing that for $50 more total I could get a pair of Magnequest TFA-204's. Has anyone ever used these in a 2A3 design, and if so how where they? I know I would have to change my operating point and everything since these are 3k and rated for right at 60mA. Otherwise to get this off the ground fast I may just pick up some Hammond 125DSE's that I can later recycle. Anyone with experience, do the higher priced Hammond's sound that much better than the cheap universals?

Just an FYI, this project may be a bit slow to progress and I may be a bit slow to respond. I am currently back in school and that gets about 90% of my awake and coherent time. Again I am sorry to those that where following the 300B build, it will get off the ground one day.

Pre-Thanks for any help / advice

Cheers

James

[kmaier]

Hi Jim,

I actually prefer 3.5K loading on the 2A3... and running them a bit higher voltage too. They sound better and you can get a solid 4-watts output. The TFA-204 are nice units... using these in a 2A3 stereo amp powered by a Hammond 302AX which has a pair of 2.5VAC CT filament windings. I wouldn't use the 125DSE for a 2A3 (maybe for a 45) but the 125ESE is decent iron for the money.

Hard to give you much feedback without a schematic regarding bias points versus intended use. I also use a 12AT7 in my design... first stage classic common-cathode amplifier direct-coupled to the second stage as a cathode follower running around 3ma of current. More than happy to send you PDFs of my design... if you think you could use it.

Sorry to hear about losing the RCA 2A3... I have numerous pairs... and some odd ones which don't match to the others.

Regards, KM

[Eli Duttman]

Jim,

I'd avoid the 'T7 as a voltage gain device in SE amps. The 'T7 has a HD character that makes it terrific in PP amps, but is likely to push things WAY over the top in SE stuff. The 6GK5 is a more linear type that also provides the highly useful combination of high mu, high gm, and low R_P.

It has become quite obvious that Edcor is a very good source of decent "iron", with modest price tags. If you can handle the $83.50 cost per trafo, Edcor's model CXSE25-8-2.5K would work well in this project and allow you to try UL/SE EL34s in the future. Unfortunately, they don't show a model in the very low cost XSE series, with an appropriate primary impedance. Still, you can make an enquiry.

[kmaier]

Hmmm, "the T7 has a HD character"... care to elaborate on that?

Regards, KM

[leadbelly]

Originally posted by Eli Duttman
Unfortunately, they don't show a model in the very low cost XSE series, with an appropriate primary impedance. Still, you can make an enquiry.

Wasn't the consensus of the thread on the GXSE series introduction that they were likely just as applicable for hifi as for MI amps? The GXSE line has 2.5K models.

[Andrewbee]

the T7 has a HD character

I think Eli is speaking of harmonic distortion (spectra), in particular the even type which with the 12AT7 as the driver would make it over the top with the even harmonics on a SE output.
I know he likes it (12AT7) with PP for the that very reason as the PP cancels the even harmonics out.
I am sure Eli will correct me if I am wrong.

[chrisb]

Originally posted by kmaier
Hmmm, "the T7 has a HD character"... care to elaborate on that?

Regards, KM

I think Eli is referring to the T7's Harmonic Distortion spectral character - it has just enough 2nd to nicely warm up the sound of a P/P design that might otherwise lean towards higher odd-orders.

Just again this weekend, I had the opportunity to play my now well beat-up "El Cheapo" inspired power amp at a DIY event. Aside from eventually running out of power ( as configured in Triode mode due to the Scott output transformers, it's in the general neighborhood of only 5wpc), it did not embarrass itself when compared to amps with much higher putative power, and certainly many times the mass and cost factor.


edit:

Andrew beat me to the send button - obviously Eli / Jim McShane's message on this particular subject has been well received.

[kmaier]

Well, maybe... I guess if you don't bias it right, still a fair number of respectable SE amps use it. OTOH, (looking at the plate curves) I wouldn't pick the 6GK5 for a class A stage that needs to deliver over 100 volts pk-pk. My preferred twin-triode in most cases is the 12AU7 but the recent DIY design (a different topology) needed more gain and costs were a factor. The 12AT7 performs well, specs well and most importantly, sounds good, but that's my $0.02.

Regards, KM

[JPeitzman]

Thanks to all so far. I will work on a schematic as soon as I can. The only good schematic sketching program I have though is a pen and paper. I can try to whip something up in paint unless someone want to put in their $0.02 for a cheap (read FREE) program.

The CXSE25-8-2.5K was actually on my list of transformers, I just wasn't sure how it compares with the Hammond. I know Tubelab did a comparison using the little and cheap versions of the Edcor's and they didn't test to bad.

I will have a look at the 6GK5. The idea was that I had most of the parts already for this one, but AES has the 6GK5 on sale for 50% off so $3.50 is tough to beat.

Off to some homework...Cheers

James

[Sheldon]

Originally posted by Eli Duttman
The 'T7 has a HD character that makes it terrific in PP amps, but is likely to push things WAY over the top in SE stuff. The 6GK5 is a more linear type that also provides the highly useful combination of high mu, high gm, and low R_P.

If we are talking about a single gain stage, then the 2nd harmonic of the T7 would subtract from the 2nd of the 2A3, as it inverts phase. Not saying that would taste better or worse than the GK5.

Sheldon