Push Pull 300B preliminary

[zigzagflux]

Hello tube lovers....

Considering a push pull 300B amplifier, something a little 'over the top', maybe. Want to try transformer coupling as well, and will probably aim for all Lundahl iron. The design I have attached is basically the K&K Audio PP input stage, with some minor changes. Would like some general comments on the schematic, along with some advice.

First, are the bias conditions for the SN7 gainstage appropriate (from experience)? I have searched for some good tutorials on choke loading, and haven't come up with much help, especially with a diff amp. The 150V will be provided through a CCS-loaded 0D3 regulator.

Second, some recommendations on a linear but very low Rp tube for the driver. My tendency is for an indirectly heated tube. Two stages of DHT is questionable for me. I hope to use AC heating throughout.

Thanks.

[Miles Prower]

Way too much iron for my liking in a design, although if you're gonna go that route, might as well get the best.

As for the driver, I don't think you can do much better than the 6BX7 in terms of g(m) and r(p). One of the most important thing about getting the best from 300Bs is to have adequate grid drive. The 6BX7 can certainly do that as well. (Although my preference for that would be power MOSFETs.)

[dsavitsk]

I don't think it's too much iron, but perhaps it is too many tubes? How about replacing the 12sn7 with a 5842, 5687, 7119, 6c45, 6688, 7788, etc. and leaving off the second tube? This set of articles might be useful: http://www.nutshellhifi.com/triode1.html

[AndreasS]

Originally posted by zigzagflux
..K&K Audio PP input stage, with some minor changes. Would like some general comments on the schematic, along with some advice.

First, are the bias conditions for the SN7 gainstage appropriate (from experience)? I have searched for some good tutorials on choke loading, and haven't come up with much help, especially with a diff amp. The 150V will be provided through a CCS-loaded 0D3 regulator.

Take a differential choke (a choke with two indentical windings and without gap) for the drivers, the CCS may be omiited (you have a input transformer).

Example:

http://www.jogis-roehrenbude.de/Lese...B4G/6B4GPP.htm

Regards Andreas

[Shoog]

.......the CCS may be omiited (you have a input transformer)

This statement is true, but it is so easy to leave the CCS in and it will help to correct any deficiencies in the input transformer.
Think of the 6H13P. Since you are driving both inputs of your LTP you will realise the full gain of the 6H13P (about 18x), with upto 40mA of drive current. With this you have a chance of eliminating some of that iron and the second driver stage. I did a similar thing driving 6080's which have similar drive requirements.

Shoog

[JimW]

I have built the amity in both 2a3 and 300b versions. If I were to do it over, I would opt for the three stage design, as the amitys, while they sound so very good, are not easy to drive. I also built a raven line stage to drive the 300b, the 2a3 I drive directly from a Mc tuner (has a low z output). The K&K driver would be good choice, as the 'n7 first stage would be easy to drive. Ken Carter can supply matched 6bx7 sections. I would be tempted to use a trioded 6v6 as a driver, as they are easy to get, sound good, even in new production. Both amity version I built are all lundahl, with the 300b being all amorphous (sp?) iron. Expensive - but sure does sound good.

[Shoog]

May I correct myself - I mean't the 6H10P.

Shoog

[andyjevans]

I just got some 300b so will be making either SE or PP. My experience with driving the 2a3 in push pull was that 01a direct coupled into a 26 run push-pull into the outputs with a LL1660 was the best sound as regards input tubes. With the 300b the 26 would benefit from being replaced with somehing like a 10Y. Admittedly these aren't indirectly heated, but they sure sound good.

[zigzagflux]

Thanks for the replies. A few posts I would like to respond to, so I'll just rattle some things off.

There are three fundamental goals I have in mind:

I am pretty well sold at this point on a three stage xfmr coupled amp, after much deliberation (internally and externally). And yes, I've thoroughly read nutshellhifi numerous times. That is most of my motivation, actually.

The input stage I am somewhat sold on with the SN7. It's reputation here, in other forums, and with Morgan Jones have convinced me I want to keep it. It is very linear, low distortion, and easy to source. The CCS in the tail may or may not stay; I think with a tube like the SN7 it could go either way; I do like the control over idle current, to be honest. I think I am more interested in some tuturials on load lines with a choke load; that's quite new territory for me.

The driver stage is probably my biggest struggle. I want to guarantee two things; linearity AND low Rp. Maybe I've automatically reigned myself into DHT territory with this request; I don't know - not enough experience with tube variety. The 6BX7 has great Rp, but Morgan Jones claims is high distortion. If I do go DHT for a driver, I will probably pursue DC heating (rest of the amp will use AC, at least for now). So I'll have to peruse some of your recommendations on driver tubes; I just don't know at this point. But all suggestions will be considered.

Another question, though, has come to mind: with the driver stage, is there any advantage to inserting the Western Electric PP capacitor from B+ to cathode, a la Amity/Karna?

Bill of Materials at this point is high, almost $4K (amorphous cores-big bucks). I'm serious about the build, just want to make the right decisions now instead of backtracking later. Keep the ideas coming!

[andyjevans]

Can I suggest that you seperate the 6SN7 into two sockets? This allows you to do several things:
- Use the 6P5GT or 2C22, both of which sound better than just about any 6SN7. Plus there are numerous good 6J5 equivalents.
- If in future you want to go DHT and leave a little space in the chassis for a supply, you can use the 1G4GT which is cheap and a lovely sound. I'd put that ahead of the 2C22 even. Octal base.

Sounds like this amp will be a keeper, so if you are able to leave some space under the bonnet you may rejoice at a future date if you need it!

As for the driver, if you use a 6SN7 type or triode equivalent, thats a mu of 20. You will need a driver with a mu of at least 5 and that's marginal. Can I suggest you make a chassis cutout of 37mm for the driver and use those big oval ceramic sockets? That way you can chose between octal, UX4 and UX5. For octal I've always wanted to use a 1626 (as in Darling amps, which you can search on) which is same pinout as 6J5. Never got round to it but it's a nice sound. Mu is marginal, and it's 12 volt.

With DHTs there's the wonderful 46 - much loved as a driver and cheap enough and plentiful. Mu is marginal. You'll get more gain with 1624 trioded. Both these are UX5 sockets. If you want to splash out, use the equally wonderful 10Y, back to UX4. 1624 and 10Y are mu closer to 10 so that solves your marginal gain issue.
These DHTs will be quieter on DC, so plan on leaving space for the filament supplies. Simplest way of doing it is the RonanReg, with three cheap LM1084s. should work fine and has been widely used. It consists of two voltage regs and a current reg. Some question the need for two voltage regs, but you may find it easier to use two because they do heat up and you want to drop about 1.5v per stage. Leave room for heatsinks on all three, and larger ones on the first two, depending on the voltage drop. You can run a differential pair in series off one filament supply, but it's better practice to use one supply for each since the voltages are more correct.

I'd be interested to see how this develops since I'm hot on your heels (and no doubt others currently building with 300b too). Andy