Now for something different...

[n_maher]

No worries.

[wink]

Make some more.....

[ujamerstand]

When I need to make something else I'll remember to throw them in.

[spritzer]

Time for an update, the BATE prototype PCB's landed here yesterday so naturally I started assembly...

http://i.imgur.com/dLaky.jpg

http://i.imgur.com/0QITM.jpg

Warning KG sized pics. I'm missing some parts so it won't be ready until Fedex shows up but it will be a neat little amp.

Edited May 8, 2012 by spritzer

[Victor Chew]

Nice work Spritzer. "Balanced all triode".....Interesting! I have always loved the sound from the triodes in particular the 300B in SE mode. Though not much of power, but when matched with efficient speakers they sound nice with plenty of dynamics to my old ears.

Please let us know how these perform.

Edited May 9, 2012 by Victor Chew

[spritzer]

This one will not sound like any 300B amp, that's way too much distortion for us. A triode is still one of the best ways of amplifying ever invented so why not cram three of them per phase for a brilliant cheap-ish amp.

I've also been brainstorming recently with Kevin about making a true Volksamp for the people. Basically, what is the cheapest amp we can build without sacrificing our level of quality while having some strict design guidelines. This just expands on my vision at the start of the Exstata project but naturally learning from that clusterfuck. The BATE is an excellent amp but it's PSU is highly advanced and requires some advanced assembly with the correct parts (i.e. use the correct insulators or you will have trouble) and it uses tubes which are out of production.

I want something which has all the parts in current production and easy to source, using a transformer which is available off the shelf at 117/230V and should just work if all the parts are in the correct spots and face the right way. Hell, I might even write up a "how to" guide for this one... Any ideas for good tubes or topologies are welcome...

Ohh and then there is that truly insane idea, making a 'stat amp with miniature tubes. It would technically be portable...

[deepak]

Cheap is good

I was thinking about building a TubeSE with 307a to power my SR-Lambda and transformer box, but a dedicated stat amp is even better

[spritzer]

Use a transformer to step down and another one to step up again? Nahh, even an Egmont would be preferable to that..

For the amp I'm leaning towards something simple like a quad of ECC8*'s feeding a pair of ECC99's. ECC99's need to be left floating so two filament windings at a minimum. Nothing else out there which fits "cheap, high voltage and reasonable power" though and we really could look into CCS loading the output tubes, hell CSS load the lot of them given the low currents involved. It just has to be weighed as a cost vs. performance issue.

[deepak]

Is the Egmont schematic up somewhere? Even the BATE comes out to more than I'd want to spend with the regulated power supply. This is pretty much something to listen to during my lunch break then maybe a few minutes at the end of the day finishing up paperwork.... at work

[spritzer]

[ujamerstand]

What's so special about this amplifier, and what is your target cost?

Edited May 9, 2012 by ujamerstand

[cetoole]

While we are on the subject of various Stax amps, I have been curious for years: have KGST schematics ever been posted?

[HeadphoneAddict]

oops

Edited May 10, 2012 by HeadphoneAddict

[FrankCooter]

The Egmont is a great place to start an all tube electrostatic "Volksamp" type project. It's as simple and cheap as possible yet offers lots of possibilities for easy tweaks and upgrades. Right now I'm running a sort of Egmont/Williamson/Bate hybrid. Because I have no need for balanced inputs, the front end is a simple 6SN7 CCS loaded voltage amp directly connected to a 6SN7 concertina phase-splitter. The 6SN7GTA/B will take 450V on the plate, so it can be run directly off the positive rail. It's not sexy, but it's not chopped liver, and gets the job done in the simplest possible manner. Output is CCS loaded 6BL7 with individually adjustable "fixed" bias. The 6BL7 plates run 12mA/400V. Four tubes, three IXYS10M90, and two coupling caps make a channel. You could do the whole thing equally well, maybe better, with with ECC99's in all positions.

The "Bate" power supply is a little intimidating for a novice. Perhaps the initial filter could be beefed up with a couple of cheap chokes and the regulator simplified to a core of a zener stabilized pass device. Maybe forgo the regulator entirely in favor of a simple, but well designed choke based

filter. This is pretty much what I'm doing except I have mirrored 6EA7 tube regulators on the outputs instead of sand.

I like the amp well enough well enough to build a case for it. It'll be a single chassis. Amp will be stacked on the power supply. Footprint is 5.25x13" uncased. Height is 10.5" This is about as small as I can go. Looks sort of like a miniature PC tower except the case is veneered zebrawood.

I'm very interested in seeing how far you guys can run with "Volksamp" concept. There's not enough of this sort of thing out there.

[spritzer]

I like that Frank, nice and simple. Have you ever messed with feedback in these amps? I used 150K and 1uf/630V in one of my Egmonts and it really helped stabilize it. I really like the idea of taking the diffamp to the max to see just how good it can be made.

If we can find cheap chokes then I'm all for it but a simple pass transistor with a zener reference would probably be cheaper and easier to deal with. Even better to use isolated parts to kill any chance of shorts.

Balanced input is a must have for the amp to be taken seriously and it really must have too much gain for the "MOAR LOUD!!!" people out there so ECC83 drivers with a bypass cap on the cathode.

What's so special about this amplifier, and what is your target cost?

The idea is to try and make a high voltage amp anybody can build and be as cheap as is possible. It would also be a nice reference point to compare against the commercial amps.

I think less than 400$ would be easy to reach with say 80$ for a transformer, 30$ for the PCB's, 50$ for the tubes and the rest for assorted parts.

While we are on the subject of various Stax amps, I have been curious for years: have KGST schematics ever been posted?

Not posted but I should have it somewhere, at least one version of it. That version is just a SRM-007t front end and a 6S4A with a 10M90 CCS and a trimmer+resistor on the cathode. I used Marc's Vulcan PCB's and modified them to become a KGST.

[ujamerstand]

Cool. Now, If we can also get a diy stat headphone kit going as well that would be awesome!

[kevin gilmore]

While we are on the subject of various Stax amps, I have been curious for years: have KGST schematics ever been posted?

http://gilmore.chem.northwestern.edu/kgst.pdf

not 100% sure if that was the final revision.

Edited May 10, 2012 by kevin gilmore

[spritzer]

That one should be close to the final version but I think we settled on +/-350V rails in the end.

[FrankCooter]

I don't buy into the whole "balanced is better" mania that pervades the headphone world , but I understand the necessity of appealing to the widest possible audience.

From the standpoint of parts count, design complexity, physical size, and perhaps even ultimate performance you're paying a heavy price for those balanced inputs. Instead of two stages, you'll now probably need three. Instead of a single coupling cap in the signal path, you now have three. Global feedback will probably be required to linearize and stabilize the circuit. To support the feedback, you'll need higher gain (but less linear) tubes.

Simple is not always cheap. Probably not practical in this situation, but the simplest balanced solution I can think of would be to use an input transformer with dual secondaries driving a pair of ECC99 voltage amps which are cap-coupled to a pair of ECC99 outputs. Very low parts count. Downside is cost.

Mid-level transformers like Cinemag ( I think this is what Craig uses in the B.A.) run about $150.00 a pair. Edcors are dirt cheap, and if money is no object, then Lundahls with exotic cores.

In answer to your question about feedback , I've not used it in any of my Egmont variants, but

perhaps I should have. My first attempt , which used 12AX7's in differential pairs driving IXYS loaded 6P14P was unstable and blew the pass transistor of my primitive regulator. Went with the Williamson front end, added gate stoppers to the 10M90, stuck protection diodes in the regulator, had no more problems. Not a very scientific approach, but that's the way I work.

I'm very interested in pushing this simple circuit to it's limits. One of these days I want to try it with 801A DHT's and 600V rails.

[luvdunhill]

Frank, I am curious if you would consider the argument that sources which offer balanced output, even if this output isn't the one used, sound better on average than their single-ended only counterparts?

[morphsci]

Frank, I am curious if you would consider the argument that sources which offer balanced output, even if this output isn't the one used, sound better on average than their single-ended only counterparts?

^^ No double-deaf testing but this is my impression as well.

[kevin gilmore]

Due to the required balanced output for electrostatics, the balanced is better thing ends up coming along pretty much for free.

Also the fact that you need a voltage gain of 500 minimum, and usually a more reasonable 800 to 1000 forces an open loop

gain of at least 2000+. Something that is virtually impossible with any 2 stage tube amplifier, without massive amounts of

distortion.

This was the main reason why most of mikhail's esx things had volume/balance problems, because even with some feedback

there was not enough open loop gain to level out the amplification.

You certainly can do the input stage with a transformer, but if you want 10 to 20db of gain from it, the input impedance of the

transformer is going to be under 1k ohms, otherwise there will not be enough output current to correctly drive the input impedance

of the first tube stage.

The hev90 was unbalanced input only, and then followed by a phase splitter anyway. So replacing the exact same 2 triodes

with a differential amplifier results in more gain, less distortion, and better balance.

For the nelson pass supersymmetry thing to work, you have to have at least balanced outputs.

[spritzer]

I'm not the biggest fan of "balanced at any cost" but it really must be balanced for the audience but also being able to be used SE. No half-assed designs here like the WES and A-10.

I'd really like to take a crack at "the ultimate" diff amp but it looks like 3 stages is the minimum we can go for. Even with a ECC83/ECC99 combo it simply doesn't have enough gain. Less volume means less good amp to a lot of people...

Edit. You wouldn't believe just how much willpower it took to keep my Baldur SE only when the balanced input is better in every way.

Edited May 11, 2012 by spritzer

[spritzer]

I've been giving this some thought and perhaps the best candidate for rework would be the ESX circuit. Three stages, mostly DC coupled and arguably, one of the finest tube circuits I've ever seen. The money we save on the various caps can be spent on a CCS for the output and even for the front end as well. With a C- supply there will be no offset issues and with a PSU of say +/-300V it will be a very capable amp. This is naturally a Stax circuit so I see no better way for the Mafia to show it's gratitude for 50+ years of great products but by reworking one of their designs. Here is the basic design:

http://www.bonavolta.ch/hobby/en/audio/hdgsta.htm

Now the issues, there is no transformer out there which meet the required specs, two filament windings, bias tap and the right HV voltages. Not a big deal though since people like SumR and Edcor could have a fixed design on their books which people can just order. Same wires every time etc.

The second is what parts to use. ECC81 and ECC99 are the obvious choices and the resistors can be what ever that fits but the CCS is a bit of a challenge. I'd really like isolated parts to make assembly easier (no high voltage pads etc, required) but I'm coming up empty. 2SA1968 is an obvious choice but there is very little stock left and I'm not dipping into my stock to supply a group build. Any ideas for a suitable part?

[luvdunhill]

I'd use a mosfet CCS, like the BHSE uses. If you want to do some R&D, then an opto-coupler biased CCS would be my choice... or, add another tube and do a pentode CCS.

Edited May 14, 2012 by luvdunhill