The Amazing Fet Circlotron by Mike Rothacher

[qusp]

I thought so and yes i bought a fair number of Antek units about 3 months ago and have been pleasantly surprised with the quality of the audio rated screened units; especially when one considers the price and the service. i initially planned to get my usual potted transformers from Richard Sumner at sumR, but was a bit unsure of what final supply voltages and currents i would end up using with a couple of experimental builds and i had a number i needed all at once, so would have been over 1000 in iron alone and the shipping from Canada blows, so i thought i'd get some temporary units from Antek and then retire them to prototyping duties and a bench supply i'm planning when i replaced them with the potted units and Japanese R-cores

that was the plan anyway, I've been so pleased with the quality that i will probably only replace half of them at the most. partially because i need slightly unusual specs with multiple secondaries, partially because i want to try out R-cores, partially because i prefer solid core leads and partially for vanity as potted custom iron looks sweet!!

good to know the cans are worth it regardless, i will probably get some and do the same.

 

[coolnose]

When looking at the datasheets of the suggested output transistors , especially FQA19N20C IT seems to me that the Toshiba 2SK1530 is a drop in replacement.

Any comments appreciated as I'd like to place my orders this weekend.

Cheers,

Max

 

[Michael Rothacher]

Looks fine to me.

 

[dnewman]

When looking at the datasheets of the suggested output transistors , especially FQA19N20C IT seems to me that the Toshiba 2SK1530 is a drop in replacement.

Any comments appreciated as I'd like to place my orders this weekend.

Also, the Vishay/Siliconix IRFP240pbf.

Dan

 

[dnewman]

Was just PM'd asking for build costs for my build. Figured I might as well post the costs here with notes. The short of it is it's about 1000 USD if you need to buy most everything (as of 9/2011). For me, it was a bit cheaper as I made my own PCBs and had the FETs and some other parts already.

Here's a rough breakdown

• 2 Heatsinks: $205 ($172 + $18 shipping + $15 tax)
• 2 Panels: $242 ($237 + $5 shipping)
• Parts from Digikey: $115 ($94 + $13 shipping + $8 tax)
• 4 PSU PCBs: $95 (estimated)
• 2 Amp PCBs: $94 (estimated)
• 2 AnTek 300 VA xformers: $83 ($72 + $11 shipping)
• 2 wooden drawers: $81 ($71 + $10 shipping)
• Parts from Mouser: $65 ($48 + $13 shipping + $4 tax)

Total $980, exclusive of JFETs (4), MOSFETs (4), insulating pads (4), fuses, wire, aluminum U-channels, spacers, nuts, bolts, and washers.

Notes:

1. Prices for PCBs "estimated" by submitting designs to expresspcb and seeing what prices were quoted (no solder mask, no silk screen). I actually made my own boards: the layout being single sided with large pads and wide traces was a no brainer. (Also, I changed the AMP PCB slightly to accomodate using 2SJ109BLs.)
2. Digikey and Mouser prices may be too low by as much as $10 each: I used some resistors from stock on hand and thus didn't order them.
3. I had the necessary FETs on hand as well as the insulating pads, aluminum, nuts, bolts, washers, spacers, and fuses. Hard for me to imagine the lot costing more than $75. Certainly under $100. (Unless you try to buy some 2SJ109BLs -- last ones I saw went for around $50 each.)
4. If you want cans for those AnTek trannys, they will run you $24 each plus shipping.
5. You can get shielded tranformers as well from AnTek: they run $41 apiece rather than $36 apiece for the 300 VA flavor.

 

[Variac]

Dan,

Did you mention where you got the heatsinks? The drawers?
The panels, just the metal cost or did you have them fabricated?
Would be good to know..

Mark

 

[dnewman]

Dan,

Did you mention where you got the heatsinks? The drawers?
The panels, just the metal cost or did you have them fabricated?
Would be good to know..

Mark

Sorry, implicit in my message was that everything not explicitly stated otherwise was as per Michael Rothacher's suggested build & chassis design in his Amazing Circolotron articles. As such, the heatsinks are those he suggests and are from Newark. The panels as are per his CAD files and from Front Panel Design. The drawers are as per his dimensions and source at Woodcraft. Further implicit was why I stated the makes of the deviated parts only. E.g., I used AnTek 300 VA transformers instead of Michael's suggested Avel 250 VA transformers. The AnTeks are less expensive and just as good IMO.

Dan

 

[Variac]

Well, a happy ending. I hadn't read the construction article. Lots of good stuff there.

http://www.passdiy.com/pdf/Amazing FET Circlotron Construction.pdf

Clearly if you scrounge around and make do with what you find, you can build it a lot cheaper, but its sure nice to have an option with everything worked out!

 

[dnewman]

Clearly if you scrounge around and make do with what you find, you can build it a lot cheaper, but its sure nice to have an option with everything worked out!

Yes. Just look at the cost of doing those PCBs through expresspcb. Given gerber files, I could easily do those boards at the same miniscule quantities for around 60-70% of the cost through other fabs (but with longer wait times). And, doing even a small group buy and/or using already available PSU PCBs (e.g., Peter Daniel) could bring those PCB costs from ~$200 to ~$75 or less. There could be savings, but not nearly as dramatic, on the panels from FPE. Etc.

So, I'm both agreeing that those are worst case costs and pointing out that there's savings to be had in addition to scrounging.

Dan

P.S. Michael used expresspcb's layout tool which outputs a proprietary file which you can use with expresspcb (and affiliates such as sunstone). However, you cannot get gerber data out of those files -- or at least I can't. But as I mentioned before, the layouts are one sided and really well suited to DIY production of the boards.

 

[steveu]

Is that what you call it! I use a 40 year old, 30KW UDCO shake table amp at work. Its a Circlotron using an array of NPN stud transistor mounted to water cooling pipes. The power transformer is as big as you can fit in the bottom of a 19" rack. Today all shake tables use class D amps, air cooling, and are 1/2 as big.

 

[Tubenstein]

Hello!
My friend has sent me this circuit since I used to interest in building a hybrid circlotron amplifier, but not now.

Refer to the circuit builder claimed that "the CIRC1 will develop about 60 watts with 12 volt supplies as long as the power supply can provide the required 13 amps current, as technically it is biased in Class AB but with a high Class A standing current. It is possible to adjust the bias on the CIRC1 to a lower level allowing a much more efficient output stage."

I do a search and the IXTH20N50D is a 500V 20A, what is the highest power to get the most from this device by not causing any harm to its "linearity"?

 

[steveu]

An important point about this topology is that you need two floating supplies for each power stage, plus another supply for the driver(s), so for stereo you would need five independant power supplies, ie only one of which can be connected to ground.

 

[tiefbassuebertr]

Andrea Ciuffoli's Circlotron

in this case Andrea Ciuffoli's circlotron (author from the famous and good known diy projects "Power Follower" and "Cool Follower") could be also of interest - go to
http://www.diyaudio.com/forums/tubes-valves/213159-thorens-tem-3200-clone.html

 

[JohanH]

I'm biting the bullet on this one and started gathering parts.
In a conversation elsewhere someone mentioned that power dissipation in R11/12 can reach peak values of 1 watt.
As I intend to build this with Vishay bulk foil resistors with a nominal power dissipation of 0,4 watts, I'm wondering I should be worried...

Thanks in advance for your input.

Cheers,
Johan

 

[Michael Rothacher]

If it's important to use those resistors, you could use two to make up the value, unless you plan on really beating the amp to death, then use three.

 

[JohanH]

Ok,

thanks Michael.
Are there other things to consider?

Cheers,
Johan

 

[Michael Rothacher]

That's the only erratum I'm aware of right now.

 

[2 picoDumbs]

I would like to see a PCB for this in the DIYAUDIO store.
Any chance of that happening?

EDIT: Actually, I would like to see at least 3 different circlotron amps in the diyaudio store.
1)This one
2) A power buffer (source follower) to attach to the output of the balanced BA3
3) And something entirely different from Nelsons crazy brain

 

[JohanH]

That's the only erratum I'm aware of right now.

Resistor values and reference of R13,R14 and R15,R16 are reversed in BOM vs schematic. Which is correct?

Thanks in advance,
Johan

 

[Michael Rothacher]

The schematic is the correct one.