JFET-only Headphone Amp "Circlotron КП903B".

[Svjatoslav]

Designed to isodynamic-orthodynamic headphones.
Maximum on the output power of 7.5 and 15 Ohm - 1,7 W, 100 Ohm - 1,06 W.
The output impedance of 3.3 ohms at 1 kHz at a voltage at the output of 2.5-3 V.
Band power at a voltage of 90% of the level of restriction is longer 200 kHz.
With a downturn in the upper frequency (200 kHz) of less than 0.5 dB.
Spectrum of push-pull amplifier is similar to a tube triode SE. )))

 

[cT equals piD]

What is a circlotron, and what advantage is there in this type of amplifier for powering headphones? What is the part shown as a diode within a 4-sided diamond? Other than that part, I might not have too much trouble in building it, although the part number of JFETs that I could easily access would start 2N.

Regards,
Pete

 

[rayma]

What is the part shown as a diode within a 4-sided diamond?

What is a circlotron? Check here: https://www.passdiy.com/project/amplifiers/build-the-amazing-fet-circlotron

That's a four diode module, and there are many kinds. A random example:
GBPC3506W - FAIRCHILD SEMICONDUCTOR - Bridge Rectifier Diode, Single, 600 V, 35 A, Module, 1.1 V, 4 Pins | Newark element14

 

[Svjatoslav]

rayma, thanks for your assistance quickly in answering.

Currently I am doing the power supply. For stereo amplifier option.
He will have low ripple, and simple.

Regards,
Svjatoslav.

 

[rayma]

Currently I am doing the power supply. For stereo amplifier option. He will have low ripple, and simple.

Good luck. I would not substitute different types of fets in the circlotron circuit without careful evaluation.

 

[Svjatoslav]

I completely agree.
Only through measurements.

In this case, the first cascade has a "compression" when the current characteristics.
Output stage, thus, increases the current. The possibility of partial compensation of non-linearity.

 

[cT equals piD]

What is a circlotron? Check here: https://www.passdiy.com/project/amplifiers/build-the-amazing-fet-circlotron

That's a four diode module, and there are many kinds. A random example:
GBPC3506W - FAIRCHILD SEMICONDUCTOR - Bridge Rectifier Diode, Single, 600 V, 35 A, Module, 1.1 V, 4 Pins | Newark element14

Thanks for filling me in. In your post #5, you are saying that substitute JFETs would have to closely match the characteristics of the JFETs making up Svjatoslav's Circlotron, is that correct?

Regards,
Pete

 

[rayma]

you are saying that substitute JFETs would have to closely match the characteristics
of the JFETs making up Svjatoslav's Circlotron, is that correct?

Yes, I would recommend using the same types to avoid headaches.

 

[bear]

What is a circlotron? Check here: https://www.passdiy.com/project/amplifiers/build-the-amazing-fet-circlotron

That's a four diode module, and there are many kinds. A random example:
GBPC3506W - FAIRCHILD SEMICONDUCTOR - Bridge Rectifier Diode, Single, 600 V, 35 A, Module, 1.1 V, 4 Pins | Newark element14

There is a rough equivalent in a non-Russian part - that is a high power JFET made in Russia, surplus stuff these days. Fairly rare, and hard to source regardless. They do show up on ebay. Probably the best way to get some, since you only need a handful for a headphone amp.

The reason for the Circlotron is that there are no complimentary parts for this device.

To find the USA made equivalent you'll need to do a bunch of online research looking for power Jfets made 30-40+ years back. I did locate the part number years back, and you can source them, but they are fairly pricey IF you can get any. The Russian parts handle a bit more power, iirc.

I think the driver circuit could be made more simply, but a bit depends on what sort of result you're going for, and what sort of voltages need to be swung.

_-_-bear

PS. DC offset might be a concern with drift? Or does the Circlotron try to self balance? Never thought about it, since most are tubes with transformers...

 

[Svjatoslav]

2П903А, 2П903В the same as КП903А, КП903В.
With the letter "Б" less current.
There are in the housing, "white", there are "yellow".

Output current (150-200 mA) keeps very stable.
Zero output - too.

Without transformer tube also holds good mode:
http://www.diyaudio.com/forums/full-range/218603-speakers-full-range-high-resistance-4.html

 

[bear]

It would be nice to see the FFT plots with the sensitivity of the measuring device/input attenuator set so that the peak of the fundamental was ~6dB of "0", so that the harmonics might emerge from the noise floor?

_-_-bear

Thanks for the datasheets!!
I was unable to find them online some time back - but I do not speak/read Russian.

 

[bear]

Guess he forgot to check back... 🙁

 

[Svjatoslav]

The spectrum of its harmonics will be measured more than once.
With the higher resolution.
After assembling the PCB.

The idea should be a little more mature, "rest in the bed". Become adult. )))

 

[bear]

I shall sit here, waiting... hoping to "mature" or maybe "ripen"?

...eating some snacks.

My comment was that IF those are actual measurements (??) and not simulations, that the level of the fundamental was ~-40dB, so the resulting information is all in the noise floor. This makes it of very limited value.

But since I have some of these devices, I am following with interest. 😀

 

[Svjatoslav]

"ripen" )))

 

[xrk971]

http://www.diyaudio.com/forums/solid-state/291767-allfet-circlotron.html

 

[xrk971]

Could this be done with all BF862's for the small signal JFETs and LU1014D for the power JFETs?

 

[sgrossklass]

Clearly not, as you'd still need some p-channel JFETs as well.

A circlotron as a headphone amp seems a bit cumbersome to me, requiring 4 independent floating secondaries and all.

 

[xrk971]

Ahh I missed the p channels in my quick look. The quad PSU's have always been a problem for circlotrons. I think in the headamp world or should I say Head-Fi world they love balanced amps and phones wired for balanced "push pull" to get double slew rates.