• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

MOSFET current source as triode load

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.

I have a Mcintosh clone tube preamp which I am going to optimise.
At the moment it sounds fine (thanks to some help from ppl on old threads) although i have only had it running into a Hitachi HA-3 power amp but i know i can improve its crosstalk and lower its distortion.

The source of the crosstalk is the tube regulated supply which can allowa 0.2V pk-pk sine wave onto the tubes HT at signal inputs of 1Khz which gives a poor crosstalk of -50.46dB at 5V pk-pk out.
I have been looking at ways to improve PSRR of anode followers and it seems that a good way is to replace the anode resistor with a current source, MOSFET's is easy to work with due to their very low gate current.

My questions are:
is this the correct way of using a Mosfet as a current source?

do i set the Drain current using the voltage accross the 1K resistor?

do i apply the same voltage to the gate from my resistor divider as i wish to obtain at the source?

Thanks a lot


  • ccs 4 tubes.gif
    ccs 4 tubes.gif
    3 KB · Views: 1,349
Well, a MOSFET current source will absolutely help the power supply rejection. But while you're at it, don't let your grounding off the hook until you're sure it's optimal. That's often a major source of crosstalk.

Gary Pimm has a very nice-looking IRF820-based CCS on his website. I'd start with his design.

Yes i have been careful with my grounds after i experienced AC hum on power up months ago.

I used star grounding and measure 0.01-0.016 ohms to ground from every point. so i should get very minimal noise from my grounds this time :cool:

p.s. in my diagram i have put no cap on my potential divider, i will put a 15uf cap there / / with 1uf or so, this will take care of my PS noise on the gate.

Joined 2003
Paid Member

In another thread I've raised a question on a ccs and no one seemes to respond yet. As I notice this thread shows competence on the function of a ccs I'd like to invite the contribuants to comment also on my question in the next thread:

Power Supply Design >Shunt-regulated 24V DC Power Supply for Preamp Modules

Don't want to abuse this thread, merely trying to stimulate curiosity.

Thanks for the bandwidth, Arjen.
These Depletion Mosfets are not that easy to find.
In addition to US Digikey.com you can search Google for all parts (won't give you everything, though, but still...).
Other Depl. mosfets are:
DN2535/2540 - TO-220, Supertex
IXTP02N50D - TO-220, IXYS
BST124 - TO-126, Philips
I have used the depletion mode devices for CCS loads in several amplifier designs. I have even used them to load a 45 triode for chokeless parafeed. My favorite device is the previously mentioned IXYS10M45S. This device has extremely flat constant current curves across a wide operating voltage.

It is not often mentioned but you should use a gate stopper resistor connected in series with the gate lead, mounted close to the fet. Otherwise they are prone to oscillation in the 1 to 50 MHz region. The oscillation can randomly appear and cause all sorts of hard to find unwanted behavior, including fried parts.
Sorry to be totally anal, but the 10M45S actually contains an enhancement mode mosfet with a built in error amp which gives it the performance characteristics of a depletion mode device.

Wonderful component, but don't trust the published current set curves too much.

I am aware that the published data calls this device a CCS integrated circuit like you describe. It does have flatter curves than any fet, or pentode that I have tried. They do act like the perfect depletion fet. I just wish they made a bigger one (more power dissipation). There is a 900 volt one coming.

I have independently measured the curves, and yes every one is slightly different. I tested several at 10 mills, and they are within .1mA from 50 volts to 450 volts.

Sandophopic tubeheads: Don't read the next paragraph!

Being slightly crazy, and always wanting to try new circuits, I wired a pair of these into a 10 watt push - pull amp in place of the 6AQ5 tubes. They sounded pretty good. In other words they make better pentodes than some pentodes do!

If you use these, put a 1K ohm resistor in series with the gate lead. They do like to oscillate. I have seen them oscillate strongly enough to be seen on a TV set in the next room. This will cause a vague cloudiness in the sound that is very hard to pin down, since they stop oscillating when you put the scope probe on them. In a high power circuit (45 chokeless parafeed) the oscillation will cause the 10M45 to fail to a short, putting the entire power supply across the 45. Not good!
Thanks for The part numbers StevenOH

I have searched Farnell and RS for any of the mentioned transistors but have had no luck (these are the best places to buy in UK).

IXYS products from these companies dont come with drain current less than about 25A and are very expensive. Ialso investigated CC diodes (Jfets with gate tied to source) but i cant find any at suitable voltages

Unfortunatly I will probably have to use enhancement like i originally said and just use resistor dividers to set them up as CCS's.

P.S. thanks for the gate stopper tip tubelab, that will save me a headache when testing :)

DN2535/2540 - TO-220, Supertex
The drawback of these things is that you need to buy them in batches of 50. (Mouser.com)

The cascade of these things measure fantastically (low capacitance) as can be seen on Gary Pimm's site.

I could sell you 4 if you like...altough my reserve is becoming dangerously low. But just because of your avatar...I'd be willing..IF you are interested that is.

My page on these critters..

By the way ...a good tip would be to use some voltage ref diodes to protect the gates...they have saved my mosfet's once..after making some wrong connections..I am so thick!

Thanks Bas but ill probably pass on your transistors, very kind of you to offer though.

I think i will go with some IRF820's from my local electronics shop, £0.50 (UK £) each is quite good i think. Ill just put up with the extra components needed to bias them.
I am CCS'ing my gain stages but i still have a resistor loaded casthode follower on the output, should i CCS this aswell while im at it?

SY yes you are right and i think they are quieter too, I just though mosfets were better because they require no bias current and i seem to have the impression that higher voltage mosfets are easier to get than bipolars.
If you know of any common HV bipolars that would do the job i will definatly check them out.

Will i notice the difference in sound from the better current regulation of the cascode?

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.