CRD diode CCS any good <1mA?

[aletheian]

I have been wanting to experiment with CRD as a simple alternative to the various plate CCS that i have been using, but don't want to waste mw time if they are no good at or under 1mA. Anyone try these at such small curents?

[SY]

I played with the CRxxx series some years ago as 12AX/AT7 plate loads. Not impressed; I never used CCS again until Morgan Jones's book convinced me that a cascode CCS might be worthwhile. I've never looked back.

[aletheian]

Originally posted by SY
I played with the CRxxx series some years ago as 12AX/AT7 plate loads. Not impressed; I never used CCS again until Morgan Jones's book convinced me that a cascode CCS might be worthwhile. I've never looked back.

Gotcha...
i have been using various cascode CCS, and was just looking for something that i could use to drop the parts count and adjustments down a bit for quicker breadboarding.

[tubelab.com]

I have used the current source diodes for plate loads but they were the 2.7mA flavor. I was not impressed with the sound quality. I have been using the IXYS 10M45 chips with much better results. You need 1 chip (3 leads) and 2 resistors. This is the simplest CCS that I have found that gives good sound. You must put a 1K resistor in series with the gate terminal right at the chip. Otherwise it will oscillate. This is not shown on the data sheet. I have not tried these below 2 mA though.

[zarniwoop]

I think they are great for a 12ax7 load. I love the simplicity and the sound is excellent. I've also tried a mosfet cascode CCS, but in a different circuit.

[SY]

I've had very good results from depletion-mode MOSFET cascodes as plate loads, at least at larger currents (8-20mA). I haven't tried them at low currents yet. Very high dynamic impedance, low capacitance, and high voltage withstand (the CRxxx are something like 50V?).

[aletheian]

I think I'm just going to crank up the current a bit and work with the fet cascodes that I am used to. I have been working with some low voltage space charge tubes, and drew up loadlines that required less then a mA of Ip, but I can't find a simple CCS that is stable and has acceptably low capacitance at voltages/currents that low.

[tubelab.com]

I've had very good results from depletion-mode MOSFET cascodes as plate loads, at least at larger currents (8-20mA). I haven't tried them at low currents yet. Very high dynamic impedance, low capacitance, and high voltage withstand (the CRxxx are something like 50V?).

I am breadboarding that one now (the top secret tubes showed up this morning).

The diodes that I used were old Motorola 1N5312 current source diodes. The Mot sales rep (20 something years ago) told me that they were really selected FETs with the gate tied to the source. Mouser still has them. These are specced to 100 volts and I have been to 110.

I have also used the old two BJT's and an LED circuit, but again not below a few mA. It would probably work OK with some low current RF transistors (very low capacitance) since you aren't running much voltage.

[aletheian]

The project that I was tinkering with tube line stages and preamps running on opamp-supply rail voltages for some hybrid experiments with a single supply and also a hybrid transformerless large cpasule condenser microphone that runs on phantom power, so I gravitated toward CRD because they take up far less space than the equivalent FET cascode CCS, and since apparently the IXYS chips lose regulation under 2mA and LM-three terminal regulators under 5mA, that idea went down too.

I figured that 1 CRD CCS on the plate + 1 diode in the cathode took up less space then a plate resistor + cathode resistor + bypass cap, as well as performing better in terms of freq bandwidth and PSRR... and $.

The regulation in the microamps range might not be an issue though, because after testing several subminis, and space charge tubes (ecc86, 12u7, etc) I found that the e88cc/6922 family performed (and sounded) better in simulations and real life... with a great amount of tweaking. The only caveat is the heater current, which blows the phantom powering idea out of the water since it would sink around 300mA or more just from the heaters! So i may scrap that project for now and concentrate on the low voltage linestage/preamp.

[tubelab.com]

Doh, I'm too blonde. I have thought up several CCS's that don't work (or haven't been tested) below 1mA, and then I looked over at the amp on my breadboard that has been patiently waiting for me to finish it for at least 6 months, and there it is a little CCS that works from 1 uA to 10 mA over 1 to 40 volts. The shunt capacitance is 15pF and the impedance is over 1 megohm to 10KHz. The LM334Z. It is in a TO92 package and requires 1 resistor. I have one of these in the tail of a LTP using a 6BQ7. In that application it is flat out past 50 KHz. These are mentioned in Morgan Jones books too.

As for the phantom powered amp, I see two choices, a SMPS or telephone tubes. It would be possible to build a buck converter to make 6 volts at 300 mA from 48 volts at 45 mA. This is a lot of work though. There were dozens of tubes made for telephone repeaters that have 48 volt filaments. Many were made by Western Electric, but they can be found surplus for cheap because no one wants them. I can't remember any numbers right now, but I know that I have seen some in my junk collection. I just don't know where they are.