constant current sources...

[cathode_leak]

This quesion is related to constant current sources... I have a triode that wants 145V on it's plate, at 45mA.. I'd like to implement it like shown here (diagram by Gary Pimm): http://home.pacifier.com/~gpimm/Image8.gif.

My headache starts when I'm trying to figure out what voltage to keep on the positive input of the CCS feeding the triode plate (wich will be the voltage rating of the Zener/VR tube), or in other words to calculate the voltage drop acoss the CCS feeding the triode plate..

Feeding the plate, I will use a fairly simple cascoded depletion mode CCS (no batteries) I found at Bas' site, based on DN2540. Diagram of the CCS crcuit can be seen here: http://basenjes.de/tubes/images/dmmosfet/dmmcccs4d.gif

[jeff mai]

Originally posted by cathode_leak
My headache starts when I'm trying to figure out what voltage to keep on the positive input of the CCS feeding the triode plate (wich will be the voltage rating of the Zener/VR tube), or in other words to calculate the voltage drop acoss the CCS feeding the triode plate.

I'm only just starting to work with these things so I recommend confirmation, but I think the answer is: "any voltage that is enough to light the VR tube and bias the CCS."

The voltage drop across the CCS *must* be the difference between its + input voltage and the working voltage of the VR tube. If this voltage drop is large, you need more heatsinking of the top MOSFET because it dissipates the power.

[SY]

"any voltage that is enough to light the VR tube and bias the CCS."

...and accomodate these voltages under the anticipated plate voltage swing.

[cathode_leak]

Thanks jeff, SY.

I'll get some russian OA2 for experiments.. They seem to be of a suitable caliber for my purpose..

[Yvesm]

Hi cathode_leak .

You don't really need to regulate the voltage that feeds the current source because the current source is, by definition insensitve to the voltage accross it as long as it is "sufficient".
By "sufficient", understand "at least equal to the expected peak AC swing".... plus some headroom !

In theory, the higher, the better providing that, as Jeff pointed, the power dissipated in the current source remains in the specs, and that the voltage accross it does not burn the FETs.

e.g. Assuming you expect 100vAC swing, the voltage at the positive end of the CSS must (at least) 145v (for the tube) plus 100v (for the AC swing be possible). Then add 45 x 1K5 (67.5v) for what is lost in the resistor at the top drain (even if don't know what this resistor if made for), and apply the same for the resistor in the lower source. It seems that you need some 350v.

As a rule of thumb (and without that resistor in the top drain) you may start with the same voltage accross the CCS that accross the tube.
That is : 145 + 145.

Anyway, the voltage accross the CCS will adapt itself at B+ minus 145v, providing the triode really draws 45mA at 145v (this being set by its cathode resistor).

Yves.

[cathode_leak]

Thanks Yves, your post are very helpful. By the way, the real purpose of the VR tube is not regulating but rather split the AC loop and the DC loop apart, keeping the PSU out of the signal path.. Although it acts like a regulator too..

[pedroskova]

...don't know what you're using the ccs for, but you could use it for the current set resistor for your VR tube and then just use a resistive plate load. The ccs will still help isolate your circuit from psu noise at that position.

[Shoog]

Gary Pimm has a range of other CCS circuits which don't use a VR tube and should perform much better than circuit sujested. The circuit you are using was published mainly for illustrative purposes.

Shoog

[cathode_leak]

Originally posted by Shoog
Gary Pimm has a range of other CCS circuits which don't use a VR tube and should perform much better than circuit sujested. The circuit you are using was published mainly for illustrative purposes.

Shoog

You got me wrong, Shoog. The circuit using a VR tube is not a CCS at all... It is an example of implementation technique of CC sources in a single ended triode circuit. If you look at Mr.Pimm's page, you'll find he actually implements CC sources and VR "zeners" the same manner in all his linestage/preamp diagrams.. I have no plan to put VR tubes in the CCS, that is MOSFET territory.

[Shoog]

My mistake.

Shoog