Greetings everyone. Longtime listener, first time caller.
I am working on some non-traditional power supply designs (electronic transformers, boost converters, and such) for a low-power/low-budget tube amp build and would like to build an adjustable constant current sink to be used as a load for testing.
I anticipate only needing to handle about 20w @ 350V which is well beyond the capabilities of my current tube stock (all 9-pin triodes) so I need to look for an octal/compaction solution, I think.
Can anyone recommend some cheap pentodes which would fit the bill here? I think the 6hj5 and friends might be suitable but I am open to all suggestions and other comments.
As for the actual circuit, I'm happy modulating grid voltages with a bench supply and monitoring current with a DMM but any pointers to a good CCS circuit would be most welcome.
Cheers
Fr0sty
EDIT: I'm adding a proper list of (somewhat loose and perhaps overly-ambitious) requirements to this now that I've had some time to think.
Must Haves:
- Voltage range: 100V-400V
- Current range: 5-75mA
- Max Power: 30W
- Maximum of two range switches (hi/lo current and hi/lo voltage)
Nice-to-haves:
- Completely self-powered (battery bias supply acceptable)
- Single knob adjustment for full Voltage/Current range
- Junkbox build friendly
I am working on some non-traditional power supply designs (electronic transformers, boost converters, and such) for a low-power/low-budget tube amp build and would like to build an adjustable constant current sink to be used as a load for testing.
I anticipate only needing to handle about 20w @ 350V which is well beyond the capabilities of my current tube stock (all 9-pin triodes) so I need to look for an octal/compaction solution, I think.
Can anyone recommend some cheap pentodes which would fit the bill here? I think the 6hj5 and friends might be suitable but I am open to all suggestions and other comments.
As for the actual circuit, I'm happy modulating grid voltages with a bench supply and monitoring current with a DMM but any pointers to a good CCS circuit would be most welcome.
Cheers
Fr0sty
EDIT: I'm adding a proper list of (somewhat loose and perhaps overly-ambitious) requirements to this now that I've had some time to think.
Must Haves:
- Voltage range: 100V-400V
- Current range: 5-75mA
- Max Power: 30W
- Maximum of two range switches (hi/lo current and hi/lo voltage)
Nice-to-haves:
- Completely self-powered (battery bias supply acceptable)
- Single knob adjustment for full Voltage/Current range
- Junkbox build friendly
Last edited:
I use a version of this CCS for a cathode current sink in the output stage of my PPP KT-120 amps (65V@300mA). They are rated to 400V in the K&K Audio kit version, but could be modified for more voltage if needed. 20W would be no problem with sufficient heat sinks. Scroll down the page to: 2 terminal cathode current sink.K&K Audio | Other kits
A power resistor might be a better model of how a circuit behaves as you vary the supply voltage. If you were really unlucky a resistor load vs. a CCS load could be the difference between stability and oscillation in a regulated supply.
Perhaps you could make a really floppy CCS by using an output pentode (or two in parallel). Run the screen grid from a potential divider across the supply (make this variable if you like - or vary the cathode resistance) - and add a decoupling cap too for realism. Put a power resistor in the anode circuit to soak up some of the power and make life a bit easier for the valve. Then you have something which will behave something like an audio circuit (as seen by the PSU) but is adjustable.
Perhaps you could make a really floppy CCS by using an output pentode (or two in parallel). Run the screen grid from a potential divider across the supply (make this variable if you like - or vary the cathode resistance) - and add a decoupling cap too for realism. Put a power resistor in the anode circuit to soak up some of the power and make life a bit easier for the valve. Then you have something which will behave something like an audio circuit (as seen by the PSU) but is adjustable.
One vote for a CCS, one vote for a power resistor and one vote for combining the two.
I agree that the power resistor is by far the simplest but that comes at the cost of adjustability. If you want to try several currents at several supply voltages it is much better to have a knob to twist than a rats nest of test leads to series/parallel your resistors for the correct load. Also, at a few bucks a pop it may actually be cheaper (not to mention more satisfying) to build a silicon or tube CCS than to assemble a suitable collection of power resistors to cover the desired range of loads.
enough 4W/7W bulbs (at least three in series since they are 120V over here) could be an option. Cheaper than power resistors at any rate.
What do you mean by a "floppy" CCS? One that is slow to react?
On the topic of incandescent CCS: are there any good resources out there to read? I know the "unlit bulb as current source" makes an appearance in many vacuum tube signal generators but I can't find any articles that cover anything more sophisticated than the "100W bulb in series with DUT" current limiters.
For high power you could use an old electric fire - the ones with the exposed resistance wire wound on a ceramic base. I guess when I was young we knew that red-hot metal and 240V electricity were both dangerous. There were Darwin awards for people who didn't know.
Those in countries which eat toast could also use an electric toaster as a PSU load. These too have exposed elements. People who drink tea and like really loud music could use an electric kettle as a load.
Those in countries which eat toast could also use an electric toaster as a PSU load. These too have exposed elements. People who drink tea and like really loud music could use an electric kettle as a load.
By 'floppy' CCS I mean one with not too high a slope resistance and also a little lag in reacting.
Thinking about this more it seems like I can accomplish this all with a 6hj5 with the grid controlled by an opamp+sense resistor on the cathode.
The only problem I see is that I will need an opamp that can swing the grid down to something like -50V for the lower end of the current range. Will it be difficult/expensive to find an opamp with that much swing?
The only problem I see is that I will need an opamp that can swing the grid down to something like -50V for the lower end of the current range. Will it be difficult/expensive to find an opamp with that much swing?
Yeah, not really good with SMD stuff yet but if that's the best option so be it. Could I get away with cheaper opamp driving a MOSFET to pull the gate down? That would require another power rail but the small currents might let me get away with a voltage divider there, right?
Or am I all washed up? Anyone have a good general opamp reference they can refer me to?
Another odd idea for a Saturday afternoon: could you do away with with the negative supply, and grid leak bias the tube?
I wish this was like FB so I could "like" some of the responses.
I use 15K and 50K power resistors.
I use 15K and 50K power resistors.
Well, the good old Art of Electronics by Horowitz and Hill is what I usually recommend.
Here's an example of simple high-voltage op-amps made with discrete transistors: https://dl.dropboxusercontent.com/u/57580004/Perveant Breadboard.pdf Just ignore the copyright statement -- this design has been released into the public domain.
- Status
- Not open for further replies.