Jfet as variable resistor with digital feedback for linearization

Thread Starter

coinmaster

Joined Dec 24, 2015
502
I'm trying to create 10 electronically controlled variable resistors.
The intended voltage range will be 0-600v at about 0-100ma.

I'm using a microcontroller to linearize the output voltage based upon voltage feedback to an ADC.

I thought about using a vactrol but from my understanding they don't have a wide range of resistances and they don't react very fast.

SIC Jfets seem to be the only other thing I can think of but I'm not sure how they will work in this situation.
 
Last edited:

GopherT

Joined Nov 23, 2012
8,012
I'm trying to create a system that will have a bunch of volt/ammeters and about 10 electronically controlled variable resistors.
The intended voltage range will be 0-600v at about 0-100ma.

I tried using a transistor as a common emitter and taking the output from the collector but the shunting of the power to ground is causing some undesired effects in my circuits.

I need a true variable resistor that takes its input from one and and outputs on the other end.
I'm currently using an arduino to linearize the voltage by setting an intended voltage value and having it bias the VCR based on what the output voltage is so the output matches the intended voltage.

I thought about using a vactrol but from my understanding they don't have a wide range of resistances and they don't react very fast.
Jfets seem to be the only other thing I can think of but I'm not sure how they will work in this situation. I know normally they are not rated for 600v but the new SIC Jfets can handle far higher.
Please read your own question and try to decipher it. From my reading of it, it seems like a mangled mess of words. Try to clarify, please.
 

ian field

Joined Oct 27, 2012
6,539
I'm trying to create 10 electronically controlled variable resistors.
The intended voltage range will be 0-600v at about 0-100ma.

I'm using a microcontroller to linearize the output voltage based upon voltage feedback to an ADC.

I thought about using a vactrol but from my understanding they don't have a wide range of resistances and they don't react very fast.

SIC Jfets seem to be the only other thing I can think of but I'm not sure how they will work in this situation.
You probably won't find too many JFETs that can handle 600V at 100mA.

MOSFETs should get you there - but you'll need a big heatsink.

Linearity could be a hurdle, I'm thinking put a current sensing resistor in the source lead and use the voltage developed across that to drive a bipolar transistor that clamps the gate.

Digital current control can get complex very quickly, you might be able to switch a bank of resistors in/out of the source circuit - using a binary code to switch in various combinations of source resistors would save a lot of wires.
 

Thread Starter

coinmaster

Joined Dec 24, 2015
502
This was the Jfet I had in mind.
http://www.infineon.com/dgdl/Infineon-IJW120R100T1-DS-v02_00-en.pdf?fileId=db3a304341e0aed001420353f03a0e4b
I have one on order now for testing.
It can handle far more than I'm planning on subjecting it to and is quite linear.

As for the linearity, I coded my arduino so if the output is greater than or less than the intended voltage then it will adjust the bias voltage to the gate/base of whatever "resistor" I'm using. It works well with this

but shunting power into ground to get a variable voltage is causing problems in my circuits so I need a true end to end variable resistor.

This digital feedback should make any non linear characteristics of whatever I use pretty much irrelevant right?
 
Last edited:

crutschow

Joined Mar 14, 2008
23,802
..................
but shunting power into ground to get a variable voltage is causing problems in my circuits so I need a true end to end variable resistor.
...............
I don't understand this. :confused:
What type of problems?
If you want to minimize ground current, add a FET source follower at the output of Q2.
Then you can make R3 a much higher value.
 

Roderick Young

Joined Feb 22, 2015
408
I have no experience with this part, but the data sheet seems to say you need a negative voltage on the gate. If so, how did you plan to generate that? Also, if you need your electronic "resistor" to be isolated from ground, perhaps an optocoupler would be in order somewhere.
 

Thread Starter

coinmaster

Joined Dec 24, 2015
502
I don't understand this. :confused:
What type of problems?
Well it's causing parts of my circuit to draw current where it shouldn't which is affecting the current of other parts of the circuit.

If you want to minimize ground current, add a FET source follower at the output of Q2.
Then you can make R3 a much higher value.
Hmmmm, maybe I'm just ignorant but wouldn't that still cause some sort of current irregularity when placed in series with a circuit?
Wouldn't a SIC JFET solution be simpler?

I have no experience with this part, but the data sheet seems to say you need a negative voltage on the gate. If so, how did you plan to generate that?.
I'll probably use an opamp or something to invert it. I'm replacing the PWM with a DAC to eliminate the time constant of the RC filter when I actually build this thing, maybe I can make it output negative voltages for me, I haven't looked into it yet.
 

crutschow

Joined Mar 14, 2008
23,802
Well it's causing parts of my circuit to draw current where it shouldn't which is affecting the current of other parts of the circuit.
How is it doing that?
Without a reason, then how do you know it's the current causing the problem.
Hmmmm, maybe I'm just ignorant but wouldn't that still cause some sort of current irregularity when placed in series with a circuit?
Wouldn't a SIC JFET solution be simpler?
That's what you feedback takes care of.
How does a JFET solve the "current problem" you have?
 

Thread Starter

coinmaster

Joined Dec 24, 2015
502
How is it doing that?
Without a reason, then how do you know it's the current causing the problem.
The current issue is based off of simulation rather than benchwork, I'm assuming LTspice can be trusted.
It's causing current to flow from B+ into the transistor in parts of the circuit where there shouldn't be current flowing.

How does a JFET solve the "current problem" you have?
Well since the current isn't shunted into ground and instead goes into one end and out the other in series with the circuit I wouldn't expect there to be a problem.
I'm guessing that having ground potential available to the circuit it's in series with is allowing current to flow where it shouldn't be.
 
Last edited:

crutschow

Joined Mar 14, 2008
23,802
The current issue is based off of simulation rather than benchwork, I'm assuming LTspice can be trusted.
It's causing current to flow from B+ into the transistor in parts of the circuit where there shouldn't be current flowing.
.....................
But where is it flowing?
LTspice will tell you.
Without knowing that, then we're basically just waving our arms. :rolleyes:
Post your LTspice .asc file.
 

Thread Starter

coinmaster

Joined Dec 24, 2015
502
Here you go.
The current from B+ through R8 should be just about non existent, but it does have current flow with the transistor and the current at the plate of U1 shifts along with the current of that section. Not by much, but enough to bother me since the circuit is supposed to maintain a stable current there.
 

Attachments

Last edited:

crutschow

Joined Mar 14, 2008
23,802
Sorry but that's a rather convoluted circuit and I can't really follow what it's trying to do.
I don't see how it can do what you said you wanted to do in your first post.
You have a PWM signal source at one spot and a sinewave at another(?). :confused:
If can explain exactly what you are trying to do with that circuit, it would greatly help.
 

Thread Starter

coinmaster

Joined Dec 24, 2015
502
It's a gyrator.
That particular gyrator is discussed here http://www.diyaudio.com/forums/tubes-valves/140699-anti-triode-sepp-how-do-best.html
but there's a zillion different ones I can choose from https://www.google.com/search?q=gyrator&biw=1242&bih=566&site=webhp&tbm=isch&source=lnms&sa=X&ved=0ahUKEwiEr8zv64nLAhUI_mMKHU03BFYQ_AUIQCgB#tbm=isch&q=gyrator+triode
The important thing is that it provides a favorable load to the triode and can be used to set plate voltage.
I'm using the variable "resistor" to set the plate voltage, I'm also going to use it in some power supplies and CCSs.


You have a PWM signal source at one spot and a sinewave at another(?). :confused:
There's no pwm in that schematic it's PWL http://www.linear.com/solutions/1814
I used it to show the voltages over time with incrementing bias voltages on the transistor.
 

crutschow

Joined Mar 14, 2008
23,802
I'm sorry, but I don't really see how that circuit will accomplish what you stated you wanted, in your first post. :confused:
Are you making some type of tube audio amp?
 

Thread Starter

coinmaster

Joined Dec 24, 2015
502
I'm sorry, but I don't really see how that circuit will accomplish what you stated you wanted, in your first post.
I don't understand it either, some well regarded people had a difficult time understanding that particular gyrator but it works in spice and in real life so whatever.

Not really the point here though, the point is the transistor is creating a current draw where it shouldn't.
If you replace the transistor with a resistor there is no current flow and the current on U1 remains dead stable.
My question is will a JFET function just as linearly across this voltage range with my digital feedback?

Are you making some type of tube audio amp?
Sort of, more like a prototyping system.
 

crutschow

Joined Mar 14, 2008
23,802
You really need to tell me exactly what you want to do, not how you think you want to do it, because I can't really follow what you want.
Giving me bits and pieces is not working.
I suspect you are making it more complicated than it needs to be.
 

Thread Starter

coinmaster

Joined Dec 24, 2015
502
I want to replace a resistor with a voltage controlled variable resistor. One that doesn't generate its own current.

There's numerous different circuits I plan on using this for.

Giving me bits and pieces is not working.
That's because there's no "official" design, I want a voltage controlled variable resistor that can handle high voltages and that won't interfere with the circuit it's in. Nothing more.
 
Last edited:

Thread Starter

coinmaster

Joined Dec 24, 2015
502
An answer to my original question would be great instead of all this beating around the bush.
Will a Jfet function linearly with the digital feedback I'm using?
 

crutschow

Joined Mar 14, 2008
23,802
An answer to my original question would be great instead of all this beating around the bush.
Will a Jfet function linearly with the digital feedback I'm using?
The beating around the bush is because you haven't shown me what's in the bush.
You've shown me a circuit but I don't see how you are controlling the JFET to get the variable resistance or using feedback to linearize it.
In general a JFET can provide a variable resistance but it's very non-linear with respect to the gate-source control voltage.
That's why I'm trying to understand your application so I can better advise you on the best way to do what you want. It's quite possible that trying to use a JFET as a variable resistance is not the optimum approach.
Otherwise if that answers your question about JFETs then we are good.
 
Top