Idea for fast variable resistive load (0.2 - 1MHz)

[Plasmateur]

I have circuit I've designed which I need to test, but in order to do so I need a variable resistive load I can control with my signal generator.

This way I can vary the resistance as function of frequency.

I was considering just using a simple transistor. The load would need to vary from around 10k - 1M Ohms at a periodic frequency up 1MHz.

Is it standard practice to use a transistor for an application such as this or is there a better method?

Thanks!

[Rerouter]

the hard part is your periodic 1Mhz requirement, as this generally means the control loop needs to have a response a fair bit faster than that to keep up, if its a sine or triangle, it stands a chance, but square would be very difficult to not ring longer than the waveform,

Might i ask what resistance vs frequency pattern / curve you need to match? if its a log curve you can likely just use a RLC circuit to match it.

[Plasmateur]

the hard part is your periodic 1Mhz requirement, as this generally means the control loop needs to have a response a fair bit faster than that to keep up, if its a sine or triangle, it stands a chance, but square would be very difficult to not ring longer than the waveform,

Might i ask what resistance vs frequency pattern / curve you need to match? if its a log curve you can likely just use a RLC circuit to match it.

Thank you for the reply.

I would apply a sine wave at a particular frequency to oscillate a restive load's value.

Or I could apply a triangle as well. A square wave would not work.

Resistance would likely be logarithmic, with resistance starting out at around 10kOhm then quickly going to "infinity" over a very short time.   

[T3sl4co1l]

Willing to take your idea at face value but let's take a quick X-Y problem examination:
What are the characteristics of your signal source?
How important is it, that the load actually be resistive, and variable?
How much reactance is tolerable?

If you don't understand these questions, don't hesitate to ask... but also, don't be surprised if their answers lead to a very different solution!

Tim

[f5r5e5d]

what I, V? I assume gnded from the word "load" - floating 2 terminal programmable is R a bigger problem

small signal long channel fets are used as variable R, some parasitic C, only ma currents

another approach would be multiplier controlled current sink, multiplying a V from a ref R and your control signal

[bson]

An optocoupler might work for this.

https://www.digikey.com/product-detail/en/luna-optoelectronics/NSL-32SR2/NSL-32SR2-ND/5039808

Edit: nvm, didn't notice you want to change the resistance at up to 1MHz, not pass 1MHz through it.  These things are very slow, on the order of ms to change.

[danmcb]

This way I can vary the resistance as function of frequency.

Is this what you actually mean?

Do you want the resistance to vary with frequency (e.g. 10kohm at 100kHz, 20kohm at 1MHz  or whatever)?

Or do you want a resistor whose resistance can waggle about as fast as 1MHz?

[Plasmateur]

Willing to take your idea at face value but let's take a quick X-Y problem examination:
What are the characteristics of your signal source?
How important is it, that the load actually be resistive, and variable?
How much reactance is tolerable?

If you don't understand these questions, don't hesitate to ask... but also, don't be surprised if their answers lead to a very different solution!

Tim

I am attempting to simulate what is attached in the picture. I need to know the current across a resistor. And that current will depend on another resistor which will be the variable resistor.

[Plasmateur]

This way I can vary the resistance as function of frequency.

Is this what you actually mean?

Do you want the resistance to vary with frequency (e.g. 10kohm at 100kHz, 20kohm at 1MHz  or whatever)?

Or do you want a resistor whose resistance can waggle about as fast as 1MHz?

Sorry. I just face palmed myself. Yes I need the resistance to oscillate or waggle or ramp over a short time.

[Plasmateur]

what I, V? I assume gnded from the word "load" - floating 2 terminal programmable is R a bigger problem

small signal long channel fets are used as variable R, some parasitic C, only ma currents

another approach would be multiplier controlled current sink, multiplying a V from a ref R and your control signal

It appears the fets might be the best option for me. I only need a few mA of current

[PointyOintment]

what I, V? I assume gnded from the word "load" - floating 2 terminal programmable is R a bigger problem

small signal long channel fets are used as variable R, some parasitic C, only ma currents

another approach would be multiplier controlled current sink, multiplying a V from a ref R and your control signal

It appears the fets might be the best option for me. I only need a few mA of current

How much voltage, though? Plasma suggests HV… MOSFETs don't go higher than hundreds of volts.

[Plasmateur]

what I, V? I assume gnded from the word "load" - floating 2 terminal programmable is R a bigger problem

small signal long channel fets are used as variable R, some parasitic C, only ma currents

another approach would be multiplier controlled current sink, multiplying a V from a ref R and your control signal

It appears the fets might be the best option for me. I only need a few mA of current

How much voltage, though? Plasma suggests HV… MOSFETs don't go higher than hundreds of volts.

It's not high voltage in the plasma for this particular regime.

The plasma's space potential is only a few volts at most.