I am working on a discrete component project design, that requires a manual potentiometer, to be replaced with active components, so as to have electronic controle to it, as well as manual controle
.
The potentiometer, in the original circuit is used to bias a square wave multivibrator, so as for it to be able to work in PWM, without to much frequency change.

The pot is set up that each end of the pot is connected to the base of each transistor, as the pot is turned, the ratio of current to each transistor remains the same, to keep the frequency stable, but allows the width of the pulses to change accordingly.

So here is my first design using active devices, to synthesize as close as possible a potentiometer.



This is a first prototype, 'proof of concept' to see if it will work, it does work, but needs lots of rework to make it less sensitive to input voltasge change, to give a wider range of input voltage.

It is designed where the right side transistor output, biased heavily bringing the voltage at its collector to around 5v. While the left side ttransistor is biaased close to cutoff, to have its collector at 11v.

Then the middle ttansistor is introduced to be biased so as to pull both the right, and left side transistors outputs close to a quiescence of 8v.

The trimmers are used to balance the bridge, at the 8v level.

With this setup the input signal goes to the middle transistoer base, which will unbakance this bridge one direction or the other, depending on the input signal; amplitude.

The middle transistor acts as the center tap of a potentiometer, while the differential transistor pair acts as the both ends of a potentiometer, when the outputs are quiescent at 8v. then this synthesizes a potentiometer whose setting is in the middle, as the signal drives the middle transistor on harder, it synthesizes, a potentiometer, whose setting is such that the right side is becoming more positive while the left side is less positive, at the same ratio.

Same concept as a differential amp, only with this the input signal is introduced to both transistor enmitters, so there is no inverting and noninveerting inputs, but one input to a center tap, that changes the output with varying signal amplitude. Just as a pot, has a center tap.

 

Have you considered digital pots?
What is the pot that you are trying to replace being used for?

 

Hi Ron,

The pot I'm replacing is used for biasing the base voltage for both transistors, in a square wave multivibrator circuit.
Based on these equations:

Fo = [1 / (1.38 * RB * C)] for 50% duty.
and
T = (1 / fo)
t1 = (duty * T)
t2 = (T - t1)
RB1 = [t2 / (0.69 * C2)]
RB2 = [(t1 / (0.69 * C1)]
where C1 = C2

To obtain the proper ratio of base current, (for stable frequency), with a manual change of the base resistors, I set it up with a single pot where, each end of the pot, is conected to a base term, and the center tap is used to apply the base voltage through each half, simulataneously, so as to change the pot setting, will allow the same ratio of base current to each transistor, which will change the duty cycle, but keep the frequency pretty stable.

The manual pot is good for direct controle of the pulse width, but I need for it to also have automatic controle when the motor is stalling, to work as torque controle.

When I built and tested this, the square wave was around 50% duty, then as I bridged my finger across the ground resistor of the bottom middle transistor then the square wave width began to go smaller on the output, while keeping the same frequaency, when I released it the wave went back to 50% then as I bridged across the supply resistor, the wave increased in width while keeping the same frequency, the input is only good for around 0.5v. change both directions, before the transistors cutoff and saturate, causing the frequency to change as well.

So now my design challenge is to logically think this through, and rework this prototype, to have less gain so as to track more reasonably the change in input voltage, with the output voltages.

Thanks.

 

Post a schematic of the circuit with the manual pot(s) in it.

 

Here is a schematic of the original circuit without the pot.
Resistors R2 and R4 are the resistors in question.



Here it is with the pot. taking the place of both R2 and R4 together.

.


Here is a video of the waveform on the osciloscope, the frequency remains fairly constant, as the pulse width is varying.

http://www.youtube.com/embed/FpFblxf4Y_U

When I get the active (synthesized pot) reworked to match the voltages of this new circuit, I'll post a schematic of the whole network, prototyped thus far.

 

Are you doing this as a learning experience, or do you just need a fixed frequency, variable duty cycle waveform? If it is the latter, there are easier ways to do it.

 

a photo-fet optocoupler such as the H11F series can be used as a 'remote-control' variable resistor. the datasheet for the H11F1 shows a very simple 'automatic gain-control' application.

 

Are you doing this as a learning experience, or do you just need a fixed frequency, variable duty cycle waveform? If it is the latter, there are easier ways to do it.

Hi Ron,

It is for a learning experiance, because I enjoy the hobby of designing circuits from discrete components, I'm sorry, for any inconveniance, I should have pointed that out sooner, that this was just a educational project.

But, I do apreciate all the time taken and the help you and ''bribri' were willing to give.

Thanks again,
very much appreciated..

 

FETs can be driven to act as variable resistors.

 

Thanks for that great idea for FET, I'll experiment with that as well.

Here is a schematic of this concept so far,
the 30K at the base of Q3 is part of the bias of Q3 to make the square wave around 50% duty.



From there I would bridge my fingers across each resistor R6 and R7 to work the square wave to around 80% and down to around 20% duty. Making it act as a potentiometer, with the wiper in the middle.

Then I changed the 30K for a 27K resistor, to put the square wave around 20% duty, then in the video, I used a wet piece of paper towel and with the jumpers connected across the R6 (9.1K) I was able to see if I could gradually adjust the waveform from 20% to 80%, with some linear consistency.

Here is a video of it demonstrating it working to this point.

http://www.youtube.com/embed/q_AFNSST4vk

Now I'll continue to work on this to get better wider range of the waveform output.

 

hobbyist Have you thought about using of differential amplifier?

 

Yeh, I tried a diff amp. but the currents through the transistors were not dividing properly, which made for a large amount of frequency change.



Thanks for taking the time to put your schematic together.

I'll build it sometime and see how it works in the stability of the frequency.

 

Yeh, I tried a diff amp. but the currents through the transistors were not dividing properly, which made for a large amount of frequency change.

View attachment 28414

Thanks for taking the time to put your schematic together.

I'll build it sometime and see how it works in the stability of the frequency.

I thought of a diff amp, but the problem is, the pulse widths are inversely proportional to current, so, as you said, the frequency will change as you vary the current proportions, even though the total current remains constant.

 

I am trying to get a larger swing in duty cycles with this concept design.

So I did this by making Q2 Vout go as low as 2v. before introducing Q3 into the mix. Keeping Q1 at 11v.out.
Then with Q3 inserted, I adjust both trimmers to get a quiescent balance of 6.5v.output at both collectors.



This worked well, I am able to get close to 90% delta of duty cycle.
Starts at a low of ~8% and goes to ~96% of duty.

With no input signal (change in resistance of either RB1 or RB2) then the balance of 6v.output of both collectors are coupled directly to the squarewave oscillator, and the~ 50% duty cycle is shown.



As I got to thinking about this, since I have a differential output balanced, this is the same as a CE amp quiescent point, so taking the Vout, of either collector, Q1 or Q2, should give me an amplified signal on the output, if I take the Vout from Q1 it would be inverted and if the output is taken from Q2 the signal would be in phase.

So I unhooked the squarewave circuit, and applied a sinewave of 50mV. capacitive coupled to the base of Q3, keeping Q3 biased as for a balanced condition at the Q1 and Q2 collectors, and got a nice undistorted sinewave Vout. of around 3v.
This gives a nice Av.~60.

When I increased the input further I got nice symetrical clipping above and below the 0v. line.

The Vout is single ended off of the Q2 collector ref. to ground.
Here is the signal voltage using the same scale of voltage on both input and output.



Here is the input scaled up to 50mV. showing the non inverted output.



This is an interesting educational study.