Signal voltage adjustment circuit

[servo]

Hey, all. I've run into a real problem with my control circuit. The circuit is simple enough. I am designing a controller for a servovalve that looks something like this.

The frustrating thing is, it works in one direction, but when I flip the switch to reverse polarity, it stops working unless I turn the pot all the way open and then I get a large current draw. I don't understand at all why this is happening. Why does it work in one direction, but not the other? When I hook up a power supply directly to the switch, it works fine, but with the regulator and pot in there it starts acting funny. Help please! I've been stumped for over a day and I need to get this figured out.
Edit - I should say that I forgot to ground the regulator in the schematic. It is grounded on the breadboard, though.

 

[yungman]

You need to give the input requirement of the device you are driving to know whether your circuit is correct. But this is just my guess:

Your negative control is nothing more than the reverse of the positive control. The positive is from 0 to 12V, the negative is nothing more than from 12 to 0 V. Is that what you want, or you really want 0 to -12V control? You don't get -12V in your circuit.

 

[servo]

The command signal is +/- 10V but it doesn't hurt anything to go a little over. The current is only a couple mA. Shouldn't the switch simply reverse the direction of the signal? Why does it only work in one direction?

 

[servo]

I don't need "negative voltage," per se. I simply need the signal to flow in the opposite direction when I flip the switch. It works when it is connected directly to the power supply, but not when I'm using the regulator and pot. I just don't understand why it will only work one way and not when switched to reverse.

 

[yungman]

The command signal is +/- 10V but it doesn't hurt anything to go a little over. The current is only a couple mA. Shouldn't the switch simply reverse the direction of the signal? Why does it only work in one direction?

No, your circuit only provide 0 to +10V, you have no -10V capability. Look at your schematic again, That is exactly what my first post said, you just reverse the direction only, you don't get -ve voltage out of your circuit. Think about it, you flip the other way, you just reverse from 0 to +10 into +10 to 0, you don't get from 0 to -10V.

 

[servo]

Forgive my ignorance, but I don't see the difference between 0 to -10V and 10 to 0V. It's all moving in the same direction, isn't it? By flipping the switch, aren't I just changing the direction of the flow through the command pins? What is the difference between attaching it directly to a power supply and doing what I did in the schematic?

 

[yungman]

Forgive my ignorance, but I don't see the difference between 0 to -10V and 10 to 0V. It's all moving in the same direction, isn't it? By flipping the switch, aren't I just changing the direction of the flow through the command pins? What is the difference between attaching it directly to a power supply and doing what I did in the schematic?

You need to put actually -10V to drive the unit, right now, your lowest voltage is 0V no matter which side you choose, you need -10V. Right now, you are only provide half the control range. You are going to need two regulators, one 10V, one -10V and then you have the range to control the circuit.

From your circuit, you only flip the direction of the control. Use the multimeter to measure the control and you see what I mean.

 

[servo]

When I flip the switch, I get nothing out the other side. It doesn't make any sense to me, but I guess that's why I'm here. I'll try what you recommend.

 

[servo]

Before I do that, though, when I disconnect the leads from the load and run it through the multimeter instead, it makes no difference. I get the signal one way, but when I flip the switch it disappears like I'm flipping an off switch. I've tested the switch and the rest of the components individually and they all seem to work fine.
Edit - nevermind. I rewired it and now I get voltage both ways on the multimeter, but it still acts funny when I reverse it. When the pot is anything other than fully open, I'm reading 20V+ off the wiper and when it is fully open the circuit draw jumps up to .5 amps and the servo starts vibrating. Now I'm totally stumped again.

 

[yungman]

Before I do that, though, when I disconnect the leads from the load and run it through the multimeter instead, it makes no difference. I get the signal one way, but when I flip the switch it disappears like I'm flipping an off switch. I've tested the switch and the rest of the components individually and they all seem to work fine.
Edit - nevermind. I rewired it and now I get voltage both ways on the multimeter, but it still acts funny when I reverse it. When the pot is anything other than fully open, I'm reading 20V+ off the wiper and when it is fully open the circuit draw jumps up to .5 amps and the servo starts vibrating. Now I'm totally stumped again.

Can you give the new drawing in more detail, also give me more info on the input spec of the servo device you are driving? I want to check to make sure you got the input correct first.

The way you described "vibrating", it could well be talking about a different problem...a close loop instability at certain operating condition. Where did you get the servo unit? What load is the servo driving and is it following the requirement of the servo driver? I design quite a few closed loop feedback control(servo), you need to be very care in doing the stability analysis. One very good example is switching DC to DC converter, I see so many of them oscillating at light load condition that the manufacturer actually tell you that you need a minimum load to get rid of the noise. It is nothing more than instability at extreme load and is easy to fix.

 

[uart]

Hi servo. The problem is that you are switching which side of the "command" input input is grounded and which one receives the input voltage. Yes this can be the equivalent to simply reversing the the polarity but only if the command+/- inputs are strictly differential (with a common mode range of at least 12 volts). You need to conform that this is indeed the characteristics of the command+/- inputs.

 

[servo]

The valve itself is powered by 24VDC. This 24V is what powers the servo, which is a pilot spool that moves back and forth depending on the command voltage. The command voltage is translated by a microcontroller inside the valve where it sends out the actual current to move the spool. When I supply the command voltage directly to the valve using a variable power supply, as I have been doing, it works just fine in both directions with no problems. It is only when I try to control it with the regulator/pot setup that I have trouble. Because of being able to run it off a power supply, I know that I do not need dual power. All I need is 10VDC flowing in one pin and out the other. I think what I have is a grounding issue, but I don't know where or why. I've attached my setup.

 

[servo]

Wow, wrong image. Here's the right one.

 

[uart]

Ok so you've definitely verified that you can switch the ground connection from command- to command+ like that, then the input must be floating differential (at least within some common mode range). Make sure you haven't inadvertently added any other ground connections to command- anywhere in the circuit (other than the ONE ground connection that is made through the center point of your two switches).

 

[uart]

BTW. What type of switches (relays?) are you using?

 

[servo]

I'm only using 1 switch. It's a standard reversal switch. The pot is a 10k .5W single turn.

 

[servo]

That's the funny thing, though. If it didn't work at all it would be easier to fix, but it works properly only when the input is command + and ground is command -. With this setup it doesn't work the other way around. It's just killing me trying to figure out why works one way, but not the other.

 

[uart]

Everything you have written tends to indicate that you have an inadvertent ground connection somewhere on the command- input.

Can you measure the voltage applied to each input relative to ground, for each of the switch positions.

 

[yungman]

The valve itself is powered by 24VDC. This 24V is what powers the servo, which is a pilot spool that moves back and forth depending on the command voltage. The command voltage is translated by a microcontroller inside the valve where it sends out the actual current to move the spool. When I supply the command voltage directly to the valve using a variable power supply, as I have been doing, it works just fine in both directions with no problems. It is only when I try to control it with the regulator/pot setup that I have trouble. Because of being able to run it off a power supply, I know that I do not need dual power. All I need is 10VDC flowing in one pin and out the other. I think what I have is a grounding issue, but I don't know where or why. I've attached my setup.

Can you give the input requirement of the servo? Is it from +10 to -10?

If you use the power supply to vary from 0 to 10V, then switch the two wires around and adjust the supply from 0 to 10V, you can exercise the full range because you are running from 0 to 10V, and when flipping the two wires, you put 0 to -10V. Of cause it will work. But your first schematic can only provide 0 to 10V only.

Give me the input spec, I think I can design a simple one for you in no time. It is too confusing talking like this. All you need to do is give me the input requirement of the servo.

 

[servo]

All of the relevant pins on the valve connector are as follows:
Pin 1 - +24V
Pin 2 - 0V
Pin 3 - +24V
Pin 4 - command +
Pin 10 - command -

The command signal needs to be +/- 10V on pins 4 and 10 for full extension, but I need to be able to adjust it through the whole spectrum, hence the pot. I was planning on putting a SPST on pin 3 once I got the command signal worked out.

 

[yungman]

All of the relevant pins on the valve connector are as follows:
Pin 1 - +24V
Pin 2 - 0V
Pin 3 - +24V
Pin 4 - command +
Pin 10 - command -

The command signal needs to be +/- 10V on pins 4 and 10 for full extension, but I need to be able to adjust it through the whole spectrum, hence the pot. I was planning on putting a SPST on pin 3 once I got the command signal worked out.

Sounds like it is a differential signal, but what is the input range. Can you give more detail on the voltage range, the document of the servo must tell you how to control the input.

I notice there is not negative input voltage, that make me think that you don't need -ve voltage driving the command inputs. Question is whether it require a true differential voltage like one end is +10 to command+ and 0V at the command- and adjust linearly to 0V fto command+ and +10V to command-. If this is the case, you need a dual pot and hook in reverse.

The other possibility is you need to bias the command- to +5V and adjust the command+ from 0 to +10V.You need to give me the input spec of the command signals.Better yet, why don't you post the link to the data sheet of the servo unit so I can just read it myself. Been going back and fore too many times already.