op-amp to drive +/-10V analog servo from old +/-5V controller

[paulus.v]

Hi all,

I am trying to control a +/-10V analog servo drive from a +/-5V controller output. The problem is that the servo is running only at half the speed.

My searching lead me to the use of a single supply rail-to-rail operational amplifier. But I'm not an electronist and couldn't find a ready to build solution/schematic.

Here is an IC, OP295 (http://www.analog.com/media/en/technical-documentation/data-sheets/OP295_495.pdf), that I can get locally. Could I use it for my application?

Has anyone here done this before and could advise me further?

Any help will be really appreciated.

[Robin Hewitt]

If you want +-10v output then you will need a split rail op-amp and a plus-minus power supply. Is that a problem?

The circuit requires only 2 resistors and a couple of decoupling capacitors for luck.

[paulus.v]

Thanks for your reply!

My only problem is that I need to build it with the components that are available locally. I need this thing asap. No matter how expensive.

[Doddy]

If you can solder, and have suitably sized iron and solder, then together with a bit of vero board (or whatever breadboard solutions you can find locally) then it's a pretty straight-forward build. Have a look at http://electronics-course.com/non-inverting-amplifier for the schematic you need (Robin's answer is correct - you'll need a couple of resistors - choose any value 1k-10k for Rf and Rg (but they must be the same as each other) - see equation (8).

Robin is also correct with the split rail supply requirement - see if you can find a DC-DC converter with a suitable input voltage for what you have available and +/- 12V through +/-15V output.

[paulus.v]

Thanks Doddy!

Now it is a bit more clear.

Luckily I found out there is a +/-12V power output from the servo drive. :thumsup:

Do I need two op-amps one for positive and one for negative? Should I look for any characteristics of the chips as I found dozens of them... Do I need any capacitors in addiction to the identical value resistors?

[Doddy]

If you're driving a single servo (I don't know the interface to these, but I'm going off your post) - and the drive is +/-10v wrt a common ground, then you need only one non-inverting amplifier.

Capacitors - pick a couple of 100nF ceramic capacitors and locate them close to the op-amp, between positive&ground, and between negative&ground.

Regarding which op-amp - how quickly will the signal change?, the one you originally picked had, I think, a gain/bandwidth product of a pretty miserly 75kHz, but if the signal is just driving an axis then I that'd be absolutely fine. If it was driving a spindle then it should still okay given the low gain you're looking at.

[AndrewMawson]

Also the gain is only 2 (ratio of feed back resistor to input resistor) so the needed gain / bandwidth is fairly modest as well

[Robin Hewitt]

I don't think we need worry about the slew rate :uncomfortableness:

[paulus.v]

I have selected some of the available op-amps but I cannot decide. I dunno why are so many of them... :single_eye:

They are ordered by price with the expensive first: LM301 (http://www.utm.edu/staff/leeb/LM301.pdf); LT1001CN8 (http://cds.linear.com/docs/en/datasheet/1001fb.pdf); LF442CN (http://www.ti.com/lit/ds/symlink/lf442.pdf); NJM4580 (http://www.njr.com/semicon/PDF/NJM4580_E.pdf); BA15218 (http://rohmfs.rohm.com/en/products/databook/datasheet/ic/amp_linear/opamp/ba15218f-e.pdf); LF356N (http://www.ti.com/lit/ds/symlink/lf356.pdf); UA747-14P (http://www.ti.com/lit/ds/slos009a/slos009a.pdf); MC33078P (http://www.ti.com/lit/ds/symlink/mc33078.pdf); NE5534N (http://www.onsemi.com/pub_link/Collateral/NE5534-D.PDF); LM318N08 (http://www.ti.com/lit/ds/symlink/lm318-n.pdf); NJM4580L (http://www.njr.com/semicon/PDF/NJM4558_NJM4559_E.pdf).

I do not get how I use only one op-amp if I have +5V and -5V input and I need +10V and -10V output? I see in the schematic only one input and one output plus ground.

[Doddy]

This goes back to my previous reply where I presented an assumption that your servo only needs a signal, variable from -10V to +10v with respect to a common ground. So, one signal, capable of any voltage between -10V..+10V. For that, a single Op-Amp, the output of which is connected to the servo, and the other servo input connected to ground.

It's possible, though without any literature we have no means to corroborate, that the servo requires a differential drive - with one signal increasing from 0..+10V, whilst the other input decreases from 0..-10V, in which case you certainly would need 2 op-amps, or a specialised op-amp that provides the differential drive. Such a signalling regime is good for noise immunity.

Can you offer any of the literature that you have for the servo just so we can be certain of the advice offered?

[paulus.v]

Yes, my bad. There is one signal +/-10V and ground.

[paulus.v]

Now the question is how accurate will be the 0V, where the motor should stand still... Some chips have some sort of balance but don't understand what it is for.

[Doddy]

Yes some op amps allow you to set precisely the 0v offset voltage, but not the original op amp that you specified. These are used when precision is required, and it's unclear whether you will need this or not. If in doubt code an op amp that supports this and reference the data sheet for how to use it (a common use is a potentiometer between the balance pins with the wiper connected to -V, but that's not due for all opamps

[paulus.v]

Yes some op amps allow you to set precisely the 0v offset voltage, but not the original op amp that you specified. These are used when precision is required, and it's unclear whether you will need this or not. If in doubt code an op amp that supports this and reference the data sheet for how to use it (a common use is a potentiometer between the balance pins with the wiper connected to -V, but that's not due for all opamps

I have linked to a dozen op-amps in a previous post. Please point me out one of them if you have the time. I need a robust one.

The 0V is important as the positive voltage will turn the motor in one direction and the negative in the other so at 0V it should stand still. But I could filter out a "dead zone" from the drive settings if necessary.

[Doddy]

I have linked to a dozen op-amps in a previous post. Please point me out one of them if you have the time. I need a robust one.

The 0V is important as the positive voltage will turn the motor in one direction and the negative in the other so at 0V it should stand still. But I could filter out a "dead zone" from the drive settings if necessary.

Chances are any of the op-amps would work well, particularly if you can set a dead-zone around zero.

If you insist on having a balance control then I'd look at the LF356, just from 5 minutes glancing at the datasheets - there's an example of how to connect the balance trim pot (I'd select a multi-turn pot for this); but I doubt that you really need it given the low gain you need.

[paulus.v]

Thanks a lot for your help, Doddy.

I'll go with LF356 just in case.

One more silly question. Should I connect the power supply ground with the controller output/ drive input ground?

[Doddy]

One more silly question. Should I connect the power supply ground with the controller output/ drive input ground?

Yes. You need a common ground.

[paulus.v]

Capacitors - pick a couple of 100nF ceramic capacitors and locate them close to the op-amp, between positive&ground, and between negative&ground.


Should I put the capacitors between the power supply pos/neg and ground or between pos/neg and ground op-amp input?

[Doddy]

Hmmm, thought I'd replied to this but was by phone. Nevermind - caps are from pin7 (+V) to ground and from pin 4 (-V) to ground. Ignore the +/- signal inputs to the op-amp as far as capacitors.

[john swift]

combining the diagram from
http://webpages.ursinus.edu/lriley/ref/circuits/node5.html

and info from the IC data sheet

20804


John

[paulus.v]

I have built the circuit with the LF356, including the offset adjustment, and everything worked perfectly while testing with a 9V battery. But in 'real world' I swapped by mistake the input/output cables and fried the chip. Had another one as backup, NJM4580, and it is doing its job as expected.

One problem solved, but the close loop servo is still "losing steps" :hororr: That's another story..

Thank you all for your help, much appreciated!

[AndrewMawson]

Good to hear a positive result :)