• Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Controlling drift on output of non-inverting amp???

Status
Not open for further replies.

Glenn

New Member
Hi - I'm involved with a project to design and devlop a fully autonomous hovering platform as part of a 4th year engineering group project.
I am dealing with the sensors, more specifically rate gyros. I am feeding the output from these gyros into a differential amp which outputs a +ve or -ve dc voltage dependant on +ve or -ve rotation. this signal then goes into an integrator which integrates the rate to give a angle of rotation. This signal is then fed through a non-inverting amplifier to bump up the signal to give +-5V for +-20 degrees rotation of the platform. The problem I have is that the output signal is drifting too much and I can't control it well enough. The strange thing is that I have managed to get the the output from the integrator to have negligible drift over minutes and also, the amplifier only has a gain of 24 (not massive eh...). I know it is pretty difficult to see without a circuit diag but has anyone got any ideas??.....I am not the best at electronic design as my degree is in Engineering and Management......thanks in advance.
Glenn Stevens :)
 

Chippie

Member
"The problem I have is that the output signal is drifting too much and I can't control it well enough. The strange thing is that I have managed to get the the output from the integrator to have negligible drift over minutes and also, the amplifier only has a gain of 24 "


Can you determine in which direction the drift is going? i.e is it positive in relation to the input or in opposition to the input ?

With a gain of 24* any change on the input is going to appear quite large, so what is the magnitude of the drift from the integrator?
 

Glenn

New Member
thanks for your swift reply.....
I have managed to reduce the drift on the integrator to a matter of approximately 5-10mV over a period of a minute or so. This should be sufficient for now as we are only really planning to have the platform controlled over a minute or so due to problems of yaw and precesion.....
The drifting output of the amp follows the slight drift of the integrator output although it kind of oscillates up and down too....

I will try to attach a circuit drawing on another reply which may be more useful....

thanks again for you help...

Glenn
 

Glenn

New Member
Hi there again....
attached is a jpeg of the circuit diagram at the moment. The 22uF cap is a massive non-leaky polyester one.....apparently more accurate for an integrator and all the op-amps are OPA277 manufactured by Texas instruments....their datasheet is located at: http://www-s.ti.com/sc/ds/opa277.pdf.....
I hope that this is more use and thanks again for your help...much appreciated....give me your details and I will reference you in my report...

thanks

Glenn
 

Attachments

pebe

Member
Glenn said:
Hi - I'm involved with a project to design and devlop a fully autonomous hovering platform as part of a 4th year engineering group project.
I am dealing with the sensors, more specifically rate gyros. I am feeding the output from these gyros into a differential amp which outputs a +ve or -ve dc voltage dependant on +ve or -ve rotation. this signal then goes into an integrator which integrates the rate to give a angle of rotation. This signal is then fed through a non-inverting amplifier to bump up the signal to give +-5V for +-20 degrees rotation of the platform. The problem I have is that the output signal is drifting too much and I can't control it well enough. The strange thing is that I have managed to get the the output from the integrator to have negligible drift over minutes and also, the amplifier only has a gain of 24 (not massive eh...). I know it is pretty difficult to see without a circuit diag but has anyone got any ideas??.....I am not the best at electronic design as my degree is in Engineering and Management......thanks in advance.
Glenn Stevens :)
I think you need a chopper-stabilised op amp such as the ICL7650 in your first stage. You'll get the details from google.

This type of amp is constantly checking its own drift and correcting for it.
 

Glenn

New Member
thanks for the advice pebe.....I have ordered samples from Maxim...

however, the op-amps which I have been using where selected precisely for the reason that they had ridiculously low drift.....They in fact seem to have the same specs as the ICL7650 however, they don't autocorrect themselves I don't think as they are not actually chopper stabalised.

Do you think that I should try to use the same opamp on for the non-inverting amp on the output of the integrator??

thanks for your advice once again - I realy appreciate it.....again, give me your details and I will reference you...

Glenn :)
 

pebe

Member
Glenn said:
thanks for the advice pebe.....I have ordered samples from Maxim...

however, the op-amps which I have been using where selected precisely for the reason that they had ridiculously low drift.....They in fact seem to have the same specs as the ICL7650 however, they don't autocorrect themselves I don't think as they are not actually chopper stabalised.

Do you think that I should try to use the same opamp on for the non-inverting amp on the output of the integrator??

thanks for your advice once again - I realy appreciate it.....again, give me your details and I will reference you...

Glenn :)
Yes, I would put them in the first two stages. They are very effective. I was once experimenting with making a miniature water flow meter. The type where you pass water through an alternating magnetic field and measure the induced current with two probes.

I couldn't get the magnetic field powerful enough, but the probes and amp worked a treat. With that amp I could actually measure the galvanic difference of different grades of stainless steel in the probes!
 

pebe

Member
op-amp drift

I've now come across my old circuit notes. I see tha amp I originally used was the LTC1051, a dual type with no external switching capacitors. Its stated typical input offset is 0.5uV, and its typical drift with temperature is zero. So it has a better spec than the other one I mentioned. But the max supply voltage is 16.5V, whereas you are using 24v - though you should be able to get around that.

There is a quad version LTC1053.

BTW, I just realised you have gain in the output stage, so on second thoughts it may be better to use these chopper types for all stages.
 

mechie

New Member
Op Amp drift

Could this be a thermal drift problem ?
You say the output amp and integrator drift in the same direction...

You can disprove this by pointing a hairdrier at the circuit - if it makes no difference then I'm wrong, if it does upset the circuit then at least we know the cause of the problem.
 

Glenn

New Member
Thanks for all your replies and help.....most appreciated.....

Unfortuantely, as I am working out of a university lab - I haven't been able to get access to it over the last few days because there are no technicians in due to holidays.....makes things rather annoying!!!

but the idea about the hairdrier is a good mechie - I'll try it.....but I dunno if it is a drift problem as the op-amps I am using are pretty reliable I think in terms of drift - specs below.....some of the lowest I could find....also - was originally looking for amps which could take +-12V due to some platform constraints....but could be got around if needs be....

l ULTRA LOW OFFSET VOLTAGE: 10mV
l ULTRA LOW DRIFT: ±0.1mV/°C
l WIDE SUPPLY RANGE: ±2V to ±18V
l REPLACES OP-07, OP-77, OP-177

thanks again - I was also given advice by a friend that I should perhaps put decoupling capacitors of 100nF across the power pin pairs of each device (pin 7 to pin 4) but I'm not sure why or how it would help??

thanks again for all your help

Glenn
 

Glenn

New Member
nah - thats what I thought....
I'm not sure why he suggested it - I think that he may have used these chips before and found that that may have helped their general performance or something.....

anyways, still can't get into the lab so will let you know how I get on next week.

thanks

G
 

Glenn

New Member
I don't think particularly - its just a normal resistor from the Uni labs...

I think that I am going to to try to reduce the time constant of the integrator by using a lower value resistor - this will mean that hopefully I won't need so much gain on the output.....

???
 
Status
Not open for further replies.

Latest threads

EE World Online Articles

Loading
Top