EEVblog Electronics Community Forum
Electronics => Beginners => Topic started by: sparkgap on October 25, 2012, 01:16:52 am
-
Hey everybody, long time lurker, first time poster here.
I'm trying to build a constant current power supply using an LTC1100 instrument amplifier to measure a shunt resistor for feedback, feeding that in to an op amp (LT1013) inverting input, with a setpoint voltage coming from a pot in to the non-inverting input. Output of the op amp going to the set pin of an LT3080 through a 22k resistor. The output of the regulator goes to an LED through a 1R shunt resistor which I'm measuring across with my DMM on mV setting to see how much current is flowing. See attached schematic.
The circuit is working mostly, I can twist the pot and measure a current flow that changes linearly through the resistor. The LED brightness varies, and it all looks good, until I connect a scope to the output, or the set pin or the op amp output, and it is oscillating pretty severely. Now, I'm still learning signal processing and control loop stuff, so I am not sure if I am understanding this correctly, but it seems that the problem might be that the current amplifier has a 180 degree phase shift straight out of the box, if I'm reading the datasheet correctly. For reference, I'm referring to the middle chart on the bottom of page 4 of this document: http://cds.linear.com/docs/Datasheet/1100fc.pdf (http://cds.linear.com/docs/Datasheet/1100fc.pdf)
Also, I'm noticing that the gain of the amplifier is not particularly close to 100 as advertised, though I am sure it is because I am doing something wrong. It seems to be stable around about 67, though since I'm intending to program the current by simply measuring it across a shunt resistor while I adjust the pot, this isn't a particularly terrible problem, but I'd love to understand why it is happening.
So, to summarize, I have 2.5 questions:
1) Am I reading the "Gain, Phase vs Frequency" plot correctly and seeing a 180 degree phase shift for signals under 1.8kHz?
1b) If so, is this likely to be a culprit in my oscillation problem?
2) Is there something wrong with my circuit that would cause this gain discrepancy?
Thanks,
Matt
-
Your amplifier(LT1030) is set up as a voltage comparator. You need a differential amplifier setup for it to work properly. Right now its just going high or low causing the regulator to swing off and on, instead of making the small adjustments needed to control the circuit, which is most likely causing the oscillation you're seeing and causing the LTC1100 not to function properly.
-
When I looked at it on the scope, the output of the op amp was definitely not going rail to rail, it was a DC biased sine wave, the DC bias depending on the setpoint but the peak to peak of the oscillation was about 200mV. Also, I verified the gain problem with the LTC1100 output disconnected and a fixed voltage across the 0.1 ohm shunt, so I don't believe the op amp has anything to do with it.
Matt
-
Maybe we could try breaking this down and going step by step, is there anything wrong with the way I'm using the LTC1100 that could cause my gain to be wrong?
Am I reading the phase shift correctly from that plot I linked to? If so, is it even possible to have a stable control loop with that chip in place?
Matt
-
The datasheet for the LT3080 http://cds.linear.com/docs/Datasheet/3080fc.pdf (http://cds.linear.com/docs/Datasheet/3080fc.pdf) states on page 10 that you need an output cap for stability maybe that is the problem?
-
I do have some output capacitance on the other side of the sense resistor, the two 47uFs. I can try adding something before the sense resistor as well, but if I bypass the current loop and provide a fixed setpoint through the set resistor, the voltage output is pretty solid, no oscillating.
Anyone got an answer on that phase plot thing? :)
Matt