EEVblog Electronics Community Forum
Electronics => Beginners => Topic started by: TMM on March 05, 2014, 11:54:56 am
-
Hey guys/girls.
I'm using an AD795 opamp (datasheet (http://www.analog.com/static/imported-files/data_sheets/AD795.pdf)) in the following circuit:
(http://farm3.staticflickr.com/2887/12947179605_b19c078a5f_b.jpg)
The problem is that i'm getting more input offset voltage than I expect. From my newbie understanding of the datasheet, at 25*C i should expect Vos to be 100uV typical, 500uV max, and through a 0-70*C range it is typically within 300uV and within 1000uV max. Is this correct?
Measuring across the inputs of the opamp (Fluke 87V HighZ + HiRes mV) i get 760uV @ 22*C ambient. Measuring across the voltage reference i get 4.999V and at the output i get 4.998V, backing up the 760uV measurement.
I realise the AD795 datasheet only characterises it for a +/-15V supply, but i have tried up to a 20V supply and the offset did not budge at all.
As far as i can tell, bias current shouldn't come into play. I tried loading the output with a 50ohm resistor and there was no change. Freezer spray on the opamp brings it up to 800uV.
I was previously using an OPA735 in the same circuit and the input offset was basically zero, so this has me stumped! (I had to get rid of the OPA735 because the auto zeroing circuitry was injecting noise into the output).
Thanks in advance :-+
-
What are the input currents doing, and what does the datasheet say about input currents?
-
My understanding is that any current at the non-inverting input won't do anything because the voltage reference is clamping the voltage. I thought it might have something to do with the inverting input seeing a high impedance to ground when the output is unloaded, that's why i tried measuring it with and without a load.
edit: I should verify that by 'the output' i mean the node marked 'out' in the schematic not the output of the opamp.
-
Is your DMM ok? (shorted leads => 0 mV?)
Is the circuit stable? (Won't show on a DMM, you need an oscope)
-
Is your DMM ok? (shorted leads => 0 mV?)
Is the circuit stable? (Won't show on a DMM, you need an oscope)
Yes and yes.
-
A couple of comments:
- The MTP3055V has a max Vgs of 4V (3V typ.). That requires a drive voltage of 8V or more to reach the 5V output. The AD795 is only specified for Vout of Vs-4V (2.5V typ). It will probably work, but more margin would be better since this may come back to haunt you with heavier/transient loads.
- I would put a decoupling cap (0.1uF) across the AD795 supply pins.
- The AD795 datasheet recommends having a minimum load resistance of 10k. The right way to do this would be a 20k to +10V and a 20k to GND.
Things to try and then re-measure the offset:
- Reconfigure the AD795 to be a dumb follower of the 5V reference, and with the 10k load (but no M1, R2, C1, etc.).
- Stiffen the 5V reference with 10k instead of 100k, although I doubt that's having any affect with this opamp.
- Vary the reference voltage from 2.5V to 7.5V.
- If you have one, go to a dual supply for +/-15V and re-home the 5V reference to the new ground. Decouple both rails to the new ground with 0.1uF.
-
I'm a bit surprised that this circuit is stable, you do have two LP-filters external to the op, R2 and Cg of M1 and any resistance feeding C1. These 2 and the compensation of the op could very well add up to 180 deg.
-
The common mode and power supply rejection ratios of the AD795 are relatively low like most JFET operational amplifiers and may be enough to push the offset far enough away from the 500uV maximum to explain this when operating at low power supply voltages.
-
Judging from your LT-Spice schematic. Could it be, that you try to operate this op amp in single supply mode? From a quick look over the datasheet it looks like it isn't a single supply op amp. And if it's not a true rail to rail type op amp it might have trouble reaching 0V without a negative supply.