I have an application in which I need to amplify an extremely small current from an ion cell. In the attached circuit I use a very low offset op amp (TI's LMP7721) operating in unit gain with 1G ohm resistors and this is coupled to a second op amp (LMP7715) to invert the signal and apply offset for presentation to the 64 x gain input of an ADS1230 20 bit AD converter. Both op amps operate on +/- 5 volt rails - the -5 volt being supplied by a switching converter LTC1144.

Everything seems to work flawlessly until I get a sudden failure of the LMP7715 where the -5 volt rail gets pulled way down. All other components appear to be working when I replace that chip and I just had a second failure today.

Am I missing something blatantly obvious?

 

The schematic has an error - the capacitor on the input to the LMP7721 is 10 pF mica - NOT 10uF electrolytic! It's there to dump any 100MHz RF which might be on the DC signal

 

Hi spottymaldo, it might be that your input voltage is going to high. It maybe a very small current from your ion cell but the voltage can get very high. You normally use 2 back to back diodes to prevent this. In the case of very low input current however its common to use 2 very low leakage JFETs with Drain and Source connected together.
I was doing something similar myself, and somewhere there is a nice application note about the JFET idea.
One thing I also came across was that using very high resistances in amplifier circuits was not ideal, and that for ion current measurements "switched integrators" were highly recomended and also gave you an enormous range of measurment. See the data sheet for the IVC102 if you are interested.

 

pmd34: Your suggestion is very useful. I'd been lured into a false sense of security by the unity gain of that first stage, forgetting that a device like that, with such high impedances, can lock up even if you just happen to sneeze nearby! Diodes it shall be - and many thanks. On the IVC102, again thanks - I shall look into it as a possibility for the next generation.

 

Hello,

Do I see some shorts by solder drops?



Bertus

 

Hello,

Do I see some shorts by solder drops?

View attachment 171191

Bertus

Hi Bertus: You have an eagle-eye and I very much appreciate you taking the trouble to point this out - yes, the right hand pair are deliberately shorted (the pin needed to be grounded) and the next one over shouldn't be, and isn't, but nearly is! Sloppy hand soldering eh? Thanks again!

 

Hi Bertus: You have an eagle-eye and I very much appreciate you taking the trouble to point this out - yes, the right hand pair are deliberately shorted (the pin needed to be grounded) and the next one over shouldn't be, and isn't, but nearly is! Sloppy hand soldering eh? Thanks again!

 

pmd34: Further to your valued input - the LMP7721 is an expensive chip and, checking the spec. I see that it is already input protected with two diodes - which is a relief. Nevertheless, I've still protected the buffer stage following your suggestion as this was the chip that actually failed and the input voltage is intended always to be low. One problem with the ion current here is that there is an offset component which has to be nulled in the buffer stage before being applied to the gain 64 input of the AD converter.

 

I've done the diodes; changed both the op amps (both were gone) and the two 10V rated capacitors which had 10V on them to 50V rating, given the whole thing a clean and it is working (for now). Thanks for the inputs.

 

Can you post a photo showing the LMP7715 on the pc board?
SG

 

Can you post a photo showing the LMP7715 on the pc board?
SG

It's on the pic I already posted - on the opposite side from the 7721 - in the SOT23-5 package