MAX4239 offset

[ArticCynda]

Hi everyone,

Due to its unavailability a while ago I've built my own µCurrent for measuring very low power microcontroller power consumption. The µCurrent is an open source design, and I sourced all parts from Digikey because the order codes are also made available in the BOM.

The only part I didn't source from DigiKey are the MAX4239 opamps because they weren't (and still aren't) available in SOT23-6 package. So instead I ordered them from AliExpress (http://www.aliexpress.com/item/New-MAX4239AUT-T-IC-AMP-PREC-LDRIFT-LN-SOT23-6-best-pirce-IC-supply-chain/1992017520.html). After assembling the µCurrent it shows a significant offset of -700 mV at the outputs when the current inputs are shorted.

When tested on a breakout board, the amplifier does behave as one would expect from an opamp, except that the output voltage is 28 mV when both inputs are grounded. For a low offset, low drift opamp I find this on the high side since this offset is also amplified by the second stage, thus resulting in a cumulative error.

Does anyone have any thoughts about this behavior? I don't have much experience with correcting opamp offsets, but the design doesn't include any compensation circuitry nor does the opamp provide offset compensation inputs. How do I get rid of this offset? Could it be that I have been scammed and sold inferior opamps in an identical package?

Thanks in advance!

[wraper]

It could be fake IC or they just sent you a different part. Check if the marking match the datasheet.

[ArticCynda]

The top marking read "ABAA" which is consistent with the datasheet. Is it possible I'm dealing with a lower quality opamp that got relabeled as "ABAA"?

[wraper]

The top marking read "ABAA" which is consistent with the datasheet. Is it possible I'm dealing with a lower quality opamp that got relabeled as "ABAA"?

Probably, but read the next post. If you buy some expensive opamp from China for relatively cheap, you will get a fake more often than a real deal. If you didn't confirm goods received yet, open a dispute on Aliexpress.
http://zeptobars.ru/en/read/OPA627-AD744-real-vs-fake-china-ebay

[wraper]

When tested on a breakout board, the amplifier does behave as one would expect from an opamp, except that the output voltage is 28 mV when both inputs are grounded. For a low offset, low drift opamp I find this on the high side since this offset is also amplified by the second stage, thus resulting in a cumulative error.

BTW, I reread this now. This is not proper way how to test opamp. To measure input offset, you need to ground non-inverting input and connect inverting input to the feedback loop (2 resistors) to set gain. If you connect both inputs to the GND, basically you will have infinite gain.

[Dave]

The non-inverting input should be connected to ground with a resistor, with a resistance equal to both feedback resistors in parallel. This cancels out the bias currents.

[ArticCynda]

I measured again following Dave's recommendations on measuring opamp offsets, and I always end up with offsets between -400 mV and 360 mV with a 101x amplification. This is still way out of range for this opamp type.