Some time ago I presented a design where I replaced the non-working chopper circuitry for a Fluke 887 differential voltmeter with an LTC1050 precision chopper amplifier. The design seemed to work fine, but I have discovered an issue ( I haven’t actually used the meter very much). The manual says to zero the meter by shorting the input terminals and then adjusting the zero adjustment for a zero reading on the meter. This works fine with the LTC1050 circuit, but I recently noticed the meter reads -0.08 when the input is open on the 1V range and the null switch is on 0.0001 (it is fine on all other ranges/null switch settings). This equates to a -8uV reading. I believe this is due to the input bias current of the LTC1050 and the 1 megohm input resistance of the 887 on the most sensitive range/null setting. I tried three different LTC1050s and they all exhibit the same problem. However, this would imply an 8pA bias current, which is much higher than expected. My question is: does anyone else have a better explanation for this issue and does anyone experience this issue with the original circuit? Thanks, Randy Evans AE6YG

Could it be current noise on the LTC1050 inputs? Charge pumping produces a current noise and bias current spikes which are deceptively large given the input bias current specification. On Wed, 26 Apr 2017 19:17:08 -0700, you wrote: >Some time ago I presented a design where I replaced the non-working chopper >circuitry for a Fluke 887 differential voltmeter with an LTC1050 precision >chopper amplifier. The design seemed to work fine, but I have discovered >an issue ( I haven’t actually used the meter very much). The manual says >to zero the meter by shorting the input terminals and then adjusting the >zero adjustment for a zero reading on the meter. This works fine with the >LTC1050 circuit, but I recently noticed the meter reads -0.08 when the >input is open on the 1V range and the null switch is on 0.0001 (it is fine >on all other ranges/null switch settings). This equates to a -8uV >reading. I believe this is due to the input bias current of the LTC1050 >and the 1 megohm input resistance of the 887 on the most sensitive >range/null setting. I tried three different LTC1050s and they all exhibit >the same problem. > >However, this would imply an 8pA bias current, which is much higher than >expected. My question is: does anyone else have a better explanation for >this issue and does anyone experience this issue with the original circuit? > >Thanks, > >Randy Evans AE6YG

Randy wrote: > but I recently noticed the meter reads -0.08 when the > input is open on the 1V range and the null switch is on 0.0001 > This equates to a -8uV reading. > I believe this is due to the input bias current of the LTC1050 > * * * > However, this would imply an 8pA bias current, > which is much higher than expected. I am attaching the relevant portion of the LTC1050 datasheet. Note that at 25C, the 1050 has an input offset current spec of +/- 20pA (both typical -- max is ~5x worse). I would expect a total input offset of +/- 20uV (typical) in your circuit. (The 0.5uV typical input offset voltage is not a significant contributor in this context.) Best regards Charles

Charles, You are absolutely correct. I misread the data sheet. It's interesting that all three LTC1050s behave exactly the same. They have the same lot code number.but I would expect more variation. Live an learn. Randy On Thu, Apr 27, 2017 at 12:18 PM, Charles Steinmetz <csteinmetz@yandex.com> wrote: > Randy wrote: > > but I recently noticed the meter reads -0.08 when the >> input is open on the 1V range and the null switch is on 0.0001 >> This equates to a -8uV reading. >> I believe this is due to the input bias current of the LTC1050 >> * * * >> However, this would imply an 8pA bias current, >> which is much higher than expected. >> > > I am attaching the relevant portion of the LTC1050 datasheet. Note that > at 25C, the 1050 has an input offset current spec of +/- 20pA (both typical > -- max is ~5x worse). I would expect a total input offset of +/- 20uV > (typical) in your circuit. (The 0.5uV typical input offset voltage is not > a significant contributor in this context.) > > Best regards > > Charles > > > > _______________________________________________ > volt-nuts mailing list -- volt-nuts@febo.com > To unsubscribe, go to https://www.febo.com/cgi-bin/ > mailman/listinfo/volt-nuts > and follow the instructions there. >

It looks like a better candidate would be the LTC2054, which has a typical bias current of 1 pA at 25C and is essentially constant up to 45C. Randy On Thu, Apr 27, 2017 at 5:02 PM, Randy Evans <randyevans2688@gmail.com> wrote: > Charles, > > You are absolutely correct. I misread the data sheet. It's interesting > that all three LTC1050s behave exactly the same. They have the same lot > code number.but I would expect more variation. Live an learn. > > Randy > > On Thu, Apr 27, 2017 at 12:18 PM, Charles Steinmetz <csteinmetz@yandex.com > > wrote: > >> Randy wrote: >> >> but I recently noticed the meter reads -0.08 when the >>> input is open on the 1V range and the null switch is on 0.0001 >>> This equates to a -8uV reading. >>> I believe this is due to the input bias current of the LTC1050 >>> * * * >>> However, this would imply an 8pA bias current, >>> which is much higher than expected. >>> >> >> I am attaching the relevant portion of the LTC1050 datasheet. Note that >> at 25C, the 1050 has an input offset current spec of +/- 20pA (both typical >> -- max is ~5x worse). I would expect a total input offset of +/- 20uV >> (typical) in your circuit. (The 0.5uV typical input offset voltage is not >> a significant contributor in this context.) >> >> Best regards >> >> Charles >> >> >> >> _______________________________________________ >> volt-nuts mailing list -- volt-nuts@febo.com >> To unsubscribe, go to https://www.febo.com/cgi-bin/m >> ailman/listinfo/volt-nuts >> and follow the instructions there. >> > >