Zero drift amplifier input bias current

[David Hess]

The JFET input noise may have a different spectral composition than the precision BJT parts.  Often one is indeed interrested in the frequencies lower than 0.1 Hz. The 0.1 - 10 Hz range is more like choosen to be easy to measure (AC coupling still works and 2 decades is large enough so that the exact filter shape is not that critical (still a factor to cause condusion). For the lower frequency noise a relatively fat DMM (to work well in 1 PLC mode) should be a reasonable choice. If not with a low of DC offset like a voltage reference the amplification can be quite large in a DC coupled configuration.

I discovered the OPA140 when researching suitable improved replacements for the old AD542, which at one time was about the best precision JFET part.  The LT1012 is the bipolar alternative that I would have used by default.  My primary interests were high common mode rejection, low 0.1 to 10 Hz noise, low input bias current, and low input offset voltage drift.

The difference in noise is apparent.  The JFET in this case has lower high frequency noise, but its higher corner frequency means comparable low frequency noise.

OPA140   140 dB cmrr   1 uV/C max   0.5 pA   0.1 to 10 Hz 230 nVpp   8 Hz   5.1 nV/SqrtHz
LT1012   132 dB cmrr   1 uV/C max   30 pA   0.1 to 10 Hz 500 nVpp   2.5 Hz   14 nV/SqrtHz
LT1001   126 dB cmrr   1 uV/C max   500 pA   0.1 to 10 Hz 300 nVpp   4 Hz   9.8 nV/SqrtHz

As you point out, it is straightforward to make a spot noise measurement over the 0.1 to 10 Hz range.  Since the corner frequency usually lies within this range, it makes for a good proxy measurement of low frequency noise.  However older chopper stabilized parts have *higher* noise within this range, so are more suitable for even lower frequency applications, where their 1/f noise is flat, and a 0.01 to 1 Hz test is required for proper evaluation which is not so easy.

I have used a high resolution DMM to make low frequency spot noise measurements before and got excellent results which agreed with the datasheet values.  I am looking to improve on this and then run some tests on modern chopper parts.

A point to whatch for is thermal stability: temperature fluctuation combined with a TC may look a lot like 1/f noise.

Yea, at low frequencies it is difficult to distinguish 1/f noise and thermally caused drift.

[Kleinstein]

The thermal part is not just drift. In some cases one even gets some thermal oscillations at low frequencies.  I got an example with some 0.1 Hz (not very stable) AFAIR.

The AZ amplifiers usually have a rather flat noise in the LF range, with more thermal effects limiting to the really low frequencies below 0.1 Hz. The chopper artifacts are usually above some 5 kHz in modern parts often higher, like 50-500 kHz.

[tszaboo]

Probably not surprising, but the Linear Technology part was like 100 times better than the equivalent TI and AD part, and you couldn't see this at all from the datasheet.

Can you tell what amp it was?

Its been a while but I believe it was the LTC2057.

"The current noise spectrum of the LTC2057 is shown in
Figure 2. The characteristic curve shows no 1/f behavior.
As with all zero-drift amplifiers, there is a significant current noise component at the offset-nulling frequency. This
phenomenon is discussed in the Input Bias Current section."

[iMo]

Based on the discussion here I added 1k resistors (R2 and R3) into both inputs of my OPA189 10V buffer.

On the first glance the output voltage dropped by some 7-10uV.

Now - what about the feedback capacitor C2 - I have not any capacitor there yet.
How the C2 affects the story?

[dietert1]

C2 is meant for opamp stability, to provide fast feedback while driving a capacitor. You could try and insert a 1 KOhm resistor between negative input and C2. I think this was the proposal of macaba. One would then expect a shift of similar size but opposite sign.
I ordered some OPA847s to try and reproduce the TI study.

Regards, Dieter

[iMo]

Adding the C2=4n7 foil as in the above schematics increased the output voltage by ~30uV.

PS:

Adding a series 1k resistor with the C2=6n8 foil dropped the output voltage by ~20uV, thus I am now at the same level I was before all those above changes..
Not sure the fine-tuning of the output voltage should work this way

[dietert1]

Yes, in order to decide one needs some criterion. Candidates are low input bias current, low input offset voltage and low frequency output noise ("stability").
From macabas tests we already saw that the common mode input voltage is a parameter. Using the 10 V output voltage as supply for the chopper amp may not be the best idea.

Regards, Dieter

[iMo]

I do not use 10V as the supply for the chopper.. See my V3 in the ADR1001 thread..

[miro123]

According to opa 189 datasheet page 22
8.3.3 Input Bias Current Clock Feedthrough
Zero-drift amplifiers such as the OPAx189 use switching on the inputs to correct for the intrinsic offset and drift
of the amplifier. Charge injection from the integrated switches on the inputs can introduce short transients in
the input bias current of the amplifier. The extremely short duration of these pulses prevents the pulses from
amplifying, however the pulses may be coupled to the output of the amplifier through the feedback network.
The most effective method to prevent transients in the input bias current from producing additional noise at the
amplifier output is to use a low-pass filter such as an RC network

SHared link by Kleinstein shows that glitches are in GHz range. poor capacitor grounding and attching of  Dc measurement equipment with long wires can create wrong conclusions
Input as well as feed-trough insulation combined with RC output filters seems the practical way to go as hobbyist /limited T&M equipment and lab facility/
https://www.eevblog.com/forum/metrology/zero-drift-amplifier-input-bias-current/?action=dlattach;attach=1842343

[iMo]

Interestingly, when looking at the schematics and designs with choppers here, even the high-end ones, such special precautions has not been provided, afaik.

Perhaps those 10-12ns/2uA needle like pulses get absorbed by always existing parasitic capacitancies in those designs.

The initial experiments here shows the resistors tame the effects, most probably the R with a parasitic C does the job. The large foil capacitors may not help much, low value ceramics would be a better choice, imho.

A nice area for experiments, as many here are using choppers in their high-end references..

[David Hess]

The designs I am familiar with which use older chopper stabilized amplifiers successfully all provided low AC impedances at both inputs, which means a large feedback capacitance and a large capacitance from the non-inverting input to ground.  They were also limited to low bandwidth extending below 1 Hz.

The designs which failed presented a high impedance to the inputs, which led to excessive low frequency noise.

I should have the chance in the near future to run some tests comparing old and new parts.  I am currently working toward having better ways to characterize low frequency noise.