LTC1052 + LTC1028 noise control ? amplifier compare to LTC1043 solution

[Overspeed]

Hello

As I search a easy ( not too costly not hard to locate IC ) solution to built an measuring amplifier DC for resistivity measurement

I have located these  schematic

Question 1 : does that a real improvement compare to the LTC 1052+LTC1043 solution ? I have never used a LTC1043 or similar circuit

Question 2: low noise measurement need skilled designed PCB , does some one have pictures ?

Regards
OS

[Kleinstein]

The combination of the LTC1052 and LT1028 is used to combine the 2 amplifiers, especially use the LTC1052 for the very low frequencies where the 1/f noise of the LT1028 is higher.
However this solution does not improve the current noise of the LT1028 and is thus only suitable for low impeadance sources (e.g. < 300 Ohm, ideally < 30 ohm at low frequencies as the current noise also has quite some 1/f part). So it is not good together with the LTC1043 switched capacitor part.

The LTC1043 part is a way to get relatively good CMRR - still not perfect due to parasitic capacitance and charge injection that depends on the voltage.
For a resistance measurement one should avoid the need for a super good CMRR and thus way get away with a simpler circuit, like a more classic INA.
The resistance range (mOhm or Mohm) can also make a difference. The low and high resistance range may want different circuits and amplifiers as different problems get limiting. So there one circuit fits all.

Low noise not necessary needs that much care with the PCB. The layout is more a thing about EMI and precision (e.g. leakage currents or thermal EMF to cause errors).

[Overspeed]

Hello

My need is to measure low value to very low resistance for a 4 points system at macro scale so I want to keep current low as possible and measuring voltage will be low . I need an ''high precision amplifier system '' as I need to high light difference between samples and not lost my signal in noise or other problems .

I work in DC only will be pulsed at low frequency and switched polarity for future trials , but first need is a true microvolt amplifier DC .

Last parameter is if possible to stay in DIP or SOIC foot print to avoid soldering problems .

Thanks in advance

Regards
OS

[Kleinstein]

When testing resistors the common mode voltage should not be very large. So there is no need for the LTC1043 type circuit for really good CMRR.

For low noise there are now low noise AZ OP-amps (e.g. OPA189, ADA4522) and there is less need to combine a AZ with a normal OP-amp. With most DMMs / ADCs one would not really care about the higher frequencies anyway.  The choice with DIP case is limited but the 2 OP-amps noted are available in SOIC8 case.

The accuracy may be limited by thermal EMF anyway. The OP-amp noise is more like a smaller problem.  A little more current can really help. Similar AC excitation or polarity reversal at the low frequency can also help, not just with the amplifier but also thermal EMF problems.

[KT88]

The main intention of combining the LTC1052 with the LTC1028 is to overcome the bandwidth limitations of AZ amps. One major issue with AZ amps is the once signal and noise exceeds fclock/2 aliassing- and intermodulation artifacts could degarde the performance at low frequencies significantly. Another one is that with higher gain (which is desired in most use cases) the bandwidth of AZ amps is very limited.
Since in this case bandwidth seems to be no issue a recent AZ amp would do the trick. The ones Kleinstein mentioned are fine.
If the noise performance isn't still satisfying, a parallel configuration could help. The noise would decrease by 1/sqrt_N.
For a low source impedance the ADA4523 might have an edge over the aforementioned AZ amps...
A handy tool collection could be found here: https://tools.analog.com/en/precisionstudio/. You can enter your design and run a noise analysis for different amplifiers.

Cheers

Andreas

[David Hess]

Thermocouple effects will overwhelm the low drift of either solution, or the LT1028 or similar precision amplifier used alone, so the noise difference over your bandwidth of interest is all that matters.  You can most likely get away using just a modern low noise chopper stabilized operational amplifier.  The LT1028 solution would be required for lowest noise when making measurements with a bandwidth of more than a few Hz.

With the 4 wire connection, there is a common mode voltage to be removed, but the highest common mode rejection will not be required.  I would pay more attention to how you plan to do this.

Long ago I extended your example of the LTC1052 correcting the offset and low frequency noise of the LT1028 to a fully differential configuration and got low noise performance which was difficult to imagine.  It worked so well that I could measure low values of resistance by Johnson noise alone to within a couple ohms.  (1)

(1)  The fun but unsettling part was calibrating the breakpoint frequency between the chopper stabilized amplifier and the operational amplifier for lowest noise.  I found a marvelous solution for this, which this post is too narrow to contain.  (2)
 
(2)  I was just kidding.  I used a high resolution digital voltmeter in sampling mode and a calculator to measure and calculate the 0.1 to 10 Hz noise, while adjusting the time constant for minimum noise.  Once the various gain factors were accounted for, the frequency breakpoint was exactly where the noise curves of the LT1028 and chopper stabilized amplifier crossed, which was reassuring.

The main intention of combining the LTC1052 with the LTC1028 is to overcome the bandwidth limitations of AZ amps.

Someone might have done that, but I largely disagree because chopper stabilized amplifier are usually pretty fast anyway, and the aliasing problem occurs when the input signal content is close to the chopping frequency.

The purpose is to limit the input offset voltage drift and rising flicker noise of the LT1028 at low frequencies.

[Kleinstein]

The older zero drift amplifiers like the LTC1052 were relatively slow. Some of the newer ones (e.g. OPA189) are quite fast. If it were just for the speed one could use the sonewhat simpler compound amplifer configuration with the 2nd amplifier just for additional in-loop gain and not also directly reading the input.

The configuration as shown at the start is mainly to get better higher frequency noise, e.g. the LT1028 down to some 0.5 Hz and the LTC1052 below.
In the old times this combination made some sense, but no longer today. The LTC1052 is more like a zero drift amplfier for low current noise and thus if at all would be combined with a a low noise FET input amplifer (e.g. OPA141).  For low impedance the low noise chopper OP-amps are already quite good on there own, like 1/10 the noise of the LTC1052. So an extra amplifier would be rarely needed.

The low side of the resistance measurement is normally also low impedance and can accept more than normal return current, e.g. from a feedback network. This can simplify the circuit and get lower noise by skipping a 2nd amplifier for the inputs.

[KT88]

Someone might have done that, but I largely disagree because chopper stabilized amplifier are usually pretty fast anyway, and the aliasing problem occurs when the input signal content is close to the chopping frequency.

The purpose is to limit the input offset voltage drift and rising flicker noise of the LT1028 at low frequencies.

Sorry... that's a misconception: First, the LTC1052 has only 1.2MHz of gain bandwidth product. At a gain of 1000 this would be a bandwidth of 1.2kHz. If that still sounds fast, be aware that this is only at -3dB. In precision applications which is undisputedly the main use for AZ amps it would require about 100x lower frequency to get fully settled... That leaves you with only 12Hz of usable bandwidth - in some cases this would be even too slow for temperature measurements...

Cheers

Andreas

[David Hess]

The LTC1052 is slow, but is one of the exceptions, and because of the typical low transconductance of its CMOS process, still has a slew rate of 4 volts per microsecond.

When did it, I used the LTC1151 which is hardly any faster, but I wanted to be careful about synchronizing the chopper stabilized operational amplifiers since I had to use two of them.  The LTC1150 would also have worked for me since it allows synchronization.

Like I said, I disagree with the premise.  They are using the LT1028 for the same reason I did, for the lowest possible noise within a restricted bandwidth with a low source impedance, and the LT1052 is only being used to remove the drift and flicker noise from the LT1028.  The wide bandwidth of this configuration is a side effect of using the LT1028 and not the intention, although some applications can take advantage of it.

[KT88]

I agree with your low noise focus which is in particular valid for these older AZ amps - the LTC1052 goes up to 300nV/rtHz @ 300Hz... A low noise amp in combination with an AZ amp as a servo for drift- and 1/f noise mitigation makes a lot of sense and was and in some cases still is a valid solution.
Just for comparison: The ADA4523 has 4.2nV/rtHz and a GBW of 5MHz which today makes life a lot easier in most cases...
...but I also came across the desire for faster AZ amps for the aforementioned reason a couple of times in the past...

[iMo]

Could you show us a practical wiring of the 1052 and a standard opamp (ie like op07) for example (the 1052 being the servo)? The 1028 opamp above has got pretty weird wiring, imho (around pins 1 and 8 )..

[KT88]

The LTC1028 has the classic offset adjustment pins which comes in handy in the given application example. It offers a summing node that is not exposed to common mode changes. Without the offset adjustment pins it would be more complex though possible...perhaps.
Edit: The wiring is not weird - the 1028 offers just simple offset compensation inputs . The resistors at pin 1 and pin 2 create an offset through the higher current into pin 8. The output of the 1052 can only operate below 5V (that's the highest voltage available). Going lower, more and more current will be conducted through the 68k resistor, evening out the current difference and going further down inverting the offset. That way the servo gets a defined range in both polarities.

[Overspeed]

Hello

ADA4523 looks to have impressive performances for the price which under 5 Euro retail in the same of price range of the OPA189 , that allow to make trials at reasonable costs

Regards
OS