Post Go back to editing

How to use a LTC1043 as a precision inverter ?


I have a high precision voltage reference and I need to build a symmetrical voltage reference from it. As the goal is to get a few ppm tracking from the reference, it is very difficult to create a way to invert the voltage using a OpAmp as an inverter as it requires precision resistors and tuning.

So I tried the LTC1043 solution that can be found in several datasheets and app notes, especially a Jim Williams one in which the precision is said to be around +/1ppm (or +/-2ppm in the datasheet).

It is a very simple and straightforward technique using a sample and hold circuit reversion the ground location between input and output.

The problem is that I couldn't achieve something in this area of precision. For a 5V input (and +/8V power supply), the error is about 2.5-3mV ! Which is 500 to 600ppm ! There is no load on the output, I'm using polypropylene capacitors (1uF) and 10n capacitor for the internal oscillator. Everything is shielded. Power supplies are very clean and linear types.

So I tried to simulated with LTSpice and to my surprise, the results are no far from what I get, even though the error is less severe (about 150uV). Simulating a non inverting S/H does give perfect results, but as soon as I invert the output, there is a serious signal degradation.

Measuring voltage at the sampling capacitor gives half the error and the other half is on the hold one. It seems that there are some leakage that discharge the capacitor that does not exist when using the no inverting way.

Whatever I try, I get exactly the same results. Have I missed something ?

If there is no way to use this method  to invert a voltage reference, what would be the best one ?

Thanks in advance,


  • Hi jipihorn,

    I tried to see if I can help but you are far ahead of me.

    It'd be very helpful if you attach your findings (i.e. simulation file (*.asc), images, etc.) onto a zip file (use Insert, Insert image/video/file command below) and post here for reference so either I, or other people can offer some help. I could not exactly follow the path you've taken so far.

    I'll take another look if you post some attachment that makes it easier to follow you.



  • Another suggestion - probe the flying capacitor with a scope to verify that the switching frequency is what you expect it to be. (If a much smaller, or even larger, capacitor were stuffed by mistake, it will affect performance in interesting ways.) Of course probing the flying cap will also affect performance, this test is just to verify switching frequency. And make sure your circuit board is clean, little things make a difference.

    Also, are you confident in the meter you're using? If you swap the leads while measuring only the original positive reference voltage, is the reading exactly the same?


  • Hello,

    As I have done lots of experiments on the PCB, it not very clean or usable anymore, especially because it is only a part of a whole system. So it not very convenient to isolate and experiment this very part.

    So I'll make a specific PCB with only this inverter and  make full measurements and behavior.

    Nevertheless, all tests I did give the same results : this is a much larger error than expected, about  two orders of magnitude. As there are only 4 components, it's unlikely due to a "not so good" implementation. In this case I would expect 50ppm instead of 2, not 500 or 600 !

    Now, I don't know how the spice model is accurate, especially with the charge balancing circuitry.

    I'll return to you when I could have a full data set.


  • Hello,

    I've tried a 3457A, a solartron 7061 and a Keithey 2410 sourcemeter as a voltmeter. I have similar results.

    Here, I don't think I have a performance degradation due to a not ideal implementation : the error is huge ! And, with so few component, I don't see any obvious errors !


  • Not sure if I've ever personally used this part as a precision inverter, I've used it as a level shifter and divide-by-two. Not sure if this makes a difference, but are you sure you're connecting exactly as shown in the datasheet, both in LTspice and your actual circuit? I get about 25ppm error in LTspice. Not 1ppm, but not 500-600.

    And of course LTspice is a simulation, and it's hard to be sure how accurately higher order effects are modeled.

    Another solution might be the LTC2053:

    See page 13 for an inverter circuit. You don't have much margin on supply voltage if you're converting +5V to -5V, but the sample and hold circuitry is integrated.