What material are you testing? If it is a common material you can probably just model it more accurately.
Hello Allan sir, Sorry, I could not make that to work because of the OPA192 which is not available to me.
OPA192's are freely available here in England - where are you?
Can you get hold of ICL7650's?
Allan
4M/50k is a huge impedance, lots of pickup and thermal noise, you don't want that
for microvolt level signals. The signal currents are 0.25pA per microvolt.
The resistors need to be as small as you can get away with to keep noise down. Also
low pass filter with a capacitor.
With a voltage divider of 100:1 or so you are going to throw 99% of your tiny signal
away - that's not great, to say the least.
There's an interesting circuit in Figure 7.27 of "Art of Electronics" (2nd edition), allowing
low voltage opamps to work with high voltage differential signal without losing different
gain.
MarkT:
4M/50k is a huge impedance, lots of pickup and thermal noise, you don't want that
for microvolt level signals. The signal currents are 0.25pA per microvolt.The resistors need to be as small as you can get away with to keep noise down. Also
low pass filter with a capacitor.With a voltage divider of 100:1 or so you are going to throw 99% of your tiny signal
away - that's not great, to say the least.There's an interesting circuit in Figure 7.27 of "Art of Electronics" (2nd edition), allowing
low voltage opamps to work with high voltage differential signal without losing different
gain.
Is this the circuit you were referring to (attached figure). Looks interesting. Thanks.
If I get low input resistance, the current will flow in my meter and not on the loads. That's not advantageous, is it?
allanhurst:
OPA192's are freely available here in England - where are you?Can you get hold of ICL7650's?
Yes, this is available.
https://www.tanotis.com/products/tanotis-icl7650scpa-icl7650-intersil-ic-opamp-chopper-2mhz-8dip-new?gclid=EAIaIQobChMIgMr5ut6d1QIVCaFoCh3aoA-kEAYYASABEgJRtPD_BwE&variant=20608822597
Yes, that's the circuit.
Voltage noise is higher with higher resistance, sensitivity to capacitive interference is higher with higher
resistance, settling time is higher with higher resistance.
The source you are measuring is milliohms so you don't have to worry about
affecting the source by loading it. With 4M sense resistors you've only got 12uA for a 50V offset,
whereas taking 1mA or so would mean 80 times the immunity to noise and interference, yet only
50mW dissipated (too much heating of precision resistors is undesirable of course, depending on
the tempco.
So several 10's of kohms seems a sweet spot for 50V or so, scale for increasing offset voltage.
By having a dual rail high voltage supply you ought to be able to balance the circuit and keep the
offsets low too, assuming the 1k-ish resistances are well matched...
OK firststep...
your requirement seems to me an x-y problem - you ask for a particular solution to a little part of the whole problem.
I got around to reading your suggestion of looking at IP - induced polarity - used in geological surveys - am I on the right track?
ie you have an exciting voltage across two probes in the ground - either a slow ac of 300v or so, or a switched +/- squarewave (with gaps) of similar voltage, and have lots of seperate sensors which pick up from independant pairs of probes placed at various distances from the exciting source and report their measurements.
From the phase change/ decaying response you can infer something of the underlying structure of the ground.
In which case your model is a bad one...
1/ the high voltage source is connected to your R1 by an unknown ( ground) resistance which is totally indeterminate.
2/ Your R1 is also totally unknown - but might be many hundred ohms - but all you should care about is the phase/decay waveform. There will be a large attenuation due to the ground resistance between it and the exciting source.
3/ you must use an ac or +/- dc source or your electrodes will become polarised and hence worthless.
The variable ( unknown) dc-only source you proposed is no good.
If you're trying to make a cheap version, that's a new ballgame. Perhaps we can help.
If for a commercial project - my rates are reasonable.
Allan
allanhurst:
OK firststep...your requirement seems to me an x-y problem - you ask for a particular solution to a little part of the whole problem.
I got around to reading your suggestion of looking at IP - induced polarity - used in geological surveys - am I on the right track?
ie you have an exciting voltage across two probes in the ground - either a slow ac of 300v or so, or a switched +/- squarewave (with gaps) of similar voltage, and have lots of seperate sensors which pick up from independant pairs of probes placed at various distances from the exciting source and report their measurements.
From the phase change/ decaying response you can infer something of the underlying structure of the ground.
In which case your model is a bad one...
1/ the high voltage source is connected to your R1 by an unknown ( ground) resistance which is totally indeterminate.
2/ Your R1 is also totally unknown - but might be many hundred ohms - but all you should care about is the phase/decay waveform. There will be a large attenuation due to the ground resistance between it and the exciting source.
3/ you must use an ac or +/- dc source or your electrodes will become polarised and hence worthless.
The variable ( unknown) dc-only source you proposed is no good.If you're trying to make a cheap version, that's a new ballgame. Perhaps we can help.
If for a commercial project - my rates are reasonable.
Allan
Sir, Thanks for taking the time to read up on my project. And yes, I am trying to make a similar cheap version for my lab use.
Your points 1 and 2 are correct. For point 3, I use relays to switch the polarity of the DC source to remove polarisation effect.
I just made this model by myself - so it can be bad. Your suggestions and help is really appreciated.
As for commercial version, the ones available commercially are way good in accuracy and precision. And I am struggling here to get a stable reading.
Read post#16 again.
The supply should be made "floating", so the measurement can be done at ground potential.
Leo..
Wawa:
Read post#16 again.
The supply should be made "floating", so the measurement can be done at ground potential.
Leo..
I don't know much about floating supplies, so I read about it in this link. My supply is a rectified DC from an inverter, and I think that it is "floating" as no connection to the "real" earth is present.
Do you want to build a real survey device, or model the ground with common components n the lab ?
In which case a simple resistance for your R1 is a very poor model, and doesn't address the induced polarity at all.
In the real world there may be several timeconstants involved through the various different possible polarising processes involved in soil/rock with unknown water content/salinity etc..
I reiterate - are you a student doing an assignment, or are you a developer of a real device?
I'd design you a real one for a fee. It may or may not use arduinos.
Allan
allanhurst:
Do you want to build a real survey device, or model the ground with common components n the lab ?In which case a simple resistance for your R1 is a very poor model, and doesn't address the induced polarity at all.
In the real world there may be several timeconstants involved through the various different possible polarising processes involved in soil/rock with unknown water content/salinity etc..
I reiterate - are you a student doing an assignment, or are you a developer of a real device?
I'd design you a real one for a fee. It may or may not use arduinos.
Allan
Hello sir, I would be happy if I could get it to work somehow by myself, even if the accuracy is a little compromised. And for that, if you can help me just measure the small voltage across R1 on a high voltage supply - i will be grateful.
As for the constraints you have mentioned - it will be good if I could calculate for all that. But currently I am focused on just getting an accurate value of the Voltage across R1 till the lowest range possible.
And for the paid help that you mentioned, can you please PM.
EDIT:
I found this circuit (PFA) in the datasheet of ads1115. It works on high voltage and measures a small voltage drop across a shunt resistor to determine current. Voltage measurement range +-50mV.
I want some suggestions like this one with a autorange feature for higher voltage ranges.
But I could not find OPA333 around.
Available Opamps : OP37G, OPA2277, OP27G, OP07, LM201ADG, TL032ID, TSX711ILT.
These listed opamps have slightly higher offset and drift, so how to compensate for those?
What happened to the clever circuit from 'The Art of Electronics' - that's just a standard differential
amplifier again with the problem of dividing down the voltages lots.
BTW the AD8628 beats the OPA333 by an order of magnitude on offset and drift.
I don't see why OP needs an opamp in front of his A/D.
There is no low impedance current shunt in his project.
Buffering, protection, offset?
Leo..
MarkT:
What happened to the clever circuit from 'The Art of Electronics' - that's just a standard differential
amplifier again with the problem of dividing down the voltages lots.BTW the AD8628 beats the OPA333 by an order of magnitude on offset and drift.
Under study sir, but I am also looking at other designs - as I don't have the components available locally and thinking of ordering from mouser. So making a good and final list before ordering.
Wawa:
I don't see why OP needs an opamp in front of his A/D.
There is no low impedance current shunt in his project.
Buffering, protection, offset?
Leo..
R1 can have quite low impedance(micro volt drop). So I am trying to use opamp for Gain.
Hello again, I think the resistor divider is the main problem when measuring in low values, as I cannot use the full dynamic range (FSR) of my ADCs. Any suggestions on that?
Correct - as I pointed out a long time ago.
Try my recommended circuit.
And if you merely want to model, rather than actually measure induced polarity effects in real soil samples , you could build a Matlab model, or PSPICE or its variants much more simply. Various models are published in the literature - have you read them?
If not, do so - I have.
Unless you're using a real soil sample, I'd go this way.
And even if you are measuring real soil in a lab, one end could be grounded, which makes things a lot easier electronically.
How else do you think the makers of these expensive geological surveying devices got their IP and resistivity values?
Allan
Hello Allan sir, I have read some - but there is a huge difference when designing the real thing. I want to get atleast some accuracy in my measurements.
Your recommended circuit looks very good - but I don't have the opamp right now. Need to be imported if I want to use it, so I am making a BOM for all such components regarding different circuits.
I will test it in real field, but for now I checked my ads1115 and the mcp3551s with the divider in a resistance box similar to this image.
https://pimg.tradeindia.com/02230255/b/1/Standard-Resistance-Box.jpg
Once my resistance tests are accurate then I will test on other materials and then actual soil on field. (Check wenner 4 point config.)
allanhurst:
And even if you are measuring real soil in a lab, one end could be grounded, which makes things a lot easier electronically.How else do you think the makers of these expensive geological surveying devices got their IP and resistivity values?
Allan
Please check the attached image. This is how I am doing the polarity switching.
Please compare my image and wenner config - and please reply what do you mean by one end could be grounded? How I should do it?
Wawa:
I don't see why OP needs an opamp in front of his A/D.
There is no low impedance current shunt in his project.
Buffering, protection, offset?
Leo..
To remove the large common-mode voltage of course. The ADC is 5V supply.
firststep:
Please check the attached image. This is how I am doing the polarity switching.
Please compare my image and wenner config - and please reply what do you mean by one end could be grounded? How I should do it?
You appear have grounded one end of the supply so its not balanced/bipolar, which means much much larger
common mode voltages and currents flowing to earth.
Wasn't the idea a balanced or floating supply so the voltage excursions at the sample were minimized?
And I think the idea was to ground one side of the sample under test to keep it at low common mode voltage.
If the ground itself has a large resistance though, "grounding" a point in the circuit isn't a well defined thing.
MarkT:
To remove the large common-mode voltage of course. The ADC is 5V supply.
I think you answered that yourself in your next post.
If you have the drive voltage floating, you can bring the measuring terminals to ground level.
Then there is no large common mode voltage.
An ADS1115 can measure differential voltage of an AC-driven circuit.
If... you bring both measuring terminals down to the common mode range of the ADS.
Leo..