Yes, you can use one of the OpAmps or PGAs for amplification.
But are you sure then 64mV are too much? This is only 2% of the supply voltage, and many regulators are less accurate than that...
What is that +/- 64mV from? Do you mean that your circuit only ouput signal of this voltage range from your current sensor?
The +/-64mV is from the power management's datasheet. It is either +/-64mV or +/-128mV of range.
True that regulators are for most less accurate but if for instance you take a worst case regulator and subtract the 64mV, it can make a difference between a working circuit and a somwhat working one.
I wish to use a lower resistive shunt; somewhere around the 1mOhm (looking at 0.5mOhm) which would drop by far less. It's also greener :)
But I agree that I'm a bit paranoid. But I have my answer and I thank you guys very much for answering. I'm still not used to the possibilities of the PSoC. It is still a strange animal to me.
If you are going to do hi side (or low side for that matter) use DelSig
in buffer bypass mode, no PGA. This will allow you to measure differentially
up to 100 mV outside supply rail.
With up to 20 bit resolution that should be enough accuracy to do the measurement.
Note plan on averaging readings and/or extra carefull bypassing, consider polymer
tanatalums for bulk cap, an order of masgnitude better ESR than conventional
tantalum. This is to take care of switching noise on main Vdd.
I am not sure to understand your solution.
The issue is the voltage drop at the shunt. I find 64mV still big for 3.3V and wanted to keep the drop to lower values.
Some questions -
1) Range of current you want to measure ?
2) Resolution of the measurement ?
3) Accuracy of the measurement ?
A big issue here is noise, averaging only works effectively on uncorrelated noise,
and a UP environment is anything but. Differential measurement helps, but limitations
are present, like BW, linearity, CMRR, etc..
Like I said: 64mV is only 2% of a 3.3V supply. If this is too much voltage drop for you: what kind of device are you measuring? Are you sure that your power supply is more accurate than that? (A LM7833 is specified with two to three times, meaning it can drop up to 160mV and is still within spec)
A shunt resistor will always have a voltage drop, no way out of that.
So the only answer will be to put the sense-input for the regulator AFTER the shunt and ensure that the voltage-levels at the shunt pins do not exceed the alowable max for the PSoC chip.
An example of diff measuring with DelSig running at 20 bits,
and your max current yields 64 mV on the Rsense. Lets say that
current equivalent is 64 mA (you can recalc for any value).
Start with Delsig, 20 bits, Vref +/- .064 V. Then resolution of A/D
in diff mode, no buff G, = .128V / 20 bits = .122 uV. The current
sensitivity of 64 mA at 64 mV = ~ 1 A /V. So resolution of A/D
in current is .122 uA.
That is so true; using a regulator with sense input would solve this.
Anyhow, thank you all for your comments, I'm just learning the possibilities of the PsOC and knowing that I can cascade blocks makes things more interesting. I thought they were tied to the pins.
Welcome to the wonderful world of PSoC. Yes, working with PSoC is more like drawing a schematic, and it seems you get the hang of it...
(Some things are more or less bound to pins, e.g. the OpAmp pins, but this is just to ensure better performance)
"That is so true; using a regulator with sense input would solve this."
Choose the regulator carefully, looking at datasheet, as the control loop
typically not very precise in power regulators, so you might not end up
with the precise Vout you are looking for. This in light your seeking very low
drop across Rsense the control loop error becomes even more dominant in overall
results. And of course the references in power regulators generally several
percent, which of course destroys the precision you are looking for.
You could use a precision reference as your regulator, thats a possibility
depending on your current/power requirements.
The accuracy for the LM7833 is referencing to the reference. Its load regulation is actually better (33mV max IIRC). So even if it had a sense connection, the regulator can introduce a greater voltage drop than the current sense resistor.
But there is still an unanswered question: what device do you want to measure which cannot live with a small voltage drop?