Is the AD5272 (or any other digiPOT) suitable for thermocouple voltage levels? I'm looking to use this to divide my thermocouple simulated voltage to about 0.940mV and as high as 3.9mV
In this case, the main contributor is the zero scale error + tempco.
The value that you are looking for is quite small....
You have two options, but the cheapest is probably current output DAC.... AD5451 should fit in your application.
Just remember that the output opamp should have a Vos lower than your minimum accuracy.
Let me know if something is not clear, please.
You can use the AD5112 (5kohms) as a Vernier DAC.
This application note may help,
Thanks Miguel, but this ap note doesn't answer my question and is not necessarily applicable to my situation. I want to use a DigiPOT as a simple voltage divider; the advantage being that I can "tune" my divider network remotely (via serial bus). My VDD going to Terminal A of the potentiometer will be in the mV range, and I'm planning on using the wiper voltage (a ration of the Terminal A voltage) as the thermocouple voltage.
My question is whether these DigiPOTs can properly handle these low range voltages
Yes, the digipots can handle this low voltage, your only problem will be the internal leakage current.
Typically is specified as 1uA max, that is why I recommended a low resistor value with a good linearity.
I need a high resistance value coz I may need a ratio of 10000:1. With 100K pot, and the divider set to 10000, the wiper (to Vss) resistance will be about 10 ohms. Does this mean the wiper voltage may be offset by +/-10uA (10ohms * 1uA)? I'm having a hard time understanding/finding concrete information on leakage current effects from the datasheet (AD5272)...
Im not fully understand your application....
first, the AD5272 is a rheostat... why you would like to use a rheostat instead of a potentiometer?
second, the leakage current is negligible as linearity error source if you consider other source of error, such as tempco.
May you provide an schematic to better understand your circuit/idea?
I don't have a schematic, just theorizing right now.
So A (Pin 2 - Terminal A) and W( Pin 3 - Wiper Terminal) are not internally tied, so I was looking at this as a potentiometer. I figured if I tie Pin2 (terminal A) to a source voltage, say 7mV, I could adjust the Wiper terminal (thru the RDAC register), to get a Wiper Voltage (Vw) that is proportional to 7mV
I guess I was looking at this wrong....
OK, I see several DigiPOTs that would work fro my application (as described above). They're not I2C as I'd hoped, but that's OK.
So to make sure I understand, the digiPOT can handle low voltages without a problem, correct? I'm just read thru the AD52932 datasheet and it looks promising. A voltage of ~900uV (microvolts) or less is achievable, as long as the Vin (at terminal A) is proportional to the steps, right? i.e. If I provide 921.9mV at VIN, I should be able to get 900uV at Vout (wiper) without any problems...
correct me if I'm wrong.
If you tell me the exact digipot I can try to measure here in the bench.
Please try the AD5292 and let me know
Why you like to use the AD5292 in your application?
This digipot is high voltage (Vdd >21V) and 1% R-tolerance.....
Answering your question and before test, you should be able to get a resolution around 900uV but you will never get 900uV at code zero, in fact, the minimum wiper voltage is probably around 3 to 8 mV.
If this is fine for you, I'll check if the flicker noise contribution affects the measurement.
3 to 8mV is too high for my minimum wiper voltage. Where did you get the numbers from?
Since you understand my application a bit more now, can you suggest a digiPOT (digital voltage divider) that would world work?
Thanks for all your help!
***that would world work?***
Is Vout on AD5451 datasheet figure 47 = Vref * D/2^10? Can Vref be small, i.e. ~5mV?
You can do Vref as small as you want but be aware of,
Output leakage = 1nA (10nA max)
Vos of your output opamp
Ios of your output opamp.
Sounds good, so the Vout equation I specified earlier is accurate?
It is Vout = -Vref * D / 2^n, where n is the resolution of the part, could be 8, 10 or 12 bits.
Miguel, can you explain the level shifting (voltage inversion) done by the 1st stage opamp (Figure 47 on the datasheet)?. How does the inversion by virtue of virtual GND work? I still find it unusual tying ADR03's GND to the opamp’s output
The voltage reference (+2.5V) guarantees a constant voltage between Vout and GND.
If the non-inverting pin is grounded, the inverting pin needs to be virtually grounded as well.
The only way to achieve this, it is when the opamp output is -2.5V, so the difference between Vout and GND is 2.5V.
Retrieving data ...