AD5290 can not reach exact 0 volt

Hello,

I am using AD5290 DPOT to generate voltage from 0 to 20V in steps of 20V/256. Through spi data is sent to dpot input. I know that due to some residual resistance we cannot reach exact max or min value . For 0x00 i should get 0V but it is giving offset value around 250mV .

Is it possible to reduce this offset???

Thank you in advance

Hi,

What does the datasheet say?

It says it thas zero scale offset of max. 3LSB.
wiper resistance is 50 Ohms max.

***

You didn´t say what resistor value your pot is, so I assume 10k.

Then the reistance step size is 10k / 255 = 39 ohms.
3LSB means 3 x 39 ohms = 117 ohms.

Now add the wiper resistance: 117 ohms + 50 ohms = 167 Ohms.
This should be the max. expectable resistance at zero position.

****
The max expectable voltage is U_fullscale * 3 LSB / 255 LSB.
You didn´t say what voltage you applied to the resistor taps, so I assume 5V.

Then the max expectable output voltage with no load is 5V*3/255 = 59mV

***
Now you say you have 250mV.
In worst case the difference voltage (250mV - 59mV) is caused by external current into the wiper.

--> I = U / R --> I = (250mV-59mV) / 167 Ohms = 1.1mA

So I expect you draw more than 1.1mA at the wiper.
(I expect a lot more)

****
To verify this you should give us the (2) missing informations and a schematic (draft) of all your AD5290 circuit, including wiper load circuit.

Klaus

250 mV with 20V input is quite near to the typical zero scale error of 3 LSB specified in the datasheet. Up to 390 mV can be expected.

Your circuit must be designed in aware of this error, possibly allowing external compensation.

Hello sir,

Initially I was using 10k pot and max volt was 20V. So according to your calculation I am getting expected offset around 250mV. Please find the below schematic which we used for ad5290 circuit.
.

We are now planning to use 100k Dpot and max supply volt 20V. But this gives ~1k ohm expectable resistance at zero position. Can this reduce the offset.??
Could you please suggest any other possibilities??

I understand that the zero/full scale error applies in addition to the INL curve and is similar to the terminal resistance of a conventional potentiometer. This means, it can't be compensated inside AD5290, except for applying a negative offset to the A terminal (and a respective positive offset to the B terminal).

Hi,

My calculatins show the worst case errors.
I don´t expect that you bought a worst case device. Therefore I expect at least a part of the error is caused by the circuit connected to the wiper.
That´s why I asked for the load circuit.
--> Please show us what is connected to the W terminal.

***
As FvM posted, it´s not possible to correct the error inside the chip.
Connecting a negative voltage (even -250mV) to the A or B terminal is beyond specification. I don´t recommend this. (Although it might be possible).


***
Using 100k instead of 10k will give no improvement. I expect larger errors. (Because i still think a part of the offset is caused by current into the W terminal)

Klaus

Hello,
Thank you all for helping. Now we are able to get 27mV offset for 0x00 input code. We changed the 10k pot to 100k pot. Before it was 250mV for 10k pot at zero scale error.

Hi,

I didn´t expect that.

Lucky you.

Klaus