Calibration question

[pgm1]

Hi, apologies if this is a silly question.

I'm using the USB-6363 BNC unit, and have it connected to an amplifier. The DAQ sends commands on AO0 and receives signals on AI 0-4.

It appears that the AOs are not precisely calibrated to zero. The amplifier I'm using (for electrophysiology) is a DC-coupled input, so the DAQ must be correctly calibrated so that zero volts
really correspond to zero.

When I connect the DAQ AO to the amplifier, it detects a small voltage.

Could anyone guide me through what I should do? I have previously disconnected all connections and run a self-calibration in NI Device Monitor.

Thanks in advance.

[oscarfonseca]

Hello pgm1,

I find interesting you find an offset in the zero volts reading of the card.

It would be interesting to know if you read the same offset when reading the AO from the AI in the same card (I presume you can use NI MAX and test panels to make this happen). If it's consistent when asking the card to output zero volts, I would assume it needs to be calibrated indeed.

How about when using a multimeter? Does it read a number significantly different to zero volts when the card is commanded to output zero volts from the NI MAX test panel?

If we do see zero volts with both the AI and the multimeter, I would focus my attention into the wiring to your amplifier, rather than the card itself.

Finally, how old is the card? I see in the specifications sheet that the calibration interval is 2 years, so if it has been more than so, it might be a good idea to send it for external calibration.

All the best,

[Henrik_Volkers]

What is the input voltage range needed for that amplifier? (If you have a lower range you can use a lower output range or a voltage devider to increase the AO resolution and compensate the offset in software)

Max acceptable zero offset? (in the +-10V range the step size is ~0.3mV )

You can always shortcut the input 😉 

but even than a real zero volt isn't easy due to thermoelectric effects 😉 (a comercial BNC cable with different temperatures at the two connectors is always good for some µV...  If the connectors are dirty (finger fat/soil) even more 😄

Do you need all AO outputs?  If not, you can use a spare output together with 2 resistors to do a compensation (or 3 resistors depending on the input impedance of the amplifier) 

If you measure the zero volt outputs of your AOs with an external DMM ... what values do you get?

Finally you can use external reference voltages and trimm it to zero at your amplifier.

[pgm1]

Thanks both - the DAQ is only a few months old.

Henrik - Thank you. The amplifier only requires a low V range - the most I'd ever need to send to it is probably 300 mV. But, how do I compensate this in software? Surely using a voltage divider in LabVIEW just reduces the amplitude of the signal that is eventually output byt he hardware.

I'm only using AO0 --- However, I have no idea how to "use a spare output together with 2 resistors to do a compensation". I'm a LabVIEW coding almost-novice and don't have a circuit board, resitors, etc.

Importantly ...If I run AO0 into AI0, the reading on AI0 is at zero.

...so perhaps it is the amplifier. However the small voltage only appears when I connect the BNC cable .... the very same cable I plugged into AI0 to test the signal. So I'm somewhat confused.

Thanks for your help.

[NIquist]

What exactly do you get when you physically short the inputs to the amp?  If it's not zero your amp has an offset you need to deal with. 

Unless your BNC cable is very long and/or in a strong magnetic field I doubt it's a problem but to test you could also try shorting one end and connecting the other to the amp.  If the amp is output is 0V shorted and shows a offset with the shorted cable then the cable is the culprit.

You may also have a ground loop of some kind.  Is the amp power supply and DAQ/PC plugged into the same outlet?  Sometimes there are mV differences between grounds even in the same room.

[Henrik_Volkers]

Is there a datasheet for the (bio) amplifier?

What is the output of your amplifier ?  (A voltage/current you can measure?)

What bandwidth do you need?  (How fast will the amplifier get new output values?)

If you have an offset with the BNC cable GND connected to the amp and AGND of the DAQ and the BNC cable shortcut at the DAQ, you can compensate this with an AO offset . 

Next is the range: Asuming you stay with a +-10V output a 20:1 voltage divider (or so) would be helpful.

I would try a 10k and 470 Ohm resistor, giving you a ~0.449 V output when applying 10V. But that is also influenced by the input impedance of your bio-amp so best would be a custom calibration curve. Set some AO voltages to cover the output range of your bio-amp. Graph it ... strait line? Bingo 🙂   or apply a curve fitting. With that curve and the increasend resolution (due to the voltage divider) you should be able to set the bio-amp output to zero.

I would place the voltage divider at the bio-amp.

AO> ---|10K|--<bioamp-input>--|470|--<GND> 

(or use trim potentiometer or ...)

If you need fast changes in the output you need to compensate some additonal capacities and a scope to match them.

[pgm1]

Thank you both, I am checking into these things.

The outputs of the amplifier are measurable voltage signals, yes.

I am capturing data from the amplifier at 100 kHz. Low-pass filter on the amp is switched off. It's an Axon instruments 700B.