NI 9203 accuracy problem

[Henrik_Volkers]

paperduck,

how does your setup looks like? 4-20mA current output from the balance? How is it wired? Shielded?

HOW DOES THE NOISE LOOK LIKE? ... (sorry, I shouldn't shout, but to help and suggest a filter some more data is really helpful 😉 )

Capture at least 300ms,  post a pic ( and nice: a power spectrum of the signal)

And also very important: How dynamic is your signal? How fast are the changes you expect? (How fast is your balance??)

General rule: As longer you are averaging, as lower is the noise....

[hrwang]

It turned out that the balance supplier didn't wire the ribbon cable correctly. It is still interesting why my old fashion multimeter gave a steady reading. Luckly we found a direct wire port from the internal board. Eventrually, I choose labjack for the convenience of my setup. Since I don't need to acquire data so often, I took the average which largely reduce the noise as you said, although the accuracy is slightly lower than the accuracy I read from the balance, anyway, it is acceptable.

Another way is to read the digital output directly to labview. But I have to connect the RS232 port of the balance to the USB port of my computer, I have no idea how to read it. Any help with that issue?

Thanks. 

[Henrik_Volkers]

Reading the balance with the RS232 port give best consistence of what you see at the balance and what you read 🙂

And noise it a minor problem... 🙂

Get your hand on a 15$ USB to RS232 adaptor, the manual of the balance descibing the serial communication and a connecting cable (have a look at the manual first, maybe you need to make your own)

If run into trouble: post a link to the manual and your vi

[AnalogKid2DigitalMan]

The DMM showed a steady reading because it performs averaging of the data before it is displayed. An oscilloscope would reveal the true nature of the signal.

-AK2DM

[hrwang]

It only showed 3 stable digits when I use labview card, however, 5 stable digits with multimeter. It is really amazing that average can do so much.

Thanks.

[hrwang]

Thanks. I was asking the Sartorius, which is my balance provider, for an adaptor cable, which would cost $250, insane. I wonder if the cable from other electronic vendor will work, I think it should. Assuming I have the adaptor, I know which pin gives a reading at certain frequency, which is sort of discribed in the manual, how can I read that pin from a USB port. The labview can read serial and parallel ports through visa? But I just can't figure out how to approach to a usb port, and how can I treat that usb port as serial.

Your answer is highly appreciated.

[Henrik_Volkers]

First you should tell us what type of scale you have and if it does already have the RS232 port. 

In the manual there should be a description of the RS232, how to configure the serial port (baudrate, handshake, parity & data bits, etc)

You can't use the USB port directly, you need to buy a USB to RS232 adaptor. Search the net for 'RS232 to USB cable'   *bay starts at about 5$ 😉

Next is a cable to connect the adaptor to youe scale, that could be a standard  cable (next 5$) or not. The scale documentation should tell. RTFM (If not , look if the scale has a CE sign, than the manufactor has to provide the information for industrial items in the EU)

[hrwang]

I didn't fix it. First of all, the supplier wire the cable loose, so I have a very large variations ( you may need to check it as well). After I fix that problem, I didn't use labview card, but labjack for my convenience. So I don't know how will the labview card perform. Fortunately, I don't need collect data frequently, and I can do as much average as I want to reduce the noise. Still I have 0.5g of accuracy instead of 0.1g claimed by the provide. Another experience is, I don't know why, but it appears that voltage reading is more stable than current reading.

Good luck.

[Henrik_Volkers]

Didn't you tell us you use a 9203?? However ... since you ignored all my questions ... how could (why should?) I help?

If you wan't the possible resolution of the scale you should use the RS232 port, all other thinks are AD -> DA (+noise) -> AD ....

[hrwang]

My last reply was for paperduck. I didn't see his/her post before. Sorry for misunderstanding.

I was using 9203, however, I am giving it up because I think the voltage/current output from the balance won't be able to get the best accuracy of the balance due to the noise of the balance itself. Another problem is that I found I have to do calibration change from time to time if I want to read very accurately from my voltage output. I don't know how did the balance resolve it internally. I am thinking to use the RS232 port now. Thanks.