USB Noise

[jmumby]

Hello all, happy new year etc etc.

I have a project on that uses a colour sensor. Usual story bugger all info about the sensor except that found from other users.

Anyway, this sensor has RGB elements you calibrate it to a white point and make readings. It is 12bit but in reality you could only really use a 100-150 unit spread to read off something like paper. It is very sensitive however.

At the moment whilst reading the same colour my readings for red (for example all colours fly around) go from 474 to 467. Whippy I hear you say but with about 100units from end to end that is a bit large at 7%.

I have heard that these sensors are quite susceptible to noise and I have found a 100khz (noise repeats every 10us) "noise" on the USB line that has .4v p2p.

Im not using the USB to program the picaxe. I use it with an FTDI chip to send readings back to telnet.

Is this interference normal for USB? Would you expect this to cause a problem?

Cheers.

 

[Dippy]

Thanks etc. etc.

I've found USB to be noisy too.

I've also found that the decoupling at PIC and device can be crucial.
I had similar problems with a PIC recently.
Changed the capacitor and bingo.
I used 4u7tant // 100nF Ceramic at PIC power pins.

Maybe you should do similar at sensor end too?

I assume noise isn't being picked up in track from sensor to PICAXE.
Are grounds good. Nice PCB or hairy old breadboard?
Have you 'scoped the ADC connection?
(Haven't got a 'scope?? Goodness, you should have had one for Christmas!!)

I've had no troubles with sensors i've bought from Ebay.

 

[jmumby]

The sensor use i2c to communicate. I have just measured directly off the USB cable and the noise is there. It reduces some what after the FTDI so coupling of some description may be on the cards.