07-16-2012 02:43 AM
this is regarding the quality of PWM signals from a NI 9476 module. I have attached my FPGA vi as a reference. Note that the FPGA VI also includes data for another module used for PWM in the same project (NI 9472). Also note that in order to keep resources used at approproate levels I had to bind most of the channels (channels 09-30) of the NI 9476 to use the same frequency..
Here is the problem.
Firstly, channels 31 and 32 of the 9476 do not work at all. Independently of the frequency or duty cycle, the output on the o-scope is a constant 24v signal, no freq, no duty cycle.
For the rest of the channels that work, the signal is not of the best quality. More specifically, for channels 1 to 30, if I set the PWM frequency up to 200 HZ, the output signal is fair and I am getting pretty much what I want. The o-scope shows a square PWM signal with the correct frequency and correct duty cycle. However, if I increase the frequency further the signal begins to deteriorate. According to the specs of the 9476 module, its max speed is 0.5 ms or 2 kHz. If I set my PWM frequency at 2 kHz, my signal is not square at all and my duty cycle is 20-25% higher than my setpoint.
Anyone who has any experience on why I get these problems, please help out a bit..
07-17-2012 04:03 PM
Reviewing what you have posted, are you able to run a the LabVIEW Example you can also find a good example here http://zone.ni.com/devzone/cda/epd/p/id/979. From your post you mentioned that you don't get anything out of channels 31 and 32, do they work normally just not with your program. Can you control them independently? If you select just one channel do you get the correct output signal and range? Also what value are your pull down resistors in your circuit http://digital.ni.com/public.nsf/allkb/C9019BC9F65
07-18-2012 01:50 AM
thanks for the reply. I have already tried using a pull-down resistor but it did not solve the problem. At the end of the day I consider channels 30 and 31 gone. They dont work in FPGA, nor in scan mode, not at all. This is a typical situation for an NI module. I mean, not working. Also, if you noticed from my message, the rest of the channels do not do what has been "described in the box". I mean that the PWM signal at the max specified frequency for the module is not a PWM signal at all. Another typical situation for an NI module.
We have found a solution here.Stop using NI.
Thanks a lot for your effort though. Nothing personal.
07-18-2012 10:10 AM
Have you checked the status lines to provide an indication of any fault? Check the manual for further guidance.
Had a 'similar problem with a 9215 module, returned to NI and they confirmed a faulty channel which was repaired.
Contact you r local sales office.
08-14-2012 02:23 PM
From the information in the post it appears that the channels "31 and 32" are damaged in some way and do need repair. As for the frequency, the period is 1 mS (500 uS high + 500 uS low) so the frequency you can achieve is 1 kHz not 2 kHz. Also the 500 uS propagation delay is a Max spec which means that the delay could be shorter than that (and typically is). The delay can also vary between turning on and turning off which is why you see the differences in duty cycle. The NI 9474 is a much faster module that would be much better suited for applications such as PWM.