xPC Target DMM-32DX-AT A/D Driver problem.
in [Matlab]
|
From: saj on 8 Dec 2009 18:10 I have a PC104 stack setup with a DMM-32DX-AT I/O card. I have a DOSloader setup with xPC Target version 4.2. As an initial test, i setup the 4 analog outputs on this card to provide constant output voltages 1, 2, 3, and 4 respectively. That seems to work fine. I checked the voltages on a scope and verified. Then i wired up these 4 analog outputs to the 1st four analog inputs of the i/o card to do loop-back testing. On the Simulink side, I added a A/D driver block to my model to read the 4 wired analog inputs. When i execute this model on the hardware, instead of seeing values 1, 2, 3, and 4 volts on the A/D channels I see some junk values that have no co-relation to the input voltages. When i stop the model, the monitor connected to the video port of the hardware shows a message that reads "Error MM-32 A/D was reset n times", where n is a number that seems to depend upon the sample time of the model and also the length of time for which i run this model. I do not see the reset message if i use the analog output channels and the digital input and output channels on this card. Only the A/D seems to have trouble. The Analog Input jumper is setup for SE 0-31 input configuration. I have experimented with the setting the jumper at location "16" on the J7 header to both 'In' and 'Out'. Did not seem to help. Any help on resolving this issue will be greatly appreciated.
From: Hans-Joachim Schulpin on 9 Dec 2009 15:15 We use the MM-32-AT and MM-32X-AT with xPC Target without any problems except that we had to disable the auto-autocalibration of the "X" model. Maybe, the new "DX" model is not (fully) supported by xPC Target. The xPC Taget documentationl says "The xPC Target block library supports this [MM-32-AT] board with these driver blocks. These blocks also support the Diamond-MM-32X-AT board." No word about the "DX". On Tue, 8 Dec 2009 23:10:20 +0000 (UTC), "saj " <sameer.a.joshi(a)gmail.com> wrote: >I have a PC104 stack setup with a DMM-32DX-AT I/O card. I have a DOSloader setup with xPC Target version 4.2. As an initial test, i setup the 4 analog outputs on this card to provide constant output voltages 1, 2, 3, and 4 respectively. That seems to work fine. I checked the voltages on a scope and verified. Then i wired up these 4 analog outputs to the 1st four analog inputs of the i/o card to do loop-back testing. On the Simulink side, I added a A/D driver block to my model to read the 4 wired analog inputs. When i execute this model on the hardware, instead of seeing values 1, 2, 3, and 4 volts on the A/D channels I see some junk values that have no co-relation to the input voltages. When i stop the model, the monitor connected to the video port of the hardware shows a message that reads "Error MM-32 A/D was reset n times", where n is a number that seems to depend upon the sample >time of the model and also the length of time for which i run this model. I do not see the reset message if i use the analog output channels and the digital input and output channels on this card. Only the A/D seems to have trouble. > >The Analog Input jumper is setup for SE 0-31 input configuration. > >I have experimented with the setting the jumper at location "16" on the J7 header to both 'In' and 'Out'. Did not seem to help. > >Any help on resolving this issue will be greatly appreciated.
From: Gordon Weast on 10 Dec 2009 14:08 The only changes to the driver we believe needed to be made were for the D/A driver. The board can be purchased with either a 12 bit D/A or a 16 bit D/A and the two versions need slightly different programming. Do you know which version you have? That shouldn't make any difference to the A/D, but I don't know for sure. The A/D driver I have works fine here with the DX board I have. Which compiler are you using? Is this a new board or one you've had for a while? What is your target PC? In particular, what speed CPU is it running? What other PC104 boards are in the stack? What are the incorrect values? Are they near zero noise readings or larger but incorrect readings? Are they stable or varying? Turning off the automatic autocal on the X version of the board became part of the shipping version in R2006b. The DX version shares that code. Unfortunately, hardware changes have been known to creep into newly manufactured boards that we don't know about. I hope this isn't a case of that happening. Gordon Weast xPC Target Development The MathWorks saj wrote: > I have a PC104 stack setup with a DMM-32DX-AT I/O card. I have a DOSloader > setup with xPC Target version 4.2. As an initial test, i setup the 4 analog > outputs on this card to provide constant output voltages 1, 2, 3, and 4 > respectively. That seems to work fine. I checked the voltages on a scope and > verified. Then i wired up these 4 analog outputs to the 1st four analog > inputs of the i/o card to do loop-back testing. On the Simulink side, I added > a A/D driver block to my model to read the 4 wired analog inputs. When i > execute this model on the hardware, instead of seeing values 1, 2, 3, and 4 > volts on the A/D channels I see some junk values that have no co-relation to > the input voltages. When i stop the model, the monitor connected to the video > port of the hardware shows a message that reads "Error MM-32 A/D was reset n > times", where n is a number that seems to depend upon the sample time of the > model and also the length of time for which i run this model. I do not see > the reset message if i use the analog output channels and the digital input > and output channels on this card. Only the A/D seems to have trouble. > > The Analog Input jumper is setup for SE 0-31 input configuration. > > I have experimented with the setting the jumper at location "16" on the J7 > header to both 'In' and 'Out'. Did not seem to help. > > Any help on resolving this issue will be greatly appreciated.
From: Sam on 15 Dec 2009 16:15 Gordon: Thanks for the update. I will take it up with support. Just out of curiosity, what kind of CPU are you using for your test setup? I can look around to see if i can get that CPU. regards, Gordon Weast <gweast(a)mathworks.com> wrote in message <hg6d8i$pku$1(a)fred.mathworks.com>... > Sam, > > This really sounds like a hardware revision issue. Please contact MathWorks > tech support and open a problem report. Mention this email as a starting > point. We may need to borrow your board to see what's different about it. > > We have our DX board in a test setup where we're doing exactly what you > are trying to do and our board works fine. > > Gordon Weast > xPC Target Development > The MathWorks >
From: Gordon Weast on 16 Dec 2009 15:36
Sam, I just checked and it seems we've been running our test on one of the slower targetbox configurations that we still have. I asked if the test can be run on one of the faster boxes. The slow one is running at 266 MHz while the faster one is 600 MHz. Still, the driver loops waiting for each channel acquisition to appear in the on board fifo in a loop that should wait up to roughly 50 microseconds for each acquisition. The IO register read time is dependent on the ISA bus speed of 8 MHz and should be independent of the CPU speed. Each acquisition should take 10 microseconds from the previous one so the loop should exit long before the 50 loops are executed. Let's continue through tech support. Gordon Weast xPC Target Development Sam wrote: > Gordon: > > Thanks for the update. I will take it up with support. > > Just out of curiosity, what kind of CPU are you using for your test setup? I > can look around to see if i can get that CPU. > > regards, > > Gordon Weast <gweast(a)mathworks.com> wrote in message > <hg6d8i$pku$1(a)fred.mathworks.com>... >> Sam, >> >> This really sounds like a hardware revision issue. Please contact >> MathWorks tech support and open a problem report. Mention this email as a >> starting point. We may need to borrow your board to see what's different >> about it. >> >> We have our DX board in a test setup where we're doing exactly what you are >> trying to do and our board works fine. >> >> Gordon Weast xPC Target Development The MathWorks >> |