|
--- In "charliem_1216" <> wrote: > > Well, you've not given enough details to make any intelligent comments. The A2D in question is http://www.diamondsystems.com/products/dmm32dx > * What's your input signal like during testing? I guess if you are calculating standard deviation, you are holding (or trying) to hold the input constant. What readings do you get, and what do you expect? We're using the A2D in differential mode with constant values on three inputs. One input is a 'D' battery with resistors to divide the voltage down to 144mv, one is a shorted channel to provide a zero reference, the third is a on-board precision 5V source that the vendor's software uses for auto-calibration. > * What sampling frequency? The A2D runs at 4us / sample. We programmer the A2D to trigger at a 5us rate, and it puts values in a FIFO. We discard the first 200 samples in the FIFO, and then sum the next 3334 (that equates to 1/60 of a second) samples to generate a sample to use. The summing is done in floating point. > * How is data taken / transferred from your A/D card? Userland or kernel driver? Userland, I suppose, since the test program is doing direct I/O to the A2D card. > * What kernel versions for x86 & ARM? glibc versions? ARM: libc-2.3.6, kernel 2.6.21-ts x86: libc-2.3.2, kernel 2.6.19.7-L31-rtai3.6 > * What do you observe about the ARM data? Random noise, HF noise? We haven't performed that type of analysis. > * Have you ruled out a bad A/D card on the 7800 system? It is the same physical card on both systems -- we just swap out the SBC and leave all of the other stuff intact. So it always has exactly the same inputs and cabling -- except perhaps for differences on the PC- 104 bus. > My first guesses would be RF interference with your analog stage, or timing jitter (your data is not sampled when you think it is), or even endian-ness issues. > > You don't really say what 'code' is running running slower. If you do any floating point at all, that can have slight to severe impacts on your program speed on ARM, depending on your system. If you are only talking of the acquisition time, you really can't discuss it without posting the code you are using. Here's pseudo-code for the way we read a sample: Tell the A2D to start sampling Loop to read 3534 samples Sleep 1ms Loop while samples in FIFO Read sample if 200 < SampleNo then Sum += Sample '' Sum is a double precision float ; EndIf EndLoop EndLoop UsableSample = Sum / 3334 Thanks for taking the time to look at this. __._,_.___
Your email settings: Individual Email|Traditional
Change settings via the Web (Yahoo! ID required) Change settings via email: =Email Delivery: Digest | m("yahoogroups.com?subject","ts-7000-fullfeatured");=Change Delivery Format: Fully Featured">Switch to Fully Featured Visit Your Group | Yahoo! Groups Terms of Use | =Unsubscribe __,_._,___ |