Hello Frederic,
After some reading up on a backlog of e-mails I have found the time to read more carefully through your specifications.
First I would like to check also if I have the right understanding of your specs.
Before doing this I would like to check if you by any chance have some pictures or log-files from a scope that shows how these signals will exactly look like?
(This would greatly help in explaining where the critical points will be)
About your 4 use cases:
1) I understand that this means that your shaft will rotate at around 2Hz. (Will this give a 2Hz signal at the input?)
How big do you expect the maximal variation to be? Do you expect it to be around 0.1Hz or bigger/smaller?
If I understand you correctly then this is what will happen:
"Normally the speed should be 2Hz. There are however case when the speed varies from this 2 Hz and these variations will most likely happen at a rate between 10 and 200Hz"
If this is correct, then this would mean that you need an Output Bandwidth of around 200 Hz to see all the variations in your signal.
This would mean you would require/choose the output bandwidth setting of 320Hz, which expects a 5 kHz minimum input frequency.
Although these (filter settins vs minimum input frequency) are conservative specifications, 2 Hz is a factor 2500 different from this.
So I don't expect this to work (I'll come back to work-arounds for "closer cases").
2) (I'll do some copy pasting to keep everything consistent)
I understand that this means that your shaft will rotate at around 20Hz. (Will this give a 20Hz signal at the input?)
How big do you expect the maximal variation to be? Do you expect it to be around 1Hz or bigger/smaller?
If I understand you correctly then this is what will happen:
"Normally the speed should be 20Hz. There are however case when the speed varies from this 2 Hz and these variations will most likely happen at a rate between 100 and 2000Hz"
If this is correct, then this would mean that you need an Output Bandwidth of around 2000 Hz to see all the variations in your signal.
This would mean you would require a filter setting of 2000 Hz, which doesn't exist (1000 Hz is the maximum).
So you wouldn't be able to detect all variations with a rate of 2000 Hz. (under the assumption that you meant that the rate of the variation of the 20 Hz ("base") frequency can be up to 2000Hz and I should consider this as some kind of frequency modulation)
3) Does you expect in this case a 150 Hz base signal that will vary at a rate of 2 Hz.
If this is the case, then you're output bandwith should be at least 2 Hz.
This means the 1 Hz filter setting will not be enough and you would need to select the 40 Hz filter setting.
This filter setting yields another requirement:
Minimum Input Frequency= 600 Hz.
In this case your 150 Hz base frequency is "too low" to be able to get the specifications of the module.
What will happen in this case is the following:
If you don't go below specs (600Hz), then we can guarantee that you will meet the expected accuracy.
If you go below specs, then how much you go below specs will also cause your accuracy to become worse.
This will be visible by a (periodic) ripple that becomes visible in your output signal.
This ripple can be countered by averaging over at least one ripple period.
If you average over more periods of this ripple, the better the result will be.
4) Does you expect in this case a 1500 Hz base signal that will vary at a rate of 20 Hz.
If this is the case, then you're output bandwith should be at least 20 Hz.
In this case you could use the 40Hz filter settings and a part of your measured frequencies (when the frequency is above 600Hz) will fall in specs.
However a big part will also be out of spec (below 600Hz) and the same reasoning as before will be valid.
For most of these cases the SCXI-1126 does not seem the optimal solution.
PS: Most of my explanation is based on an incorrect 1-1 relation between encoder and shaft, but the explanation about the effects of low frequencies will still be the same.
Kind Regards,
Thierry C - CLA, CTA - Senior R&D Engineer (Former Support Engineer) - National Instruments
If someone helped you, let them know. Mark as solved and/or give a kudo. 😉
