- This topic has 3 replies, 2 voices, and was last updated 6 years, 4 months ago by
.
- Posts
We’ve got a project here that we are about to embark on and are looking hard at the Mooshimeter. It’s not clear to me what the sample/averaging time is. I’ve seen user definable update rates of 1 second etc. So here is what we will be doing – We will be observing very low range current changes on a small DC electric motor as the mechanical load changes . We may have to use the SD card because we need to get the data into a spreadsheet eventually. I understand the streaming data is still in test. It might be OK if the data record rate is 1 second as long as the sample/averaging time was a second as well. Any help out there for me? Thanks
Thanks Ville. That was helpful.
What happened to my post? (#18096)
I tried to edit it couple times, as underline-tag screwed it, at least on chrome based browsers. Editing did not remove the problem, but it seems this forum gets cached on its way to europe/finland anyway, so i might just got my first version from cache, even if there was that fixed version available. I tried to repeat that edit couple times, did it screw something?
Hello Ville,
I received a copy of your post and response to my question in my email notification. I’m going to put it up as an addendum to this post because it was complete and well written and may be helpful to others. Once again – thank you
ville wrote:
Mooshimeter can do one thing at a time (apart from sampling two channels, but i don’t know how simultaneous that really is).
It goes something like this:
step 1: take samples. Default for dc seems to be 64 samples at 125 Hz. So it takes roughly 500 ms. You could play with those settings and set amount of samples to 32/64/128/256 and sample rate to 125 – 8000 Hz in doubling increments also. That means anything from ~4 ms to ~2 seconds for just the sampling.
step 2: averaging etc that sample buffer, probably takes more time for larger set of samples. That takes some unknown amount of time, as it is just one slow processor doing it all.
step 3: saving that one value that step 2 produced. Again, takes some time. I’m quite sure that memory card speed has nothing to do with it, as those can take tens of megabytes in second and we are only writing hundreds of bytes..
optional step 4: transfer value from step 2 over bluetooth to your phone. You guessed right, it will take some time. But disconnecting would give you little faster logging to memory card in “no wait” mode.
step 5: wait for next set time for taking sample. “no wait”/1s/10s/min/10min.
step 6: go to step 1 and start again
So, constant sampling and saving the results is not really possible, but for example with dc-default-sampling-options, ~500ms sample of every 1 second is. You could get little bit more with “no wait” option, but still, processing takes some time and you could lose some tens of milliseconds long spikes. And with 125Hz/64smpl those would probably hide in the averages anyway.
And then there are errors in sampling.. James told that sometimes whole sample buffer might get discarded and sampling would start again.. That would appear at least in “no wait” mode logs by longer interval in timestamps.
What was the question? When we got here? :D
