Home › Forums › Mooshimeter Support › High frequency logging unreliable
- This topic has 1 reply, 2 voices, and was last updated 7 years, 7 months ago by
admin.
- AuthorPosts
Stefan Thiel
GuestIntroduction: I’ve bought the mooshimeter to perform the mechanical transfer function of my airwings saddle post of my new bicycle during real cycling on the road. 8000 samples per second were mentioned (!!), I could get along with 1000 sps. I have a lot of experience in similar data acquisition tasks on running presses in my job and I took two acceleration sensor over weekend to do it.
Problem: The highest data rate achieved (32GB class 10 card) was ~ 300 Hz. This would have been nearly acceptable, but the rate was unstable and changed after various logging times to arbitrary values between 300 and 3000 ms delay from sample to sample. (setting war 4000 /s rate, 1 sample). And the (really not) funny thing is, that the sample rate stayed high as long as no real signal occured and fell down with the first peak – when I started the ride. After arrival, some minutes later, the rate went high again.
I’ve made three rides, than gave up. The EXCEL-File with data is here
What can be done to get this working?admin
KeymasterHi Stefan,
First off, very sorry for the slow response, some web server settings were changed and forum notifications stopped being delivered. It’s fixed now and responses should be faster.
To answer your post requires digging in a little about how the Mooshimeter does its sampling. Because BLE has very low bandwidth, the Mooshimeter builds up a sample buffer of up to 256 samples, runs analysis on them (mean, RMS), then delivers only the result of its analysis over the BLE link to keep the amount of data transferred low. So that internal buffer can be sampled at up to 8kHz, which means events at that frequency are reflected in its RMS analysis, but the rate at which it delivers notifications to the next level is lower than that.
Why does this affect SD card logging? The SD card logging feature was designed to capture whatever would be sent over the BLE link, so it stores the calculated data as well. Because it was designed with the idea that someone would be putting the meter somewhere and leaving it for a long time, the way it’s written it assumes the logging will be slow, so after every sample it writes the data out to the SD card (which takes a few ms). So your description of maximum 300Hz sampling sounds about right.
There’s a “buffer mode” in the app that lets you look download the raw sample buffers and examine the data at its full bandwidth, but only in bursts, so it’s not exactly what you’re looking for. Unfortunately there’s no way to log the raw sample buffers to SD right now. I’ll put it on the wishlist though, I can see it being useful in other circumstances.
I’m really sorry the meter doesn’t support your use case right now, please let me know how else I can help. Best
~James
- AuthorPosts