jokubasver
commented
3 years ago Git didn't commit a few changes to my fork, so here they are: https://github.com/jokubasver/OxiVis/commit/cf6bcb0f298ddf9399ea905089665737574084d6
Closed jokubasver closed 2 years ago
jokubasver
commented
3 years ago Git didn't commit a few changes to my fork, so here they are: https://github.com/jokubasver/OxiVis/commit/cf6bcb0f298ddf9399ea905089665737574084d6
jokubasver
commented
3 years ago I think I found all the values.
I'll repeat myself to be clear:
Writing 0x9b011c to characteristics UUID 0000ff01-0000-1000-8000-00805f9b34fb turns on data streaming/notifications.
Data is streamed on 0000ff02-0000-1000-8000-00805f9b34fb characteristics UUID.
There are two separate hex strings:
---8 byte stream - SpO2, BPM--- Hex: eb 01 04 53 63 7f 00 25 Dec: 235 1 4 83 99 127 0 37
1st byte - 235 (Constant) 2nd byte - 1 (Identifier that the byte stream is for SpO2 and BPM) 3rd byte - 4/6 (Probably a finger detector - 04 when no finger, 06 when finger is in device) 4th byte - 83 (BPM) 5th byte - 99 (SpO2) 6th byte - 127 (Doesn't change) 7th byte - 0 (Doesn't change) 8th byte - 37 (Unknown, seems to be BPM but values are shifted down by 46 units)
---6 byte stream (PPG, bar graph)--- Hex: eb 00 03 48 09 3f Dec: 235 0 3 72 9 63 Hex: eb 00 43 3c 07 71 Dec: 235 0 67 60 7 113
1st byte - 235 (Constant) 2nd byte - 0 (Identifier that the byte stream is for PPG) 3rd byte - PPG peak detector? (When PPG is at the highest peak, this 3rd byte goes high) 4th byte - PPG 5th byte - Bar graph? (Basically looks like a low resolution PPG when graphed to a line chart) 6th byte - PPG, but with exaggerated peaks and different scale
Graphs were made with extracted data from Android Bluetooth logs and Wireshark.
Some data points are missing from my own data extraction to Google Sheets, as sometimes the hex string arrives two or three times in a row (for example instead ofeb000348093f it's sometimes eb000348093feb000348093f)
Now I need to see what data is streamed when sending 0x9b001binstead of 0x9b011c
tobiasisenberg
commented
3 years ago Hi @jokubasver! Great that you try to make this work also for another device. I vaguely remember that there was also a low-res PPG value in my data. But since I do not have this specific oximeter I am afraid I cannot help much, but if you have a pull request that makes it possible to run the code on either device I would love to see it. Great work!
tobiasisenberg
commented
2 years ago Closing issue.
I am trying to adapt your code to support this Contec CMS50D-BT oximeter. But I am having issues.
You can see the changelog here: https://github.com/jokubasver/OxiVis/commit/35c159698ef362c76972c0efaa8cdb7894d65c26
Thanks to Android Bluetooth logging and Wireshark, I was able to read data from it - the main thing it needed was to send
0x9b011cto0000ff01-0000-1000-8000-00805f9b34fbcharacteristics UUID for the other characteristics UUID (0000ff02-0000-1000-8000-00805f9b34fb) to spit out data.0x9b011cneeds to be sent often, as otherwise the meter only sends data for less than a minute and then stops, so I had to move thewriteValuecommand into thevoid loop.Also,
pRemoteCharacteristicOximeter->getDescriptor(BLEUUID((uint16_t)0x2902))->writeValue((uint8_t*)notificationOn, 2, true);didn't work in my case, so I did it a bit differently, but I am not sure if I did it properly.One issue I was having is an "Incorrect packet length" error on ESP serial. Not sure if this is a correct fix, but since the hex length in my case is 8:
I changed
int packetCountto equellength / 2;.With this change, the error goes away and I am able to see data via serial print commands, but it takes around 10 minutes for the data to be saved to csv, which seems too long. The data also has some garbage in the csv file too, so I'm not sure what's going on. Maybe there is an issue with the buffer while using this Contec meter.
I'm attaching an example csv file. oximeter-20211008-203743-12079.csv
Some data from MQTT Explorer:
I am also not quite sure yet where the PPG data is. It could be the last byte of that hex stream, but I also found that sending
0x9b001bgives a another stream of data, which might be the PPG stream.Any help would be greatly appreciated.