An rpi based DIY Vibration meter - Page 6 - AVS Forum | Home Theater Discussions And Reviews
Forum Jump: 
 103Likes
Reply
 
Thread Tools
post #151 of 321 Old 05-21-2017, 02:59 PM - Thread Starter
AVS Forum Special Member
 
3ll3d00d's Avatar
 
Join Date: Sep 2007
Location: London, UK
Posts: 4,149
Mentioned: 267 Post(s)
Tagged: 0 Thread(s)
Quoted: 2482 Post(s)
Liked: 2701
Quote:
Originally Posted by MrGrey View Post
Thanks allot
updated build with the adjusted levels - https://github.com/3ll3d00d/vibe/rel.../0.4.0-alpha.2

didn't have time to do the other bits, will be later this week probably
MrGrey likes this.
3ll3d00d is offline  
Sponsored Links
Advertisement
 
post #152 of 321 Old 05-21-2017, 05:12 PM
AVS Forum Special Member
 
MrGrey's Avatar
 
Join Date: Oct 2014
Posts: 3,043
Mentioned: 149 Post(s)
Tagged: 0 Thread(s)
Quoted: 1737 Post(s)
Liked: 6490
Thank you for all you have already done. i will try this build out later tonight.
MrGrey is offline  
post #153 of 321 Old 05-23-2017, 10:34 PM
AVS Forum Special Member
 
MrGrey's Avatar
 
Join Date: Oct 2014
Posts: 3,043
Mentioned: 149 Post(s)
Tagged: 0 Thread(s)
Quoted: 1737 Post(s)
Liked: 6490
i tired the new version and it works well
3ll3d00d likes this.
MrGrey is offline  
Sponsored Links
Advertisement
 
post #154 of 321 Old 05-27-2017, 07:32 AM - Thread Starter
AVS Forum Special Member
 
3ll3d00d's Avatar
 
Join Date: Sep 2007
Location: London, UK
Posts: 4,149
Mentioned: 267 Post(s)
Tagged: 0 Thread(s)
Quoted: 2482 Post(s)
Liked: 2701
I now have resizing charts working as well as custom styling, still a few more options to add (like titles)

Attached Thumbnails
Click image for larger version

Name:	chart.jpg
Views:	183
Size:	142.7 KB
ID:	2157858  
MrGrey likes this.
3ll3d00d is offline  
post #155 of 321 Old 05-27-2017, 02:42 PM - Thread Starter
AVS Forum Special Member
 
3ll3d00d's Avatar
 
Join Date: Sep 2007
Location: London, UK
Posts: 4,149
Mentioned: 267 Post(s)
Tagged: 0 Thread(s)
Quoted: 2482 Post(s)
Liked: 2701
some weirdness going on with translating the canvas size to the exported image size but it works apart from that, an example



Attached Thumbnails
Click image for larger version

Name:	chart.jpg
Views:	177
Size:	131.6 KB
ID:	2158250   Click image for larger version

Name:	style.jpg
Views:	190
Size:	66.8 KB
ID:	2158258  
MrGrey likes this.
3ll3d00d is offline  
post #156 of 321 Old 05-29-2017, 08:41 AM - Thread Starter
AVS Forum Special Member
 
3ll3d00d's Avatar
 
Join Date: Sep 2007
Location: London, UK
Posts: 4,149
Mentioned: 267 Post(s)
Tagged: 0 Thread(s)
Quoted: 2482 Post(s)
Liked: 2701
another release -> https://github.com/3ll3d00d/vibe/rel...e-analyser.exe

I added the ability to save chart format options as presets.

There is still some oddity in the size of the output, the actual image size is some fixed percentage different to what I request. It seems to vary by machine too (so it's 33% smaller than I request on one laptop, 25% smaller on another). I have no idea why this is and I'm not going on a wild goose chase to find out

I think this one will be the final one for this round, just need to run a few more tests to be sure.
MrGrey likes this.
3ll3d00d is offline  
post #157 of 321 Old 05-29-2017, 03:14 PM - Thread Starter
AVS Forum Special Member
 
3ll3d00d's Avatar
 
Join Date: Sep 2007
Location: London, UK
Posts: 4,149
Mentioned: 267 Post(s)
Tagged: 0 Thread(s)
Quoted: 2482 Post(s)
Liked: 2701
published a full release - https://github.com/3ll3d00d/vibe/releases/tag/0.4.0

exe is at the bottom of the page
MrGrey likes this.
3ll3d00d is offline  
post #158 of 321 Old 05-29-2017, 05:23 PM
AVS Forum Special Member
 
MrGrey's Avatar
 
Join Date: Oct 2014
Posts: 3,043
Mentioned: 149 Post(s)
Tagged: 0 Thread(s)
Quoted: 1737 Post(s)
Liked: 6490
MrGrey is offline  
post #159 of 321 Old 05-29-2017, 06:47 PM
AVS Forum Special Member
 
MrGrey's Avatar
 
Join Date: Oct 2014
Posts: 3,043
Mentioned: 149 Post(s)
Tagged: 0 Thread(s)
Quoted: 1737 Post(s)
Liked: 6490
A compassion between Vibe and Speclab.
i wish there was a way to make the lines thicker and a brighter red and green in vibe.


SPECLAB



VIBE


Last edited by MrGrey; 05-29-2017 at 06:56 PM.
MrGrey is offline  
post #160 of 321 Old 05-30-2017, 02:37 AM - Thread Starter
AVS Forum Special Member
 
3ll3d00d's Avatar
 
Join Date: Sep 2007
Location: London, UK
Posts: 4,149
Mentioned: 267 Post(s)
Tagged: 0 Thread(s)
Quoted: 2482 Post(s)
Liked: 2701
easily done, will post an exe later

Attached Thumbnails
Click image for larger version

Name:	chart.png
Views:	167
Size:	139.6 KB
ID:	2162506  
MrGrey likes this.
3ll3d00d is offline  
post #161 of 321 Old 05-30-2017, 02:58 AM
AVS Forum Special Member
 
MrGrey's Avatar
 
Join Date: Oct 2014
Posts: 3,043
Mentioned: 149 Post(s)
Tagged: 0 Thread(s)
Quoted: 1737 Post(s)
Liked: 6490
Awesome thank you

Sent from my LG-H901 using Tapatalk
MrGrey is offline  
post #162 of 321 Old 05-30-2017, 02:53 PM - Thread Starter
AVS Forum Special Member
 
3ll3d00d's Avatar
 
Join Date: Sep 2007
Location: London, UK
Posts: 4,149
Mentioned: 267 Post(s)
Tagged: 0 Thread(s)
Quoted: 2482 Post(s)
Liked: 2701
Quote:
Originally Posted by MrGrey View Post
Awesome thank you
try this one - https://github.com/3ll3d00d/vibe/rel.../0.4.1-alpha.1

hopefully the options are obvious, some of them have tooltips to explain further. Note that if you save a preset named default then that preset will be applied automatically whenever you view a chart.

one other thing you might find handy is that the analyse page lets you show the difference between any measurement. This could be handy to show the exact size of the difference between 2 tracks in particular frequency ranges.

To do so, just create a target first (the bullseye button on the upload screen) then load the targets into the analysis screen and then set one of them as the reference

e.g.

Attached Thumbnails
Click image for larger version

Name:	compare.png
Views:	187
Size:	154.8 KB
ID:	2163322  
MrGrey likes this.
3ll3d00d is offline  
post #163 of 321 Old 05-31-2018, 12:40 PM
AVS Forum Addicted Member
 
Marc Alexander's Avatar
 
Join Date: Aug 2001
Location: Inland Empire, CA
Posts: 13,518
Mentioned: 240 Post(s)
Tagged: 0 Thread(s)
Quoted: 5472 Post(s)
Liked: 4972
Quote:
Originally Posted by Marc Alexander View Post
I'm planning to get mine built this week.
I can't believe I am an entire year late!

3ll3d00d likes this.
Marc Alexander is online now  
post #164 of 321 Old 05-31-2018, 01:17 PM - Thread Starter
AVS Forum Special Member
 
3ll3d00d's Avatar
 
Join Date: Sep 2007
Location: London, UK
Posts: 4,149
Mentioned: 267 Post(s)
Tagged: 0 Thread(s)
Quoted: 2482 Post(s)
Liked: 2701
FWIW I haven't got round to updating my rpi's to jessie (which I assume yours will start on) so it's not impossible that some versions will differ from what I've tested, I wouldn't expect any particular issues but let me know if you encounter anything odd.
3ll3d00d is offline  
post #165 of 321 Old 05-31-2018, 02:53 PM
AVS Forum Addicted Member
 
Marc Alexander's Avatar
 
Join Date: Aug 2001
Location: Inland Empire, CA
Posts: 13,518
Mentioned: 240 Post(s)
Tagged: 0 Thread(s)
Quoted: 5472 Post(s)
Liked: 4972
Quote:
Originally Posted by MrGrey View Post
A compassion between Vibe and Speclab.
i wish there was a way to make the lines thicker and a brighter red and green in vibe.


SPECLAB



VIBE

Your graphs look identical? How did you measure to achieve these results?
Marc Alexander is online now  
post #166 of 321 Old 05-31-2018, 03:09 PM
AVS Forum Special Member
 
MrGrey's Avatar
 
Join Date: Oct 2014
Posts: 3,043
Mentioned: 149 Post(s)
Tagged: 0 Thread(s)
Quoted: 1737 Post(s)
Liked: 6490
Quote:
Originally Posted by Marc Alexander View Post
Your graphs look identical? How did you measure to achieve these results?

I extracted the lossless track from the Blu-Ray or UHD into their individual untouched channels L,R,C, LFE, ect and then used Audacity and then SpecLab Version 2.77b22 to give the final results. all done on the PC. https://www.avsforum.com/forum/113-subwoofers-bass-transducers/2763785-ultimate-list-bass-movies-w-frequency-charts-123.html#post55171238
MrGrey is offline  
post #167 of 321 Old 05-31-2018, 03:21 PM
AVS Forum Addicted Member
 
Marc Alexander's Avatar
 
Join Date: Aug 2001
Location: Inland Empire, CA
Posts: 13,518
Mentioned: 240 Post(s)
Tagged: 0 Thread(s)
Quoted: 5472 Post(s)
Liked: 4972
Quote:
Originally Posted by MrGrey View Post
I extracted the lossless track from the Blu-Ray or UHD into their individual untouched channels L,R,C, LFE, ect and then used Audacity and then SpecLab Version 2.77b22 to give the final results. all done on the PC. https://www.avsforum.com/forum/113-subwoofers-bass-transducers/2763785-ultimate-list-bass-movies-w-frequency-charts-123.html#post55171238
I get that part. I'm surprised how similar your Vibe graph is.
Marc Alexander is online now  
post #168 of 321 Old 05-31-2018, 03:34 PM
AVS Forum Special Member
 
MrGrey's Avatar
 
Join Date: Oct 2014
Posts: 3,043
Mentioned: 149 Post(s)
Tagged: 0 Thread(s)
Quoted: 1737 Post(s)
Liked: 6490
Quote:
Originally Posted by Marc Alexander View Post
I get that part. I'm surprised how similar your Vibe graph is.

O, all i did was copy the settings over from Speclab and drop the same file in into Vibe. it's all thanks to the OP for making those settings in Speclab available in Vibe so it was pretty straightforward.
MrGrey is offline  
post #169 of 321 Old 05-31-2018, 03:38 PM
AVS Forum Addicted Member
 
Marc Alexander's Avatar
 
Join Date: Aug 2001
Location: Inland Empire, CA
Posts: 13,518
Mentioned: 240 Post(s)
Tagged: 0 Thread(s)
Quoted: 5472 Post(s)
Liked: 4972
Quote:
Originally Posted by MrGrey View Post
O, all i did was copy the settings over from Speclab and drop the same file in into Vibe. it's all thanks to the OP for making those settings in Speclab available in Vibe so it was pretty straightforward.
Ok, so this was not data from a RPI?
Marc Alexander is online now  
post #170 of 321 Old 05-31-2018, 03:45 PM
AVS Forum Special Member
 
MrGrey's Avatar
 
Join Date: Oct 2014
Posts: 3,043
Mentioned: 149 Post(s)
Tagged: 0 Thread(s)
Quoted: 1737 Post(s)
Liked: 6490
Quote:
Originally Posted by Marc Alexander View Post
Ok, so this was not data from a RPI?

Nope, the same file i analyzed with Speclab is the same file i put into the Vibe application.
MrGrey is offline  
post #171 of 321 Old 06-08-2018, 07:08 PM
AVS Forum Addicted Member
 
Marc Alexander's Avatar
 
Join Date: Aug 2001
Location: Inland Empire, CA
Posts: 13,518
Mentioned: 240 Post(s)
Tagged: 0 Thread(s)
Quoted: 5472 Post(s)
Liked: 4972
Quote:
Originally Posted by 3ll3d00d View Post
FWIW I haven't got round to updating my rpi's to jessie (which I assume yours will start on) so it's not impossible that some versions will differ from what I've tested, I wouldn't expect any particular issues but let me know if you encounter anything odd.
I don't know if any of this is related to Jessie.

It does not appear that the analyser or recorder is accessing the MPU6050 registers. I am read/writing the 6050 registers from Raspbian via i2c. I set them to match the Vibe analyzer default targets.

Here is what is reported by the recorder (http://10.10.10.199:10002/api/1/devices):
Quote:
[{"accelerometerEnabled": true, "status": "INITIALISED", "name": "mpu6050", "fs": 500, "failureCode": null, "gyroSens": 500, "samplesPerBatch": 125, "accelerometerSens": 2, "gyroEnabled": false}]
Here is what Vibe shows (Update Device has no effect):


Otherwise everything appears to be working as designed. Excellent work @3ll3d00d !!
3ll3d00d likes this.
Marc Alexander is online now  
post #172 of 321 Old 06-09-2018, 12:28 AM - Thread Starter
AVS Forum Special Member
 
3ll3d00d's Avatar
 
Join Date: Sep 2007
Location: London, UK
Posts: 4,149
Mentioned: 267 Post(s)
Tagged: 0 Thread(s)
Quoted: 2482 Post(s)
Liked: 2701
Quote:
Originally Posted by Marc Alexander View Post
It does not appear that the analyser or recorder is accessing the MPU6050 registers. I am read/writing the 6050 registers from Raspbian via i2c. I set them to match the Vibe analyzer default targets.
probably need to see the analyser & recorder logs, possibly with debug mode on. Can you post some of the logs to cover the time when you make the change? The analyser should be repeatedly attempting to send the command to update the config so need to see whether the recorder receives that and fails to respond or whether the analyser can't send it to the device.

The command it sends is handled in https://github.com/3ll3d00d/vibe/blo...devices.py#L36 so is like

PATCH http://recorder/api/1/devices/mpu6050 {"accelerometerSens": 2}

if it returns a 500 from this then it is failing to apply the change (which means it can't be talking to the device correctly)

something like https://www.getpostman.com/ will let you send a PATCH manually

debug logging will give you lots more detail on what is going on so might need that too

can you make measurements?
3ll3d00d is offline  
post #173 of 321 Old 06-09-2018, 12:39 AM
AVS Forum Addicted Member
 
Marc Alexander's Avatar
 
Join Date: Aug 2001
Location: Inland Empire, CA
Posts: 13,518
Mentioned: 240 Post(s)
Tagged: 0 Thread(s)
Quoted: 5472 Post(s)
Liked: 4972
Quote:
Originally Posted by 3ll3d00d View Post
can you make measurements?
Yes.
Marc Alexander is online now  
post #174 of 321 Old 06-09-2018, 12:41 AM - Thread Starter
AVS Forum Special Member
 
3ll3d00d's Avatar
 
Join Date: Sep 2007
Location: London, UK
Posts: 4,149
Mentioned: 267 Post(s)
Tagged: 0 Thread(s)
Quoted: 2482 Post(s)
Liked: 2701
curious, if you post some logs then I can have a look
3ll3d00d is offline  
post #175 of 321 Old 06-10-2018, 01:56 PM
AVS Forum Addicted Member
 
Marc Alexander's Avatar
 
Join Date: Aug 2001
Location: Inland Empire, CA
Posts: 13,518
Mentioned: 240 Post(s)
Tagged: 0 Thread(s)
Quoted: 5472 Post(s)
Liked: 4972
I'm going to capture some logs as well as reduce the sensitivity today. This Vibe sensor is extremely sensitive at the moment. On a level service (our dining-room table), it picks up the vibrations of conversation as well as my son lightly tapping the iPad.

I'm very happy with the kit I purchased in Amazon. I mounted the MPU6050 upside-down inside of the case with 30lb Gorilla Tape. This allowed me to contain the wiring and allow a space for the backside of the pins, otherwise you can't place the sensor level. The sensor knows its orientation at all times so it doesn't matter (z-axis values will just be negative).

I am powering with a 8000mAh power bank with a lightweight but high quality USB cable. I have not been weighing down the sensor. It is solidly held between the seat and back cushions of my seating. Placing it up-side-down puts the sensor right-side-up. If I need a weight I could attach the power bank. I'm going to take groundplane measurements with just a thin piece of foam between it and the ground to protect the case from scuffs.
3ll3d00d likes this.
Marc Alexander is online now  
post #176 of 321 Old 06-11-2018, 02:51 AM
AVS Forum Addicted Member
 
Marc Alexander's Avatar
 
Join Date: Aug 2001
Location: Inland Empire, CA
Posts: 13,518
Mentioned: 240 Post(s)
Tagged: 0 Thread(s)
Quoted: 5472 Post(s)
Liked: 4972
I was able to look at the logs and determine my issues are/were network related. The vibe-analyser is communicating with the recorder and sensor more reliably now.

Is it possible/advisable to run the Analyser and Recorder on the same RPI?
Marc Alexander is online now  
post #177 of 321 Old 06-11-2018, 03:24 AM - Thread Starter
AVS Forum Special Member
 
3ll3d00d's Avatar
 
Join Date: Sep 2007
Location: London, UK
Posts: 4,149
Mentioned: 267 Post(s)
Tagged: 0 Thread(s)
Quoted: 2482 Post(s)
Liked: 2701
It is possible for sure, there were no precompiled arm binaries available for some of the python dependencies though last I checked so expect installation time to be extremely long (overnight at least).

A quick Google says piwheels is now a thing though and stretch uses it by default so it should install quickly. Just use a separate venv as per the "running on Linux" docs and you should be up and running.

Last edited by 3ll3d00d; 06-11-2018 at 03:29 AM.
3ll3d00d is offline  
post #178 of 321 Old 06-15-2018, 07:46 PM
AVS Forum Addicted Member
 
Marc Alexander's Avatar
 
Join Date: Aug 2001
Location: Inland Empire, CA
Posts: 13,518
Mentioned: 240 Post(s)
Tagged: 0 Thread(s)
Quoted: 5472 Post(s)
Liked: 4972
Observations & lessons learned so far.

Network connectivity must be rock solid. The RPI 3b does not have the powerful WiFi and does not recover well from lost connectivity. I have now dedicated an AP for a dedicated WiFi network I can move with the Vibe meter. I felt that hardwired Ethernet could potentially restrict motion.

I have a USB ammeter and the Pi 3b consumes 0.3A on average and 0.6A max. Any 1A 5V supply should be plenty.

I was getting failed measurements with "Measurement overflow detected". Simply reducing sampling rate from 500Hz to 400Hz has eliminated these.

The noise floor on my setup is <40dB per axis (stable, level surface w/no vibrations of significance).
Marc Alexander is online now  
post #179 of 321 Old 06-15-2018, 08:06 PM
AVS Forum Addicted Member
 
Marc Alexander's Avatar
 
Join Date: Aug 2001
Location: Inland Empire, CA
Posts: 13,518
Mentioned: 240 Post(s)
Tagged: 0 Thread(s)
Quoted: 5472 Post(s)
Liked: 4972
I do have some mysterious behavior.

With the default settings, I get measurement overflow for any measurements 1s or over.

[{"name": "mpu6050", "gyroSens": 500, "accelerometerSens": 4, "fs": 500, "accelerometerEnabled": true, "status": "INITIALISED", "failureCode": "Measurement overflow detected", "samplesPerBatch": 125, "gyroEnabled": true}]

2018-06-15 19:47:39,878 - analyser.measurement - WARNING - put - Failing measurement 20180616_024736_null for mpu6050 because Measurement overflow detected
The errors go away if I reduce the fs (I've used 250 and 400 without issue).

However, when I do schedule measurement length under a second they fail unless fs: 500. I either get a yellow icon or stuck at "measuring" in the browser. There doesn't seem to be any measurements actually attempted. I don't see anything in the log. If I restart the recorder I don't even see the "metadata does not exists" message that normally shows for failed recordings that have yet to be deleted.

At fs: 500 these measurements always at 0.25s regardless of the duration scheduled.
Marc Alexander is online now  
post #180 of 321 Old 06-21-2018, 12:03 PM - Thread Starter
AVS Forum Special Member
 
3ll3d00d's Avatar
 
Join Date: Sep 2007
Location: London, UK
Posts: 4,149
Mentioned: 267 Post(s)
Tagged: 0 Thread(s)
Quoted: 2482 Post(s)
Liked: 2701
short reply as avs forum software ate my previous one for no obvious reason

re overflow on a longer run, it means you must be gradually filling up the FIFO on the chip, i.e. not reading fast enough. Solution = produce less data which means reduce sample rate or turn the gyro off. If you turn that off it means the FIFO can fit in 2x the samples so this should be sufficient to sort it out.

re errors on subsecond duration measurements, don't recall ever trying such measurements but I assume it is a bug. If you collect debug logs from both apps then I can take a look at what is going on. What is the use case for such measurements though? it seems very short for vibration data.
3ll3d00d is offline  
Sponsored Links
Advertisement
 
Reply DIY Speakers and Subs

Thread Tools
Show Printable Version Show Printable Version
Email this Page Email this Page


Forum Jump: 

Posting Rules  
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off