Brew Bubbles: Web-Enabled Airlock

Homebrew Talk - Beer, Wine, Mead, & Cider Brewing Discussion Forum

Help Support Homebrew Talk - Beer, Wine, Mead, & Cider Brewing Discussion Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
Just a little update: I am switching name registrars, DNS and web hosts so I am SURE I will screw that up somewhere. :p

If you notice any issues getting to any of the hosted or DNS options, post here.
 

Because: $8.99+ No Import Fees Deposit & $8.57 Shipping to Canada
Arrives: Jan 21 - Feb 10

That would make the food colouring the most expensive part :) . I've been doing a little research about what liquids might be opaque to IR and can't find a lot. Food colouring isn't mentioned much. So far water seems to be mentioned the most ... but it depends on what wavelength your sensor is sensitive to. The recommended sensor measures at 900nm. Here is what I could find on food colouring:

1610416061522.png
 
I can tell you this much about IR: my Swissflow SF800 flow meters have zero problems metering my 60+ SRM chocolate stout (or my very red raspberry hibiscus wheat). Impressive AF to me - before I ran my first keg of stout through one I had my doubts whether I should have gone with Hall Effect meters and sacrificed the accuracy, but they really work just as accurately as with my clearest kolsch...

Cheers!
 
Maybe I should try beer in the airlock. From the data I can find, it does look better than food colouring. I did find food colouring in town yesterday, but there was a choice of two colours so I went home to research what colour would be best, only to find that, in theory, they all should be useless. I found a spectrum that goes into the IR and all food dyes should be transparent to the sensor. .

You can't tell by looking what might be opaque to IR. I saw a photo of red wine that was transparent in the near infrared. I tried a couple things I had in the kitchen yesterday including soy sauce and nothing seemed to make a difference. Unfortunately I also knocked the sensor out of alignment while adding stuff so it is hard to tell since aligning the sensor seems to be the biggest factor.

Update - OK, looking a little closer at the specs for the GP1A57HRJ00F sensor (can't double check - access to the brew bubbles documentation is currently "403 forbidden" for me - maybe related to the change Lee mentioned above?) the emitter is 950nm and the peak sensitivity of the sensor is 900 nm, however, the sensitivity is pretty broad, 400-1200nm, so blue(?) food colouring might help.
 
Last edited:
@James_42 if you look back over some early parts of this thread, a few of us did some testing with food colourings but were unsuccessful in producing consistent results. Personally, I also found that some colouring stained my fermentation lock!

What we did find was that restricting the view the receiver has, by adding a pin hole collimator, was the most consistent way (so far?) of bubble counting. I've been using both a piece of card with a pin hole stuck with double sided tape and also heat shrink, with great and consistent effect. Yes, you have to get the allignment correct, but once the pin hole is fixed in place, in my experience it hasn't moved.


Chris
 
A mix of red and blue would block most of the light - but as Chris says - don't do it to your favorite airlock. :) Water actually blocks IR very well, the coloring helps to block ambient. That said I also used gels and I've yet to do anything that works as well as the cardboard (or heat shrink) and pinhole.

I wish I had a box of the sensors I used when I developed this - they all work just fine with ambient light and no coloring.
 
Thanks, yes, I have a piece of shrink tube with a pinhole in it which has helped a lot. I think my issue is getting the pinhole lined up with the airlock. I don't have a bracket, just elastic bands in an arrangement that allows some adjustment, but also stays in place. With all the fiddling I have done I haven't been able to find an arrangement that gives less than about 3X the actual BPM.

I'm trying starsan now and it isn't any better than water - worse because it is bubbling a little like soap.

Lee - I think it is working fine for me. Once I get some way to fix it in place the relative changes should tell me all I need to know. I did find the specs on the sensor I have and it reads a little higher than the one you reccomend 940nm instead of 900nm so maybe that is the difference. I do have the recommended parts coming in the mail so hopefully those will work better. I thought maybe the information might be usedful to you and others but maybe I should stop posting about this.
 
Last edited:
And all that said - I was having a conversation with another unnamed co-conspirator (right @UncleD ?) about how the platform really does not need to be tied to the board and sensor I have here. I wanted to release a functional proof of concept and I was hoping someone would figure out a way to maybe meter the bubbles evenly (and emulate SG like Pl**to), or use a flow meter like @day_trippr is talking about. Anything could work, Brew Bubbles just counts pulses. I did not use a more expensive flowmeter because this was $4 and those are ... well, more. :)

James are you in Canada? Maybe there's someone semi-local who can print a bracket for you? Not sure what shipping is to Canada from here.
 
And all that said - I was having a conversation with another unnamed co-conspirator (right @UncleD ?) about how the platform really does not need to be tied to the board and sensor I have here. I wanted to release a functional proof of concept and I was hoping someone would figure out a way to maybe meter the bubbles evenly (and emulate SG like Pl**to), or use a flow meter like @day_trippr is talking about. Anything could work, Brew Bubbles just counts pulses. I did not use a more expensive flowmeter because this was $4 and those are ... well, more. :)

James are you in Canada? Maybe there's someone semi-local who can print a bracket for you? Not sure what shipping is to Canada from here.

That's what someone needs to build, is something that goes between the fermenter bung and the traditional airlock that has the brew bubbles sensor on it. This "something" could be 3D printed and "regulates" the size of each bubble, so the counts can be used to calculate the volume of CO2 expelled...
 
...With all the fiddling I have done I haven't been able to find an arrangement that gives less than about 3X the actual BPM...
For me BrewBubbles is an "indicator" as how the fermentation is progressing rather than a "meter" for how the fermentation is progressing. For example, I have a WineXperts 23l Chilean Malbec fermenting away in a plastic fermentation bucket and plastic lid. You can see the lid move as the pressure is relieved! The bubbles come along in bursts and sometimes no (or very few) bubbles pass the sensor when it's really active and sometimes I get multiple edges of water as the gass moves pass.

In a 1 gallon glass demijohn I find the counting more consistent.

For me it's just something else to confirm that the fermention is, well, fermenting. Like the oil pressure light on a car for me, I don't drive watching the light just in case?


Chris
 
James are you in Canada? Maybe there's someone semi-local who can print a bracket for you? Not sure what shipping is to Canada from here.

Yes, I'm in Canada. I have a 3d printer, but I'm using a sensor module like the one below and I don't think it will fit in the standard Brewbubbles bracket. I will print one anyway since I have a standard PCB and sensor on the way now I hope 🤞

1610481589085.png

Edit: Everybody is different. For me making the gadgets and looking at the data is as much fun as making and drinking the beer :)
 
Last edited:
So yesterday I finally received my PCB and parts from jangevaa (thanks!) for the stock Brewbubbles arrangement. I soldered it all up and printed out the bracket with the collimator and added more sugar to my test fermenter. It seemed to work perfectly right away, with its counts matching mine. In the middle of the night, however, the counts suddenly dropped from about 12 to about 6. Investigating this morning the actual BPM is around 10, but Brewbubbles shows about 7 bpm. I watched closely it is missing bubbles now.

I'm not too concerned that the exact number isn't right, it's that it isn't consistent. Any ideas what might be going on? The second drop at 8am was when I tried adjusting it and I forgot to check the uptime before that. There was a cover over it so it would have been dark all night.

1610733896005.png
 
A dropoff like that would make me wonder if something moved slightly.

I did not find the bracket with the collimator worked any better, as a matter of fact, it was harder to work with than the pinhole.
 
A dropoff like that would make me wonder if something moved slightly.

I did not find the bracket with the collimator worked any better, as a matter of fact, it was harder to work with than the pinhole.
I was asleep, alone in the house, and the boards are wedged pretty solidly in the bracket. Also moving it on purpose this morning didn't make much difference so I didn't think it was movement. I was kind of hoping it would be less sensitive to movement in the bracket, but I will try moving it some more.

I tried to use my original heat-shrink collimator first, but it didn't fit and I didn't think shrinking it on the plastic part was a good idea so I printed out a new bracket with the collimator, which initially seemed to work perfectly.

I've been delaying brewing a batch of beer until I have this working reliably - I think I'm close 🤞
 
Looks about the same and I used water this time so it wouldn't froth. I was already thinking of putting it on some kind of blowoff jar arrangement so maybe I should work on that. My thinking was that once set up you wouldn't have to touch it, even for dry hopping.

After fiddling, it looks like I have it back to one count per bubble. Have any other methods worked for keeping the bracket from moving other than the elastic band? Stronger elastic?
 
I use a set screw for testing. I have to be honest though: I can't fit an airlock on my fermenters in my fermentation fridge so I use a blow-off tube most of the time.

I looked at using one of those CO2 harvester setups to allow me to sort of have a remote airlock. The larger tubes/bubbles created a scenario where I blew the liquid out of an airlock pretty quickly. I was going to play with using aquarium-sized tubing but never got back to it.
 
@James_42 I found a similar behaviour when I first used it in anger. I put it down to a missalignment, as after settling the PCB into a different position within the bracket I got much more consistent readings.

Whilst I haven't fixed the PCB in place I use some elasticated webbing to hold everything in place. Although I might have to adjust the relative positioning of the bracket, PCB and fermentation lock over the intial fermentation. Yes it is a bit finicky but it does work.


Chris
 
Has anyone had their just suddenly stop counting? Mine went to zero abruptly and fiddling with the position and changing to the bracket without the collimator as well as rebooting hasn't fixed it. There are still regular bubbles the sensor just stopped detecting them. Putting something like a pencil in front of the sensor is detected but bubbles aren't. Below the points on Jan 15-16 are bubbles and on Jan 17 it is just me trying to get it working again.

1610938679740.png
 

Attachments

  • 1610938646189.png
    1610938646189.png
    9.6 KB · Views: 12
@James_42 I'd say it's an alignment issue. The sensor has moved relative to the fermentation lock. How are you fixing the sensor/PCB relative to the fermentation lock?

I use a wide elastic webbing and it's quite tight. When fffirst connected it often takes a bit of adjustment to get the alignment right. Yes, it also sometimes moves and it needs to be adjusted.


Chris
 
I just had a regular rubber band and I thought I had tried all possible adjustments including going back to the non-collimator bracket, but going back to the collimator bracket and adding a thick elastic band seems to have it working again :oops:. Thanks!

edit: no, it stopped again without being touched and I couldn't find a way to adjust it to get it working again including rebooting it a few times
 
Last edited:
Are we still talking about the sensor that came mounted on the board? Or one of the Sharps that get soldered onto the shield I designed?
 
@James_42 Does it still detect something solid? If it doesn't then the probblem is with the sensor, if it does then it's an alignment issue.

Alignment can be troublesome. All of my fermentation locks are of the same design, but I've found that some are better/easier to align and remain aligned.

Chris
 
This is the full package Lee designed, custom PCB with sensor, resistors, and capacitors soldered on. It does still detect something solid so it does seem like it is an alignment issue, except, it has stopped counting twice now without being touched/disturbed. Not only that, but it seemed pretty insensitive to alignment while it was working and also insensitive to any attempts to realign it.

I think I have found the problem. The bubbles were actually sneaking by the sensor. Somehow a 'permanent' bubble had formed in the bottom U tube and the bubbles were forming at the bottom of the second chamber and floating to the surface as soon as they got big enough. So the sensor was just always seeing a bubble. I tried cleaning the airlock and have added sugar to my test fermenter so hopefully I will know if this worked as soon as it starts bubbling again.
 
Quick question about OTA. It should work from a 2.1.1 to 2.2.1 ?
I have the Firmware Update in Progress page, waits for more than 5 minutes but nothing happens ...
Any idea ?
1613134358199.png
 
This is an awesome project, I'm ordering stuff now. Random thought after reading your explainer article. Would load sensors work for this since the weight is decreasing?
 
Would load sensors work for this since the weight is decreasing?
Technically speaking, weight can be used for this. In a practical sense, load cells drift and this would be a challenging application for a home project. Even with a well-known commercial product intended to track keg contents (much larger delta), we have noted statistically significant drift.

Finally, no. ;) As written, this counts pulses. It would require a re-write to track something that was analog. If a good application requiring an analog input can be devised, I'm willing to have that conversation. It's just not possible the way the code is written right now.
 
Last edited:
2.2.2. Bugfix Release
(aka @ZeSlammy release)

This feature release contains bug fixes. Before this release and subsequent to a hosting provider change, users upgrading to 2.2.1 via the Update function in the web UI received a quiet failure. This has been fixed.
  • New users may install this version via BrewFlasher
  • Current users may use the Settings > Advanced > Update to update without losing settings
Bug Fixes
  • bedbc99 Add a nodrd flag to config to prevent popping portal on programmed reboots
  • e9a9d1f Move IP stack back to 1.4 High Bandwidth in order to get around an upstream bug
  • (Hosting Provider) Add .htaccess rules to avoid rewriting the firmware requests
 
The connection refused messages are normal during an update. That's because the page is constantly "pinging" the controller to see if it's alive again. Here are a vid and a log of a successful upgrade from 2.1.1 to 2.2.2. If you continue to have issues I'll need to see a serial log:

 

Attachments

  • upgradelog.txt
    7.1 KB · Views: 6
There's a bug upstream, as a result, I needed to downgrade the IP stack. Annoying and it took a while to figure out for sure. That + new hosting provider and I had an interesting day yesterday. :)
 
Super weird. It ended up working but I did not had the message saying it did.
Check your "about" page (and make sure you do a shift+refresh to get the latest version):
1613311055005.png

If it looks like the above and says v2.2.2, you only got half the update. If it says 2.2.2, of course, you got none of the updates.

A version 2.2.x "about" page will look like this:

1613311189510.png

If it's not correct, do the update again. A serial log will help you zero in on any update issues.
 
Back
Top