iSpindel - digital WiFi hydrometer

[lupinehorror]

giving brewspy a go with the brew i started today (SJW all apple).
initial impressions are good.

I cannot find a means to refresh the displayed data, other than shutting Brewspy down and restarting it

pulling down on the screen doesn't refresh it? it does for me if the set interval has passed.

 

[UKSkydiver]

What is your formula for tilt to sg? It should be a second-order polynomial.

I think I'm struggling a bit here:
The values that come out of the ispindel calibration page are

Degree 1: 0.9565120241162155 + 0.0017848221633354573 *tilt
Degree 2: 0.9607417230501502 + 0.0015844445141719621 *tilt + 0.000002139335687479827 *tilt*tilt
Degree 3: 0.9398113280454689 + 0.003115574971749137 *tilt-0.00003267912058701393 *tilt*tilt + 2.4883863132850617e-7 *tilt*tilt*tilt

The video I referenced earlier says to remove all the spaces and change 'x' to *tilt, however, the text out of the calibration page does not have 'x' and already has *tilt included (albeit multiples at times) Is it possible they have modified the code to output the *tilt instead of 'x' ?

So given some of the info here, I have input the following into the the ispindel, having removed the spaces and following the 'power of' notation:

0.9398113280454689+0.003115574971749137*tilt-0.00003267912058701393*tilt^2+2.4883863132850617e-7*tilt^3

That doesn't work and when I go back into the ispindel, it seems to have changed the above to (note the loss of a 't' and the addition of three zeros to the very end):

0.9398113280454689+0.003115574971749137*tilt-0.00003267912058701393*tilt^2+2.4883863132850617e-7*til000

 

[Harry Bloomfield]

pulling down on the screen doesn't refresh it? it does for me if the set interval has passed.

I'm using an old Nook HD+, which I rooted to install Android, to enable me to run Brewspy. It is rather slow to respond, making it difficult to find my way around Brewspy, but I will try your suggestion. I'm aware the graphs are not updated instantly, they sometimes seem to lag behind by up to an hour, probably processing delays at the server?

[EDIT] Yes it does refresh pulling the screen down, so thanks for that great hint..

 

[lupinehorror]

I'm using an old Nook HD+, which I rooted to install Android, to enable me to run Brewspy. It is rather slow to respond, making it difficult to find my way around Brewspy, but I will try your suggestion. I'm aware the graphs are not updated instantly, they sometimes seem to lag behind by up to an hour, probably processing delays at the server?

nice use of an older device!
unsure about server delays. that's one thing i'm not too keen on with brewspy. the lack of transparency in that department. there's not even a contact link on the site (though maybe that can be done via google play) and the 2019 © BrewSpy doesn't really fill me with confidence.

 

[Jim2607]

I too am having wifi trouble with my iSpindel. It will send to Brewspy if its on 10 second transmit intervals, but when I revert to the 300 or 900 second intervals I dont get any results at all. Have also tried to configure with Brewfather without success. Im wondering if there is a dry solder joint in the wifi transmitter somewhere, causing intermittent transmission. Have spent a lot of time trying to make it work but not having much success so far.

 

[UKSkydiver]

Whoa - just found this from the Facebook group:

I'm about 113 characters - investigating. . . .

 

[kelper]

That doesn't work and when I go back into the ispindel, it seems to have changed the above to (note the loss of a 't' and the addition of three zeros to the very end):

Here is my iSpindel 'configuration' page (the token is mine and you must use your own)

The formula I used is

2.4950708612059392E-6*tilt^3 -3.597130446990714E-4*tilt^2 +0.018498247948015193*tilt +0.694611987372000

try copying this as a trial

 

[UKSkydiver]

I'm about 113 characters - investigating. . . .

Well - that's a bit annoying. It does seem that if you use the Degree 3 value and it's over 100 103? characters, it craps out as I've described. I used the (shorter) Degree 2 value with spaces removed and ^2 / ^3 notation where necessary - and it seems to work.

Seems a bit of an oversight. Why would you go to the trouble of going through the calibration process only to end up with a value that you can't use, only then to have to use a seemingly less precise one?

I'm surprised this information isn't in the documentation somewhere.

Further, is it just me, but it's a bit of a faff having to input the formula and save, which then exits and connects to your home wifi, and then you have to reset it and connect back to the ispindel to check the "ispindel Info" page to see if it's working correctly.

Anyway, seems to be working now.

 

[kelper]

Did you get a graph showing how well the formula fits your measurements? My second third order polynomial is a good 'fit' But I used Google Sheets.

 

[UKSkydiver]

Did you get a graph showing how well the formula fits your measurements? My second third order polynomial is a good 'fit' But I used Google Sheets.

I did. Not sure what I was looking at, but it seems pretty straight and close to the line throughout

EDIT: Also, your formula is 100 characters with the spaces removed, and it works. I added a trailing zero and it also worked, so it may be that it craps out somewhere between 101 and 113 characters.


EDIT: I'm going to guess it's 103 characters:

 

[kelper]

I think your calibration looks odd. I would not expect the relationship between tilt and gravity to be linear!

 

[UKSkydiver]

I think your calibration looks odd. I would not expect the relationship between tilt and gravity to be linear!

Not sure what to say - the values in the Tilt and SG columns are what I recorded when calibrating and the graph is what the calibration web page squirted out.

 

[kelper]

Can you list the values you used so I can cut and paste into my own spreadsheet. I don't want to copy one by one from your screenshot!

When you made each successive sugar solution, did you measure with a hydrometer or were the solutions just calculated from the weights?

 

[UKSkydiver]

Can you list the values you used so I can cut and paste into my own spreadsheet. I don't want to copy one by one from your screenshot!

I don't blame you

This post has them as text

When you made each successive sugar solution, did you measure with a hydrometer or were the solutions just calculated from the weights?

I did it the other way around. I started with 1080 and diluted down using a dilution calc and checking with a hydrometer. Clearly It won't be mega-accurate, but it's what I have and I'd be using the hydrometer to measure anyway. The temps were there or thereabouts 20c.

 

[Harry Bloomfield]

I too am having wifi trouble with my iSpindel. It will send to Brewspy if its on 10 second transmit intervals, but when I revert to the 300 or 900 second intervals I dont get any results at all. Have also tried to configure with Brewfather without success. Im wondering if there is a dry solder joint in the wifi transmitter somewhere, causing intermittent transmission. Have spent a lot of time trying to make it work but not having much success so far.

Similar happened to me and right or wrong I developed a theory.....

At the time I had two routers/ wifi access points, the main one connected to my FTTC broadband and a second one wired via LAN to the first. Lots of items on my network, some with static IP's most relying on DHCP allocated by the main router, so their IP's would vary. I had a certain range of IP's reserved in my router for static addresses, the rest for dynamic items, which connected occasionally.

My theory is that if the router doesn't see a transmission from the iSpindel for a long time, it might allocate the IP it gave it to another device. The iSpindel will then continue to try to send its reports as though still on its original IP and fail. A shorter transmission interval improves its chances of retaining the same IP to the end of the brew. The iSpindel seems to want to stick to using the same IP, even after it is power cycled, the way to force it to request a usuable IP from the router, is to go into the iSpindel's configuration, make a slight change (300 to 301 second updates) and save it.

I have suffered this three times now, but it is only a theory.

I have now revised my system, so almost everything on my LAN uses a static IP, allocated by the router, including the ispindel. This brew is my first brew since doing that - So far, all is working well. I have it set to report every 10 minutes and no reports have been missed in the four days since I started my present brew.

 

[kelper]

This post has them as text

I put your numbers into my spreadsheet. The best fit trendline is linear, with formula 1.78E-03*x + 0.957

just change the 'x' to 'tilt'

https://docs.google.com/spreadsheets/d/1uulbXMJXaGDUKZhv4KgyUTBIMTsNt-TQkcIUFC9lHAA/edit?usp=sharing
Make your own copy of this spreadsheet as anyone with the link can edit.

be patient, it takes a minute to load fully.

Right-click in the chart and select 'series' and then 'SG' scroll down to the trendline settings.

I set it to linear, but you can set it to polynomial and try changing 'polynomial order'

 

[lupinehorror]

Similar happened to me and right or wrong I developed a theory.....

At the time I had two routers/ wifi access points, the main one connected to my FTTC broadband and a second one wired via LAN to the first. Lots of items on my network, some with static IP's most relying on DHCP allocated by the main router, so their IP's would vary. I had a certain range of IP's reserved in my router for static addresses, the rest for dynamic items, which connected occasionally.

My theory is that if the router doesn't see a transmission from the iSpindel for a long time, it might allocate the IP it gave it to another device. The iSpindel will then continue to try to send its reports as though still on its original IP and fail. A shorter transmission interval improves its chances of retaining the same IP to the end of the brew. The iSpindel seems to want to stick to using the same IP, even after it is power cycled, the way to force it to request a usuable IP from the router, is to go into the iSpindel's configuration, make a slight change (300 to 301 second updates) and save it.

I have suffered this three times now, but it is only a theory.

I have now revised my system, so almost everything on my LAN uses a static IP, allocated by the router, including the ispindel. This brew is my first brew since doing that - So far, all is working well. I have it set to report every 10 minutes and no reports have been missed in the four days since I started my present brew.

interesting. is it showing up on your LAN as ESP-73E6B8?

 

[kelper]

I, too, though static IPs might stop all my gadgets losing their internet connection but I don't use them today as it did not eliminate the problem. I guess the DHCP server in the router is clever enough to avoid the problem unless the firmware is buggy. My phone (Android) never has a problem, but our laptops (Windows) often fail to connect. When rebooting the laptop does not work I reboot the router and range extender and that does always fix it. I think it is a Microsoft issue.

 

[lupinehorror]

i use static IPs for NAS and the many mini servers around the place but the rest i haven't set. seems like my DHCP is pretty solid though and the devices that use it seem to be served the same IP pretty consistently.

 

[Harry Bloomfield]

interesting. is it showing up on your LAN as ESP-73E6B8?

It did, once appear, but generally it doesn't appear at all - problem is it connects and disconnects very quickly. Too quickly to even be logged by my router. What I did was check its MAC in the configuration page, then add that MAC with an allocated IP.