[Initial Release] RaspberryPints - Digital Taplist Solution

[skemp45]

Glad to see so much excitement!

 

[tssgery]

Excellent. I have a pi just waiting and I can replace my existing kegerface clone/mods

 

[de3isit]

Will the Raspberry Pi Model A work for this project?

 

[Adeering]

So when I try it on my samsung 10.1 it loads the page, but doesnt show any of your taps and just have the "Nothing on Tap" for each one

 

[thadius856]

Model A *may* work for a basic taplist, but I may not be able to support future advanced features, as the Model A has a different GPIO pin layout.

This prevents the preferred breakout board we have from direct mounting on top of the Pi. Not only does it provide 5v -> 3v3 conditioning to prevent frying the Pi, but it also provides I2C polling support for flow meters, solenoids, temperature probes, RFID card readers, etc. that we plan on implementing.

If you're buying new, go with the Model B. It's only a couple bucks more.

 

[thadius856]

So when I try it on my samsung 10.1 it loads the page, but doesnt show any of your taps and just have the "Nothing on Tap" for each one

This is intentional. That message is there to tell you that you haven't completed setting up your taps/kegs.

Edit: Oh, on my live site? I'm noodling around. Gimme a few minutes, I'll fix it up for ya.

 

[thadius856]

There ya go. Site is up and live again. Noodle around as much as you like.

Admin panel can be accessed by clicking the "Your Logo Here" image, then using the l/p admin/admin.

 

[EmDubeu]

There ya go. Site is up and live again. Noodle around as much as you like.

Admin panel can be accessed by clicking the "Your Logo Here" image, then using the l/p admin/admin.

OMG this is AWESOME!

 

[thadius856]

Annnnnd... a build list is now active on the [post=5866692]2nd post[/post] of this thread, for anybody who wants to get a head start.

Please feel free to ask questions, should you have any.

 

[FoxPaw]

Thadius, job well done by all in such a short time, will be looking to build after keezer is set up and running. Brew on.

 

[KindofBlues71]

I'm just getting into kegging and have been excited for the delivery of my first two kegs tomorrow. Then this morning I see this thread and now my focus is wavering from kegging to displaying my tap list digitally. DAMMIT! Time to start dropping hints to SWMBO...

 

[Josh1275]

Thadius and team, looking good so far! Just bought my package of stuff from amazon (using your links on the prep page to help support the project). Keep up the good work, and I can't wait to see this thing go live!


Sent from my iPad using Home Brew

 

[Adeering]

Just went back and tried it on my tablet and it looks awesome. The admin area is awesome. Now Im really excited

 

[Lost_Arkitekt]

So, I've been checking out the thadius856 website every so often, but I haven't noticed the keg volume column in quite some time. Is that not part of the first release? Just curious over here...

 

[mcangeli]

It's hidden. The flow meters and code to update them are planned in the v2 release. They decided to hide the keg for the time being.

Sent from my SAMSUNG-SGH-I337 using Home Brew mobile app

 

[mattd2]

It's hidden. The flow meters and code to update them are planned in the v2 release. They decided to hide the keg for the time being.

Sent from my SAMSUNG-SGH-I337 using Home Brew mobile app

This is intense! Was there any request to implement weightscales instead of flowmeters?
Also with the GNU licence is it available for comercial use at all? Seems like this could be a good option for many small taprooms!
Good work!

 

[mcangeli]

There was discussion about weight pads in the original thread.

Sent from my SAMSUNG-SGH-I337 using Home Brew mobile app

 

[Face Eater]

Sweet! Gotta build this.

 

[thadius856]

It's hidden. The flow meters and code to update them are planned in the v2 release. They decided to hide the keg for the time being.

Sent from my SAMSUNG-SGH-I337 using Home Brew mobile app

Indeed. The code's still there, and you can display them using the admin panel. But you'll have to manually update keg levels in the SQL database if you want to use them. If you don't know how to do that, you probably don't want to turn the feature on.

 

[thadius856]

Updated build list to remove affiliate links.

HBT requires a Sponsor membership to post affiliate links (or link to a website with affiliate links). With only taking in 4% of the purchase price as a referral fee (of which we currently have a 0% conversion ratio), I doubt we'd ever break even on the cost of the membership (whatever that is).

We'll continue to host the site, pay for the domain, and buy test hardware out-of-pocket. If you meet any of us in person, feel free to buy us a beer.

 

[kerber]

I'd encourage everyone to hold off on buying the Slice of PI/O right now. We're still very much in the process of hashing out hardware specifics. With the RPi being a non real-time OS, I'm hesitant to dump a bunch of functionality on it and still trust it to read flow sensor pulses at the speed we need.

The SF800's are absolutely going to work, it's just a matter of how we interface them. We've got things in the works and affordability is still very much at the forefront of our decision making.

 

[rossi46]

Updated build list to remove affiliate links.

HBT requires a Sponsor membership to post affiliate links (or link to a website with affiliate links). With only taking in 4% of the purchase price as a referral fee (of which we currently have a 0% conversion ratio), I doubt we'd ever break even on the cost of the membership (whatever that is).

We'll continue to host the site, pay for the domain, and buy test hardware out-of-pocket. If you meet any of us in person, feel free to buy us a beer.

Add a donate button on the raspberry pints website.

 

[rgauthier20420]

Add a donate button on the raspberry pints website.

And 1 of those graphics that shows what's due for the year and how far along you are. I, along with many many others, will gladly support this project!

 

[CadiBrewer]

and 1 of those graphics that shows what's due for the year and how far along you are. I, along with many many others, will gladly support this project!

+1

 

[bad coffee]

I'd donate

 

[thadius856]

Mods, any rule against have a donate button somewhere on our website and a link to the front page of our website here?

We intend for the software and installation instructions to stay free. We don't plan on selling anything physical either.

 

[Lost_Arkitekt]

The SF800's are absolutely going to work, it's just a matter of how we interface them. We've got things in the works and affordability is still very much at the forefront of our decision making.

I'd hope so since probably 10 to 20 of us bought 4 to 10 of the things. :fro:

 

[EmDubeu]

tick-tock, tick-tock, tick-tock.....

 

[AgentHubcap]

I'm back in the game. I got sick for a bit and am finally feeling better.



Sent from my Nexus 5 using Home Brew mobile app

 

[thadius856]

For anybody that bought flow meters, good news and... potentially mixed news, depending on your outlook.

First, the good news. Initial flow meter tests have been promising. We're seeing 99% accuracy through 500 pulses/sec, and 97% accuracy through 600 pulses/sec (approximately one pint poured in 5 seconds).

The potentially mixed news is that we're having to veer away from the Slice of PI/O, so hopefully not many people have bought them. There's some good reasons why...

While the Slice of PI/O excels at what it's meant to do (provide additional, conditioned I/O) , once we pass the 600 pulses/sec mark, the Pi can't keep up with the polling speed and accuracy drops drastically. The effect gets worse when the Pi is under any meaninful load. We need dedicated hardware designed to poll these pins.

To that end, we're locking our development path (for flowmeter users) toward Arduino for the polling. Because we want assembly to be absolutely elementary, we're gearing up to use the Alamode (an Arduino with swapped layout to fit right on top of a Pi) and a Centipede (a Slice of PI/O on steroids). This expands and conditions your I/O to be able to hook up a retarded amount of sensors (64 in fact, unless you stack a second Centipede on top!) with dedicated hardware polling.

The cost difference is about $45 more, should you decide to use flow meters. It will improve accuracy and prevent the Pi from spamming false data when under load. It will remove most of the soldering you'd have faced previously -- if we can figure out how, it may remove all need to solder.

That said, we're dedicated to providing a digital tap list for under $100. To that end, none of these parts are required (and never will be!) unless you want real-time keg level monitoring.

 

[CadiBrewer]

For anybody that bought flow meters, good news and... potentially mixed news, depending on your outlook.

First, the good news. Initial flow meter tests have been promising. We're seeing 99% accuracy through 500 pulses/sec, and 97% accuracy through 600 pulses/sec (approximately one pint poured in 5 seconds).

The potentially mixed news is that we're having to veer away from the Slice of PI/O, so hopefully not many people have bought them. There's some good reasons why...

While the Slice of PI/O excels at what it's meant to do (provide additional, conditioned I/O) , once we pass the 600 pulses/sec mark, the Pi can't keep up with the polling speed and accuracy drops drastically. The effect gets worse when the Pi is under any meaninful load. We need dedicated hardware designed to poll these pins.

To that end, we're locking our development path (for flowmeter users) toward Arduino for the polling. Because we want assembly to be absolutely elementary, we're gearing up to use the Alamode (an Arduino with swapped layout to fit right on top of a Pi) and a Centipede (a Slice of PI/O on steroids). This expands and conditions your I/O to be able to hook up a retarded amount of sensors (64 in fact, unless you stack a second Centipede on top!) with dedicated hardware polling.

The cost difference is about $45 more, should you decide to use flow meters. It will improve accuracy and prevent the Pi from spamming false data when under load. It will remove most of the soldering you'd have faced previously -- if we can figure out how, it may remove all need to solder.

That said, we're dedicated to providing a digital tap list for under $100. To that end, none of these parts are required (and never will be!) unless you want real-time keg level monitoring.

To a guy who is just happy that you guys are making the efforts to move this forward, none of this is bad news. I bought the meters but not the Slice of PI/O. I'm on board with whatever solution you feel works best, even at a few bucks more. Thanks again for your efforts.

 

[mattd2]

What was the reasoning behind going for the SF800 flow meters? over the cheaper hall effect type? Since both are rotor based types would both need some sort of calibration and accurancy would vary somewhat depending on actual flow volocity? With the cheaper hall effect the frequency would be much much lower and potentially avoiding the IO issues.
I'm not meaning to sound critical, just trying to understand the reasoning behind going with the SF800s
Also I see the SF800 medium type is clear/translucent liquid that does not block IR - Has dark beers (porters/stouts) or Cloudy beers been considered and trialed with the sensor?
Cheers

 

[kerber]

For anybody that bought flow meters, good news and... potentially mixed news, depending on your outlook.

First, the good news. Initial flow meter tests have been promising. We're seeing 99% accuracy through 500 pulses/sec, and 97% accuracy through 600 pulses/sec (approximately one pint poured in 5 seconds).

The potentially mixed news is that we're having to veer away from the Slice of PI/O, so hopefully not many people have bought them. There's some good reasons why...

While the Slice of PI/O excels at what it's meant to do (provide additional, conditioned I/O) , once we pass the 600 pulses/sec mark, the Pi can't keep up with the polling speed and accuracy drops drastically. The effect gets worse when the Pi is under any meaninful load. We need dedicated hardware designed to poll these pins.

To that end, we're locking our development path (for flowmeter users) toward Arduino for the polling. Because we want assembly to be absolutely elementary, we're gearing up to use the Alamode (an Arduino with swapped layout to fit right on top of a Pi) and a Centipede (a Slice of PI/O on steroids). This expands and conditions your I/O to be able to hook up a retarded amount of sensors (64 in fact, unless you stack a second Centipede on top!) with dedicated hardware polling.

The cost difference is about $45 more, should you decide to use flow meters. It will improve accuracy and prevent the Pi from spamming false data when under load. It will remove most of the soldering you'd have faced previously -- if we can figure out how, it may remove all need to solder.

That said, we're dedicated to providing a digital tap list for under $100. To that end, none of these parts are required (and never will be!) unless you want real-time keg level monitoring.

As anxious as people might be to go out and buy those boards, it might not hurt to hold off until the feature is closer. While this is the current target hardware, there is still a lot of testing to be done. For now, the only 100% known hardware is the RPi.

What was the reasoning behind going for the SF800 flow meters? over the cheaper hall effect type? Since both are rotor based types would both need some sort of calibration and accurancy would vary somewhat depending on actual flow volocity? With the cheaper hall effect the frequency would be much much lower and potentially avoiding the IO issues.
I'm not meaning to sound critical, just trying to understand the reasoning behind going with the SF800s

While a lot of the cheaper flow meters are probably food safe, these are known to be. They are also really accurate sensors and we got them for a good deal

As long as the flow sensor uses 5v and has a pulsed data line, it should be interchangeable. The only real difference is the number of pulses per volume of liquid, but that can be handled in calibration.

 

[seeGarzz]

I have seen a couple references to the SF800 flowmeters but I have not seen a source for purchasing, except for the manufacturer site. Is there another way to get these flow meters? At a discount from the manufacturer price? Thanks.

 

[rgauthier20420]

I believe eBay was where people were buying them and making an offer. Try here.

Maybe someone can confirm that's accurate?

 

[Lost_Arkitekt]

The potentially mixed news is that we're having to veer away from the Slice of PI/O, so hopefully not many people have bought them. There's some good reasons why...

The cost difference is about $45 more, should you decide to use flow meters. It will improve accuracy and prevent the Pi from spamming false data when under load. It will remove most of the soldering you'd have faced previously -- if we can figure out how, it may remove all need to solder.

That said, we're dedicated to providing a digital tap list for under $100. To that end, none of these parts are required (and never will be!) unless you want real-time keg level monitoring.

Not sure what you mean by $45 more. Does this mean RPi, Keyboard, Dongle, Cable, Wall wart, SD Card....is necessary. (Not really $45...more like $60). That is around $60. Then, the Slice of Pi I/O was about $30 plus shipping. Now, is that $45 more on top of what that would be? That is 50% more. I'm no math genius, but I'm just sayin'. (Actually, I used to be a Math major! Haha.)

I'm just trying to get my ducks in a row. For some of us, $45 here and there is no big deal. For others on a tight budget, an extra $45 pays for a year of HBT membership and a batch of homebrew, which is why we are all here to begin with.

Also, I'm not a tech guy, but if 500 pulses per second is 99% accurate, why couldn't you just set the interface to have the flow meter to run at that frequency?

Again, I'm not trying to upset the apple cart, but I'm trying to understand the switch from Slice of Pi (relatively inexpensive) to Arduino and Centipede combo...which is essence, is 3x the Slice of Pi option. Maybe I'm misreading this and the Arduino/Centipede combo is $45 total?

 

[Marquez]

Killer project for beer tech voyeurism.

Conceived and birthed while the world watched.

I wish the projects at my work went this smoothly...

 

[kerber]

Ok. This same confusion has come up in our dev communication, so I want to address it. The only required part of this project is a web server. The target platform being used is the Raspberry Pi. If you have an old laptop around, that would work perfectly fine. That is well under $100 for the cables, dongle, keyboard, mouse, etc.

All other discussion of hardware (Alamode, Centipede Shield, Flow sensors) are for the completely optional feature of keg level monitoring. This is purely an add-on and in no way required to run the digital tap list. There has even been a lot of discussion on having a button on the interface that you can use to tell the tap list that a beer was pulled.

Please please please don't get caught up in the cost of this single feature. It is the cheapest way we can do get monitoring, but by no means necessary. It won't even be part of the first release.


Main Project Hardware:

• Raspberry Pi ~ $35
• Cables ~$10
• SD Card ~ $4
• Phone Charger ~ $5
• Wireless Dongle (optional) ~ $11

Optional Hardware for Optional Keg Level Monitoring(PLANNED - NOT SET IN STONE):

• Alamode Arduino-Compatable Board ~ $35
• Centipede Shield I/O Board ~ $25
• Flow Sensors ~ $10-$30

 

[rgauthier20420]

Ok. This same confusion has come up in our dev communication, so I want to address it. The only required part of this project is a web server. The target platform being used is the Raspberry Pi. If you have an old laptop around, that would work perfectly fine. That is well under $100 for the cables, dongle, keyboard, mouse, etc.]

So wait, I DO have a spare laptop lying around and the display has burnt out. So that's perfect for use with this because I'll need a new display for the tap list anyway.

Will there be a guide to get this set up on a Windows device then also?

 

[kerber]

... if 500 pulses per second is 99% accurate, why couldn't you just set the interface to have the flow meter to run at that frequency?

Here's the problem. The faster your system pours a beer, the faster the flow meter is going to pulse. It just does some number of pulses per liter. We were originally trying to read the pulses from the Raspberry Pi and an I/O expander chip. The problem is that the RPi can only read so many pulses/second before it trips over itself. It was fine reading up to 600 pulses per second, but anything over that and it would miss up to 50% of the pulses.

To make the problem doubly worse, the RPi is a non-real time operating system. It may run 600 pulses/second fine now, but as soon as it's under load with a webserver and mysql, that could drop dramatically. Even worse, it could spend so much time on other tasks that it completely misses large chunks of pulses. I tweaked a bunch of thing and set the process to a high priority, but it still didn't provide any kind of reliability for a fast pouring system.

What I want to avoid is everyone going out and buying hardware for a feature we've only done initial testing on. Obviously we can make the flow sensors work, but the hardware to interface them with the tap list is still in limbo.

Again, I'm not trying to upset the apple cart, but I'm trying to understand the switch from Slice of Pi (relatively inexpensive) to Arduino and Centipede combo...which is essence, is 3x the Slice of Pi option. Maybe I'm misreading this and the Arduino/Centipede combo is $45 total?

The Slice if Pi/O was simple a guesstimate on Thadius's part. Once we got into testing, we realized we would need something more reliable and real-time to read the pulses. This hardware change has no effect on the main part of the project, only the keg monitoring. If someone is already spending the money for 5-10 flow sensors and the Pi, an additional $45 to make it work reliably should not be the end of the world. If that is too much, then don't do keg monitoring or contribute code to make it work without the Arduino.