Kegbot Home » Kegbot Forums
KegBB
Kegbot Kegerator Project Forums
- You are not logged in.
#1 May 16, 2010 13:56:06
- josh.dgill
-

- From: Tucson
- Registered: 2010-01-01
- Posts: 11
- Profile Send e-mail
New Build!
Hey guys,
I'm kinda new here, I posted a few times a few months ago, but got distracted by my senior design project. Now that that all is over, I can get back to the more important things, the kegbot.
My project has a few goals that I see common between most systems here-
Here's some details:
Pics!
RFID access board
Main DAQ
Since I'm making the boards myself, some traces are creatively routed to avoid as many vias as possible. I'll be making the boards probably next week, and the components should arrive by then. I will post more pictures later!
Josh Gill
josh.dgill@gmail.com
BSME, University of Arizona
I'm kinda new here, I posted a few times a few months ago, but got distracted by my senior design project. Now that that all is over, I can get back to the more important things, the kegbot.
My project has a few goals that I see common between most systems here-
[*] Integrate the SF800 for volume monitoring, provided by SwissFlow
[*] Use a 1N4148 for temperature monitoring
[*] RFID access using Innovation's ID-12 (1.5" read range)
[*] Host keg's status on my website, krewarmy.com for mobile phone access, that way I can know when to buy beer if I'm at the store
Here's some details:
[*] Processor: Freescale MC9S08JM16 (I love Freescale, I'll never go back to another processor, I've used both AVR and uChip)
[*] USB compatibility for uploading status to webserver
[*] Temperature amplification: INA114U
[*] INA114U amplifies difference between sensing diode and offset diode; Calibration will be the tricky part
[*] PCB is same size as my HD44780 20x4 LCD screen, will piggy back on the screen so the whole box is tiny
[*] RFID access board is attached via USB to the main board
Pics!
RFID access board
Main DAQ
Since I'm making the boards myself, some traces are creatively routed to avoid as many vias as possible. I'll be making the boards probably next week, and the components should arrive by then. I will post more pictures later!
Josh Gill
josh.dgill@gmail.com
BSME, University of Arizona
Edited josh.dgill (May 16, 2010 13:56:23)
Offline
#2 May 20, 2010 20:15:43
- mike
-

- Registered: 2003-12-03
- Posts: 297
- Profile Send e-mail
New Build!
Thanks for sharing, Josh! It is always fun to see new builds. I'm not familiar with that Freescale part so it will be nice to learn something new. I'm curious about your RFID access board, too (parts, interface?)
I've always hoped that people will interface new controllers with the pykeg core & backend, though so far it hasn't happened; hop onto #kegbot on freenode if you'd like to chat -- I'm usually around on weeknights..
mike
I've always hoped that people will interface new controllers with the pykeg core & backend, though so far it hasn't happened; hop onto #kegbot on freenode if you'd like to chat -- I'm usually around on weeknights..
mike
Edited mike (May 20, 2010 20:15:43)
Offline
#3 May 21, 2010 13:21:04
- josh.dgill
-

- From: Tucson
- Registered: 2010-01-01
- Posts: 11
- Profile Send e-mail
New Build!
mike
Thanks for sharing, Josh! It is always fun to see new builds. I'm not familiar with that Freescale part so it will be nice to learn something new. I'm curious about your RFID access board, too (parts, interface?)
I've always hoped that people will interface new controllers with the pykeg core & backend, though so far it hasn't happened; hop onto #kegbot on freenode if you'd like to chat -- I'm usually around on weeknights..
mike
You should definitely look into Freescale. I thought I would never switch from microchip, and was very hesitant. I ended up switching because my senior design project had four stepper motors, so I needed 8 pwm peripherals. Microchip can't satisfy this requirement unless you move to 32 bit which was overkill for my project. Their pulse steering stuff is worthless, as most people generally use a driver chip of some kind and don't need to control all four corners of an h-bridge. In my case, I used TI's SN754410, so I only needed 2 signals per coil.
My dad is a EE and suggested to check out Freescale, and I'm so glad he did! The data sheets are very clear, and all peripheral initialization is done through a GUI in the IDE so you don't need to worry about incorrectly setting a bit and trying to debug the malfunctioning peripheral. Also, the in circuit debugger is cheep ($80), and works for all their processors, except the 32 bit processors. I had some trouble with some of microchips 8 bit processors, specifically the PIC16F1934. I am using it for a prototype MEMS intraoccular pressure sensor and need the 16 channel capacitance module. This processor has been a huge pain- I bought an adapter for the ICD2 because apparently not all PICs program at the same voltage. This still didn't work, apparently its max supply voltage is 4, not the typical 5, so I was burning the chip. Basically what I'm getting at is that Freescale's continuity is a breath of fresh air. Also, National Instruments (LabVIEW) uses Freescale, so that should speak for their exceptional quality. Freescale also has some hybrid chips, with say an H-bridge ASIC and a micro-controller ASIC all in one package.
For the RFID board, the reader/antenna is Innovations ID-12, its I think $25 bucks, very compact compared to Parallax. Uses a serial interface, I think max BR of 9600bps. Sparkfun did a review of this series if you want to check it out. On the board I also have an LED and buzzer, driven by a n channel mosfet.
As for connectivity, there's one USB port on the main board that connects to a host PC that will upload the status to my website using the USB 2.0 peripheral of the MC9S08JM16. I haven't determined if I want to rely on the host computer for authentication or store RFID numbers in the EEPROM. Probably will do it with the EEPROM. The second USB port isn't USB at all, the pinout is as as follows- 1- 5V supply 2-Interrupt from flowmeter 3-Tx from connector on RFID board 4- Voltage divider sense from 1N4148 to INA for temperature measurement 5- GND
As for interfacing it with your backend- I had no plans to do it, I was just going to write my own software on the host computer. Maybe I should look into yours, could save me some time.
Enjoy!
Josh Gill
josh.dgill@gmail.com
BSME, University of Arizona
Edited josh.dgill (May 21, 2010 13:21:04)
Offline
#4 June 24, 2010 01:00:28
- josh.dgill
-

- From: Tucson
- Registered: 2010-01-01
- Posts: 11
- Profile Send e-mail
New Build!
Well, I finally got time to etch my board. I made quite a few changed, ditched the INA so I will be using a LM34 temperature sensor instead of a diode so there is no need for the INA. I was able to get all the traces on one layer, and a ground plane. Board turned out great, going to solder it up in the morning and begin coding the Freescale processor. More to come!
Josh Gill
josh.dgill@gmail.com
BSME, University of Arizona
Josh Gill
josh.dgill@gmail.com
BSME, University of Arizona
Edited josh.dgill (June 24, 2010 01:00:28)
Offline
#5 June 27, 2010 03:31:51
- josh.dgill
-

- From: Tucson
- Registered: 2010-01-01
- Posts: 11
- Profile Send e-mail
New Build!
Well its dead in here, but I'm assuming someone's reading! I finally got started coding, I finished my LCD routines and layout for the screen today. Here it is! The data is made up to be placeholders for how it will actually look
Josh Gill
josh.dgill@gmail.com
BSME, University of Arizona
Josh Gill
josh.dgill@gmail.com
BSME, University of Arizona
Edited josh.dgill (June 27, 2010 03:39:19)
Offline
#6 June 27, 2010 10:36:29
New Build!
What's up, Gotta say nice work on etching your own board, and it looks like the LCD display is coming along. Have you had a chance to work on the flow sensor data?
Edited geoffculp (June 27, 2010 10:36:29)
Offline
#7 June 27, 2010 12:14:30
- josh.dgill
-

- From: Tucson
- Registered: 2010-01-01
- Posts: 11
- Profile Send e-mail
New Build!
Thanks! Making your own boards is actually really straight forward. I haven't calibrated the flow meter yet, but I will gladly do plenty of tests
What calibration factors have you all used? Default out of the box for water is something like 6k pulses per liter of water.. And how much of an effect does foam have on accuracy? I'm sure it has some effect.
Josh Gill
josh.dgill@gmail.com
BSME, University of Arizona
What calibration factors have you all used? Default out of the box for water is something like 6k pulses per liter of water.. And how much of an effect does foam have on accuracy? I'm sure it has some effect.
Josh Gill
josh.dgill@gmail.com
BSME, University of Arizona
Edited josh.dgill (June 27, 2010 12:14:30)
Offline
#8 June 27, 2010 14:46:22
New Build!
I'm using 6100 pules per liter with my swissflow sensor. When your kegerator is flowing good, you shouldn't have any foam passing though the beer line(and flow sensor). Most of the times foam in the beer line isn't an issue for me.
Edited geoffculp (June 27, 2010 14:46:22)
Offline
#9 June 27, 2010 21:06:52
- mike
-

- Registered: 2003-12-03
- Posts: 297
- Profile Send e-mail
New Build!
josh.dgill
And how much of an effect does foam have on accuracy? I'm sure it has some effect.
Most/all of the foaming typically occurs when the 'excited' beer mixes with the air, eg after it comes out of the tap & well after it passes through the meter. So, the meter tends to report the fluid volume accurately.
It is pretty easy to confirm this -- just pour a foamy beer into a measuring cup, and compare the liquid volume to the meter-reported volume once the foam has subsided..
Edited mike (June 27, 2010 21:06:52)
Offline
#10 Aug. 10, 2010 18:43:03
- josh.dgill
-

- From: Tucson
- Registered: 2010-01-01
- Posts: 11
- Profile Send e-mail
New Build!
The KegBot's moving along! Got the flowmeter working, but for some reason the calibration factor was way different than what is calibrated for water. I used a graduated cylinder and measured it out and got 9 pulses per .1oz of beer using shock top. Instead of counting pulses, I ended up using a 2 byte counter for ounces left, and a byte counter for the tens place of remaining beer. So every 9 pulses, I sub 1 from the tens place byte, etc etc. That seems to be really repeatable with the graduated cylinder, so I guess it doesn't matter that the factors different as long as it works.
I changed the GUI a bit,
Login's not done yet, instaed of RFID I'm going to use a touch PCB keypad I'm etching in a few days with a Microchip PIC16f1934. I don't want to use Microchip's stuff, but they have a great 16 channel cap sensing module built in. That will communicate with the Freescale on the main board using RS232.
Josh Gill
josh.dgill@gmail.com
BSME, University of Arizona
I changed the GUI a bit,
Login's not done yet, instaed of RFID I'm going to use a touch PCB keypad I'm etching in a few days with a Microchip PIC16f1934. I don't want to use Microchip's stuff, but they have a great 16 channel cap sensing module built in. That will communicate with the Freescale on the main board using RS232.
Josh Gill
josh.dgill@gmail.com
BSME, University of Arizona
Edited josh.dgill (Aug. 10, 2010 18:43:03)
Offline
