Prototyping Project: TEC

Biowarejak

Maker of Awesome | User 1615
Original poster
Silver Supporter
Mar 6, 2017
1,727
2,205
Currently On Hold - 04/11/18
In order to do justice to this project, I need to actually finish a few of my other ones. In particular #77 and Radian. I love to work on this stuff, but my plate is full right now and I have very limited time. Thank you, everyone, for your support and patience. This will be awesome.

Introduction:

This project started out as a simple mod, really. The 7-segment LCD for my cooling loop dimmed to the point of being unusable - prompting me to search for a way to replace it with a more robust and reliable solution. That search led to my purchase of a nifty little display compatible with my Arduino UNO R3. I took some time getting around to playing with the example code, but it’s presence was invaluable during the early days of this project. Ultimately, however, I am going to change displays later on; Adafruit sells a few of them that I would love to try. That would enable me to make much better use of the libraries they developed - as well as unlocking more pins on my Arduino.

With the current display, I’ve discovered that opting to not use the SD Card reader frees up 4 pins. When I discovered that, I had to make sure any sensors I used stayed in that range. After some research into temperature probes, I decided upon the purchase of these waterproof DS18b20 modules, rather than the thermistor type commonly used in water loops. Not only are they single-wire data devices, they provide far superior accuracy for an idea I had - what this project has ultimately become.

You see, here on SFF Forum we constantly compromise cooling performance for size. Some of our water loops are beaten by high-end air coolers due to the lack of radiator space! Or, so I’ve anecdotally noticed anyhow. I decided that wouldn’t be acceptable when I finally downsize my rig. So I took the skills I developed for monitoring the water temperature and set to work controlling it.

I plan to accomplish this by using a thermoelectric cooling module, or several, to pump out the excess heat from the loop. I’m well on my way already, but I decided with the impending start of college classes this fall and potentially a job change on the horizon that I should begin public documentation of my project.

So, hey everyone! :)


Project Outline:
Key:
Blue: Completed
Orange: In Progress
Yellow: Future


  • Controller:
  • Acquire Preliminary Components:
    • Arduino Uno R3
    • Display
    • Waterproof Temperature Sensor
    • Ambient Humidity Sensor
    • Protoshield
    • Molex Extension Cables
    • Crimping Tool And Dupont Connectors
  • Design Sensor Breakout Shield:
    • Decide On Whether Or Not To Include TEC Module Related Components At This Stage
    • Decide On Connectors
    • Order Batch
    • Order Stackable Pin Headers, Connectors, and Misc. Components
  • *Deep Breath* Design My Own Microcontroller Board...

  • Operating System:
  • Sensor Readings Over Serial Connection
  • Sensor Readings On Display
  • UI Design:
  • Get The Letters In The Boxes
  • Align All The Things
  • User-definable Text Size
  • User-definable Color
  • Proper Alignment Post-scaling
  • Proper Alignment Once Sensor Reading Hits 3 Digits
  • Implement A Dew Point Calculation
  • Touch Screen Implementation:
  • Work Out Kinks In Display Rotation
  • Figure Out Why Preliminary Test Failed
  • Code For Resistive (Current Display!)
  • Code For Capacitive (Adafruit Display!)
  • Improved Use Of Constructor Functions
  • SD Card For User-images (Requires Adafruit Display!)
  • Ability To Display System Information
  • Implement Control Scheme
  • Implement and Benchmark PID Library
  • Implement and Benchmark Direct Control Algorithm
  • Implement Scheduling Algorithm
  • Case:
  • Model Early Drafts
  • Post and Poll
  • Pending….
Gallery:


We were at 360+ views before I fixed the gallery! :D
 
Last edited:

AleksandarK

SFFer
Moderator
May 14, 2017
703
772
WOW, I am really amazed.

I have planned to do TEC loop my self, but i may not end up doing it. How do you plan to cool the hot side of TEC? What TEC you are going to use?
 

Biowarejak

Maker of Awesome | User 1615
Original poster
Silver Supporter
Mar 6, 2017
1,727
2,205
How do you plan to cool the hot side of TEC?
Separate water loop :) It will require a separate pump and waterblock, but my research would suggest it's a reasonable way to run things when certain precautions are taken.
 
  • Like
Reactions: AleksandarK

Biowarejak

Maker of Awesome | User 1615
Original poster
Silver Supporter
Mar 6, 2017
1,727
2,205
How much of a TDP would you have to cool?
That's going to depend on a few things honestly. I'm not sure the specifics about how TDP measurements line up with my current understanding TEC thermal transfer rates.
 

Biowarejak

Maker of Awesome | User 1615
Original poster
Silver Supporter
Mar 6, 2017
1,727
2,205
Yo dawg. How would this save space though if you need another loop to cool the TEC?
I asked myself that question too before realizing I could fit it all in well under 20 liters o_O actually I think 11 was a comfortable enough fit. Since it's essentially cutting a larger loop in half it honestly doesn't take much additional space provided the pump is on the smaller side of things.
 
  • Like
Reactions: AleksandarK

Aibohphobia

aka James
Feb 22, 2015
4,955
4,681
Maybe just use an AIO for the TEC? Should be plenty of surface area and super low temps aren't necessary, just enough for the TEC to function so that'd save you quite a bit of space.
 

Necere

SFF Guru
NCASE
Feb 22, 2015
1,175
2,365
My research into TECs some years ago led me to conclude that they only really make sense if you have an abundance of excess rad capacity. I.e., if adding more radiator to your loop gets you negligible temperature gains under load, then a TEC might start to be useful. This is generally antithetical to SFF, however.
 

Biowarejak

Maker of Awesome | User 1615
Original poster
Silver Supporter
Mar 6, 2017
1,727
2,205
Maybe just use an AIO for the TEC? Should be plenty of surface area and super low temps aren't necessary, just enough for the TEC to function so that'd save you quite a bit of space.
Precisely why I aquired a Kelvin T12 :) I'll be trying things a few different ways.
My research into TECs some years ago led me to conclude that they only really make sense if you have an abundance of excess rad capacity. I.e., if adding more radiator to your loop gets you negligible temperature gains under load, then a TEC might start to be useful. This is generally antithetical to SFF, however.
That is certainly interesting. My goal is to augment a small radiator, getting the heat out of the component loop and into the rad loop. Hopefully a steady-state condition is achievable before the pump goes into a temperature fault. I'll be doing the math on that soonish.
 
  • Like
Reactions: AleksandarK

EdZ

Virtual Realist
Gold Supporter
May 11, 2015
1,578
2,104
My research into TECs some years ago led me to conclude that they only really make sense if you have an abundance of excess rad capacity. I.e., if adding more radiator to your loop gets you negligible temperature gains under load, then a TEC might start to be useful. This is generally antithetical to SFF, however.
The other way to use them is to allow the use of a much smaller rad, by keeping the 'hot' loop hotter and the 'cold' loop colder, and making the radiator more efficient (the hotter your radiator the more efficiently it can conduct heat to the air). As long as your working fluid does not boil or you exceed the maximum temperature for your TECs, you can tolerate a very hot loop temperature. The downside is you need at least as much power pumped out by your TECs as the actual parts you want to cool, so whether you make any net gain in performance is mainly determined by the ambient temperature (hotter room = bigger gain from raising the radiator's delta-T).
 

Necere

SFF Guru
NCASE
Feb 22, 2015
1,175
2,365
The other way to use them is to allow the use of a much smaller rad, by keeping the 'hot' loop hotter and the 'cold' loop colder, and making the radiator more efficient (the hotter your radiator the more efficiently it can conduct heat to the air). As long as your working fluid does not boil or you exceed the maximum temperature for your TECs, you can tolerate a very hot loop temperature.
The pump might be a limiting factor in that scenario, though. IIRC the DDC has a rated maximum operating temperature of 55-60C. You'd also want to be careful about tubing selection, I'd think.
 

EdZ

Virtual Realist
Gold Supporter
May 11, 2015
1,578
2,104
Tubing definitely, but there is plenty of CoTS high temp tubing available. A small industrial/lab pump (e.g.) should be able to handle a high loop temperature - though at additional expense if bought new rather than second hand - or even the DDC or D5 if given adequate external cooling (the rotor will be fine, but the motor will need an alternative heatsink to the working fluid it would normally use).
 

Biowarejak

Maker of Awesome | User 1615
Original poster
Silver Supporter
Mar 6, 2017
1,727
2,205
What TEC you are going to use?
Sorry I just saw this part of your post. That's up in the air at the moment as I'm primarily focused on implementing the OS side of things for my Arduino right now.

I have gotten Horizontal and Vertical alignment post-scaling accomplished today. I need to add a few more lines to get it to adjust based on the number of characters on the screen. I'll add a picture of that when I have a chance.

-EDIT-

The other way to use them is to allow the use of a much smaller rad, by keeping the 'hot' loop hotter and the 'cold' loop colder, and making the radiator more efficient (the hotter your radiator the more efficiently it can conduct heat to the air). As long as your working fluid does not boil or you exceed the maximum temperature for your TECs, you can tolerate a very hot loop temperature. The downside is you need at least as much power pumped out by your TECs as the actual parts you want to cool, so whether you make any net gain in performance is mainly determined by the ambient temperature (hotter room = bigger gain from raising the radiator's delta-T).
Yep, that's what I was trying to get across earlier :)
 
Last edited:
  • Like
Reactions: AleksandarK

iFreilicht

FlexATX Authority
Feb 28, 2015
3,223
2,331
freilite.com
So this is a super awesome project, but I do have one nitpick: Could you put all the imgur images you linked into an imgur gallery and then embed that into your post? Opening spoilers for each picture is pretty 2010.
 

Biowarejak

Maker of Awesome | User 1615
Original poster
Silver Supporter
Mar 6, 2017
1,727
2,205
So this is a super awesome project, but I do have one nitpick: Could you put all the imgur images you linked into an imgur gallery and then embed that into your post? Opening spoilers for each picture is pretty 2010.
I can, I honestly hadn't even thought about it. Derp.

-Edit-

Fixed it :) Glad you like the project.
 
Last edited:
  • Like
Reactions: AleksandarK

Biowarejak

Maker of Awesome | User 1615
Original poster
Silver Supporter
Mar 6, 2017
1,727
2,205
Update!

Got the display to scale properly! Updated the OP Project Outline to reflect that :)


Now I'm focused on getting the sensor-shield finalized. After taking some measurements it looks like I will have to use right-angle connectors in order to fit the display on top of the shield stack. It's also going to cost ~$60 in components to populate the boards if I order through Digikey - which means I'll be making a trip to Fry's to grab a handful of Arduino shield headers for about $2 a set. Much cheaper. I'll also test and see if I can get away with only using a single pull-up resistor for the two different sensor types.