my first banana project - controlling ventilation by banana

10 5697

I finally have managed to get my banana working it is that bloody time, it seems that there never is enough of it. I hope you people dont mind that I post it on several forum pages as I try to find people who want to create something similar and maybe combining our efforts what can make more .

In short I want to make my apartment kind of like it is in star trek, you come in say lights and voila you got lights or set the ventilation, heating whatever just by sitting down on computer or just watch at your phone to see what the current temperature in your kitchen . In my vision everything is monitored and logged, I havent really figured out why it would be good but I think it just is cool - maybe in future I figure out what good is in that whole punch of data. First I was building my smarthouse project around raspberry but now I have found banana and I like banana more - no reason at all just do.

So I plan to use my first banana as brain of ventilation system, I want it to control 12 fans, I want it to monitor temperatur and humidity in garaga (where ventilation system is installed), in box that will contain banana, outside and also in two channels for air to move inside my apartment. Also I want it to monitor how much current my ventilation system uses at any moment(in my vision everything is monitored, I even think about monitoring radiation levels in my apartment, I would know instantly if somebody would come to kill me with polonium ).

so picture of the box with banana in it banana_box.jpg

any suggestions about my setup or anything else are really welcomed .
* on it you see banana getting power from yWRobot breadboard power supply(http://www.petervis.com/Raspberr ... d_Power_Supply.html). As said in this review AMS1117 voltage regulator gets really how and I had some small headsinks with clue on back, so it seemed like good idea to stick one on 5v regulator one (as everything runs on 5v, I do not use 3.3 at all), is it good idea ? Also I have worries about AMS1117 capacity, I might use more power than it can regulate for me but I tend not to worry about it yet, it is always trial and error method for me .
* then you see dht22 temperature and humidity sensor, there will be few more connected at the end but for testing purposes, one is good enough.
* then you cant really see but there is mcp3002 ADC connected to banana over SPI. ADC is connected to breakout board with acs712 current meter - this part in future should take care of measuring the current used by ventilation system. I have little code that I created for raspberry once some long time ago, to use same setup (mcp3002 and acs712) and measure current but I never got it really working, I mean it all works and numbers almost make sense but something is bit off . I believe something is wrong with my math. if somebody knows more about how it all should work, maybe you can look at my code and tell what's wrong :). the code is located at torch.bumpclub.ee/thisandthat/acm.c
* also there is small computer fan, I was thinking that if it is in closed box and on summer it might be bit hot so it needs some ventilation ... but 5v from breadboard barely make it run and I havent figured out really how to make it work after all it is all work in progress. any ideas ?

I want to put 12 - 16 relays(to control relays) into box similar like this and connect it with banana box over i2c cable. I want to use mcp23017 gpio expander in this other box - signal comes from banana over i2c to mcp23017 chip that then turns on or off relay that controls fan - if it makes sense :). I have build one box like this before but I left it unattended for a moment and it got stolen, true story, I believe nobody really have any use for it but well.

picture of the relay box that got stolen old_relay_box.jpg

last time before the box got stolen I had idea to connect it to raspberry inside over 8m cable. I had problems with this setup, relays used too much power - system worked on border line ... I believe it was mostly the fault of 8m cable now first I have different sort of relays(solid state relays), I will experiment which ones will take less power and use these and now relay box will be connected with less than half meter cable, right next to banana box ... I think or rahter hope it solves my power problem :)

people :) I wait for you suggestions, comments and ideas ... it is kind of force making me move more :)
House automation is a famous subject and if I had more time I also would do something similar.

However, if you need "kind of force making me move more" then imho most probably it never will come to a happy end. It is a major task, it is very very annoying if it does not correctly work, so be warned

Great job.

Edited by tkaiser at Mon Dec 22, 2014 08:24

How do you feed the yWRobot PSU? Which voltage? And do you know that you can read out current/voltage the PSU supplies to the AXP209 PMU? Have a look into the network/servers subforum for details. The same applies regarding thermal issues. If you use convection correctly you might not need a fan. If you don't... heatsinks won't help if you operate the components under higher loads all the time (heatsinks help in situation with short load peaks when there's not enough airflow around hot components).

I would be really interested in the AXP209's values regarding voltage/current since the yWRobot MB102 can just deliver 700mA according to its manual (would be a nice combination to use passive PoE injectors/splitters together with the MB102 to provide power over Ethernet in some setups -- thanks for mentioning the MB102, I've two of them lying around: http://kaiser-edv.de/tmp/ZoqfG1/BPi_RPi_OLinuXino-LIME2.jpg)

Edited by destroyedlolo at Mon Dec 22, 2014 21:10
tkaiser replied at Mon Dec 22, 2014 15:24
If you use convection correctly you might not need a fan

If you're not using the GPU, I don't think you need a cooler mechanism at all. My "production" PI is on my basement, in a candy box, and it never goes above 40°C or 45°C in the summer, even during heavy compilation period.

Edited by tkaiser at Mon Dec 22, 2014 14:47
destroyedlolo replied at Mon Dec 22, 2014 14:08
If you're not using the GPU, I don't think you need a cooler mechanism at all

Would be interesting what you experience regarding thermal values when you use 'stress' instead of just compiling. We've already talked about this stuff lengthly http://forum.lemaker.org/forum.p ... =8137&pid=42347

In my opinion the best idea is to simply have a look at temperatures rather than guessing something. Because 'overheating' depends on so many factors that it makes more sense to monitor this stuff instead of wasting thoughts on it

tkaiser replied at Mon Dec 22, 2014 08:24
How do you feed the yWRobot PSU? Which voltage? And do you know that you can read out current/voltag ...

I use 7.5v 1000mA PSU but I ordered myself 9v 2000mA PSU for it - somehow I feel myself better if numbers are bigger but is it better or is it just in my head. This 7.5v PSU is some old switch PSU, just something that was laying around (of course I found it after I ordered the 9v PSU ).

About voltage, in 5v rail I have 4.92v when nothing but banana is connected and 4.91v when everything was connected(2x dht22 sensors,  mcp3002 adc, acs712 current sensor, mcp23017 gpio expander and 8 relay solid state relay board with one relay switched on).

I am feeling pretty positive about AMS1117 regulators, when I played around with raspberry and I powered raspberry with 2000mA usb charger and everything from raspberry then whenever I did something I always noticed fluctuations in 5v rail but now doing same voltage remains steady. It is my first time of using MB102 but when I finish my project I have several raspberries/bananas controlling and monitoring different things and I think MB102 is good way to go. I do like the idea that all the current do not go through banana/raspberry. again it is just my thinking but it seems like a good idea .

I still do not know will 700mA be enough power to run 14 relays and the rest needed but I was thinking that if it is not enough then I can replace 3.3 volt regulator with another 5v one and this way in theory I could have 1400mA or is there some other bottlenecks ? I heard that the word around the campfire is that actually AMS1117 can handle 1000mA, so I keep wondering myself why MB102 is marked as 700mA.

LinAdmin replied at Mon Dec 22, 2014 01:06
House automation is a famous subject and if I had more time I also would do something similar.

Howe ...

I know time is one funny thing one moment you seem to have like so much of it and next moment doing something you find out that you actually have no time at all

I been on and off with my project over a year now, it is that sometimes you dont have time then you dont have resources and then there are also days that pure laziness takes over but I still believe in beautiful future and I know one day it will be finished just I will not shout out the date yet after all Rome wasnt built in a day

by the way meanwhile I have started my work on new relay box picture of proof: IMG_20141223_202127.jpg

Also this time I use different relays, solid state relays, dont know much about them but power consumption seems smaller than the ones I used before, of course old ones were connected over 8m of cable, now it is just 1m. Also I have only 2 relays connected(one of them switching on and off tablelight at my workbench and second one doing nothing but just is connected), so mayor testing is still on the way - at moment I am really positive about them .

tkaiser replied at Mon Dec 22, 2014 14:45
Would be interesting what you experience regarding thermal values when you use 'stress' instead of ...

I plan to design it so that it could operate on extreme conditions, maybe just watching too many sci-fi movies and I think it is fun of it's own to design the system that way

also I plan to station banana and relays in garage, in small space under roof and there's also sealing, so the space is rather small and I believe on hot summer days temperature there will rise even uncomfortably high and on day like this it would be pretty much one of the worst things to happen if ventilation system would give up (pretty complicated sentence, sorry time to time I have habit of making sentences that even myself find hard to understand )

so I think there should be some servo or something, controlling the fan - turning it on and regulation rpm as needed. servos and everything related to them is pretty far from me (never touched them). so I was wondering is there some easy to learn way to get to know it and find bits and bobs needed ... or is there some other solution maybe that I havent considered yet

Also one question off the topic, but just curious ... I monitor among other things my raspberry and banana cpu temperatures, I have a whole tables of different temperatures and 7days history of them ... but as I run banana and raspberry next to eachother with similar workloads, how come that banana runs about 15C lower that raspberry. I am thinking myself that A20 chip do not have temperature sensor inside ? I have cubieboard 2 and it have no temperature sensor, I dont think it was just software issue. So is this temperature sensor added somewhere next to chip and in case of raspberry it is in the chip and that's why the difference ? or do A20 just run that much cooler ?

One more cool thing!

You have to log in before you can reply Login | Sign Up

Points Rules