Arduino controlled automatically actuated grill block
 

(NOTE - Nov, 2012 - the original grill block described in the first part of this thread didn't work well. Jump to post #94 to see the redesigned version.)

-----

I've been poking around with the Arduino for a little while trying to figure out how to get this project to work. I've compiled a bit of info on it, but Darin finally bugged me enough to get me to post the info as some help definitely wouldn't hurt. Also, this is by far not an area of expertise (electronics) for me. I know just enough to start getting me into trouble. I've only gotten this far through a lot of web surfing and reading up on things and some help from dcb (thanks again). So, I'm throwing out my initial designs for you to all look at, tear apart, and come up with something good we can all use.

Anyway, this is the general idea. Use the arduino to monitor a temperature sensor. This is an additional sensor (no splicing into existing car wiring) will be monitoring coolant temperature. It will be mounted in a radiator hose coupling that will need to be spliced in after the thermostat. Monitoring the temperature from this sensor, the arduino will output commands to a h-bridge which will control a simple two wire electric door lock actuator. This actuator will close and open the grill block door. Sounds pretty simple eh? Well, its not, at least not for me. :)

The current problem I'm having is finding an acceptable h-bridge. Almost everything I seem to find is surface mount, and I'm not even exactly sure what specs to look for in this thing. I think I found an acceptable one. Its the TLE 5205-2. Its in a P0-220-7 case so it can actually be worked with. The down side is its (I'm guessing) overkill as it can handle 5A continuous, and its also a bit pricy at $7.50. As I was browsing mouser and digikey there were a lot of cheaper chips, but as I said, mostly surface mount. I'd obviously like to keep the price down where ever possible.

Here is the list of necessary items:
Arduino of your choosing
LM61 temperature sensor in a TO-92 case
Electric door lock actuator
h-bridge
radiator hose coupling (local automotive store, mcmaster)
wire
grill block of your design


I have most of the arduino coding done, but I'm not going to bother posting it as it may change depending on the h-bridge.

All in all, I think this can be put together fairly cheap. I've looked into other ways of doing this and this is easiest, most customizable and cheapest way I've found.

Awesome idea!! I'll be watching the progress of this one...

Agreed - fun project.

Remind us again what the h-bridge does?

The h-bridge is a neato chip that will help me control the actuator. To understand why I need it I'll explain how the actuator works. The actuator is a really simple and dumb device. It only has two wires. You apply voltage to lead 1, and ground lead 2 and the actuator extends outward. If you apply voltage to lead 2, and ground lead 1 the actuator tries to go back inward. Its not even smart enough to cut power after its already fully extended, it just keeps trying to go out (or in) even once it has hit its internal stops. Anyway, the h-bridge lets me take two output pins from the arduino, hook them up to the h-bridge, and control the actuator. The h-bridge does the polarity reversing for you based on the inputs it gets from the arduino. It also handles the amperage needed to power the actuator kinda like a mosfet or relay would. If you tried to power the actuator with the arduino it would burn it out.

Gotcha. Thanks for clarifying.

Interesting idea. You need to plumb your sensor not only after the tstat, but after the radiator as well. You want to know how hot the coolant is before it goes in the motor, not after the fact. This will give you a little safety factor. You also need to allow for overlap in the temps. In other words, door opens at 220 degrees and closes when at 210. Otherwise, hysterisis will take place and door will get "confused".

Interesting... that's actually what I thought the "h" in "h-bridge" stood for... hysterisis... Tells you what I know... :p

How about just have it open when the electric fan comes on? If you have an electric fan of course.

My guess is that you would end up having the fan coming on and off constantly. It would cool as normally, due to the grille being open, but things would heat up rapidly upon the grille closing. Having the motorised grille block open at a temp below the cut in temp of the fan would help level out the temps before the fan cuts in. What would be really "cool" would be to have it on a servo motor so the grille could close by an amount that would hold the temps at a pre-determined level. As you'd drive through cold pockets of air, it would close slightly and vice versa. Awesome!:D

ollie

Good point and cool idea.

I don't think just any grille block design would work with a door lock actuator. At some point I imagined having just a cover that hinges forward to open, but I think a door lock actuator wouldn't have any holding force once the voltage is removed. Maybe a grille block with a sliding slotted door would be good.

Actually shows what I know. Didn't know what an h bridge is. LOL

Ideally, you would either use the engines existing coolant temperature or an additional one very close to it just to measure the actual engine coolant temp. I chose to splice in the sensor after the thermostat as the best easy option from an installation and universal programming standpoint. All the controller needs to know is, is the engine coolant hot enough to open the grill? Yes/No. I don't see why you'd need a second sensor to answer that question. The hysterisis is programmed into the arduino code.




In addition to having the fan come on and off constantly, you don't want the coolant fan running. It draws a fair amount of electricity and we are doing this for efficiency's sake. I currently have my arduino code setup to open up the grill just before the fan comes on, and closes the grill just above max thermostat temp.

I agree, the servo controlled grill block would definitely be best. But, try finding cheap parts for it. Plus its more complex and this is already straining my electronics knowledge haha. If you know of any good solutions feel free to share! :thumbup:

I think either one would work. If you wanted a hinged door you could just make an L shaped pivot arm and connect it to the actuator and door. My main concern is what design is superior when it comes to snow and ice. I tend to side with the hinged door for this reason since there is less surface area to have snow/ice build up on. Also, most h-bridges that I've seen have a 'brake' setting which would allow the actuator to hold in position. Not ideal as it still sucks electricity. However, I think most grill doors will be made of something fairly lightweight and it shouldn't be a problem holding them in place. But, this issue would suggest a sliding door would probably be better.


I never even knew h-bridges existed until about a week ago when dcb enlightened me to the fact that the arduino could not even come close to powering the actuator alone. Lucky for me, I never tried it. :)

that door lock actuator is just what i've been looking for... i've been messing with a motor an pulleys, but i can't get it reliable enough.

i think i'll stick with the fan trigger though. i've noticed about 90% of the time i can get away with a full block, but sometimes in summer if i happen to get stuck intraffic, things get to hot and the fan will come on, even without a block... the fan will stay on untill i get moveing again, so while the block will have the fan come on sooner i don't think it'll cause a lot of "flapping", especially since the fan already has a cutoff delay build in i'd imagine.

nothing but respect for people who will attempt something more sophisticated, but i don't think in practice it will add any benefit... i wouldn't like to build an overly complex system that doesn't do anything most of the time.

does anyone know what the wires on the door lock do? i think i'm going to get one that looks rather similar, what are the wires? is that ground and a 12v line for on/off or are their two temporary imputs of on/off?

If you can get it to run off the fan temp switch that would definitely be a much cheaper way to go. However, I'm not sure how you'd go about doing that. The actuator does require a little bit of smarts to drive it properly. Check out post #4 in this thread to see how it works.

By the way, I sure could have used this thing today! It was a frigid -13F out this morning. I noticed as I got off the highway, with the heater on, my coolant temp just plummeted down into the 150s.

You may be able to do the Fan On / Fan Off method by hooking up two actuators, one that opens it when the fan comes on, one that closes it when the fan turns off... Not sure how to get it to work, but if it costs a few extra dollars up front to save a lot of headache, I would say it's worth it.

That's why I suggested putting the temp sensor after the radiator, or at least right before the exit of the lower hose. You want some type of feedback for the coolant coming out of the rad - it keeps you from finding out after the fact that the engine is getting too hot.

That is the way that most aftermarket electric fan assemblies use - temp at the rad. Ideal placement of the probe is within 4-6 inches vertically and 1-2 inches horizontally of the radiator outlet.

Putting it after the rad is perfect, as it gives you right then temps on whether or not there is sufficient airflow through the rad. Not after it makes a trip through the motor and heats up even more. I think this would make for more accurate control.

There was a great write up on stuff like this from an engineer, but alas the website is gone. That same engineer markets and sells aftermarket fan controllers that work with pulse modulation/infinitely variable speeds (I have one on my mustang).

Granted, you could pull the temp information from anywhere and figure in the difference, but there will still be some time delay involved.

My two cents worth. :-) Hopefully it makes decent sense, as I typed rather quickly and didn't review it before posting. :rolleyes:

this is how i'd do it
http://i77.photobucket.com/albums/j5...y/text3293.png

the idea is:


fan switches relays... two which switch the position of the ground and the power line to the actuator. changeing its direction (fan on=extend, off = in) and one wich determines wich controle switch (fan on = monitor pistion full out, fan off = monitor piston full in) is monitored

this controle switch operates a 4th relay that switches the main power off if the pistion is at the end of its travel.

maybe not the most refined system but the logic could be used for something more advanced.

Interesting setup lunarhighway. Let us know how it works out! :)

I ordered a few arduino's yesterday (I have a few projects in mind).

I'm still waiting on some input on my h-bridge selection! I'd really appreciate it.

I went ahead and ordered the rest of the components. Once everything is here I'll setup a test setup and make sure things work. Ironically, this weekend at the EV meet, there was also an amateur robotics club meet afterwords. I stayed for that as they were doing a presentation and lab on h-bridges! It was perfect, I stayed and messed with them and asked a bunch of stuff.

I also think I'm going to go with a single hinged door. If I can fit a single door on the car, it will be hinged on the top to try and keep snow and ice off the hinged area.

My concern with mounting it after the radiator would be that it would be very hard to adjust for the cooling effect of the radiator when blocked. I want the grill to open before the fan comes on. If I wait until the coolant goes through the radiator and cools somewhat (which will vary depending on a bunch of things) I'll have no idea what temperature to set the switch open temp to. I'd like my sensor to know the engine coolant temp as soon as the radiator fan sensor, not somewhere farther down the chain. It would seem that putting the sensor after the radiator would create too much fluctuation and lag (vs the radiator fan's temp sensor) than if it was put further upstream.

What kinda temp switch are you planning on using?

i'm not sure i understand the question. the idea is i hook up an extra wire to the fan power line... there's a socket on the fan, so if i can find i male and female connector i can literally "hook up" my addition without any wire cutting. i let all the factory stuff do the temp detection,so when the opel engineers decide things get to hot and the fan comes on the doors are kicked open. the idea is to combine the simple blunt effectiveness of a full grillblock with the "death vally" proof stock setup. allowing a little airflow trough at all times will keep the temp below the treshold 90% of the time so the fan doesn't come on every 5 minutes

that's the theory :)

Ah yes, I forgot... Too much looking at circuits lately haha. Do you have any idea when you'll build your setup?

i've been working on and off on the thing for the last few months, but other stuff has kept me from tinkering with the car to much.
basically i've got most of the stuff, i've had the door installed on my car to check the fit so now i "just" need to breath some life into it.

so far i've got a more or less working door mechanism for the top grill but the pulley system i'd build looked a little messy... if i had found something to use as pulleys to route the wires i might have gotten it to work, it's not a dead end but if i can find a simpler actuator like that great central locking motor i think that's the way to go

but since i'll start moving next week to my new place wich needs some major (eco) modding i'm a little worried it'll be some time before i get some time to work on it.

btw an old idea just popped back in my head! i'll just mention it for what it's worth.

when i first started thinking of a grillblock i considdered a gradual opening and closeing system...
although i never got very far with the idea the motion of the temp gauge was inspiring... there it was right before my nose something that gracefully moved in relation to the engines temperature!... i assume this is controlled by an analog changing voltage so taping into this signal could be interesting.

mapping voltages to temperatures can give you key points where something should happen (opening angles for the door perhaps)... some relatively simple electronics could trigger a relay at a certain voltage (basically something like that makes the fan kick in...that's maybe how i arrived at that idea)

unfortunately my knowledge of electronics are a little rusty but i remember in school we once build a simple voltage meter where leds representing preset values lit up when that voltage was reached... there was not much more on the print than a couple of resistors and a few transistors...

like i said...just an idea

That may be a future step, and the arduino would definitely provide that flexibility. Its a good idea and it would also be a superior grill block IMO. Baby steps for me though. I'm all new to this. :)

Alright, I have the physical block and door pretty much designed. I kept it as simple as possible. Its still a bit rough (yes I know the control arm in the 2nd pic isn't attached), but I'll tweak it as it's built. I just wanted a rough idea of how long the L shaped arm needs to be.

The block will be quite easy to mount behind the lower grill on the Matrix. There is plenty of room down there. I'm planning on making it out of coroplast. If reinforcement is needed I'll do that with some .030 aluminum sheet I have.

Anyway, pics!


Closed:
http://ecomodder.com/forum/attachmen...1&d=1232564425


Open:
http://ecomodder.com/forum/attachmen...1&d=1232564425

Is the L arm in the same plan as the grille or am I seeing things out of perspective? Something looks wrong in your linkage. That L arm is going to want to bend sideways during operation. I also don't see how it could possibly close the grille shut at speed.

EDIT: Ok, the arm is probably not in the same plan as the grille. It's lined up with the actuator. I still doubt that little actuator will be strong enough, especially since the L arm gives leverage against the actuator.

There is going to be a bending action there at the L-bracket's pivot point. It being a problem should become apparent with trial and error. Perhaps bending the pivot bolt upwards by about 20 or 30 degrees might strike a balance in terms of the pushing force on the door. Or raising the pivot point as far as possible from the door will also lower the bending action enough to let it work (invert the L-bracket to have the pivot point even higher?). A spring might also come in handy to counteract the weight of the door although the door lock is probably plenty strong.
Best of luck with this. I tried this with bike brake cable but it didn't work out.

ollie

No, you're looking at it correctly. The L arm will be pulled backwards when the grill door is open. I'm not too worried about it as it'll be made of something hefty and its lifting a really really light door.

Closing it at speed is a question that I don't have an answer to yet. I don't know how much psi will be building up on the door. I also don't know how much torque the actuators put out, but its not a weak device. I could always add another actuator if need be.

Another concern is keeping it closed at speed. Once the actuator powers down, is the motor and gearing in the actuator enough to hold it there? I'm not sure. Me and Metro were just talking about this actually. There are multiple solutions though. The simplest would probably be to use some magnets to hold it closed.

If all else fails, a sliding door would be the next step.

I'm taking the sliding door route. Since my actuator is a shape memory alloy wire, and I don't intend to keep it energized the whole time, the sliding door seemed like the most workable solution.

I agree that the door will not want to close at speed. You might be able to hinge it at the center (equal pressure on top and bottom for easier opening), or I wonder if a sliding design would work... he said there is plenty of room...

EDIT: Two of my key points were already made by others while I was making my drawing.

With the sliding door, even if there's an "heavy" lateral load, worst case scenario you mount it on ball bearings and longitudinal load will be irrelevant.

How about a rotary type valve. Have your opening as normal but install a length of, say, 4" drain pipe cut in half along its length. Fit some end plates with a pivot hole so the half pipe rotates around its centre. Kind of like a ball valve but with pipe instead. That would overcome the wind pressure opening the door at speed. Not unlike the sliding door so whichever is easier i suppose. Just throwing ideas into the pot!

ollie

It's been done.
 

Exactly what basjoos did with his car... I think he operates his with a lever from inside his cabin, but the rotary idea has been done.

ultimately i think many designs are valid, most of the time it will come down to the design of the existing cooling setup and the restrictions thereoff, and the type of actuator you have.

here's some idea's i've been looking into... in my experience rotateing doors are easier to make than slideing ones, but they require a bit more depth, if you're going for a slideing desing a slotted setup as shown below might have some advantages... especially you'd only have to slide the door a fraction of the full distance to make the same opening...thus an actuator with a fixed throw (like the door lock) can be used to provide more torqe, as the same foce has to move the same load over less distanse.

since i was originally planning to use a small motor and gears proved problematic i figured pulleys could work... at first it may look a little mideval, but i figured that if they could put something like that together 700 years ago with simple hand tools, and make it reliable and strong enough to entrust it with lifting a massive wooden bridge that was one of the main safety features of the castle, i should be able to pull off somethign similar that could move a little coroplast rectangle a few cm's... the idea is that two wires are wound around one axle, propperly spaced and maybe separated by a central plate to pervent tangling. useing a relative thin axle it's also an very easy way to convery the fast spinning motion of a relative weak motor into a slow slideing motion with enough torque to get things done

one wire is already wound around the core while the other is just attached to it.

as the core spins the unwinded wire will wind, provideing pulling foce and the othe wire will unwind..this will pull the door open, and if the direction of the motor is reversed it the other wire will pull.

since there's a wheel or lever at the other end the unwindeing wire will be pulled from that end so there shouldn't be to much slack at any time.

i imagine small pulley or running the wire trough flexible tubes could enable the motion to be reliably transmitted over quite some distance, and around curves...

it could be used for both sliding and rotateing doors and in the case of rotating door they can be stacked on top of each other

although i don't think it uses pulleys, the stacked rotary shutter design is currently used on some bmw models

http://i77.photobucket.com/albums/j5...way/pulley.png

Finally got my order from moderndevice.com that I ordered on the 15th. I got an arduino duemilanove and a RBBB arduino kit. I'll program and test with the duemilanove, and then transfer the chip over to the RBBB board and run the final tested and true version on that board.

I already have had got my sensor and h-bridge from digikey a while ago, so we're ready to go. Soon, I should be able to get everything wired up with a prototype board and hook up the actuator to see how this is gonna work.

http://ecomodder.com/forum/attachmen...1&d=1233013461

I gotcha, you are going to use the duemilanove to program the chip and swap it. :thumbup:

For development purposes, you can also pop the chip out of the duemilanove, and make a cable to connect its rx and tx and reset and 5v and gnd to the rbbb versions of the same and serial program the chip on the rbbb directly (they should be right next to each other on the rbbb).

Much more expedient for development purposes once you get it set up. If you are like me, you will have to recompile and upload many many times before you are happy with the software.

Sorry, back to aerodynamics and mechanics... :)

Ohhh, I didn't know you could even do that. Thanks for the tip!