Electronics Complicated custom keypad system, need your input.

what prog was that done in zap?

 

As with all my renderings: trueSpace3, search for truespace3 free download on google.

 

Zap:

Have you considered the fact that a CAR is ISOLATED completely from GROUND because of the 4 RUBBER TIRES???

You may have some real issues getting your ground loop to work due to that. Does QPROX have any datasheets on getting this to work with a vehicle??? As far as I can tell, all of the units that they utilize or show as functions are kiosk type units, or stationary systems (like buildings) which have an earth ground of some sorts.

In addition, have you considered *how* you are going to get the unlock circuits to work? You might have to pull some of the schematics on your vehicle to find a point in the 'unlock' circuit that you can tie too effectively. Otherwise, you'll have to build yourself a relay system, since most unlock mechanisms that I'm aware of, draw anywhere from 1-5 Amps at 12v to work. However, if you can find a Chilton or a Mitchell manual on your particular vehicle, you can probably find the relay circuit closer to the computer which you can drive directly instead.

You may also want to consider temperature variations. Since this would be working in a vehicle, the MC68HC11 might need some help in the cold. That or if you can find one of the wide-temp variants of the 68HC11 that would help (they are difficult to get ahold of however). Something that has worked for me in the past is to do a final layout on the board, and then encase it in epoxy resin. It makes repairing the board impossible, but it prevents thermal expansion from breaking solder joints, popping chips, and other general junk. On the flip-side, you may want to consider how you're going to COOL the circuit, since during the summer, the inside of the vehicle can get to 120+ degrees...your system would need to still work under these conditions...

Just more to think about I guess...

Hmmm... some other things I thought of too. Zap, you may want to add some type of feedback mechanism. Like an LCD screen of some sorts, to show you what mode the car CURRENTLY is in, or what mode you have just activated. Considering the multiple pages to the functions, it might be pretty easy to get lost and have to start over from scratch if you misstep.

You may also want to add some type of 'back' or 'backspace' button, so you don't have to go through the WHOLE code entry just for one mistyped button.

 

Yes, I have thought and asked about that.
The chips self calibrate to the enviroment, and it doesn't matter if the device you are running is battery powered, handheld, or in a building, it will still work.

I thought I already explained how I am unlocking the car.
I am tapping into the existing unlock button on the driver side door.
It needs only two signals: 0ohms to GND for unlock, and 1000ohms to GND for lock.

As far as the microcontroller and heat, my prototype will be useing a cheap old basic stamp1, so if it can handle the heat then I am not worried.
(As far as cold, it doesn't get cold here very often)

I don't plan on encasing the first circuit in epoxy, but mabe later I might.
The circuit board I had designed uses 0.06" traces through out the entire design, and is only single sided, this should be very durable.

In all reality failure of the circuit will not prevent me from getting in the car, as I can still use keys, or the keyless entry remote.
But the keypad is handy if you happen to lock you keys in the car.

 

You're referring to the QPROX chips? Or the HC11? From experience, I have never seen an HC11 work over 100 degrees F. Granted, these weren't wide temp variants, they were just std HC's, but after about 100 the swtiching get's real erratic, and it starts flipping out...literally. Considering that heat might be a problem, there are supposedly some wide temp variants of the HC's that can work up to 120 without issue. I'm just concerned that the inside is going to get too hot, bake your chip, and you'll end up losing units too often (at least the HC's).

You may also want to ensure that you get a solderable pin package instead of the all-to-common PLC package. Because of the temp changes, the PLC packages can pop out of the socket, unless you use some type of screw retention mechanism.

You didn't really explain where you were getting the connection point from. Have you measured the current draw at these points? I know my Pontiac requires about 2Amps to unlock the vehicle. I didn't see any relays in your circuit, so I wanted to point that out to you...

Oh yeah.. that's right, you're in TX! Hehe... I'm from WI, so I generally have to worry about working at minus 30 to 120 F...hehe.. Yeah, I wouldn't worry during the proto stage, but once you get that HC11 in there, they can be REAL buggers with heat.

Nah, I wasn't worried about that, just trying to help you design something that wasn't going to fail-out every few weeks or so. Otherwise even if it *did* work, the constant replacements would make the thing useless....

You didn't answer my first question though (edit: yeah you did, I just read the answer wrong...sorry)... are you sure the ground loop will work? The Kirchoff effect uses a common ground principle to establish the different potentials. However, the vechicle is isolated completely from ground. You may have to have one hand on the outside of the vehicle, and then use your finger to touch the glass, since the physcial earth itself is isolated from the backside of the sensor (that is why Vechicles carry their own ground, which is generally the body - in Uni-body construction - or the frame).

The handheld situation works, because one hand is holding the unit, while the other uses the screen, thus providing a full ground loop. However, they tend to STOP working, if you wear tennis shoes, and then put the handheld down on table, and try to press the buttons. This is exactly why most mfgs of PDAs went to the capactive plate technology for PDA screens, or resistive block.

I've personally used Ultrasonic systems and they work *great*, but they tend to be very expensive (lots of backend dsp processing). These would probably be the most reliable in a vehicle (in fact, they are as both BMW, Cadillac, and Lexus utilize ultrasonics for all their touch screens in their cars). I happened to know the original designer from Zenith who worked on the design for GM.

I'm just kind of wary on these QPROX units... I guess we'll see when you build it!

 

Never mind about the Kirchoff question Zap. I read through the chip-specs a bit, and it states that you should be able to use the vehicle body as a ground plane. The capacitive coupling only needs to be in the nano-farad range....Hmmm... I wonder, the difference in potential between Cx and Cs needs to be managed a bit it seems. I wonder if the vehicle might make the pads TOO sensitive... as the value of Cx will increase dramatically, thus making it smaller than Cs.

This may become a sticking point on getting this working right. If you couple the unit to the ground plane of the car, you get a good loose couple with earth ground, but then you might have too large of a Cx value, as the vehicle would have a huge capacitance. However, if you don't couple it to the vehicle ground plane, then your Cx would be smaller, but there wouldn't necessarily be enough coupling between *you* and the plate, which would require high-gain values, and therefore possibly increase the potential for false trips.

I can see why most vendors have stayed away from these touch panels... there's *way* too much playing around to do.

This should be interesting to see. I might pick up a few of these chips myself to play with.

 

Well I won't be building the large circuit for quite sometime, I can choose the appropriate microprocessor at that point.

I am not hacking the motors, but the switch.
The switch is a low current signal that runs to the Lock Control box, then the lock control box handles the high amps for the actual motors.

I am using transitors to duplicate the circuit that the keyless entry uses to unlock the doors.





----
I am sure the ground loop will work on the QProx chips.

I already have the touch part of the circuit built and is working great.

 

Well, you're car is a bit nicer than mine. My Pontiac has this incessant condition in which certain modes of operation of the central computer override the switch responses, depending on what is happening. Leaving me in a situation in which I had a real bear of a time getting the car alarm setup. It's cool that your's uses low voltage. I ended up having to fudge 4 lines for each switch...

What distributor did you end up gonig with the find those QPROX chips? I searched through Newark and some other std places, and I can't find anyone that has stock on these guys. Did you have to special order those?

 

Normal stock stuff at Digi-key.
(Or Farnel in the U.K.)
Digi-key even stocks the evaluation boards.

The PCB for this is beautiful, single sided, all 0.1" spacing, and less then 4" square. (All dip parts)

The device could be compacted into a double sided PCB behind the actual keypad it's self.

 

Sweet, somebody else actually knows about digikey. You're the first person to even mention it bnesides me that i have seen.

I'm curious about the tough sensors ability to function through glass, let alone the 3 layered windshield. are you sure it will function?

 

Well Digikey is US, Farnel is UK, these forums are a mix of people from all over.

I have tested a mock-up of the sensor already, it works through two layers of 1/4" acrylic, a layer or 1/8" lexan, and a final layer of 1/8" glass.

The sensor I am using is built for multiple key's so it can handle the sensors being close together, yet still can tell them apart enough that you don't get false triggers or "fat fingering"

 

wow that is the most advanced and complicated mod i have ever seen or heard of. programing logic chips and such

i wish i could do stuff like that

why dont hook up your car with neons, hydraulics and all the nice shiny stuff so you could program a command in the keyboard to do something cool with that

but i have an idea and i dont know if it was mentioned above
lexus is now offering a remote key that opens your locks and stuff when you come next to it with the remote key, no button no nothing, just walk up to it and the doors will open, and start your car and such
soooo i'm thinking you can do something like that so that your car will be able to detect when you are around and unlock the keypad itself that way
just an idea...
the penguin

 

i don't like to necropost, but i'm dying to know whether you went through with making this thing, ZapWizard...

 

Wow. Zap, move to South Dakota so I can become your apprentice. Your ideas are original and you follow through. Your mods are awesome.

 

ZapWizard: You might just want to take a look at this, maybe you can take some ideas.

http://www.gideontech.com/content/articles/212/1

 

That is cool, but I want my system to be completly inside the car, so that it is weatherproof, and also can't be ripped off, or hacked.

I am still building this, I have all the electronics parts ready, and even a prototype (Not in my car yet)
I just don't have the time just yet.
I should after QuakeCon

 

this thing is going to be great.. and when it comes to the dasing of it, I like it.. (could it be an idea to have the red leds flashing until a botton is pressed? (like a alarm diode)

 

K then, waiting for pics.

 

As a thought for lock/unlock have you thought about linking it to a bluetooth device and when the device detects a specific paired device (i.e. your phone) it unlocks....when your phone goes out of range - it unlocks?

Obviously you could override this but it might be a nice auto-lock thingy.

Good idea though - let me know how you get on with the keypad design as I'm currently planning on using the QT60161B-AS for a touch sensitive keyboard

 

I sense quite a few more modding awards in zap's future
First off, he gets my award for most interesting forum post of the day