Society of Robots - Robot Forum

Electronics => Electronics => Topic started by: kwik on February 18, 2012, 05:56:33 PM

Title: great idea, need help with the circuit
Post by: kwik on February 18, 2012, 05:56:33 PM
i have been asked as a university product design task to design a artifact that relates time and people.

so i have come up with a design for a clock that is this.

there will be 12 petals arranged in a circle to represent the hours (bare with me...) and 4 inner petals to represent 15min, 30min, 45min and 0min...

to tell the time the petals will curl up (using flexinol) 1 at a time every hour, so for 3:15 the 3rd petal on the outside will curl up and the 1st petal on the inside will curl up.

i would also like 12 LED's around the outside (1 at each hour) to turn on when it is that hour to aid in telling the time at night etc.

i have come to a robotics forum for help as i do not know how i will control the flexinol, i know you need to pass a current through it of say 0.4A but then it could easily overheat, so was considering having it pulse so 0.5sec on 1sec off etc... so it wont overheat and it will stay curled up, maybe move a bit but it would look cool.

i had an idea of using a servo controller? 1 petal per output on the controller, something like the ssc-32, im not sure, there might be a really simple way of doing this that i am completely oblivious to....

thankyou for any advice though :)

chris
Title: Re: great idea, need help with the circuit
Post by: kwik on February 19, 2012, 04:18:58 PM
ok, after doing research for the day as i have been given no information, i have found out 2 things but 1 needs confirming.

the first is i have thought of a way to control the 12 pieces of flexinol using a 12 way switch and a mechanical clock (obvious really).

the second being is i read that you can run current though a 0.006" or less diameter wire constantly without fear of overheating, if this is the case then it will be a massive help. can anyone confirm this?
Title: Re: great idea, need help with the circuit
Post by: Soeren on February 19, 2012, 04:55:58 PM
Hi,

i have come to a robotics forum for help as i do not know how i will control the flexinol, i know you need to pass a current through it of say 0.4A but then it could easily overheat, so was considering having it pulse so 0.5sec on 1sec off etc... so it wont overheat and it will stay curled up, maybe move a bit but it would look cool.
The current should be selected for the particular wire, so posting a link to the wire, or even better, a link to the dataheet for the wire you'll be using wouldn't be a bad idea (the wire diameter/resistance and max current as a bare minimum).
Using PWM (Pulse Width Modulation) is just one way of controlling it and your 1/3 scheme will be equivalent to a third of whatever current you use. If using PWM, it can be used as a current limiter based on eg. the temperature of the wire.


i had an idea of using a servo controller? 1 petal per output on the controller, something like the ssc-32, im not sure, there might be a really simple way of doing this that i am completely oblivious to....
A servo controller would be gross overkill.


ok, after doing research for the day as i have been given no information, [...]
What?
Doing research because nobody gave you the answers???


the first is i have thought of a way to control the 12 pieces of flexinol using a 12 way switch and a mechanical clock (obvious really).
You mentioned 16 pieces of wire?

It would need to be a strong clock to drive a switch (probably not practically possible). a sequencer circuit driving transistors would be one step up, but a microcontroller would really be the best solution, and it could control everything - 16 wires, 12 (or 16?) LED's and the timing itself


the second being is i read that you can run current though a 0.006" or less diameter wire constantly without fear of overheating, if this is the case then it will be a massive help. can anyone confirm this?
You can run a certain amount of current through any Flexinol wire and the reason for using a higher current (and then back it down if it need to hold it actuated for some time) is to speed up the actuation, but in your case, it will probably be more "organic looking" to let the petals curl up in a gentle fashion.


"to design a artifact that relates time and people."
How does a "flower clock" relate time and people?
(Sure, the time aspect is clear, but the people part eludes me).
Title: Re: great idea, need help with the circuit
Post by: kwik on February 20, 2012, 04:32:45 AM
Thank you for the reply, the second post probably did confuse things a bit.

I said 12 as there will be the 12 petals on the outside with 4 on the inside, sorry about the confusion.

The wire i want to use is the 0.006"LT http://www.dynalloy.com/TechDataWire.php (http://www.dynalloy.com/TechDataWire.php) and roughly 3inches long so 4.8ohms/inch and 410mA for a 1 second contraction.

Again thank you for your help so far
Chris
Title: Re: great idea, need help with the circuit
Post by: Soeren on February 23, 2012, 07:24:02 PM
Hi,

The wire i want to use is the 0.006"LT http://www.dynalloy.com/TechDataWire.php (http://www.dynalloy.com/TechDataWire.php) and roughly 3inches long so 4.8ohms/inch and 410mA for a 1 second contraction.
Do you have the wire and a lab power supply for some experiments?

If you need to lower the current (if for no other reason than that it's on 24/7, so it would be the "green" thing to do), you need to find the lowest current that will safely hold the wire contracted (under all possible ambient conditions).

A current "back-down" PWM circuit that gives full power for 1s and then a lower current until turned off, can be made by a single CMOS Schmitt trigger NAND gate (of which there's 4 in a 4093 chip), 1 diode, 2 capacitors and 3 resistors.
You'll need this for each wire, so for your 16 wires, it sums up to four 4093 chips, 16 diodes, 32 caps and 48 resistors.

If you use any sort of mechanical timer with  that routes to each of the wires, a voltage boost could be added at each switching over and the normal running current set t the low current (found by experiments) by using an appropriate voltage and a resistor.

But before I confuse you with more options... Do you have any experience in building electronics circuits?