touchRead Sensor Size

Status
Not open for further replies.
jwatte said:
This might work: http://www.digikey.com/product-detail/en/abracon-llc/AIRD-02-221K/535-11415-ND/2660709
7 Amps rating, 220 uH, through hole.
Note that the capacitors are important, too -- without the "H" of capacitors bracketing the inductor, you don't get the filter steepness you want.
And if 1000 uF is too big for you, 100 uF instead will raise the rejection frequency by little over 3x (sqrt(10)) -- if 800 Hz is still good for you, that'll work, too.
Click to expand...

Thanks, I look forward to testing this. Good learning experience.

I'm going to check out one other thing this evening. To tell a bit more about the project, the 24V is for some "muscle wire" actuators controlled by the Teensy via mosfets. The rest of the system can run off a smaller 12V wall-wart transformer supply. But I think I probably need to share ground between the two in order for the mosfets to do their thing. It might be that I can share the ground only without getting the noise injected into the Teensy/capacitive circuits. Anyway I'll play with that tonight. If it works then all I need is a couple of $5. wall-wart adaptors, which I probably have a bucket full of somewhere:)
 
Interesting!

With the 3.3V out of the Teensy, if you use N-channel low-end switches, you're going to need MOSFETs with quite low Vgs thresholds to get good turn-on. Depends on how much current you're pushing through them, I guess? Or use P-channel high-end switches and a small-signal N-channel for pulling its gate down (with a pull-up for turn-off.)

How much current does the Teensy use? Why does it need 12V at all? It seems like a 7805 or similar linear regulator could power the Teensy straight out of the 24V into Teensy 5V, and with very small capacitors in front and behind that regulator, it's pretty excellent at rejecting transients. (80 dB transient rejection, the data sheet claims.) The current draw of the Teensy might be dozens of mA, which should keep you way below the watt or so that a TO-220 can dissipate without heatsink even going from 24V to 5V.
 
jwatte said:
Interesting!

With the 3.3V out of the Teensy, if you use N-channel low-end switches, you're going to need MOSFETs with quite low Vgs thresholds to get good turn-on. Depends on how much current you're pushing through them, I guess? Or use P-channel high-end switches and a small-signal N-channel for pulling its gate down (with a pull-up for turn-off.)

How much current does the Teensy use? Why does it need 12V at all? It seems like a 7805 or similar linear regulator could power the Teensy straight out of the 24V into Teensy 5V, and with very small capacitors in front and behind that regulator, it's pretty excellent at rejecting transients. (80 dB transient rejection, the data sheet claims.) The current draw of the Teensy might be dozens of mA, which should keep you way below the watt or so that a TO-220 can dissipate without heatsink even going from 24V to 5V.
Click to expand...

12V powers a T-class audio amp and also gets regulated to 5V to feed the Teensy (and an XBee wireless module) which has its own 3.3v regulator. I have some logic level mossiets that turn on at something like 2.5V and they seem to be working fine. Right now I'm only switching 2A (calculated).

I have a 4A power supply both for margin and also because I may need more muscle than 2A as this project develops. The Teensy itself uses very little current. I found some very inexpensive and small adjustable chinese regulators that are working great both to drop the 24V to 12 for the amp and to 5 for the Teensy, both direct from the 24.
 
mark_m said:
I'm going to check out one other thing this evening. To tell a bit more about the project, the 24V is for some "muscle wire" actuators controlled by the Teensy via mosfets. The rest of the system can run off a smaller 12V wall-wart transformer supply. But I think I probably need to share ground between the two in order for the mosfets to do their thing. It might be that I can share the ground only without getting the noise injected into the Teensy/capacitive circuits. Anyway I'll play with that tonight. If it works then all I need is a couple of $5. wall-wart adaptors, which I probably have a bucket full of somewhere:)
Click to expand...

Disappointed to find that even with only the shared ground, the noise contaminates the system and messes with the capacitive sensing. I can't switch the mosfets without the Teensy ground being referenced to the 24V ground that's being switched by the mosfets (tested just to be sure). I tried joining the two grounds with a 1K resistor and that seemed to help; with a 10K resistor it stopped working.

I have inductors and caps ordered. Funny it's getting hard to find power supplies that aren't switchmode. I'm sure some are a lot quieter than others and the ones I have are cheap but I don't have time/money to buy a bunch of different ones and test. I realized too that in putting these systems together I probably cut off an over-the-cable inductor on the power supply's output cord since I was hard-wiring it into the system and didn't need the plug. So in addition to the other parts I ordered some cheap clamp-on inductors, every little bit probably helps...
 
I do also have a drawer of small through-hole relays, good Omrons. I guess I could switch them with the mosfets (overkill) and have isolation that way. I guess I can try that too...

The tough stuff always comes up at the end of projects it seems, much as I try to catch it all early on. I guess it's really only tough stuff because it is at the end of my schedule. This is a sculpture project that's going into a show at the end of the month, and I'm taking the family to Panama for a week to visit one of our kids who's in the Peace Corps, the week before I'm installing the piece. And physical construction's not even done yet:) Trying to remain calm...

Will a mosfet switch low currents ok (like 25 mA) for the relay coils?
 
Last edited:
Yes, you need common ground to switch MOSFETs. If you add a resistor, you have to do the math on how much current you get through that, and what voltage that loses in difference between the two circuits.
It should, in general, be perfectly safe to hook up the ground between the two supplies (assuming they're both properly isolated from mains.)
Also, doing so should not magically "transmit" any signal, assuming your ground is a solid connection with very low impedance.
Well, there is a small risk of ground loops transmitting a small signal, if the resistance of the ground is measurable, and the return on one of the supplies is marginal.

Another option is to use an optoisolator to drive the MOSFET. Some complications arise because optoisolators typically invert voltage levels (they transmit current relation) and the Vgs of most cheap power MOSFETs is only 20V so you'll kill yours if you try to tie it to 24V with a pull-up.
 
Well I'm low on time so if I can make it work I think I'm going to punt with these relays for the immediate need. After this one-month show I'll re-do my mosfet board which is hacked already since I'd originally designed it for TIP120's. So I'll redo a board maybe with optoisolators (never really even heard of them:) or some other strategy to separate the power supplies completely. Or just get 24V power supplies that are quieter, or implement the chokes and caps, which I still want to try and have ordered... Guess I'll wait to test that before I tear things apart to put in relays. Would be a little less medieval:) Though part of me likes that aspect... I used to live in a town called Crockett that had the last running physically switched phone office with rows of operator stations with patch bays and I assume lots of relays one way or another. Kind of steam punk.
 
Last edited:
I received and tried the choke & capacitors. Against logic and odds it didn't solve the problem, and still even with only the ground shared, capacitive sensing was badly disturbed. I think because I'm sensing at distances of 10-12", the noise in the ground wire even without a circuit is affecting the measured capacitance which becomes extremely noisy. As soon as the ground is lifted the measurements become relatively consistent and useable.

I set up a relay to only bring the 24V & ground into the circuit when I'm using the mosfets to power the muscle wire, then disconnecting before going back to capacitive sensing. That seems to be working great. The coil for the relay draws 35 mA, nominally 5V, but is switching fine at 3V. I ganged two pins on the Teensy so that they are splitting the 35 mA load. So far so good... I found some cheap wall-wart power supplies to run the Teensy's, XBees and Class T amplifiers. I don't know why but that doesn't seem to create any problem
 
Sorry to resurrect an old thread, but have been using a Teensy 3.2 recently developing an application that uses the touch pins. Got to this thread because I'm looking to adjust the touch sensitivity in order to make the touch pad work within the physical constraints of the housing. The standard touchRead function doesn't provide enough sensitivity, and based on the comments above, I'll experiment with the values in touch.c.

My question is how do I copy the touch.c library into my project (standard Arduino IDE) an leave the standard library unchanged? Cheers!
 
2moto said:
Sorry to resurrect an old thread, but have been using a Teensy 3.2 recently developing an application that uses the touch pins. Got to this thread because I'm looking to adjust the touch sensitivity in order to make the touch pad work within the physical constraints of the housing. The standard touchRead function doesn't provide enough sensitivity, and based on the comments above, I'll experiment with the values in touch.c.

My question is how do I copy the touch.c library into my project (standard Arduino IDE) an leave the standard library unchanged? Cheers!
Click to expand...

I hope someone can answer your question. I can't remember enough as I have not touched this project since approximately last post. I had an accident that included a brain injury and am just now getting back to it and may start this project again almost from scratch just so I can relearn what I was doing. Biggest problem was noise in the sensing and lack of good response from large triangular sensors. Turns out triangles are not a good shape so I need to re-think a lot of things:)
 
Status
Not open for further replies.
Back
Top