Full disclosure, I'm a mechanical engineer with a (relatively new) electrical engineering hobby i.e. I'm a noob in the world of serious electronics. I'm also the kinda person that likes to figure things out for themselves and I'm usually successful. So the fact that I'm here seeking help means that I've been trying pretty hard to solve my problem for a while and I'm REALLY stuck.

So, I have a project I'm working on where I'm using several homemade touch sensors fabricated out of ITO film and affixed to the back of a piece of plexi glass. I'm using ITO because I need the sensors to be transparent. The front of the plexi is covered with a sheet of vinyl with some printed graphics that correspond to the touch sensor locations. The sensors are controlled by a raspberry pi hat from Adafruit which is based on the MPR121 chip. The whole thing is intended to be backlit by an LCD TV. I chose to use a TV because there are many graphic elements that need to be illuminated independently and interactively based on what the user touches. This allows me to display graphics on the TV screen directly behind the graphics that are being illuminated. I thought this would be easier than wiring a bunch of LEDs which wouldn't produce the consistent lighting that i need (among other reasons). I have video of this in action if anyone is interested or needs clarification.

It works pretty well on it's own, without the TV. However, when placed near the TV screen, the sensors go crazy. Now, I've discovered that it isn't all TVs that cause this. In fact it seems to be exclusive to LED TVs, which may only be a coorelation, because I've also noticed that LED TVs (at least mine) aren't electrically grounded at the outlet (they only have two prong plugs). Whereas the CCFL TVs I own (which have three prong plugs) do not affect the sensors at all. Also interesting note is that the LED TV produces interference when plugged in only, but not turned on (power supply EMI?).

So the problem I have is that my project requires a specific size TV (39"), but with either CCFL back lighting or proper grounding (which I don't think are produced anymore) OR I need to find a way to mitigate the noise. I've considered previously owned TV's but this project will be used in a commercial environment so it requires some degree of reliability. I've tried adjusting the filter registers on the MPR121, which help but can't quite get there. It's entirely possible I don't know what I'm doing though.

All suggestions are appreciated (and sorry for the long post), thanks!

 

Was it something I said?

 

Was it something I said?

Most likely, no one could come up with a good idea on what might be causing the problems you described. That does happen here, from time to time.

 

Most likely, no one could come up with a good idea on what might be causing the problems you described. That does happen here, from time to time.

Well I guess that's somewhat reassuring, I assumed it was just my ignorance that was the problem, but maybe the issue isn't as trivial as I had hoped. Thanks for your response OBW.

 

Can isolate the problem, regarding just the LCD or the backlit is creating havoc.
You need to cutoff the backlit or the LCD to confirm the issue.
Not an easy thing to do some LCD's

 

Can you get an ITO shield to put between the LCD and the touchpanels? Also, try disconnecting the backlights -it can be electrical noise that is getting into the touch sensors (Once had an unshielded backlight supply that kept a touch sensor from working. )

 

Can isolate the problem, regarding just the LCD or the backlit is creating havoc.
You need to cutoff the backlit or the LCD to confirm the issue.
Not an easy thing to do some LCD's

Well, I can tell you that I have reproduced the problem with two different LCD TVs with LED backlights, and confirmed NO interference with two other LCD TVs with CCFL backlighting. I think this makes a strong case for at least eliminating the LCD surface as a source of interference. But as I stated originally, the backlighting type may be irrelevant since I also noted that the LED TVs are equipped with two prong plugs while the CCFL TVs have three prong plugs. I also observed interference with the LED TVs even when the TV was off, but plugged in, which I think points to a power supply issue and again, my theory is improper grounding. My (uneducated) guess is that CCFL TVs demand more robust grounding because the lighting method inherently produces a significant amount of EMI and/or have higher power demands.

 

Can you get an ITO shield to put between the LCD and the touchpanels? Also, try disconnecting the backlights -it can be electrical noise that is getting into the touch sensors (Once had an unshielded backlight supply that kept a touch sensor from working. )

I didn't see this till after my last post, but looks like you might be affirming my suspicions. I've considered attempting to shield the sensors in this way (with another piece of ITO) but wasn't exactly sure how to execute it. I've actually experimented placing a sheet of aluminum mesh screen (same stuff used to repair screen doors) between the TV and the panel, and then connecting the screen to ground. This helped quite a bit but not enough. I could use guidance regarding the proper shielding procedure. Thanks!

 

Two choices to try.
1. Drive the TV's back lit with a CC driver ( home made ), thus elimination the PWM switching noise.
2. Make a ground trace around the touch panel and earth it.

The second choice might work as I have seen in microwaves which uses touch panel. The touch switches has an additional trace around the switch panel like a 2mm border around the complete switch assy and it is connected to the mains earth which is the earth of the microwave chassis.

 

The tiniest amount of noise can throw off a touch switch or capacitive proximity detector, and I'm talking microvolts depending on method used.

As per rif@@ ground the ITO or mesh to the touchscreen ground. You might have to try several places before finding the optimum grounding points, but I would start with a ground near the output of the power supply.

If I understand correctly that you have multiple circuit boards, the common connection to the power supply would be a good place to start but you might need to run a low impedance ground strap from each panel to the shield.

Make sure that any switching power supplies are well shielded and that the power supplies to the touchpanels are as free of noise as possible. Though you might notice the problem when the LCD is present, noise from other sources that were present might be contributing to the problem.

Your intuition about noise sources, ground loops, etc. should help guide you to a solution. If you don't have intuition about such things now you probably will by the time you solved the noise problems.

 

Thank you very much for the advice! I will continue experimenting based on your suggestions and report back

 

Update:

After countless hours of experimentation, I'm admitting defeat and moving to plan B. I either don't have the knowledge to solve this or there just isn't a practical solution. I disassembled the TV and gained access to the control board. From there I found an available pin that was tied to ground and tied my touch controller system to that. I also tied everything to earth ground and discovered that this (along with the other ground connections) actually made significant improvements. At one point I thought I had it licked...until I gave it time. It turns out that the longer the TV operated, the worse the interference got. I hooked up a DSO to one of the touch sensor leads and observed. From the moment the TV is turned on to about 1 minute after, everything is nice and clean. But after that, the signal gradually degrades and by 10 minutes the noise is overwhelming. I'm attaching a few screen captures for reference, shown sequentially with time. The two channels shown are two different sensor inputs, channel 1 has no sensor attached and channel two has a sensor pad attached.

 

This is very strange. One must ask what is changing during the first 10 minutes after power is switched on.

If you run this for 10 minutes, switch it off and wait a few seconds before switching it back on, are the signals clean when it comes back on are they still noisy? That would tell us whether the effect is thermal or something similar.

 

It's thermal I beleive.

The TV panel is heating the touch panel

 

This is very strange. One must ask what is changing during the first 10 minutes after power is switched on.

If you run this for 10 minutes, switch it off and wait a few seconds before switching it back on, are the signals clean when it comes back on are they still noisy? That would tell us whether the effect is thermal or something similar.

I had a feeling someone might ask this, and I never focused on this aspect specifically so I'm not sure, but I think it's still noisy based on my memory. I'll confirm later tonight after work.

It's thermal I beleive.

The TV panel is heating the touch panel

The touch pad of channel 2 is not in direct contact with the TV panel, and the signal on channel 1 of the scope doesn't even have a pad attached to the controller, which is position about 12" off to the side of the TV. While I do agree this is somehow thermal related within the TV itself, I'm 100% confident it is not due to heat transfer to the touch controller or its sensors.

 

Keep the touch away from panel and scope it

 

I've never fooled around with touch sensors. What's the shortest human interaction? Can you filter to ignore shorter pulses?