TTP223 Erratic Behavior

[ThomsCircuit]

Yes. Totally up to you. Radial, Ceramic would be good, like MLC or X7R type caps.

What about doing this. of course i would relocate C5,6,7 just before PAD IN. i could then just add 10pf to C5-7 to get desired results.

 

[eTech]

What about doing this. of course i would relocate C5,6,7 just before PAD IN. i could then just add 10pf to C5-7 to get desired results.View attachment 132194

The three footprints look SMD. Are you using thru hole or SMD caps?
For testing I would breadboard as a test platform to find the right value. Then just use one cap.
The cap value isn't critical, just need to find out which value solves the issue.

 

[ThomsCircuit]

The cap value isn't critical, just need to find out which value solves the issue.

ok got it. C5-7 they are SMD 10pf. those i can find. it was the 20, 30pf i could not locate.
i will re-design for one cap. ill use lead type. i can breadboard those.
thank you ETech. talk to you soon. Go Bolts!

 

[Tony Stewart]

To eliminate the EMI problems consider;

1. A 3 uA series resistor to attenuation relay spike noise. Consider >=1k and perhaps add a cap
2. Use only twisted pairs >4 twists/cm for everything and keep at right angles to noise source.
3. Add a low ESR cap for Vcc near the Relay , again with twisted pair
4. Consider a ground plane under the touch pad.
5. follow the self-calibration for best results.
6. If the touch pad is not directly connected, consider coax or STP cable grounded.
7. If PS common mode noise is an issue,, consider an RF cap 0V to earth ground !

“
After power-on have about 0.5sec stable-time, during the time do not touch the key pad,
And the function is disabled
Auto calibration for life
At low power mode the re-calibration period is about 4.0sec normally,
When key detected touch and released touch, the auto re-calibration will be redoing after about 16sec from releasing key.

 

[KeepItSimpleStupid]

Tony Stewart Did you crawl out from under a rock?

 

[KeepItSimpleStupid]

Stray inductance can be minimized in SMT caps by mounting them on edge.

 

[Tony Stewart]

Stray inductance can be minimized in SMT caps by mounting them on edge.

Is that the best you can say to solve this EMI
problem? That might apply to UHF but not here with a snubber on the flyback, unless return paths are being shared by the chip to the power source.

Did anyone highlight the note for the touch sensor cap must be a low microphonic type like NP0?

It remains for me to understand what physical layout he needs with adequate sensitivity with not too big a load cap and how much margin needed from interference for reliable operation.

 

[ThomsCircuit]

Ive replaced the CMD caps with quality lead types. I was able to locate PF values 2.5,10,15,22,25,and 30

 

[ThomsCircuit]

It remains for me to understand what physical layout he needs with adequate sensitivity with not too big a load cap and how much margin needed from interference for reliable operation.

my thoughts exactly. lol
The touch pad is 16-24 inches from the ttp223 circuit. The lead wire to the pad is 18 gauge. the pad itself is a cast metal plate 2"x1"
I apologize that this design of mine is may not be all that clear. I hope this helps.

 

[ThomsCircuit]

BTW- Keep the wire twisted together from the TP to the PCB.

if there were 2 wires id understand but there is just but one wire from the pad to the pcb. I used to run, hookup, and certify CAT5 cables. Runs sometimes near 300 feet. Twisting the pairs of wires controlled the interference.

 

[Tony Stewart]

if there were 2 wires id understand but there is just but one wire from the pad to the pcb. I used to run, hookup, and certify CAT5 cables. Runs sometimes near 300 feet. Twisting the pairs of wires controlled the interference.

I assume the casting is connected to preferred ground, otherwise it makes a great capacitive antenna to EMI. Better work on that.

 

[eTech]

if there were 2 wires id understand but there is just but one wire from the pad to the pcb. I used to run, hookup, and certify CAT5 cables. Runs sometimes near 300 feet. Twisting the pairs of wires controlled the interference.

Can you post photo's of the top and bottom of the touchpad?

Nevermind - I See you already did that.

Twist the ground wire (VSS) and Pad In/out wire together.

 

[ThomsCircuit]

otherwise it makes a great capacitive antenna to EMI. Better work on that.

im going to show you pics of the pad im using and you guide me on how to solve or reduce the EMI
the circuit board is about 15" away from the button.

 

[ThomsCircuit]

Twist the ground wire (VSS) and Pad In/out wire together.

hmm. how would i do that? pin 2 VSS is part of a copper pour. (pad in) comes from one direction (the plate) and pad out goes another (the circuit board we designed last year)

 

[ThomsCircuit]

i need to point out that this interference only happens in this circuit. It uses spinning magnets (attached to fan) i can turn the circuit on and off without issue as long as i do this within 3-5 seconds of turning it on. Once the fan reaches full speed the circuit will turn back on after i turn it off. Is it because the fan has reached full speed? dont know. that is what i assume. i can turn it off in this state if i hold my finger on the pad until the fan stops.

 

[eTech]

Fan? What fan? Isnt that a touch pad in post 53? I think part of the problem is that we're slowly getting bits and pieces of the whole picture. So far, I've seen two TTP223's, one on the PCB your making and one on the Touch switch module. Are you using multiple touch circuits?

OK...So lets go back to the beginning....
What you have described so far isn't erratic (inconsistent) behavior. It's consistent, just not what you expect. Correct?

 

[ThomsCircuit]

OK...So lets go back to the beginning....

Post #1 outlines the issue i am having.
It morphed into this other circuit because of a comparison I was making to a similar circuit. Some users gave opinions on the morphed circuit and I figured (newbie excuse) that they were the same so here we are. But not all is lost. We did re-create the same touch pad circuit. Only now I understand it and its limitations and how to overcome them. Well sort of.
Now that we are back on track I believe the magnets are creating the interference. Both circuits have many similarities. They use the same power supply, have a buck transformer, same relay, and an external pad. Only difference is this one with the magnets and fan. And you are correct. The behavior is not what I expect. Ive tested the pre packaged (sonrom) with up to an 18" lead on the "External Touch Pad" and a 10 foot lead on I/O with no issues. As I increased the length of the External Pads wire the sensitivity degraded. I concluded that 20" was maximum distance.
Now this swimming pool circuit with the fan has just a 3" lead on the external pad and it triggers with a single touch. Turning it off is, well, difficult. Im sorry if I caused any confusion. I will stay on topic.

 

[ThomsCircuit]

There is some information on post #44 that may help i just don't understand how to implement it.

 

[ThomsCircuit]

Swimming Pool Circuit.
Im using this PCB model while referring to the sonrom module #1. As I make changes to the design it is easier to see the changes.
I have researched and applied some changes as per post #44.
I added a 1k resistor between pin3 and PAD. i found this was also suggested in an article at electroschematics.com
I added a ground pin. this to be used with the pads leads. Both soldered to the plate (penny) #55 using one twisted pair from a CAT6 shielded cable. One lead to IN, the other to ground. I will rework the wires in the BOX. Ill twist the power in, Twist the fan lead wires, and re-route so they do not cross each other or more importantly run parallel.

 

[Tony Stewart]

Auto cal requires no touching in the 1st x seconds. That will affect the results. perhaps the fan needs to be forced to full speed during this period to normalize the gain of background noise below touch threshold. Consider every magnetic switching device a source of noise That needs twisted pair and a ferrite CM clamp with high mu and a cap