CPS, CVD and CTMU

[Pret]

Hi there,

I have made a numeric touch keyboard using CPS and so far everything works fine. I've made it manually, not with "#use touchpad" because I wanted high power range. The touch pads work through a 7mm glass cover. In my application, the CPS seems to have better results in comparison to CVD. The concern I have is that I keep seeing over the net that CPS is not that reliable. I keep hearing that its noise resistance is rather low. In my application, I even have an RFID next to pads (about 1.5cm) and it doesn't seem to influence whatsoever. I can see some variation in time, but with a simply compensation algorithm everything works fine. The CVD seems to be deranged by the RFID, though...

Now, what would be the risk of using CPS? What are the scenarios where CVD or CTMU are more viable than CPS? For instance, my device will be placed outdoors and I know that I should concern about the rain. But I think any adopted solution would have same problems, right?

I know that Microchip pushes its new CTMU technology, I've never tried it yet. I've head that pad cover cannot exceed 2-3mm. Since mine has 7mm, it's possible not to work at all?

Thanks,
Regards.

[Mike Walne]

You've already concluded that your experience is at variance with the "internet received wisdom".

Maybe it's time to perform some definitive tests.

Do some experiments with:-

1) Different size pads.
2) A variety of gauges and material cover, glass, acrylic.
3) A watering can and/or hose.
4) Anything else you can think of.

Let us all have the benefit of your findings.

Mike

Just in case you thought otherwise. I am being serious

[Ttelmah]

The 'point' about the CTMU, is it allows very small time intervals to be accurately measured. Advantage therefore small sensors, and you can re-sample very quickly. Simple circuits can be made to work, but it is not great in variable conditions. The CSM module has the more stable current sources, and seems itself to be more stable over time.

It is worth looking at 'non Microchip' application notes, like:
<https://www.silabs.com/Support%20Documents/TechnicalDocs/AN447.pdf>
Which gives a lot about the advantages of different shapes/sizes for the pads, and improving RF performance etc..

The short sampling time of CTMU approaches and small pad sizes, make them easier to provide some RF immunity, but a large pad gives more available signal change, which can outweigh this. The big problem will be tweaking for things like humidity changes, and keeping the response 'in range' as conditions change.

CVD, is just an AC voltage divider. It should be possible to still give the same 'delta' in the presence of RF fields, but this ignores the tendencies with modern high frequencies to 'hit' a tuned wavelength with almost any sensor track, and as a result the advantage vanishes....

They are all workable technologies, and in all cases to get reliable operation outdoors is going to involve some careful design.

Best Wishes

[Pret]

Thanks guys.

The Silabs application note looks very neat. I think I'll try the ground shield (that 30% hatch density), hopefully I would get a better immunity. Still, surprisingly, the 'problem' that I have is that my CPS method works too well. I don't know how should I stress it to confirm that works in all (most) scenarios. I mean, I need a some reasons to shift to different controller (18Fs) and try the CMTU. Since most users says that CPS is not reliable, it gives me the feeling that I'm missing something, or there is an obvious scenario where CPS rather fails...

Except spraying water, what would consist in a relevant test? I've been told that CPS is vulnerable when the power source is noisy. I'm powering it with a trivial 7805. I think I will connect a small brush motor too see how it handles..