Fingerprint reader susceptibility to EMF interference?

[Tony's X60s]

I may be going over old ground here, but I have just worked out that some of the unreliability of the fingerprint software/CSS7 combination on my X60s could be due to interference from my Netgear DG834Gv2 ADSL rounter and wireless access point which (until I moved it) was about 15 inches from my Thinkpad. Checking the Amtel site (they make the fingerprint reader chips that IBM uses) I see the chip is a capacitive type which relies on EMF fields to detect skin ridges on the finger - so EMF interference seems a likely explanation. When I unscrewed the aerial from the router, print recognition performance improved a lot.

It would be nice if the CSS7 software gave some indication that the fingerprint reader was having problems - it just seems to not respond at all to multiple swipes, leaving me to puzzle over what the problem might be.

Has anyone else noticed this issue? How far away from WiFi WAP's does the Thinkpad need to be for the reader to work properly?

[christopher_wolf]

That is quite the possibility; except that UPEK makes the Fingerprint Readers on the Thinkpads last time I checked.

See; http://www.upek.com/custpart/pcnetworking/ibm.asp

It might not only be the WiFi, but it is also related to certain CCFL lighting from what I have noticed. There are certain areas where the Thinkpads take some swipes to actually "catch" the fingerprint.

[Tony's X60s]

My aplogies Christopher_Wolf, it certainly is Upek. I'm not sure how Amtel crept in there.

Your comment about CCFL's is of interest as I also use a compact Flouro desk lamp above my Thinkpad, so it's copping interference from two potential sources - no wonder I'm seeing reliability issues!

Now, if IBM had only chosen thermal scanning instead of active capacitance scanning .... (fade into wishful thinking reverie).

[christopher_wolf]

=
Now, if IBM had only chosen thermal scanning instead of active capacitance scanning .... (fade into wishful thinking reverie).

Then we would have to deal with much more calibration and possibly more "re-scans" to boot! Thermal recognition is difficult at best; given what is to be worked with, active fringe capacitance was probably the best way to go. Granted, we do get weird things like this from time to time.

The drawback is primarily that you are using the skin of the surface to generate a pattern of capacitances, you also have to deal with the body acting as an antenna and a giant resistor. Hence, the body will pick up a steady 60Hz coming from the CCFL and that has to be dealt with.

Usually, the FPR and its associated filtering circuits do a pretty good job of it, but there are some situations where they could run into some trouble. I am still trying to figure out the Whats and Hows of each.