RFID odd behavior when on battery power

Check this out. A simple hello-world app with the RFID module on it and NO functionality whatsoever. I can tap, tap, tap the tags, the reader responds pretty snappy (has its own LED’s).

Works GREAT on USB power from PC, via USB Client SP, on Cerberus.

Now plug it into battery (via USB Client SP). Battery has a good charge (batt type is same as used to external power/charge cell phones). Tap once, it responds great. Tap again… nothing, again… nothing.

If that’s not weird enough, it gets more weird. If I tap with a DIFFERENT tag and alternate them, tap A, tap B, tap A, tap B… no problem! Works great!

Why on earth is it refusing to read the same tag more than once??? And how in the world does being on battery vs PC power have anything to do with it? It’s this kind of issue that makes it very disconcerting whether your app and device will work outside of the dev enviro!!!

Clearly not a power issue, as it reads [em]alternate [/em]tags quickly, just fine.

More info:
non-issue on the Spider, works fine on same battery. Only happens on the Cerberus.

So I tried a different port on the Cerb. Same issue. So lame. It appears to be something about the Cerb and RFID combo, something is not getting along.

BTW, off topic - I’ve never experienced BSOD reboot of PC with the Spider. Cerberus seems great at BSOD. Happened twice today when pulling the USB cable or hitting reset when things appear locked up. Cerberus also gives me grief deploying in Release mode. So not impressed with Cerberus right now.

Narrowing down:

Spider + RFID + USBClientDP + PC USB Power = Good
Spider + RFID + USBClientDP + 5v USB Battery = Good
Spider + RFID + USBClientSP + PC USB Power = Good
Spider + RFID + USBClientSP + 5v USB Battery = Good
Cerberus + RFID + USBClientDP + PC USB Power = Good
Cerberus + RFID + USBClientDP + 5v USB Battery = Good
Cerberus + RFID + USBClientSP + PC USB Power = Good
Cerberus + RFID + USBClientSP + 5v USB Battery = Bad

What is wrong with this combo? Should not be an issue.

Did you measure the 5v coming from your battery?

I do need to do that. However, even without that knowledge factored in, consider the following:

The same ~5v battery powers the Spider with the RFID, even with the T35 screen also plugged in! And bear in mind, it also works fine on the Cerb as long as DP (instead of SP) is used, as indicated in my little grid of tests here. I’m having a hard time blaming voltage level, but I will check it nonetheless. Such a strange situation!

This does not sound like a voltage issue, but more like a current supply issue.

Most battery packs you talk of limit the output current to around 500mA as they are designed to emulate USB power but in reality this has been around 450mA for the ones I have tested.

Some can give out 1000mA.

Are you able to measure the current drawn by the Cerberus + RFID + USBClientSP when running from the PC USB?

It is a current issue for sure. RFID uses induction for communication and may ask more current when reading a tag (there might even be a different current requirement while reading a new tag or reading a tag twice). Check the spec sheet of the rfid reader for the requirements.

Mystery goes deeper.

Referring to my earlier post with the chart of tests, check this out:

Cerberus + RFID + USBClientSP + 5v USB Battery = Bad (when battery is “full” and I get 4.85v (under idle load) from the SP)

Cerberus + RFID + USBClientSP + 5v USB Battery = Good (when battery is partially depleted (exact voltage not measured).

This is just getting so WEIRD! Works fine when the battery is NOT as full???

Sorry, I’m just not seeing how this is a current problem. How is it that the same battery works fine in this more current hungry setup:
Spider + RFID + USBClientSP + 5v USB Battery = Good (even with T35 screen plugged in!)

…but not the “Bad” setup mentioned?

Furthermore, how is it that a battery capable of dropping 500ma or more struggling to produce a wee 100ma for this setup?

Wow, how would anyone explain this! We will see if we can repro on our end.