Reset Loop

Sal Ammoniac

I've built a data acquisition system to collect and transmit environmental data to a base station. It runs on a 6v gel cell battery, which lasts about 3 weeks before it needs to be swapped out for a fully charged battery.

Whenever I swap out the battery with a charged one, the system gets into a mode where it's resetting itself continuously for around five minutes until it settles down and starts transmitting data to the base station. Once it gets out of the reset loop, it works flawlessly until the battery dies or until I replace it with a charged battery.

I can't figure out why it keeps resetting itself for the first five minutes of operation after replacing the battery. The board is a custom layout using the DIP version of the Propeller and all through hole parts except for the 4.096 v reference, which is surface mount.

I've attached the schematic--anyone have any ideas why it gets stuck in a loop resetting itself for five minutes every time I change the battery until it settles down and runs normally for weeks?

BTX

Hi
Maybe trying with a 0.1uF cap in the input of IC3 (very close the input pin).
If this wont work, maybe could be usefull take a look at your PCB.
Also, what about the pchip decoupling ? I always use 1.0uF close to the pchip Vcc. (Pin 8)

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Regards.

Alberto.

Envio editado por (BTX) : 5/9/2010 2:42:05 AM GMT

Mike Green

The only things that could cause a periodic reset would be something connected to the Reset line or a power fault of some sort. It looks like you can connect a PropPlug, but there's normally not one there so that doesn't make sense. Given that this behavior appears when a fresh battery is connected, it may have to do with the fact that a freshly charged battery has a relatively high terminal voltage that drops rapidly as it's first used. You may have inadequate heatsinking on your 5V regulator and it's shutting down with the additional heat load until the battery voltage drops enough. This is kind of odd since your device doesn't seem like it would use a lot of current, particularly with it running off the battery for weeks. There may be some initial condition that results in higher current drain until the Prop program completes its initialization.

Sal Ammoniac

Thanks for the comments.

When I have a chance, I'll remove the cover of the device (it's mounted in a plastic box which itself is installed in my weather station enclosure) and check the temp of the 5v regulator with an IR thermometer when I connect a fully charged battery.

The entire circuit draws 17 mA from the battery, and even after it's been running indoors for several hours neither the 5v or 3.3v regulators have ever been even warm to the touch. In fact, when I was testing the device in my lab, I was running it from a 9v bench power supply, which is almost 2.5v higher than the battery voltage when fully charged (6.5v).

When I replaced the battery this evening the ambient temperature inside the box was 60F (there's another LM34 inside the box). The ambient temperature inside the plastic box has gotten up to 90F here in the last month or two and it's continued to work flawlessly. It didn't go into a reset loop under these conditions, which seems to only happen immediately after replacing the battery.

Toby Seckshund

Does the current draw stay consistant, at the nominal 17mA, when the fresh battery is put on?

I was just wondering if the quiesent current of the reg is very high and this flattens the battery until the reg turns into a piece of wire. The Nat Semi datasheet I looked at shows a very high spike of quiesent until the input gets to about 8-9V. If that is true, then so much for its low dropout spec !

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Style and grace : Nil point

Sal Ammoniac

Well, I've also had another problem with this setup -- data corruption on the XBee data link. The data stream to the PC inside the house would often become seriously corrupted or drop out entirely for minutes to hours at a time. I tried everything to fix it, including replacing the XBees on both ends of the link, using better antennas, etc., but nothing seemed to cure the problem. I eventually attributed it to the hostile RF environment in my neighborhood.

I finally bit the bullet and got out my cordless drill and masonry bits, drilled a hole in the wall of the house, and ran a cat-5 cable between the weather station enclosure and the PC in the house. Since I was running a cable anyway, I used one of the unused pairs to run 7v DC to the box and eliminated the need for the batteries entirely.

It's been running 100% reliably ever since--no data corruption and no strange reset loops. Now it starts immediately (after the one second delay while the Propeller reads the EEPROM) without needing a five minute settling down period.

So it looks like replacing the battery with a wired power connection fixed something, although I have no idea what.

Phil Pilgrim (PhiPi)

Sal,

Did you have a Prop Plug in place, but disconnected from USB, when the reset problem occurred?

-Phil

Sal Ammoniac

Nope. The Prop Plug was only in place to program the EEPROM and then immediately unplugged the Prop Plug from the board.

Post Edited (Sal Ammoniac) : 5/18/2010 7:10:59 PM GMT