Relay eblock drop out

Hi,

I am using the GHI-RELAY-192 eblock and the Panda II to switch between a battery charger and a load for testing Li-Ion batteries. The test pulls the relay in to apply the load, and this can take up to 3 hours to drain a battery. I have 3 test systems, and two seem to have an intermittent relay. I wrote a separate app to test the relay, there is only 1 line of code to pull in the relay (in an interrupt routine) and no code to release the relay. What I see is that the relay will release randomly (usually after an hour) and then pull back in after 10 minutes. (I log the data to an SD card). I am using a digital pin to control the relay, and I have moved to different digital pins with no improvement in reliability. When the relay releases, I can see the status LED flash on the relay module.

Since I am pretty sure I have eliminated the software as the cause, I wonder if anyone else has experience a problem similar to this.

What is the DC current rating of this relay? Switching DC loads is different than switching AC loads and if your switching a higher DC load than it can handle you will be breaking down the relay contacts and getting intermittent behavior.

So what is your SD card logging showing you ? Does your Fez reboot? Why is 10 minutes relevant, do you do periodic checks? If your code only ever fires the engage-relay once, how does it do so after 10 minutes? Do you have watchdog enabled?

The relay us rated at 10 A @ 125 VAC. my battery sources 1350 mAh @ 7.4 VDC. the resistor is 18 ohms, so the calculated current is approx 410 ma, which I think is within the tolerance of the relay.

The test code samples the state of the relay every minute, and logs the time and state to the SD card,
and since there is only one statement to engage the relay, (no release statement in the code) I can eliminate the software.

My question is, has anyone had problems keeping these e-block relays engaged for long periods?
I have tried moving to different digital pins to control the relay, with the same results.

sounds like you are drawing too much current from the voltage regulator, or dropping too much voltage, and it goes into thermal overload. After a cool down period the regulator reactivates.

What do you have connected to the internal 5V power supply?

How are you powering the device?

Check if the regulator get ver hot.

[quote]The relay us rated at 10 A @ 125 VAC. my battery sources 1350 mAh @ 7.4 VDC. the resistor is 18 ohms, so the calculated current is approx 410 ma, which I think is within the tolerance of the relay.
[/quote]

Depending on the contact material the relay uses .4A may not be enough to keep them clean, switching DC currents tends to erode contacts much more than AC.

[quote]The test code samples the state of the relay every minute, and logs the time and state to the SD card,
and since there is only one statement to engage the relay, (no release statement in the code) I can eliminate the software.
[/quote]

Sampling it how?

The easiest test would be to hard wire the relay coil to the on position and see if it ever drops out. A picture and schematic of your set up would also be very helpful.

If your only turning on the relay that has 7.4Volts across it and running only 410 mA, then the relay contacts should not be overloaded in anyway as they can handle 10A (about 50X your current amperage) and 125Volts (about 15tims your current voltage).

I’d also (as MIKE suggested) look at the heat output of the Panda board voltage regulator. If it is very hot, then the relay is drawing too much current from the Panda board and the voltage regulator is shutting down until it cools off.

Finally, can you measure the current going/coming from the Panda board to the relay.
Measure the current on the control pin (which should be very small) and measure the current on the RED power line (which is what powers the solenoid in the relay).

Is there a part number on the Relay - go to findchips.com to locate a datasheet and determine current draw of solenoid on relay. If the relay solenoid is drawing too much current, then may need to alter Panda power design to supply more power or find another relay that draws less current.

If the control pin from the FEZ is trying to sink/source too much current, then the port pin may be overloaded. If that is the case, then you’ll need to put a small NPN transistor between the control pin and the control input to the Relay.