Relay returns to open within moments of closing

I have a relay control board that is computer controlled through serial com ports (RS232) The relay is connected to a 12v Battery and a 12V pump. Unfortunately, about 90% of the time the relays will shut off (return to open) as soon as they close. They don't stay on long enough for the pump to spin up. The 10% of the time the relays close properly everything works fine.

So far my best results happen when I have the amp meter in series with the pump to measure running amp draw. The pumps start up around 75% of the time without incident. I'm not sure what could be causing this behavior but it is repeatable.

Do you have any suggestions as to what could be causing this problem?

Relay rated 5amp
Pump draws less than 2.1amps running
12V DC Large capacity battery

I have tried putting 2 ohm resistance inline with pump to reduce startup amp spike. I have tried parallel capacitors to suppress induction voltage spike.

First thoughts:
An electric motor draws a much higher current than normal when it is starting up, even more so if it is starting under load conditions.
The battery voltage will drop significantly when its starts to supply current to a motor, particularly if the battery is old or weak or has little charge; this reduction in voltage may then quickly cause the relays to "drop out" from under-voltage to their coils.

Does the control program have a function that detects a low voltage condition and then cuts the power ?

There is also the distinct possibility of a simple loose contact or wire that is being affected by movement or vibration, producing the (difficult to trace and utterly maddening) intermittent fault.

If the system is old, the there may be corrosion between parts, causing additional resistance and voltage loss in the circuit.

Frank Bushnell is right,
To isolate the problem apply the power to relay control board from separate power source and see if the problem exist..

I agree with Frank, i would say either your battery Ampere hour capacity is less or discharge rate is high. since the load is a pump motor, it consumes a large inrush current on closing of relay. If the load is duty cycled, the current consumption will be more. Hence i would say you cross check your battery capacity.And yes as Jawed said , its better to excite the relay from another source, even if the problem exists, its better you replace it................

Thanks for the responses! I've added a regulated power supply to control the relay board which was previously being powered off the battery.

The relays are still dropping out. This is all new wiring so corrosion shouldn't be a problem.

The battery I'm using was taken from an emergency jump starter for automobiles so I'm hoping it has sufficient amperage to supply the pump. Currently it's reading 13v so it is close to a full charge.

I've uploaded a 15 second video to show the problem.
http://www.youtube.com/watch?v=YBI9sFMy0w8&feature=youtu.be

Here are some silly questions:
* What is the nominal relay coil voltage (i.e. the voltage on the coil at which the relay is specified to operate)?
* Is that voltage being applied to the relay coil?
* Is it being applied consistently -- i.e. every time and for the proper duration? In other words, have you used a meter or scope to monitor the voltage being applied to the relay coil?
* Have you tried applying the rated voltage directly to the coil of one of the relays that is failing to hold? In other words, have you tried removing one of the relays and testing it separate from the system?
* Are there flyback diodes across the relay coils (to prevent back-emf from getting to the driving circuitry)? If not, then something may have been damaged, so if the problem persists even after adding the diodes (one per relay coil), then you may need to replace one or more drivers.
* What about the RS232 data. How is the integrity, there? Are the voltage levels consistent between the transmitter and the receiver? In standard RS232 a '1' is defined as a negative voltage between -3 and -15, yet often somethin closer to zero volts is used instead. If the RS232 data is coming from, for instance, a PC, but is being sent to a Microcontroller, then there will, likely, need to be some voltage translation from something like -12volts to zero volts, and probably from around +12V to +5V or less.

It's more about internal resistance in the battery than Ampere-Hour capacity -- though one may follow the other (but not necessarily).

Hi thanks for the good questions! I hadn't thought to ask some of these. Some of these I wont be able to answer till Monday and some of these I can't easily try at all. The control board is manufactured by a 3rd party so I don't have a ton of insight into the wiring of the relay's themselves.

The manufacturer did tell me that there is no over voltage / over amp protection on the relay controls. They suggested that the voltage spikes from the collapse of inductive magnetic fields can interfere with their controller logic.

The relays board is powered by a 12v DC power supply so I'm assuming the relays are 12vdc. All the relays work consistently when there is no power being passed to the pumps. This makes me think it's something to do with the interaction between the pumps and the relay.

The RS232 data is actually being transferred through a bluetooth module built into the board and seems to be working correctly since everything works without load.

There is 1 feature of the board that will retry a given command N number of times. When I set this command to 10 the pumps do turn on and start running after a few attempts. I suspect this is my best clue as to what is preventing them from working first time.

On Monday I will be able to connect my scope to the control side of the relay and to the load side and I'll post back my results.

You say, "All the relays work consistently when there is no power being passed to the pumps."

Now that you have a separate (presumably reliable) 12V supply for the relay board, this suggests the problem could be that an overload cutout is functioning.

This is possibly because the safety cutout (over-current relay) is set to cut out at a current level very close to the maximum current drawn by the pump (which is when it starts up).
Therefore its operation would be erratic, and it may need to be adjusted to a slightly higher current level.

Also. observe how much the battery voltage drops when the pump starts, to get an idea of whether the battery condition is sufficiently good for its current purpose. (Pun intended

I hooked up my scope to the pump circuit and there are some pretty high voltage spikes. I couldn't get this graph to export with the units but the V-max is 95V and the V-min is 200V.

The control board tries to close the relay 10 times. Looks like it took 6 or 7 tries before the pump finally started up. What do you think I should try to smooth out these spikes?

Here's another potential issue. This is the voltage measure across an 1 ohm power resistor. I'm trying to see how much current the pumps are drawing but I'm not sure if this makes sense. It looks like the peak load is 11 amps which isn't completely unreasonable. But 0 amps while running doesn't seem right.

Screen capture was dropped.

Before you commence further on this topic You need a circuit of your serial control bourd

Thats the most critical part of the setup

One

You want opto isolators on the ports of the Rs232 reciever

Two

You want to ensure that tracks in and out of the relays to motors,pumps etc are kept as short as possible

Three
Your power supply to recieving logic controller needs to be highly stable

Four
You need to check for noise on the comms side that means any spikes etc which can and does disrrupt serial communications btw Dte and dce

FIVE
Driver transistors for relays etc must also be decoupled from collector to grounding depending on type of transistors etc used In some cases you might be better off using Fets or Igbt dependant on switching speeds of relays etc

Six
OYou should if possible implement for important design a return protocol that should implement error correction, parity checking

SEVEN

You also need to decide on synchronous or asynchronous rs232 communications and advantages, Dis advantages plus of course propagation time

I. E How quickly can you get control messages from DCE to Dte relative to switching speeds required by relays and motor start circuitry and back again

Add with this web control or internet control and this becomes a vast and complicated topic requiring Network engineering plus network programming, Network traffic analysis

Eight
What is the maximum permissable length of an rs232cable just as important Shorter the better other wise you get data corruption

Nine

Power supply driving logic control bourd Emf from transformer etc is enough to corrupt data to and from DCE to DTE Ive had that problem as well

Ten out of Ten

Your mains into power supplies motors logic power supplies etc also needs to be filtered otherwise with high emf present You get data corruption in otherwords message not delivered retry retry retry retry
system hangs and buffer not empty or has corrupt data pushback or skip character

This is why I build design myself This is the only way you will ever learn and you learn purely from mistakes

Those are just a few points that i can relay to you which ive found by shear hard horrible experiance and thats with small design which can take hours and hours to debug nevermind coding or Controlling front ends which should also be multi threaded

Finally is Rs232 the right answer to this and in some cases your answer may well be no its not and requires a more robust guaranteed answer that requires your controlling messages be always delivered with ack an nack That may well mean Rs434 or I2C for example In extreme cases TCP/Ip dependant on your criteria so its not all down to just serial coms It may well be that Usb may be the answer too

Very difficult Ive built two so far and its by no means easy
But yes like all subjects it can be done