Picaxe behaving badly............

[alphamike27]

I recently completed a solar powered LED lighting (for a house) project. I will post the project after I get time to clean up my code so I don’t embarrass myself.

Basically the project involved monitoring the presence of mains power and the ambient lighting conditions to switch LED modules off and on automatically.

However, I began to experience some very erratic operation of the Picaxe (a 08M) whereas outputs would randomly switch on and off and inputs would change to outputs and vice versa.

Fortunately, I included some “smoke preventing” 330 ohm resistors on the pins which prevented any damage

Suffering from much self doubt, I labored through my coding to see if I could ascertain the problem. However, the few small changes that I affected made no difference, and, as the Serin pin was pulled to ground through a 10k resistor to prevent interference, I was at a loss to explain.

To cut a very long story short, I am running a PLC controlled hydroponic control system in the same cabinet (on a completely different power supply). Apart from lights and water pumps, there is also small air compressor (as used in home aquariums) in the cabinet.

This was obviously broadcasting some pretty strong hash, because after installing an earthed perforated steel shield around the compressor, the problem disappeared.

Just thought I would share this in case others have had similar problems………………………….

 

[westaust55]

Many of those aquarium air pumps are really just a vibrator coil driving a diaphram with reed valves to get the pumping effect.
Long time since I had one apart but guess as they are an AC coil in a plastic housing little barrier to RF noise emission as you have found.

 

[hippy]

Thanks for the feedback, identification of the problem and solution. Good advice on the current limiting / protection resistors on pins which can switch from being input to output. This case demonstrates just how potentially catastrophic a failure can be and is why posters on the forum are usually very cautious about PICAXE and other home-built systems used in automotive environments or in other safety critical situations.

These days most commercial products should be 'EMC friendly', neither generating nor being affected by radiated energy, interference or electrical noise but the real world is rarely perfect and an exposed PICAXE circuit can be affected by such interference. That said, I would expect it has to be pretty extreme interference to cause the sort of problems you were having with the PICAXE.

Shielding is always useful but doesn't in itself prevent pick-up through power and connection cables but it's always a good idea, around the PICAXE circuit, and particularly around whatever is causing the interference. Better to mute the source of interference than have to protect everything from it.

I'm pleased you've tracked down the culprit and hopefully your PICAXE will continue to have reliable operation.

 

[gengis]

As westaust55 says, common aquarium pumps are just a coil and metal or magnet that vibrates with the AC. That should not cause interference unless perhaps the magnetic field was inducing voltage into nearby wires or high impedance inputs?

Were you able to ascertain the transmission route of the noise?

A long time go there were 12VDC pumps that used a weighted metal reed (to lower the resonant frequency) and set of contacts that interruped the magnetic field to create a pump using a bellows and check valves. They were unreliable and I haven't seen one in ages. But those did cause interference. I'm wondering if you have something like that.

OT: some of the "solid state" automobile fuel pumps here use switching transistors and a kind of "Bendini Motor" to produce low frequency mechanical movement with a high excursion (large pump volume - relatively speaking). A power transistor is biased "on" causing a solenoid to pull a magnet against a spring - the motion of the magnet induces a voltage in a separate coil that switches the transistor off allowing the spring to push the magnet back again. It replaced the old electromechanical contact type of pump cars some cars used.

 

[alphamike27]

Thanks for the feedback guys.

The power supply was well filtered, 15V zener across the 12V input to the LM7805, 250 mfd on the 5V load side. Only the Picaxe is connected to the 5V supply with a ULN2003 sinking the 12V load for the LED's

I was not really able to track down exactly where the interference was was getting into the system, but as the screen fixed the problem, it was obviously some sort of radiation

I am pretty sure that the pump is the type that Westy referred to, does not move a lot of air, but it was cheap (or so it seemed at the time........lol)