Line Noise from Electric Diaphragm Pump

I would think that a cap would smooth out the signal for you. But I am not the one to tell you what cap to use. Good luck.

Yea, I'm going to contact the mfg. of the pump this morning to discuss that. I installed a line noise filter, one that we normally use on our power supply mains, and that eliminated ALL the noise. However, that is an expensive option. It will work for the prototype, but when the system goes into production, a cheaper method is desirable.

Where the AC goes to the pump wind both Live and Neutral through a ferrite core as many times as you can, 3 or 4 generally works well, add an X rated capacitor say 0.01uF or similar across L and N on the pump side of the ferrite coil, or one on each side.
We use ferrite cores about 10mm long 8mm dia with a 5mm hole through, but any one you can get your hand on will do.
You will find one in the 'bulge' in a laptop powersupply cable, or in any of the electronics catalogues/stores... digikey etc.

This is just 'wet finger' to design as special filter would be rather complicated and a waste of time, as often the theory and practice don't coincide with this sort of problem.
(Some RF bod my want to take a pot shot at me for such heresy...buy I call it as I see it)

Del

Thanks Del. I actually ordered a few yesterday. We use them in other applications, but the ones we stock were too small for a few wraps of the 18GA wire to go through. When they arrive, I'll compare them to the results I got with the mains filter I mentioned earlier.

You can stick the same sort of filter on any DC lines going near the pump.
Some people will tell you to add extra 'C' onto the 5v rail and suchlike.
My experience is that, it's no good trying to quash the noise once it's on the board. You have to stop it getting on the board in the first place.
This is a bit of a simplification of course, but without an EMI lab to play in, it's your first/best shot.
Good luck...you may well need it as this sort of prob can be a real *&^%$@~#
Del

I tried similar measures with caps and re-routing DC lines, using sheilded cable away from AC, but it gets on the board directly from the pump hot line. I agree, you need to keep it from getting onboard and so far that's what the mains filter inline with the pump has done. We'll see how the ferrite works also.

add an X rated capacitor

x rated capacitors are not widely available here. They are only sold in "adult " shops.

You can always go "across-the-line" to Georgia to pick some up. They delibrately mislabel them as "coupling capacitors" there to fool the law.

I have had good luck using common line filters like these: http://www.corcom.com/Series/PowerLine/B/ on the ac feed to noisy devices. The filter should be located as close (electrically) as possible to the load and a good ground must be provided to the grounding terminal on the filter.

If you look around, you will probably find one or more in some junk equipment to steal for a test.

I have even used these line voltage filters to filter 24VDC noise from a motor in a printer drive to keep it from interfering with the sensitive recording electronics signal amplifiers.

What can I say?

I was desperate!

It worked!

We had a ton of them in stock from a discontinued product!

I wuz a HERO!

Another thing I have found to help in noise issues with AC powered loads is to use a solid state switch with a zero crossing function. A lot of SSRs (solid state relays) use zero crossing detectors to allow all switching to occur on the zero crossing. This minimizes voltage and current spikes on switching. This only works if it doesn't matter if the starting or stopping is delayed 1/2 cycle of the AC. Usually, it doesn't matter.

One caution on zero crossing SSRs is to be very careful about the loads you try switching with them. If the load has a diode in series with the power in and a capacitor hash filter on the input, It may never release after it picks up. This can lead to premature gray hair and irritated scalp from excessive head scratching. So if you see problems like the above with your new design, remember: I told you so, before you go nuts trying to figure out why, not after.

Have FUN!
TT3

Thanks, that's the exact type I installed earlier this morning and it worked like a charm. We also adjusted the debounce in the software and now it works flawlessly.

Del, I tried the 3 or 4 wraps on the ferrite, but it didn't do the job for me. Cleaned it up a bit, but the line noise filter was the way to go, although the cost is not desirable.

To all, thanks for the input, PROBLEM SOLVED!!!

Hi Aces,

i see you have your problem solved but i would ask the question, Why have you used a pump with a brushed motor when an induction motor would do?

Induction motors don't generate electrical noise, usually cheaper and they are far more reliable. You have 115Vac available. Usually the choice of a dc motor is only because there is no ac available in mobile applications. (I appreciate that the torque curves are different on these two types of motors)

regards

Chas

Why the company I work for uses these particular pumps, I do not know entirely. Currently in the industries we serve, peristaltic and AODD are the primary means of pumping. This gives our customers another choice as far as I can see. Through extensive testing, I have found there to be virtually no gain when using this type of pump. If you know of any sources for electric pumps with induction motors with a pumping range between .4gpm and about 3.5gpm, let me know. We are always looking to improve our products.

Hi Aces,

I'm in Spain, so what I get here wouldn't make sense for you but I see alot of american boats and there are plenty of pumps with Baldor and other brands of US spec motors on them.

regards

Chas

That's funny, FloJet is a direct competitor of ours.

Hi Aces,

Sorry! It's jut that it fitted the bill, i.e. induction motor & diaphragm pump. What are you producing? I supposed the pump wasjust a small part of your product

Chas

The system pumps a 2 part pre mixed chemical out to dispensing device, can't really go into detail about that. The pumps respond to various inputs. The first being a simple toggle switch which enables the system. Once enabled it waits for a trigger signal from a remotely mounted photo-optic sensor. Once the trigger signal is recieved, the user programmed dispensing routine begins. It runs sort of 2 programs at once. It will either dispense based on a time function or respons to a second photo-optic sensor. During the dispensing routine, the "enabling" switch is constantly monitored. If the pin goes low, the system goes into a "pause" state, once high, it resumes the operation. The noise induced by the pump onto the board, when measured on a scope, disrupted the +5V to the pin just long enough to give a false "pause" signal. This would interrupt the routine and cause the control relays to sort of chatter rapidly on/off - not desirable. The line noise filter worked great, but really the adjustment to the de-bounce of that pin during the routine eliminated it completely. We use these pumps because one, they are a stock item in our inventory, second, they have smoother delivery and higher capacity than our peristaltics, and third, we couldn't rely on the air supply sources of our customers to use AODD's.

So in short, yes they are a small part of the overall system, but most critical to the operation. This unit is just a prototype which is scheduled to be installed for field testing the first week of March. If the noise impacts the operation for some reason out in the field worse than in our lab, then we might be forced to use and induction motored EDP. I'll come back to this topic the second week of March to fill you guys in on how the solutions suggested worked in "the real world".