Fuel pump contact breaker points

[Trevor Rymell]

Hi
My motorcycle uses an old style 12VDC fuel pump with a solenoid and a set contact breaker points.

The points are subject to arcing which can shorten their life and there is no suppression built in such as the condenser (capacitor) that was used on distributors before contactless ignition systems became common.

I've read that the arcing can be suppressed or reduced by putting a diode in parallel with the solenoid. See here for example:

**broken link removed**

The explanation sounds ok but I would like to hear from automotive electronic experts if they think this would do any good. Why I have doubt is that if the solution was so simple, why wouldn't the manufacturer put a 10 cent diode in in the first place?

Thanks in advance.

Trevor

 

[MikeMl]

The creation of the pump might predate the invention of Silicon?

At what rate does the pump pulse? Putting a snubber diode around the coil might greatly change the pulsing rate. Will that effect the fuel delivery rate?

 

[Trevor Rymell]

Thanks for the reply, Mike

Well, it's a 1996 model so not as ancient as that! Old enough to be using pre-historic fuel pump and carbs though, but modern solid state TDI ignition system.

A "snubber" diode? Is that the correct term? And when you say "around" does that mean in parallel? Sorry, I'm a bit out of touch.

The pump pulses every couple of seconds or so when it has filled the float chambers of the carbs depending on engine RPM and load. If for any reason, the pump is empty it will pulse rapidly about 30-40 pulses/sec I'd guess.

Why I mention that is because every so often, I turn on the ignition and the pump will race away at full speed because it's empty or has sucked in air from somewhere. If I disconnect the line to the carbs and switch on, the pump will prime itself and all's well till next time. I haven't found the cause of this yet. I believe that these incidents are shortening the life of the contacts as I've just had to put in my third pump in 98,000 miles. I'm told this shouldn't be necessary.

I've got the old pump on the bench on a 3A bench PSU to watch for arcing across the points. There was a small blue spark as you'd expect. I tried adding a 1N5402 across the solenoid but could see no visible difference. Maybe I'd see more variation if I was running the pump off the 16Ah motorcycle battery(?)

The creation of the pump might predate the invention of Silicon?

At what rate does the pump pulse? Putting a snubber diode around the coil might greatly change the pulsing rate. Will that effect the fuel delivery rate?

 

[MikeMl]

...
A "snubber" diode? Is that the correct term? And when you say "around" does that mean in parallel? Sorry, I'm a bit out of touch.
...

Inverse parallel, meaning that if the breaker points are on the negative (Ground?) side, then the cathode of the snubber diode would be connected to the plus end of the pump solenoid, and the anode would connect to the end of the coil that is also connected to the points.

 

[debe]

What you propose to do will work fine. I used a 400V 1Amp diode, It didnt alter the speed of the pulses that i could notice.

 

[Trevor Rymell]

Thanks, debe, Mike

I connected an 1N5402 (200v, 3A) diode in parallel with the solenoid and ran the pump off the vehicle battery. There was no observable difference in the size of the spark between the points. Should there be?

Thanks again.

Trevor

 

[MikeMl]

...I connected an 1N5402 (200v, 3A) diode in parallel with the solenoid and ran the pump off the vehicle battery. There was no observable difference in the size of the spark between the points. Should there be?...

Just to make sure, the Cathode of the diode is soldered directly to the end of the coil which ultimately gets fed from +12V? The anode of the diode is soldered directly to the end of the coil which ultimately gets connected to ground?

You know that the diode is good as checked with an Ohmmeter?

 

[KeepItSimpleStupid]

You can do a couple of things:

1. Put a 0.1 uf to 1 uf ceramic cap across the contacts.
2. You can put a bidirectiional ZNR rated for about 18 VDC across the contacts.
3. You can convert the high current contact to a low current contact although you want contact to switch at least 10 mA.

 

[Trevor Rymell]

Mike, yes, thanks I got that correct but as I said, I can't tell from looking at the points in operation if it makes any difference.


KISS, thanks for the suggestions. I'm sorry to keep asking more questions but would the cap be instead of the diode or in addition?

And again, what observable difference should I see if any in terms of visible arcing? Or is there any other way to tell if this is actually prolonging the life of the pump contacts?

Regards

Trevor

 

[MikeMl]

Something is fishy. The diode should almost completely eliminate the arc at the contacts. However, adding the diode should also change the rate at which the pump pulses because it slows the release of the pump's armature after each pulse. Since you are not seeing any change in the sound or pulse rate, I am still thinking you do not have it in the right place, or that the diode is open....

Can you post a picture of what you are working on?

 

[KeepItSimpleStupid]

The cap would be in addition to.

The diode we are talking about in parallel does nothing for the making of the contact. I does, however do something when the contacts break.

The contact has to be attacked at the contact. A ZNR would limit the voltage across the contact. If sized right (about 18V).

One think you have to remember about a capacitor is that the voltage across it cannot change instantaneously. Generally the reason to use small ceramic capacitors across contacts or even motors is to help eliminate RFI which is actually caused by the arcing of the brushes.

 

[MikeMl]

Look at the attached sim of three identical coils which have a series resistance of 4Ω and an inductance of 0.02H. I have recorded similar traces from real relays and solenoids. Compare V(nodiode) to either V(diode) or V(hiside). Note that without the snubber diode, the voltage across the coil reaches 15kV, which is enough to make an arc. With the snubber across the coil, the voltage is clamped to one diode drop above the supply voltage, or one diode drop below ground.

Note the current I(L2) which shows how the coil current keeps flowing (through the diode) after the switch opens.

For KISS's benefit: there is no arc or transient when the switch closes, only when it opens. The inductance of the coil forces the current to build up slowly [see the I(L2) trace]. Without the snubber diode, it is the inductance of the coil that creates the 15kV voltage spike when the switch is opened.

 

[Trevor Rymell]

Hi Mike

More likely I'm just doing something wrong. Here's a pic of my pump showing the contacts assembly. I hope it's useful. Sorry, I haven't learned how to embed a JPG in a post yet.


**broken link removed**

I think I have the diode connected correctly ie cathode connected to +ve side of solenoid (RH black wire with barely visible blue tracer), anode to the points side of solenoid (LH black GND wire). Diode is a fresh 1N5402.

Thanks again.

Trevor

 

[Trevor Rymell]

Mike, KISS

Re: your last two posts. Many thanks for your continued assistance.

Mike, the screen shots of the sim are above my head but I understand it shows results with and without the diode. I don't have access to anyway of measuring the benefit of adding the diode but can certainly see no change in a) size of blue spark at points when run off the vehicle battery, b) no discernible change in pulse rate. Bear in mind that the pump is pulsing flat out as it's empty of fuel. When pumping fuel it only pulses once or twice a second or so.

The intended outcome of all this is simply to extend the life of the points whenever the pump is running flat out as if empty. This is I believe the reason I'm getting through pumps so fast as detailed in my original post.

The advice you've given so far is

1) a properly connected diode should reduce/eliminate arcing when the points break. The diode is "snubbing" the high voltage spikes as the field in the solenoid collapses(?)
2) I should use a zener diode (as the one in my pic in parallel with the solenoid?) AND a cap in parallel with the points(?) I assume the cap would function the same way as the old condensers used in 70's/80's car distributors which makes me wonder why a mid 80's designed bike wouldn't have one fitted in the first place. The question is academic of course.

Thanks again to you both.

Trevor

 

[MikeMl]

Looking at the pix on photobucket, I have some questions. Let me ask this. You used an Ohmmeter to measure the resistance between the two wires that go into the body of the pump? What is the resistance?

How does the pump sense fuel pressure? Is there a diaphragm in the pump which somehow pushes on those contacts at the top of the picture?

If there is no push rod, is the timing of the pump determined by some sort of bi-metal thermostat with a heater under it? If so, then the contact arcing is likely being caused as the bimetal heats/cools, the contact points dont snap, they open/close sloooowly, so repeatedly make/break during the transition. If this is the case, then the arcing is likely being caused by the current which heats the heating element.
Since the heater is a resistor, that cannot be suppressed by a diode or capacitor. The arcing is caused as the bimetal pulls the contacts apart, the area though which the current flows through a progressively smaller cross sectional area, burning it... Same problem as the small non-snap action bimetal thermostats used with fish tank heaters....

 

[KeepItSimpleStupid]

I gave another piece of advice, but it's more complicated to do, but possibly the best way in the long run.

The switch from the pic looks like it switches the (-) lead, which could make things more complex.

Would you happen to know the amount of current the pump draws?

It might be possible to isolate the switch or just have the contact closure to ground and replace the actually switching element with a solid state switch. You might have to add an external waterproof box and heatsink the solid state device.

Being able to isolate the switch somewhat. Contact closure to ground. A contact closure to (+) would be preferable.

This https://www.electronics-diy.com/electronic_schematic.php?id=1012 has the wrong sense, I believe, but you get the idea.

 

[KeepItSimpleStupid]

I did some looking around, and a device such as this one: https://www.electro-tech-online.com/custompdfs/2012/10/mic4451.pdf (TO220 version) Probably the inverting variety would provide a nice interface to the existing switch. It handles 12 A peak and is about $3.00. But still a lot has to go into using the device and it would be better to find a similar part that is automotive qualified. I don't know if this is. I'd add two diodes for spike suppression, a recommended bypass cap, a metal flanged gasketed enclosure (and a place to mount it, probably a big problem), then you need a thermopad and/or heatsink grease, a TO220 mounting kit. When your done, you won't have any issues. You would need the diode across the motor as well.

I mounted a similar packaged device, but no extraneous parts on a lawn mower, but you don't operate a lawn mower in the rain. The package was attached to the lawn mower, wires were soldered and heat shrinked to the package and away we go. That's still possible here.

 

[Trevor Rymell]

Hi Mike
Sorry for the delay in responding. The solonoid measures 3 Ohms on both the old pump and the brand new one I'm about to put in. It's a simple push rod and diaphragm pump and as far as I can see there's no by-metal strip as part of the contacts assembly so have no idea if or how the pump is timed. I just assumed it doesn't need to have any kind of timing system as it pumps on demand as the back pressure from the carbs goes down. In other words, the float chambers of the carbs fill, the float valves close off and the back pressure makes the pump stop. It only runs free when there's no resistance from the fuel flow. Apologies for my poor technical explanations.

trevor

Looking at the pix on photobucket, I have some questions. Let me ask this. You used an Ohmmeter to measure the resistance between the two wires that go into the body of the pump? What is the resistance?

How does the pump sense fuel pressure? Is there a diaphragm in the pump which somehow pushes on those contacts at the top of the picture?

If there is no push rod, is the timing of the pump determined by some sort of bi-metal thermostat with a heater under it? If so, then the contact arcing is likely being caused as the bimetal heats/cools, the contact points dont snap, they open/close sloooowly, so repeatedly make/break during the transition. If this is the case, then the arcing is likely being caused by the current which heats the heating element.
Since the heater is a resistor, that cannot be suppressed by a diode or capacitor. The arcing is caused as the bimetal pulls the contacts apart, the area though which the current flows through a progressively smaller cross sectional area, burning it... Same problem as the small non-snap action bimetal thermostats used with fish tank heaters....

 

[Trevor Rymell]

KISS, many thanks for your detailed help and suggestions. I'll reply properly when I've had a chance to look at this.

Trevor

 

[KeepItSimpleStupid]

Trevor:

Here **broken link removed** is another tutorial to look at. This one uses Bipolar transistors. The transistors used in the tutorial can only handle about 0.2 Amps. Bipolar transistors are controlled by a current injected into the base. They have a current gain (Hfe) The collector current = Ib*Hfe and because they are a silicon diode they have a 0.6 V or so voltage drop across them. They comein N and P channel versions which affects the polarities. Theire are "digital transistors" which have a couple of embedded resistrors in them, but they are used for low current.

FETS (generic term). MOSFETS would be used to switch power also come in N channel an P-channel. Ggs or the Voltage from the source lead to the gate determines how much the FET conducts. LOGIC FETS are designed to turn on with a much lower Vgs. MOSFETS have an rds (Resistance Gate to source) term. This is the resistance of the source to drain lead when the FET is totally on. Hence MOSFETS generate less heat when fully on and are used for higher currents.

Suppose the contact in your fuel pump was connected to ground and the other end could be free. It would be ideal if a pull-up resistor would draw 10 ma, so R < 12 V/0.010 Amps would be a nice value. When the contacted closed, there would be at least 10 mA of current flowing through the contact. At lower currents, it's possible that the contacts won't stay clean enough to reliably actuate.