SNGB submersible pump fault

[Heuer]

I recently installed one of new SNGB submersible fuel pumps and was quite happy with the way it behaved. The other day I went out in the car and before putting it back in the garage filled it right up with fuel for the first time. Went to the garage this morning to find the over powering smell of petrol. Petrol is leaking past the gland where the power cable comes through and a large amount is in the boot area and garage floor. Fortunately I did not operate the electric garage door or switch on the lights!


I carefully removed the pump and tried to adjust the rubber sealing gland to no avail - it seems like a very poor design. Rang SNG but the person there was (unusually, I have to say) particularly unhelpful saying I had to send the pump back for inspection and replacement or repair. I did point out this would leave the car with a full tank of petrol with the top plate off which wasn't a particularly safe thing to do, but them were the rules! I offered just to buy a new pump, swap them over and return the other for a refund but they are out of stock. Any suggestions as I am living with a fully restored concours bomb in my garage at the moment .

I thought of putting some sort of sealant in there but that would screw up any warranty claim.

[Heuer]

Had a brain wave - I could reinstall the original Lucas unit. It works and has a new gauze filter but the car has been converted to -ve earth and I am not sure if the pump is polarity sensitive. Anyone know?

[Heuer]

One step forwards, two back After checking the Lucas pump was operational I went to fit it but I could not get it far enough in to the tank. Now bear in mind I am working with a full tank of fuel here but there was no way it was going to find its home. I then dipped the tank with a stick to find the depth from top of tank to bottom of sump and it was 11" - unfortunately the pump is 13.5" deep.


Does anyone have a Lucas 2FP they can measure for me? Either the sump is too short or the pump is not correct for the E-Type (although it came with the car). Can't find a date stamp and it is only marked as Model 2FP but the parts catalogue says it should be D-2FP, or A-2FP for the very early cars. Maybe it is off another Jaguar?

Spoke to the ever helpful Andy Rayner who unfortunately did not have a submersible pump in stock. He did suggest I use Petro Patch liquid to seal the gland and screws - they have resorted to using it everywhere on the tank because the Ethanol in standard petrol is playing havoc with anything rubber. Always fill up with high octane fuel is his advice.

I have now syphoned off 2.5 gallons of fuel out using a big Mityvac and reinstalled the SNGB pump because the problem only occurs when the tank is brimmed so it gives me time to work on a solution and at least the garage is safe.

[Mark Gordon]

David,
Are you sure that you are getting the pickup to drop down into the sump? The 2" or so that you are falling short of seating sound like the pickup is not going into the sump.

[Heuer]

Mark

It is a bit like fishing with a spear, the refraction of the liquid alters perspective. But having dipped the tank with a stick several times I am certain it is only 11" to the bottom of the sump. Ideally I would have an empty tank and could fit the pump with the sump off and see how much room there was. lt is very odd!

[Dave K]

David,

That 2FP pump looks exactly the same as mine and I'd have to agree with Mark something is stopping it going home into the sump.
I don't have mine at hand to measure it at the moment.

Dave

[Heuer]

Yes, but rattling metal against metal in an open full fuel tank is one of those things that qualifies for the Darwin awards: http://en.wikipedia.org/wiki/Darwin_Awards so the standard bullish approach of keep twiddling until it fits is a bit risky! I could understand the odd 1/2" but 2" is unfortunate. The tank appears to be a new Robey one but even so I can't think they have got it that wrong. :?

[stef]

David,

I also installed a SNG pump 2 weeks ago. (No leaks yet)

My old Lucas pump measures 11 1/4"

[PeterCrespin]

Your old Lucas pump is bent from being badly installed by somebody.

The depth from the tank pump flange to the base of the sump is a hair over 13" The pump overall depth should be about half an inch less to allow a bit of room at the base of the pick-up filter. The 3.8 pick-up is very easy to bend if shoved down, (the 4.2 is a strong steel tube). The foolproof way is to install the pump in a dry tank with the sump off to check pick-up placement, before installing the sump through the boot floor over the protruding pick-up mesh, Failing that, having the pick-up at the right depth and feeling it into the sump from above in a partly full tank should still work.

[38E]

David,

I have a new NOS Lucas E-Type fuel pump right here. It Measures 11-1/2" from the underside of the plate to the bottom of the pick-up tube. The gauze filter is missing so this is to the end of the copper tube.

IIRC, the pumps are polarity conscious but they are easy enough to swap over just by reversing the connections in the rubber box in the boot. I know I changed my own car to negative earth years ago and I think that's what I did at the time to keep the pump working.

[PeterCrespin]

David,

I have a new NOS Lucas E-Type fuel pump right here. It Measures 11-1/2" from the underside of the plate to the bottom of the pick-up tube. The gauze filter is missing so this is to the end of the copper tube.

IIRC, the pumps are polarity conscious but they are easy enough to swap over just by reversing the connections in the rubber box in the boot. I know I changed my own car to negative earth years ago and I think that's what I did at the time to keep the pump working.

This is true. Although not a permanent magnet motor, the 3.8 pump does only pump in the right direction when hooked up as per the factory polarity. If you switch to negative earth you have to swap the wires round too.

The depth to the floor of the tank is about 11 1/2", so if your pump is only that depth I think maybe the banjo got nudged round or the copper pipe got distorted slightly - both of which are feasible over the years (in transit or torage even). If not, then maybe it was set up at the facotry on the short side of tolerance, to leave loads of room for the filter sock to go down into the sump?

[abowie]

It should be easy enough to replace the seal with something that doesn't leak. You need something rated to IP67 or IP68 that's ok for hydrocarbons. Any good electrical supply place will have something that will fit.

Have a look at the McMaster Carr website to get some ideas.

[Heuer]

Just had a call from SNGB and they have found a maintenance spare which should be with me tomorrow. They apologised profusely about the fault and knew exactly what the problem was but preferred to send me a tested known good replacement rather than me attempt to reseat the rubber gland. Fantastic service!

As regards my original Lucas 2FP it measures exactly the same 11-1/2"as Clive's with the filter removed and Stef's picture confirms it is correct. The copper 'pig-tail' looks to be unbent, the sump looks to be correct at 4-1/2" so the problem seems to be the filter but it measures the same as another I have. Pete's suggestion is probably correct - the pump ideally needs to be fitted before the sump is attached especially as it is all too easy to bend/break the mesh filter. All academic because I will fit the replacement SNGB version and keep the 2FP as part of the car's history.

[Heuer]

Update on this. Replacement fuel pump arrived as promised early on Friday morning last week, fitted immediately buy unfortunately within 1 hour:


Called SNGB who were understandably puzzled and organised a UPS pickup of my original pump so they can investigate further. Based on Andy Rayner's advice I bought some Petro Patch (the liquid/paste version not the putty) and I am going to experiment with some bits of brass, rubber and shrink-wrap sleeve to see if it bonds them together and then leave them in a jar of petrol.


I was wondering if pouring the PP epoxy down the hole might effect a permanent seal? Have to wait and see what SNGB come back with first though.

[abowie]

David is the leak around the cable gland or where the gland fitting is held into the top plate? I have one of these to fit to my car as I rebuild it; I would prefer to fix the problem before filling the fuel tank...

[Heuer]

Andrew

The fuel is coming up though the rubber gland that is supposed to seal the cable. No problems with the top plate fixing. From what I can see it is not a very good design. Basically the cable is sheathed in heat shrink sleeving and goes through a lower brass fitting containing a thick rubber grommet. The brass fitting screws onto the other brass fixing (that is part of the bit going through the top plate) trapping the grommet and squeezing it when tightened. Unfortunately if the fitting is tightened too much the grommet squeezes out of the end distorting it and allowing fuel past. If not tightened sufficiently it will leak. Ideally the lower fixing should have a hole no larger than the cable so the grommet has nowhere to escape.



Unlikely any fuel will escape under normal filling/driving but if the car is left to stand with a full tank it becomes pressurised due to evaporation and changes in ambient temperature forcing petrol out through the gland. SNGB say the gland pressure can be adjusted, with experience, to maintain the seal integrity but my original and their carefully fettled replacement make that seem a difficult task. Three solutions I can think of:

1. Redesign the bottom fixing to have a smaller hole to match the cable diameter so the grommet can't escape and distort. A slightly thinner grommet would also be sensible to allow it room to expand and it would make assembly easier.

2. Fundamental redesign of the cable management system perhaps mimicking the original Lucas system of a braided sealed pipe but that may be impractical.

3. Apply a few drops of Petro Patch down the cable exit to permanently seal it.

The problem only came to light because I was low on fuel and filled the tank before putting the car away - not something I usually do. So until we have some guidance from SNGB best to leave the tank not quite full - I took out 8 litres from a full tank and the leak stopped. Unfortunately you will only know if you have a problem once the pump is installed and the tank is full as there is no other way to test it. You then have to remove the pump to attempt to fix it and that is a nerve racking thing to do especially when you have your wife angrily telling you of your selfishness in potentially blowing up the house and yourself Should never have told her! Incidentally is is safer to work on an almost full tank or an empty tank?

[PeterCrespin]

Almost full. The gland nut assembly will not be designed by SNGB any more than they would design a set of points - it will be a standard bought-in part sold as fit for purpose I expect. So either the choice was iffy or the assembly procedure needs tightening up (pun intended). A post-assembly test presumably needs developing - maybe attaching small batches of plate assemblies with wing nuts to a tank of water or some such? I'm not sure where the heat- shrink tube comes into it though. With appropriately-specced cable and grommet why introduce an added leak possibility?

I'd mark all tested/passed units with a dab of tamper-proof paint for sure, to confirm warranty failures....

Pete

[tinworm]

Looks like they use an electrical brass TRS gland for the pump plate / cable junction - I bet the rubber disc is not made of nitrile rubber which is what you would need for sealing petrol. Incidentally I had a similar problem with an original pump on another 3.8 coupe of mine - I found the braided original cable conduit had a rubber internal tube and this had perished allowing petrol in and up the cable as you are getting. I replaced this with a normal piece if petrol hose and a couple of 'O' clips and all was well. The anti static washer takes care of earthing out any static build up. I have also built and fitted a pump and mounting similar to the Barratt one to the 3.8 Roadster I have now, using a V8 landrover pump which is the same as they use (but only 30 quid) instead of ?165 (is it?) .

regards and best of luck Barrie

[Heuer]

I assumed the heat shrink is more fuel resistant than the PVC covered cable and did not investigate further. The rubber grommet had not been affected by the fuel as far as I can see so can't comment on the type of rubber being used. Interestingly there is a piece of fuel resistant rubber hose (says so in white writing) cut open between the cable ties and the cable to act as a buffer. Parts of it were crumbling off after a month in the tank! I told SNGB about it and included the bits I recovered from my bench in a small bag for them to analyse.

Pete - not sure you could use water as a test liquid given Petrol has an SG of 0.737 and is more likely to find its way through crevices, I assume. Petrol's volatility is what causes a full tank to 'pressurise' speeding up the leak. They would need to set up a test rig of a petrol tank full of fuel and run various solutions to the problem through it. Suggest you get your SNGB order in quickly before they start!

[Heuer]

I was rummaging around my workshop looking for a suitable bit of brass, rubber and heat-shrink to run a test of Petro Patch and happened on this:


It is a plastic electrical cable gland for outside use and seems well designed. The white ring is hard plastic, the same internal/external diameter as the rubber grommet and the screw on end piece. With a cable inserted the white plastic ring compresses the rubber which cannot escape through the end piece. I am going to drop this gland into a jar of petrol and see if it survives. Any thoughts? Meanwhile here is my Petro Patch test rig:


Brass, heat shrink and rubber. Needs to harden overnight.