MG-Cars.info

Welcome to our Site for MG, Triumph and Austin-Healey Car Information.

Parts

MG parts spares and accessories are available for MG T Series (TA, MG TB, MG TC, MG TD, MG TF), Magnette, MGA, Twin cam, MGB, MGBGT, MGC, MGC GT, MG Midget, Sprite and other MG models from British car spares company LBCarCo.

MG MGB Technical - Fuel problem?

Grateful for your thoughts on a peculiar problem. My engine suddenly falters, and struggles at exactly the same spot on a hill on my way home! I am not kidding, it's + or - 20 yards from the same spot every time. The hill is a long steady incline. The car goes up short steeper hills without any problem. From the feel of the car, I think it is a fuel starvation problem. I have already fitted a new fuel pump, and new tank to pump fuel pipe (which was rotten.) When stationary I have a decent flow of petrol going into the carbs. I'm wondering if I have a perforated draw pipe within the tank. Can anyone think of any other things I should be looking at? Can anyone think of a test which would show whether the draw pipe is the problem? The car is a 66 roadster, with engine in standard trim save for electronic ignition, inside a 45D distributor. The car starts on first pull, and otherwise runs well. Grateful for any help.
I Hodges

What do you have to do to get it running again to get home? Does it make any difference how much fuel is in the tank?

Charley
C R Huff

Check the fuel flow from the pump by pulling off the supply pipe to the carb, placing a suitable container to catch fuel under it and turn on ignition. Should be in excess of an imperial pint per minute.
Allan Reeling

Charlie, I pull in to the side and wait 30 - 40 seconds until the spluttering stops and the engine runs smoothly again. If I stop half way up the hill before the loss of power occurs, and wait a minute, then I can get to the top without the problem arising. I experimented last week by filling the tank as full as I could get it, and I still had a problem on the hill, but I thought it might not have been as severe as previously, but that might have been wishful thinking.

Alan, I do have a very decent flow of petrol to the carbs.

Thanks, both.
I Hodges

If it does it with the tank full, I think that makes it less likely that the pick up tube is perforated.

How about that new line you put on between the tank and pump? How stiff is it. If it is rubber and is not stiff enough, it can get sucked flat and stop the flow. If this is the case, it would have to be borderline to only do it when fuel demand was increased to climb the hill. But, I have had this happen. If the run from tank to pump is a significant distance, a metal pipe with two short rubber connectors would be preferred.

Then, of course, we have to question your fuel tank venting. Is it through the cap or through a charcoal canister? Is it working okay? Again, it would have to be borderline to only cause a problem on the hill.

Is the fuel level in the float bowls okay? Since it is sputtering for that long to get caught up, how about pulling off the road when it falters and shut it off. Then pull the top off the float bowls to see the level. (Oh, you didn't say if you have HS or HIF carbs.)

Charley
C R Huff

The new pipe from tank to pump is copper, and I am fairly confident about the connections either end. Good point about the fuel cap. I can't see any obvious vent hole, but there does seem to be an open airway through the locking assembly. I don't get any hiss when I take the cap off, and I rather think the problem would be more frequent than just at the very precise spot on this hill where it fails. I've changed the needles in the float chambers without any effect. (They are SU carbs, HS4.) Where the car falters on this hill, there is no space to get the car off the road to get the bonnet up, that's why I have taken to stopping in a recess half way up to avoid being stuck in the road. (The hill is a narrow country road.)

If there is an air leak on the suction side of the pump which only shows under the load of this long hill, I should be able to replicate this on a flat dual carriageway, shouldn't I? If I floor it for half a mile or so?

Sherlock Holmes said when you have eliminated all the alternatives, the remaining option, however implausible, must be the correct one. I'm wondering if I'm now getting to that point, and the time has come to change the tank. It is the original and 50 years old. But it would be nice to know I' m solving the problem before I shell out 300 notes.
I Hodges

The first thing to do with anything like this is look at the tach (which as a 66 it should have). If that is jumping around with the misfire then it is an ignition LT problem. If not then HT or fuel.

I remember this problem from someone in Australia, going over the same bridge every day to work. I never saw a resolution to that.

There was also the 'vanilla ice cream' problem some years ago - non MG. That was someone going to the local supermarket to buy ice cream, and if he bought chocolate chip the car restarted, but if he bought vanilla it wouldn't. That turned out to be choc-chip was popular so was at the front of the store, but vanilla was at the back. The extra time it took to buy vanilla was enough to allow the greater heat-soak to cause a problem.

The first thing I'd do is drive home via a different route, at the very least to drive for at least as long as you normally take to get to that point on the hill, before tackling the hill.

Secondly while stopped on the hill I'd check the fuel flow there and then by removing a pipe from a carb while the problem is still occurring - although narrow road etc. noted.

Thirdly I'd remove the petrol filler cap before tackling the hill, to see what happened. That would need you on at least one other occasion to stop and go to remove the filler cap but not actually do it, to see if you got the problem or not. However given your other comments it does seem unlikely to be the cap.

If there was an air-leak on the tank side the pump would be chattering. If it's not clicking or chattering it isn't trying to deliver fuel. That could be because the fuel line is pressurised, or it could be because the pump isn't even trying to deliver fuel. The symptom for the latter case would be that immediately before the engine DID run normally again, the pump would chatter to refill the carbs.

If it's increased flow rate up the hill in 4th gear then it would also occur if you increased the flow by other means, such as maintaining a given speed in a lower gear which would need higher revs of course - but that would be much the same as flooring it on a dual carriageway.

But if the only place it happens is on a narrow road with no passing place, and you are avoiding that by stopping beforehand, it's going to be difficult to diagnose further until you can reproduce it elsewhere.
Paul Hunt

I had a very similar problem with my 70B which was great fun climbing to Shap on the M6. Car just ran out of push and slowing down didn't really help. Pulling over to the hard shoulder with a call to the AA in mind, I found everything returned to normal after a few seconds. Off it would fly for another few hundred yards only to repeat. Dipping the clutch for a few seconds improved things and Shap summit was successfully reached, For the life of me I can't remember how or even if I found the fault before taking the car off the road for bodywork! I know that doesn't help but.....
If it is fuel starvation and the pump is delivering enough to the carbs, I would ask; did you check the float levels, clean any crud out of the bottom of the bowl and blow through the jet when you changed the needle valves? Have you checked all the small pipes around the carb's to make sure the fuel is actually getting to the carbs and not just up to the first connection?

Best of...
MGmike
M McAndrew

Paul, thank you. I'll get out tomorrow and give all your suggestions a try. But yes you are right. If my theory about a perforated lift pipe within the tank is correct, then there would have to be extra activity by the pump on the hill before the engine falters. Not conclusive if it does, but I'm on the wrong tack if it doesn't.
I Hodges

You say that you have a 45D distributor fitted to your engine. The problem may lie therein. Try disconnecting the vacuum advance line and plugging it. Then try to get the car to duplicate the problem while going up that hill. One more thing to eliminate from the Holmes's equation. RAY
rjm RAY

"You say that you have a 45D distributor fitted to your engine. The problem may lie therein. "

I don't see why the 45D4 should be any different to the 25D4, and really I can't see why the vacuum capsule on either would cause this, given that when stuck on the hill, waits 30-40 secs for the spluttering to stop, then carry on i.e. up the hill.

Having said that a problem in any type of distributor could cause this, the as-yet missing information is what the tach does when the problem occurs.
Paul Hunt

Very grateful for all your responses. Did a lot of testing today. First things to rule out: no crud in the float bowls, no problems with the pipe work round the carbs, nothing happening with the tachometer, no difference going up the hill with the fuel cap off. I found another long hill to climb and the problem re-occurred, with the fuel pump becoming very busy. Also found a straight and level piece of road and floored it. Didn't take very long before the engine faltered, and again the fuel pump was chattering away. I put as much fuel in the tank as I could get, and I did get a little further up my homeward bound hill before it faltered. Not much, but previously it has always stopped at the very same spot. One thing I did notice was that there was very little time between the pump becoming audible and the stuttering starting. I guess I was expecting it to be longer. I'm still inclined to believe it is a supply side fuel problem. I think my next step is to re-do the tank to pump connections, just to be sure, and then I will have to bite the bullet and order a new tank. Idly thought of putting an inspection camera into the tank, but that's probably a very bad idea.
I Hodges

Because it's "a new fuel pump" doesn't mean it is immune from fault. (see many previous blogs). If the pump is chattering with a full tank, then it's either drawing a lot of air from somewhere or valves in the pump aren't sealing and the diaphragm is just moving fuel backwards and forwards. Quite often poor pumps will move enough fuel when there is little demand but falter on high demand. What make is the pump?
Allan Reeling

Maybe partially stuck valve, in the pump, if an SU.

Herb
H J Adler

Alan, Herb, thanks, but the problem was there before I changed the pump. I take your points, but a bit reluctant at the moment to think I've swapped out one faulty pump for another. The new pump is a German aftermarket model from Moss-Europe.
I Hodges

The pump chattering does show it is losing it's supply of fuel. So it does sound like a supply-side problem i.e. before the pump - unless you have two pumps doing exactly the same thing which seems unlikely ... unless they are identical pumps!

Earlier, Allan recommended a delivery check, which should deliver a minimum of one Imperial pint per minute from a pipe removed from a carb and in practice double that, and you said you had a 'decent' flow to the carbs. You really do need to do that measurement, and it should deliver in a continuous and regular series of pulses with minimal bubbling.

If that is OK, then I can't see it being the pump. Neither can I see it being a partial blockage from tank pick-up to pipe, as that would show up in a static test. I can't see it being a perforated pipe inside the tank, as that would get worse as the fuel level dropped.

On the face of it I can't see it being orientation either as you can reproduce it on the flat ... although that will tend to move any debris in the tank backwards, like climbing a hill, so it could be something in the tank swilling around and blocking the pick-up.

You say there is very little time between the pump chattering and the stuttering. Exactly how long could be indicative. In normal driving there should be quite a noticeable gap between the pump stopping and the engine spluttering, it will be less at wide throttle openings, but 'very little' could mean that the float chambers are not being kept full even when the spluttering isn't happening. At a suitable location I would drive with some gusto, but short of getting the spluttering. Then switch off and declutch to stop the engine, pull over, and remove the lids of the float chambers to check the levels. Compare that with having switched on but not starting the engine, then switching off when the pump stops clicking.
Paul Hunt

Allan,

If you have an auxiliary tank, like for an outboard boat motor, you could plumb it to the pump intake and eliminate or condemn your fuel tank system as the problem.

Charley
C R Huff

Paul, clever ideas, thanks. I'm going to have to wait for a dry day to try them out. (There is a reason why Ireland is green!)

Charlie, thanks, I've been puzzling over the same thing for a few days now. I don't think I've got the means to do it safely, but it would certainly answer the question.
I Hodges

I Hodges - We all may be going down the primrose path because of the statement that the pump is chattering away when the problem occurs, at least that was my thought and Paul's since a chattering SU fuel pump indicates an air leak on the pickup side of the pump. A pump that stopped could indicate a clogged pickup. You got me thinking when you stated that the pump is a German after market pump. When I think of German pumps, I think about the older Harting pumps that were nearly exact clones of the SU pumps. I am not familiar with the Hardy pumps available from Germany today and I am wondering if they may have a bypass valve like the Facet pumps available on this side of the pond. The Facet pumps chatter all the time, but if the inlet side of the pump becomes clogged, the chatter becomes slightly different, but it is still chattering. I may have missed it in your narrative, but if the fuel filler cap is not venting properly, that could very easily be causing your problem. Removing the filler cap prior to going up the hill would tell you if that is the problem. Cheers - Dave
DW DuBois

I H,
Just picked up your reply as regards the origin of your new pump. I had one of the German, blue cap, pumps on my V8 for a time. I got incredibly noisy but always delivered but in the end the incessant chattering got on my nerves and i bought a "proper" SU pump. Maybe, as Dave intimated above, we are being led astray by the noise thing.
Try syphoning the tank, jack up the drivers side and extract the sender unit to inspect. Should only take an hour if the bayonet hasn't corroded, but buy a new one and a seal first.
Allan Reeling

Apologies for the delay in coming back - been distracted by other matters. I've listened more carefully the last couple of times up the hill and I now know that the pump starts to become active before the engine starts to falter. It does chatter noisily whilst the engine loses power, but recovers to its normal silent running when I stop. This new pump doesn't behave any differently from the SU I took out. I've now tested the flow of this new pump and it filled a half litre bottle in 30 seconds, so almost 2 pints in a minute. No problem there. I did notice there was a little pressure in the hose ( float chamber inlet) when I took it off to test. Paul, I haven't yet had chance to try out your before and after scenario, but I've been wondering where it would lead if there were to be a difference. Alan, it's a nice idea, but I think the pick-up passes over a baffle, so only a part of the pipe would be visible. I've found a good deal on a tank in the UK and so I think the time has come to get one shipped over and see what happens. If it solves the problem I'll take a can-opener to the old one and report back what I find! Thanks all.
I Hodges

" If it solves the problem I'll take a can-opener to the old one and report back what I find!"
Just be sure that the can opener doesn't produce any sparks :) Cheers - Dave
DW DuBois

I H, The gauge sender unit and fuel pick-up come out as a unit.
Allan Reeling

Only on the later models (at least that is true on this side of the pond). I don't know just when they started combining the two. Cheers - Dave
DW DuBois

Dave is right. According to the Moss on-line catalogue, cars up to number 415000 had a tank with a separate feed pipe attached near the fuel gauge sender. From 415001 the pipe was incorporated in the sender unit. The change-over happened some time in 1976 I think. I used this fact to advantage on my fuel injected V8, using an older style tank with the later sender unit meant I had a ready-made feed and return for the fuel rail.

Mike Howlett

Indeed, combined sender and pickup only provided for the 77 model year on.

Residual pressure in the hose when you take it off a carb is normal, if the ignition has been on in the previous few minutes. The normal state of an SU pump with the ignition on, even when running, is with the diaphragm pulled back, and return spring pressure pressurising the line to the carbs. Only when that has fully returned, and the points flipped over to close, does the solenoid operate to pull the diaphragm back again (and open the pojnts). That is the only time the fuel line isn't pressurised, and is a very brief period compared to the length of time the fuel line is pressurised.

2 pints per minute in a static test is exactly what you should betting, which makes the chattering under heavy load shortly before it starts to falter even more strange.

About the only other test I can think of is to park the car at the angle of your hill, or higher if possible, and try again!

Other than that its a case of jury-rigging a separate supply to the pump inlet, and seeing what happens then. It is looking like something very odd with the tank, but it's rather a large expense in time and money to test out, when you don't know it definitely is the tank that is the problem.
Paul Hunt

Incidentally, as a non-standard pump how is it mounted? If directly to the body try insulating it on rubber bushes or somesuch.
Paul Hunt

I H,

When you did the flow test, were there air bubbles in the fuel stream? If so, it would indicate sucking air somewhere.

From what Mike said, if there is a problem with your pickup tube, it could be corrected by installing the new style pickup combined with the sending unit. That should be a lot easier/cheaper than changing the tank.

I like Paul's idea of doing the flow test with the car inclined.

Allan and I H, sorry for getting crossed up with the names a few posts back.

Charley
C R Huff

No worries Charley, I got crossed up on pick ups and sender units!!
Allan Reeling

Thanks for all your comments, I like the idea of static testing of flow rate on an incline. I'll do that today, sometime before the courier arrives with the new tank! It looks like it will be a week or two before I will have the time to fit it, but I will report back.
I Hodges

Well the new tank is fitted, and I still have the same problem. I have to say, I have run out of ideas. The whole supply side has now been changed, including the pump. I did a flow rate test on an incline as suggested, and it took 55 seconds to fill a half litre bottle, almost twice as slow as previously. As I only had a gallon in the new tank, I put in another two gallons and retested at 45 seconds, still slower than before. I then retested on the level, and twice recorded 45 seconds. On each of the four tests there was some small but suspicious initial spluttering of the flow (and busy pump activity) before settling to a constant flow. Not sure what to make of this. Have I introduced a new problem to replace the one that was there before I touched anything? My pipe joints are clean, dry, and satisfactorily tight. But I'm wondering if I should remake them with PTFE and plumber's goo. I've assumed that any delivery side problem would either be continuous and/or would be visibly wet. The only other fuel factor not previous disclosed is that my immobiliser isolates the fuel pump, but I cannot think of any scenario where an immobiliser problem would kick in on my hill , or under load on the level.

If not a fuel problem, what else? In all other respects, the car starts and runs very well. As usual, very grateful for any suggestions.
I Hodges

"I did a flow rate test on an incline as suggested, and it took 55 seconds to fill a half litre bottle"

Was there any bubbles in the stream of fuel being pumped? If so, start looking for and air leak along the fuel path. If no bubbles, look for some kind of a blockage along the path. I would start by removing the fuel line at the output of the pump and use a new piece of fuel hose and rerun the flow rate. If it is still slow, look at the voltage to the pump, while the pump is running. Cheers - Dave
DW DuBois

56 seconds per half litre is right on the minimum for the later AUF300 pumps, really it should be around half that.

"On each of the four tests there was some small but suspicious initial spluttering of the flow (and busy pump activity) before settling to a constant flow."

Definitely not right. It implies the pump may have emptied/filled with air somehow, it will chatter until the pump chamber refills with fuel. Were two or more tests done immediately one after the other? Or was each of them done some time after standing with the ignition off?

A leak on the intake side may well not leak fuel, but will let in air, as it is usually at negative pressure, both when pumping (from suction) and standing (from gravity). However that doesn't explain the problem - pump chattering shortly before engine splutters - developing on a hill or under hard acceleration.

Checking the voltage right on the pump spades is valid, but as the pulse of current taken when the solenoid is pulling back the diaphragm is very brief compared to how long the solenoid is de-energised and taking no current during fuel delivery. I won't go into the maths, but you need to connect a constant load to the pump terminals in order to get an accurate indication of supply voltage. A spare headlamp bulb is best, a 60W filament takes about 5A at 12v for example, and while not as much as the solenoid anything more than half a volt or so drop with the bulb powered may well be indicating bad connections. Having said that the WSM says the minimum starting voltage is 9.5v, so you would have to get down to that before maybe getting problems. But again, that doesn't explain why it only happens on hills or hard acceleration on the flat - unless something in the boot is joggling the earth wires where they bolt to the rear panel - much less why it starts chattering shortly before engine spluttering. If there were insufficient voltage it would simply stop, which it would also do if your immobiliser had kicked in.
Paul Hunt

Pondering since the previous post, I think you need to go back to the beginning.

Very early on Charley Huff and subsequently I suggested taking the float lids off to inspect the levels, and I see no indication that this has been done yet.

1. Turn on the ignition, turn it off when the pump stops clicking, remove the float lids and do a dipstick check of the fuel level in each float chamber.

2. After replacing find a suitable road and drive it hard in 2nd or 3rd until you get the problem (as you mentioned you had been able to) and as soon as that happens, switch off, knock it out of gear, and bring the car to a halt. Remove the float chamber lids again and do another dip-stick check.

I'm beginning to doubt it is fuel, but the above is really the only way you will prove it one way or the other. If the level is way different in both carbs, then it is a supply problem. If only in one carb then it is that carb. But if the levels are the same then it is something else, i.e. ignition, and not fuel.
Paul Hunt

Paul, Dave, thanks, apologies for the radio silence but I've been away. You are right Paul, the float chamber levels is a test still to be done. Thinking back over the whole saga, the only fault I found was the rotten tank to pump pipe. The likelihood it was admitting air is probably quite high, and therefore also the likelihood that the continuing problem is still associated with that area. Unpalatable as it is to my ego, I think I have to question the soundness of my connections fore and aft on that pipe. I think before anything else, I have to remake those, if only to say definitively that it is not a supply side problem. Thanks again both for your continuing support. Ian H.
I Hodges

Just discovered one of the washers with the banjos should have been a rubber o ring, not fibre. Nothing on the supplier's website to indicate this.
I Hodges

Several years ago, Burlen switched from using a fiber washer, on the inside of the banjo fitting, to a rubber O ring. The outside, however, continued to use the fiber washer. Go figure. RAY
rjm RAY

Ian - See the article, New SU Fuel Pump Fittings Using ‘O’ Ring Seals in the SU
Fuel Pumps Articles section of my Homepage at: http://homepages.donobi.net/sufuelpumps/ for the 'O' ring size. Cheers - Dave
DW DuBois

"one of the washers with the banjos should have been a rubber o ring"

As both ports are in the same casting either both should have an O-ring or neither should. The indicator for O-rings is a small rounded recess immediately above the threads in the pump casting. If the threads come all the way to the end then fibre washers should be used.

It seems the banjos themselves may have changed at some point to gain a recess on one side - deeper and with flat side and base compared to the recess in the casting. However that is for a smaller (i.e. smaller than the one on the other side if there are no recesses in the pump casting) fibre washer which butts up against the hex end of the through-bolt.
Paul Hunt

Well, very nearly fixed it! I remade the connections on the supply pipe, using an o ring under the banjo ( thanks Dave for your clear article. ) Car now sails up the hill, when driven normally. If I push it hard there is a very slight recurrence of the spluttering and loss of power a lot further up the hill, just as I get to the top. I've checked and reset the float levels in the carbs with no different outcome although one was lower than the other. The flow rate has improved to half a litre in just under 40 seconds. (Interesting that the flow rates with the new tank are slower than with the original.) I can probably live with this now, as the problem occurs after driving a lot harder than I normally do, though it would be nice to clear it up completely. I think the story of the whole event is I had a perforated supply pipe (it was rotten) and I exchanged that air leak for a new one at the banjo connections. Purchase of a new pump and tank may not have been necessary, but that is the price of car maintenance by replacement in the absence of a trained mechanic! Thanks all for your assistance.
I Hodges

Just a thought. You haven't changed the gas cap lately have you? If so, could you have replaced a vented cap with a sealed cap?
Richard Smith 1

This thread was discussed between 28/03/2016 and 14/05/2016

MG MGB Technical index

This thread is from the archives. Join the live MG MGB Technical BBS now