MODERN FUEL VAPORS AND A '54 BENTLEY

[Guest JWPATE]

Today I decided to open a thread, concerning an old issue I have had with fuel vapor problems on an R-type Bentley. I would not really describe it as vapor lock, for the engine has never stopped running, even in hot summer driving. Rather, what I have is a fairly common condition, of running problems after restarting during hot weather conditions. I never noticed the issue before about 2006, when E10 fuel came along. That corn alcohol in the blend may be at the root, or, more likely I believe, it is simply that the modern fuels are being blended with an expectation that they will be always under pressure, as in new cars. Even the tanks are under pressure and the pump is often inside the tank running 30 pounds plus pressure to the engine injector rail.

Anyway, the issue is one of heat soak after hot weather shutdown. After 15 minutes or so, the uncirculated coolant in the engine will show a 5-10 degree rise in temperature. The engine will always fire up, but now the fuel seems to have boiled out of the float chambers and there is no throttle response. Sometimes it will just tick over on that lean mixture long enough for coolant flow and engine fan, etc. to allow the floats to refill and all is normal, after say 15-20 seconds. On really hot days, one simply must leave it for 40 minutes or so to cool down.

It's an old story, and many have already used one or more tricks to get around it. For myself, I plan to move in small steps, attempting not to change more of the original engine setup than is necessary. Because it seems to be just a matter of ten degrees or so between problems and normal operation, I have started by doing small things to insulate and improve cooling.

First off, the fuel feed is about four feet from the tank to the frame mounted electric pump. This car has dual exhaust and both tail pipes run less than a half inch distance along side the fuel tank. I started the exercise by insulating that section of exhaust pipes. We cannot easily find asbestos these days, but I used a two inch wrap of fiberglass tape, and gave it three heavy coats of a high temp clear to help hold the fibers together.

Then for the unpressurised run of metal fuel line, from tank to filter, and on to the pump, I used this sleeve seen below.

It too uses fiberglass as the insulation, inside a tough outer coat. This stuff is by Earl's, and I believe he markets it under the name - Flame Guard. As with the exhaust pipe wrap, I just used stainless safety wire to secure the ends, as seen below on the filter lines.

Below is a view from under, with both the fuel line insulation and the tailpipe treatment, to illustrate how close the right side exhaust pipe runs to both the tank and the filter.

I really don't know exactly what the problem is on those hot days. Really though, it just might be partially a vapor issue here, in the unpressurised pump feed. So I start here. I really don't want to go too far with any change until its possible effect has been tested. The weather is not there just yet.

I am at 75 degrees ambient now, and on my run the other day the engine coolant gauge was reading 75 C. which is the opening range of the thermostat. No problems during this kind of weather. It is later, when the coolant is "off the thermostat" on the high side by five degrees or so that it gets interesting.

This is only the beginning, and I welcome any suggestions or criticisms. I have already thought about fuel doctoring with kerosine, and even fitting a return loop for fuel back to the tank. But one step at a time. Would love to manage it without changing things too much.

[Guest JWPATE]

That is as far as I intend to go with insulation, at least for now. So the attention turns to engine cooling and questions as to whether anything might be done there to reduce even a few degrees of temperature at shutdown.

So today I pulled the engine fan out, along with the water pump for inspection.

The image below is of the brass coolant distribution tube, which runs inside the coolant chamber, right back to the rear of the engine. There are various size holes along its length, intended to direct the coolant first of all at the chamber faces immediately inside the exhaust valves. They are not easily seen in the photo, but my main concern here was to make certain none of the distribution holes were clogged or blocked. They were all clear.

Next, the water pump comes under inspection. I thought something might be gained by using a new impeller, but upon closely checking the thing - I don't any room for improved flow here. It looks just fine to me.

Behind the pump, are two versions of the pump pulley. Both were offered by Rolls-Royce when the cars were new. The standard production pulley is seen here on the right.

On the left, is what was called the high speed pulley, or more commonly, the hot temperature pulley. It was available from the factory, where the cars were to be operated in such places as where I now live. The smaller pulley is 4 inches diameter at the belt bottom, while the standard pulley is five and a half. I already have been using the small, high speed pulley - and the improvement in cooling was noticeable enough. But of course, there is still a vapor problem.

So now, I am wondering about the fan. As seen below, this was always a thing of curiosity for me. Notice the irregular spacing of the five blades. The company claimed that this pattern was used only after extensive testing to achieve the lowest possible wind noise from the operating fan. I always had doubts about that claim.

None the less, one great feature of it is the large gap between blades there at one point. Through this gap it is possible to wiggle a hand in and remove all four of the mounting bolts without problem.

Still, I wonder about another blade - and have placed on order a six bladed after-market fan. It has not yet arrived so I wait to see whether I can even manage the mounting bolts, not to mention how different it may look. It is intended to have a vintage style, so we shall see.

[nickelroadster]

I would think that the flame guard that you are installing is more of a fire shield than insulation although it does seem to have some insulation qualities. As last resorts you might try an insulator spacer under the carburetor or a bypass line back to the tank. A spacer should help with heat soak after you shut the engine off. Alcohol strikes again.

[Guest JWPATE]

I would think that the flame guard that you are installing is more of a fire shield than insulation although it does seem to have some insulation qualities. As last resorts you might try an insulator spacer under the carburetor or a bypass line back to the tank. A spacer should help with heat soak after you shut the engine off. Alcohol strikes again.

Yup, I do expect that that flame guard sleeving material was intended for another use. I don't know of a better insulation though, for these 3/8 inch copper fuel lines. The feed lines are at ambient air pressure, and so will vaporize at even lower temperatures than the pressure lines. Just didn't want to leave any stone unturned, before considering more drastic measures.

I also have thought about insulating spacers for the carburetors. Were this a '54 Ford setup, for example, that is one thing I would experiment with first off.

Here though, it is a different matter I think. Take a look at the image below, which is a view across the engine from the port side. The round object in left foreground is the oil filler cap, and just left of it I am holding a pointer with its end on a section of coolant tubes. That is the intake manifold, and it is mounted on the cold side, opposite the exhaust headers. The coolant distribution tube seen there is part of the thermostat bypass loop, intended to warm up the intake manifold, even before the thermostat starts to open. This was done in an effort to better atomise the fuel, particularly during cold-start warmup. Specifically, the warm coolant is directed to three cavities. One is the automatic choke control spring, and the other two are at each carburetor flange. It seems to me, that even the modern fuel blends will need this, particularly considering that the flow from carb to intake valve requires two 90 degree turns in flow pattern for most cylinders (Number two and five cylinders are a straight shot, and always run richer than the other four. One and six are at the end of the line, and those spark plugs always show evidence of a leaner burn mixture.).

So, I am now thinking that changes here will be last resort attempts.

And below is an image of the carburetors used. Notice that, while a serious effort was made to warm up the flange area, and by conduction also the venturi portion of the carburetor, the fuel bowl (7) is somewhat insulated from that effort.

I have started a little chart of operating temperatures for these items already, though the ambient temps are too mild for troubles. It is clear already though, that the fuel float bowl runs about 30-35 F below the temperature at the flange (3).

My thoughts about carburetor insulation spacers have been along the lines of using a thicker insulating washer, there at item 44. The float bowl though must not be altered in its relationship to the carb, of course, because it controls the level of fuel inside at the jet orifice. Simply put, this doesn't look very promising. To thicken that washer, would require milling an equal amount from either the float bowl arm or from the carb body. Neither have enough extra metal there for it to interest me. At least not at this point.

For one thing, I still have not isolated the problem to fuel boiling from the float bowl. After all, if I simply leave the car without running for a week or so, that bowl will be empty every time, its fuel having evaporated out through the overflow pipe, which attaches to that banjo bolt atop the float bowl. This happens always, in any weather. Turn the key on, and the pump can be heard running for 10 seconds or so, as it fills the bowls. If I didn't hear that I would know something was wrong with the pump.

Why should it be different if the bowls boil out from hot conditions?

I think, it would be interesting to first see what the pressure is inside those lines. Don't have a pressure gauge in that low range yet, but one is on the way. I plan to temporarily plumb it in there near the carburetor feed line and it may help sort this business out.

Edited April 10, 2013 by JWPATE (see edit history)

[DavidMc]

I am enjoying reading your methodical approach to this subject that has been discussed so much. You may be interested in the attached link, even though it deals with the very specific problems associated with the earlier Autovac fuel supply system.

[DavidMc]

Forgot to attach the link: <!--[if gte mso 9]><xml> <w:WordDocument> <w:View>Normal</w:View> <w:Zoom>0</w:Zoom> <w:PunctuationKerning/> <w:ValidateAgainstSchemas/> <w:SaveIfXMLInvalid>false</w:SaveIfXMLInvalid> <w:IgnoreMixedContent>false</w:IgnoreMixedContent> <w:AlwaysShowPlaceholderText>false</w:AlwaysShowPlaceholderText> <w:Compatibility> <w:BreakWrappedTables/> <w:SnapToGridInCell/> <w:WrapTextWithPunct/> <w:UseAsianBreakRules/> <w:DontGrowAutofit/> </w:Compatibility> <w:BrowserLevel>MicrosoftInternetExplorer4</w:BrowserLevel> </w:WordDocument> </xml><![endif]--> Vapour Lock Blues

<!--[if gte mso 9]><xml> <w:LatentStyles DefLockedState="false" LatentStyleCount="156"> </w:LatentStyles> </xml><![endif]--><!--[if gte mso 10]> <style> /* Style Definitions */ table.MsoNormalTable {mso-style-name:"Table Normal"; mso-tstyle-rowband-size:0; mso-tstyle-colband-size:0; mso-style-noshow:yes; mso-style-parent:""; mso-padding-alt:0cm 5.4pt 0cm 5.4pt; mso-para-margin:0cm; mso-para-margin-bottom:.0001pt; mso-pagination:widow-orphan; font-size:10.0pt; font-family:"Times New Roman"; mso-ansi-language:#0400; mso-fareast-language:#0400; mso-bidi-language:#0400;} </style> <![endif]-->

[nickelroadster]

It is a Rolls. They can't make it too easy or it wouldn't be right. Alcohol just makes this problem worse. How open is the breather? I can see you having a problem with either the fuel boiling out or just evaporating. I am also enjoying your systematic approach to the problem. Maybe you just need to find alcohol free gas if that is possible in your area. keep us posted.

[Guest JWPATE]

David, thank you for that link. I did enjoy reading it, and it brought back memories of a '33 20/25 about 35 years ago. It had also that same vacuum tank fuel delivery. Never had a problem with it, but that was before modern fuels came along.

nickelroadster, your question makes me realize that I used the wrong term when referring to the overflow pipe as a breather. I will go back and edit that part. Below is the kind of thing I was thinking of. It doesn't show up in that illustration of the carb, because it was considered to be part of the fuel lines group. However, the banjo bolt holding the float bowl top down is double length, and fits up this overflow pipe in addition to the one shown in the illustration.

This is made from 1/4 inch brass tube and is intended to carry any overflow (resulting from a leaking needle valve) safely below the engine, and especially below the generator (which is operating directly below the forward carb).

In effect, this overflow pipe represents an always-available path out for vapors in the float bowl whether the engine is running or no, hot or cold.

As to alcohol free fuel, no - the nearest such pump is up in the Fallon farm country. Only E10 is available within about a 300 mile radius of my location.

[nickelroadster]

You mention that the fuel will disappear even if no running of the engine takes place. It sort of sounds like you have a two part problem. Your efforts seem to have gone through most of the steps that anyone has suggested. Other than installing a special refrigeration unit with fuel recovery features I think that it may just be that you have to run the fuel pump for a few seconds to fill the bowl. Do you think that anything has improved with all the things that you have done? Keep on trying. Maybe the solution will come as a flash inspiration some time.

[Rusty_OToole]

The fan is indeed made to cancel out noise. Each blade creates a sound as the fan turns. If they were all the same distance apart they would "sing" together and make a loud noise. By staggering them each produces a different note and they cancel each other out. The tricky bit is to have all 5 blades different distances apart, yet have the fan in balance.

I would leave the fan alone. If you really need more air flow, try an electric fan ahead of the radiator. Many cars have these auxiliary fans in addition to the engine driven fan, especially German ones, and Lincolns.

An easy way to add a return line is to use a fuel filter made for certain Japanese cars of the seventies and eighties. It has a small line coming right off the filter. This bleeds off air, and pressure, also allows a certain amount of fuel to circulate for extra cooling. The return line must go all the way back to the tank to do any good. You may want to add a restriction, such as a carburetor jet, in the return line if it bleeds off too much pressure.

[Guest JWPATE]

Nickelroadster - Yes, the fuel evaporating from the two carb float bowls is absolutely normal. On a cold startup, I just listen a few seconds to the pump clicking over. When it stops I know the carbs are ready for me to start the engine.

This normal sequence though, does not take place if the engine has just recently been shut down hot. The reason why not is still a mystery to me, and I think the next step will be to fit a fuel pressure gauge and watch it for clues.

Rusty - Thanks for the suggestions. As regards the electric fan, that idea has certainly been through my thinking, but I have ruled it out. I have no problem with spin-on oil filters, K&L air filters, and such improvements, but for me, hanging on an electric fan is just going too far. Since I will never actually add one, there is no point in considering it further (however logical it may be).

I am very interested in that filter arrangement, whereby some bypass fuel can form a loop back to the tank. That sort of thing I will certainly look into, if the fuel doctoring with kerosene fails.

And, since I am going to stick with only an engine driven cooling fan, it is time to take a closer look at this six-blade model. It arrived after yesterdays post, and looks to be a well made unit. Below, I have placed it beside the R-R fan (this one from spares locker) for comparison. I could learn very little about this fan, except that it is made in USA. The only stamping on it is 4M and so I will refer to it by that name. Notice also the tips of 4M are curved, reminding me of modern aircraft wing tips. Is this done to reduce blade tip noise or to increase efficiency?

The 4M fan weighs four pounds and four ounces. That is one pound and nine ounces heavier than the R-R five blade unit. Both are stamped from almost identical thickness sheet steel, both the blades and the support bodies. The R-R fan is about 3/4 inch larger in diameter, but the five blades have an effective surface area of 94 inches compared with 116 inches for 4M.

The pitch angle is 35 degrees for the R-R fan and 28 Degrees for 4M.

So which should be expected to move the greater volume of air? Just looking at them side-by-side, I know where I would place my bet.

Rusty, pointed out how the R-R fan required careful positioning of the five blades, in order to achieve a weight balance. The 4M fan has equal spacing, but of course balance is none-the-less of highest importance. Enough importance that I thought it best to test it. Below, I broke out the old tire balancer and tested both the above blades. Both are as perfectly balanced as I can detect.

The mounting slots on 4M will work just fine on the R-R water pump adaptor, but the center hole is too small, at only 9/16 inch.

So below, I have it mounted atop a 4x4 timber in preparation for drilling it out to the required 1 1/8 inch. Had to order a bit that size, and after it arrives I shall launch my attack.

[nickelroadster]

Where is your pump located? What style of pump is it, positive displacement, vane type or what? After a hot shutdown are you saying that running the pump for a few seconds does not fill the carbs? Is it boiling as soon as it pumps fuel in? Let us know where you take the fuel pressure. I guess you will have to look at it as a learning experience at this point it is not clear what all you have learned. I hope you are at least having a good time.

[Guest JWPATE]

Where is your pump located? What style of pump is it, positive displacement, vane type or what? After a hot shutdown are you saying that running the pump for a few seconds does not fill the carbs? Is it boiling as soon as it pumps fuel in? Let us know where you take the fuel pressure. I guess you will have to look at it as a learning experience at this point it is not clear what all you have learned. I hope you are at least having a good time.

What have I learned, at this point? The short answer is - nothing!

But I didn't expect to have learned anything at this point nickelroadster. All I have done so far is to check that I am getting all the performance from the cooling circuit that is reasonable to expect, by the high speed pump and a higher volume radiator fan. Plus insulating the exhaust pipe at the rear, where they do pass close by the fuel tank and the pump feed line. This is only preparation for an attempt at actually pinpointing the problem. All I know at this point is that it is fuel related, and not some other cause, such as a heated ignition coil breaking down.

To really start isolating the problem cause, I will have to wait for the ambient temperatures which bring it on. That will not happen in April. Probably it will be late May before I can get seriously into it. This so far, has only been a few steps in getting ready.

The fuel pump used on these cars was a late version of the Skinner Union (SU) pumps. These pumps are electrically operated by means of a solenoid, which is switched off and on using breaker points. There is a spring loaded diaphragm, to which is attached a round iron disc, and check valves to control the flow in one direction only. As the solenoid energizes, it pulls the diaphragm fully home against the spring pressure. The sound of the iron disc hitting the solenoid can be easily heard on this, the suction stroke. The points then open, allowing the spring to move the diaphragm fully in the other direction, and this is the pressure stroke. The points now close and the cycle repeats itself until the fuel pressure reaches a PSI at which

the spring can no longer fully extend and close the points. If there is no background noise, this process can be easily listened to as a clicking of the pump until the floats are full, the needles close, and the pressure builds sufficient that the spring cannot close the points. Then the pumps stops running until the pressure in the line falls, either by using fuel in the engine or by it simply evaporating out the overflow pipe.

The pumps are beautifully suited to the SU carburetors they feed, because they don't run at all until needed, and then only up that little spring pressure, which is supposed to be on the order of 2.75 PSI. Testing this is the reason I am sourcing a low pressure fuel gauge.

They did not have a great reputation for reliability, even in their day, largely because the breaker points were never cleaned or changed by owners. On this particular model used on the R-R cars, there were actually two pumps in one, sharing a center fuel chamber with one set of inlet/outlet check valves. This was not done to increase the output, either in PSI or in flow rate (which is on the order of 24 gallons per hour unloaded), but rather is was for reliability. If one side fails the other pump would keep up without causing a problem. (Of course, the usual result of this situation was, since nothing was ever checked by the owners, whenever the pump breakers on the last pump failed - that was the end of the line. The other pump half had been out of service for years). A decade or two back, solid state components became available to replace the breaker points. Since that improvement, they do serve quite well and the clicking sound of the SU pump is just part of operating an English auto of this vintage.

Below is the pump in this Bentley, located inside the starboard side frame member and just aft of the B pillar. Seen here is the forward pump half and the common fuel chamber, with the insulated feed line and the plane pressure line. The rear pump is hidden behind that parking brake lever.

And oh yes, I am having fun. There wouldn't me much other reason for the effort - sure doesn't pay much.

[nickelroadster]

I am still rather curious as to how the thing acts with a hot shutdown. How do the fuel pumps act with a hot shutdown? Are you unable to fill the float chambers up? I guess it may be hard to observe this until the hot weather gets here. I too am having a good time reading your extensive posts. It is almost like being there.

[Guest JWPATE]

Well, I drilled out that six bladed fan today. Carefully and coming at it from first one side and then the other, attempting to keep the 1 1/8 inch hole exactly concentric with the smaller one already there. Then used a 1 1/8 inch bore hone to smooth out the inside surface of the hole. I realize that that center hole is critical to the balance, for it locates the fan on the coolant pump adaptor. So the balance was again checked before fitting it to the engine. It was no real problem to start the four set screws and tighten them down.

It doesn't look out or place in there, to my eye. In fact, it looks right at home; and provided it performs noticeable better, it just may be at home for a while.

Also started to gain a little data on those carburetor float bowls. I always can hear the pump filling them before a cold startup, but just how much has evaporated out?

So today, after exactly two weeks since the last fair weather shutdown, I took the bowl top off from the forward carburetor. After lifting out the brass float, I would estimate there was about a tablespoon of fuel remaining.

Most of the float bowl is, of course, taken up with the brass float itself. But, I wondered just how much fuel is in there when the pump has again filled it up? So I removed that last spoon or so of fuel and replaced the float. Then I spooned in fuel until the float was at the level needed to close the needle valve. It took five spoon-fulls. That was with a measuring table spoon, so it needs, say, two and a half ounces of fuel to each carb.

[Guest JWPATE]

I took the Bentley out for a test drive today, and the observations are encouraging. Todays ambient temperature is 70 F, exactly the same as my last temp readings on the five blade fan. Upon shutdown, the carburetor mounting flange was a full ten degrees cooler and so was the float bowl. After a 15 minute heat soak, both those temperatures were five degrees cooler than any previous testing.

Should help, but will the change be enough? Waiting now for warmer days to arrive.

[nickelroadster]

Righto! You are making progress on a failure to proceed. Incremental progress is better than no progress.

[Guest JWPATE]

While waiting for warmer weather, I have slowly come around to thinking about a way to test out some sort of return loop for the fuel to tank. Probably I would take it from the forward carb fitting (somehow) and return it to the tank, first just using a clear flex hose. Then if it provides enough improvement, I could go back and plumb it in proper.

I still am encouraged by the improved cooling, but it seems doubtful that will be enough.

So, if a loop is to be considered, I first must come up with a return fitting at the fuel tank. There is nothing in place that I could just T into. So I believe the best approach would be below, using the fueling inlet pipe. Before taking it out, I marked a location which would allow access to it, once back in place.

There is one fitting already present, and that is the vent pipe connection. Too small for this purpose, and they left most of the fitting outside the pipe here, because it is so near the fuel nozzle inlet. For my purpose, I plan to enter the pipe downstream of that bend, so I would rather have most of the threaded section inside the pipe.

Couldn't think of anything better, and so will start out with this length of 3/4 inch mild steel rod.

In the case of that vent fitting, it was brazed directly to the filler pipe. But that is a thin wall pipe, and for the larger 1/4 inch fitting I will need, it seems better to spread the load a little. So I cut out this little plate of 1/8 inch sheet steel.

Then drill it for the 3/4 inch rod, and hammer it sufficient to match the curve of the pipe surface.

About a 5/8 inch length of the steel rod will do for the threaded portion of our fitting. Best to drill a pilot hole now, as it will not be easy to see the center of the thing, once the welding is complete.

Weld it in place, with none of the rod extending above the surface. Grind off the welds, leaving a flat spot where the rod is located. Then open out the hole to 7/16 inch, the correct size for a 1/4 inch taper tap.

And this shall be the new order of things, back at the fueling pipe. I have cut the 1/4 inch British Standard Pipe (taper) threads, A 3/4 hole is now drilled in the fueling pipe, where I had my mark. The fitting is not yet attached, but just dropped into the hole to check the fit.

I could just weld it in, for the pipe is also steel. But I believe I can make a better job of it by brazing it in place, or maybe even silver solder. I will think on it, as my time is up for today.

[Guest JWPATE]

The mental debate was won by the silver solder team, so I will get on with it. First, build up some jury-rig clamp to hold the thread-fitting firmly in place for the operation.

Then I am using soft silver solder here, and using only a butane torch. It flows in beautifully, with only 450 degrees or so needed. The brown matter seen below is the heated flux, and must be thoroughly cleaned, it being very corrosive if left on the metal.

And after cleaning as thoroughly as I know how, the thing was primed and painted. Now I will put it back in place, with only a pipe plug in the new threads. Who knows, that may be all that ever goes in those threads. But now I feel better prepared to consider a fuel loop return project, knowing this hurdle is behind me.

[Guest JWPATE]

Well, today the ambient temperature got up to 90 F, so I took it out for another test run. This time, the heat soak problem is back in evidence.

After a 15 minute soak the carb float bowl increased to 130F, and troubles began. On start-up it was the same symptoms as described earlier, maybe not so pronounced as before, but essentially the same. Running too lean, and not responding as it should to throttle movement.

This time though, I shut it down before it had time to fully clear itself and opened the hood. Then I turned the electrics back on, but didn't start it up. In the quiet of my garage, I found it interesting that the fuel pump was not running. No clicking sounds, even I know that the engine is not getting sufficient fuel.

Now I depress the override button on one of the carbs (item 25 in the photo) and immediately the pump starts to run, and continues after I release the button for another few seconds until the bowl is again full. Strange?? It behaves as though somehow there is enough vapor pressure inside the float chamber, so that the float doesn't fall as it should to open the needle valve. But if it is forced to do so by the button, then all behaves normal???

I repeat the exercise on the front carb. Same story exactly.

Get back inside and start it up, and everything is as if nothing ever happened. Runs and responds just as it should.

Going to have to think on this one. At a minimum, on the next hot day, I shall repeat the testing. Only next time I shall go directly to the carbs and listening for the pump, before ever starting the engine.

[DavidMc]

It behaves as though somehow there is enough vapor pressure inside the float chamber, so that the float doesn't fall as it should to open the needle valve.

I have spoken to the owner of an older car who discovered that this was exactly what was happening when his car vapor locked. He had insulated everything and fitted a heat shield but the vapor lock persisted in hot weather. He claimed to have eliminated the problem by fitting a large vent line to float chamber, even though the float chamber was vented with a very fine hole.

[Guest JWPATE]

THANKS DAVID, for that story, and I will certainly keep it in mind. I also remember seeing a U-tube video some while back, where a fellow placed a glass jar of gasoline on an electric hot plate and turned up the heat. There was a thermometer in the gas, and at sea level pressure that jar started to boil at between 150F and 160 F.

I witnessed a carb flange temp yesterday of 155. Now the carb bowl was cooler at 130, but what about that passageway from the bowl to the jet? It would be somewhere in between those temps, and would certainly be throwing off vapors. The jet is .1 inch from memory, but at rest with the needle fully down, most of the jet is closed off. So I imagine a trapped expansion of vapors there in the internal drilling, possibly causing the float to remain up/needle valve closed, even while the jet is getting no fuel except for the vapors getting past.

Anyway, the next thing which keeps returning in my thinking is this - what is the fuel pressure doing? I want to know more about that question, so have ordered a low pressure gauge. It has now arrived. With all the market focused upon fuel injection systems, it is not that easy to find a gauge reading 0-5 PSI as this one does. This gauge came from Switzerland.

Tomorrow I shall plumb it in, right there at the end-of-the line, that is, at the inlet to the forward carb. We shall see what the actual fuel pressure is in this system, for I have never before tested it.

[Guest JWPATE]

Ok, so today it was time to fit up a fuel pressure gauge, and see if that tells a story or not.

In order to tap a source of fuel-under-pressure, I located this replacement for a standard SU carb banjo bolt. The main shank is the same as a stock banjo bolt, but there is a threaded extension in 1/4 inch BSPP thread. That will require also that a nut be in this thread, and here from England is the nut and solder type nipple shown below.

Then it is necessary to cut out a mounting plate for the gauge itself. I used a scrap sheet of 21 gauge steel to make it up, with the mounting holes matching two existing 1/4 inch studs at the base of the air intake horn. From this location the gauge is free from any interference and will be very easy to see.

I decided it would also be worth the time to connect the two with that transparent fuel hose. It is polyurethane fuel line, as normally used on light aircraft. Here, it may be helpful to see any vapor bubbles go past (if I later add a line to the fuel tank). For now, the gauge is the end of the line, so we don't expect to see much at this point.

Had to add extra length to the flex line so that it forms a loop downward, thus leaving room for the hood support rod to stow without fouling.

As it is, there are fairly close quarters for the attaching end of this setup. That one electric control in the photo had to move down a couple of inches, and is temporarily fitted to a stud at that location.

Then the gauge is mounted as explained above. I have fitted a "T" with one end plugged, in the event that I later decide to run a vapor/fuel return line back to the tank.

Here I am testing the setup out for the first time. The fuel pump is on, but the engine is not running, as I check for any leaks.

Notice the gauge reading, indicating that the SU pump is outputting exactly its rated pressure of 2.75 PSI. I checked both pumps, and the readings are the same.

With the engine started up and running, the pressure will vary between 2.5 and 2.75 PSI as the pump cycles. All is exactly as it should be when in this cold start situation.

We are having record heat this week, setting new records for April. So, later in the day, with the OAT at 90F I took it out for the usual run, and bring everything up to the same temperatures noted in the last test run. Upon shut-down the temperatures were exactly the same as on the previous run. The fuel pressure after shutdown was reading 2.5 PSI.

So I go for a cup of coffee and wait for the heat soak, which I already know peaks after 15 minutes. So what would you expect the fuel pressure to be after the heat soak. From the earlier cold start, I know that it normally bleeds down to less than half a pound after about 10 minutes.

But, when I return after 15 minutes the fuel pressure gauge is PEGGED OUT AT 5 PSI. WOW, now we are getting somewhere. No matter what is taking place inside the carbs, I now know for absolute certain that there is vaporization in the lines between the pump and the carb inlets.

By about 30 minutes after shutdown the temps have cooled to about what they were at shutdown, and likewise the fuel pressure is back at 2.5.

I believe therefore, that the next step will involve routing in a return vapor line and experimenting with various size restrictors, to attempt and relieve the vapors without creating too much volume demand from the pump.

[Guest JWPATE]

To get started with experimenting various restrictions for a vapor return-to-tank line, I made up this little system today. The barbed hose fitting is silver soldered to a short length of 5/16 copper tube. Then the compression fitting will be used to connect into "T" fitting seen in the previous photo.

I made the five restrictors from 1/4 inch BSF hex screws, first drilling them, then cutting off the heads, and finally cutting a groove for a flat screwdriver. The five holes used here, are the common drill sizes 1/16, 5/64, 3/32, 7/64, and 1/8. I cut threads into the ID of the copper tube, and have the 1/16 inch restrictor mounted in this image.

With zero experience to fall back on here, it is all simply guesswork for me. I imagine that the 1/8 restrictor would easily carry away all the fuel vapors, and the flow of that much volume would eliminate any trace of vapor except for what is actually inside the carb drillings. But, I doubt whether the SU fuel pumps would be up to that kind of volume flow demand. Seems likely to me that the 24 gal/hour rating of those pumps might be insufficient, leading to short pump life and possibly even failure to keep the carbs fully supplied under high engine power settings. I don't know of course - only thinking it through.

The smallest restrictor, 1/16 inch may just be too little to quickly flush out the vapors, which I now know are present when the outside temps are at 90F and above. Or, it just may be effective. Since it is all a questionable approach, I shall start by testing with the smallest hole.

This is all just experimenting for me. If it turns out that one of the larger holes is effective, but too much for the pumps, then perhaps I will even consider fitting a modern rotary style pump, but that all remains to be seen.

[Guest JWPATE]

Ok, the weather has cooled again so that no effective road testing can continue for a while.

Meanwhile, I got to thinking about the subject of flow rate verses the various restrictor sizes. For that matter, what is the flow rate on the engine, and how much of the rated pump output is already needed for the engine without any vapor bleed line? It can generally be stated, I believe, that the average fuel burn for an R-Type bentley is, say, 12 miles to the gallon. Then if we are driving along at a speed of one mile per minute, that works out to be approximately five gallons per hour. Not full throttle, mind you, just cruising along at 60 miles per hour. So five gallons per hour against a pump rated at 24.

Understanding this, it is easier to guess at just what bypass fuel flow to the tank will be acceptable. I should think on the order of an additional five gallons per hours through the bypass circuit would be just about right, and would still leave the pump adequate reserve.

Therefore I thought it worth while today, to see just what the flow rate would be for those five restrictors. Starting with the smallest orifice of 1/16 inch, I set up the scene below, allowing me to accurately measure the time required for the bypass circuit to flow a pint of gasoline. This is done without the engine demand. Just the pump output and the 1/16 inch bypass line in play.

Oh! The flow was strong indeed. Stronger than I expected, and far more than should be needed to prevent the vaporization problem. Actually, I repeated the exercise several times to be certain of the results, and consistently the one pint line was reached after only 45 seconds. Do the math, and that works out at 10 gallons an hour, or about twice what I would like to start the experiment with. So, can I get a 1/32 hole through one of those bolts without breaking the bit? Dunno, but that is where I am headed, with the understanding that it may well be necessary to switch to nylon restrictors.

Forget it with the other four sizes I made up, they are all out of the ballpark high.

Oh yes, if anyone else is running one of these SU double pumps, todays testing has proven one of the old beliefs I was told by R-R "experts" to be utterly false. The old story was that because the springs inside the two pump ends will always be slightly different, only the strong-spring pump will actually be working, and therefore it is best to use only one at a time for longer life.

False - On both the pressure output and the flow rate. The 45 second fill time to the pint line was with both pump ends active. With only one it took another eight seconds or so. The pressure held fine at about 2.3 PSI with both ends working, but could only maintain about 1.5 PSI on a single pump.

[Guest JWPATE]

It did seem best to me, to switch the material for the little 1/4 inch restrictor plugs from steel to nylon. I had at first thought I would use mild steel for initial testing, and make the final restrictor from either stainless or brass. But neither of those two are particularly easy to machine, especially now that we are using such tiny drill bits.

Nylon cuts and drills nicely and will actually form a better seal around the threads. It is also unaffected by either gasoline or ethanol.

Here I have fitted a 1/32 inch restrictor, as that seems about right after the initial tests. The other is slightly larger at 3/64. The 1/32 specimen flows a pint of gasoline through in two minutes, twenty seconds. Now that is more like the rate I am looking for. It works out to 3.2 gallons an hour, the pumps remain at full rated pressure and the total fuel flow with engine operating still leaves generous reserve capacity.

Later, I decided to drill an even (slightly) smaller restrictor of .024 inch, and test its flow rate. It was 2.8 gallons per hour and I like this even more, as that is only 12% of rated pump output. So below is the test results for the three nylon restrictors I now have, along with the percentage of rated pump volume they would bypass.

.047 6 gallons/hour 25% of pump output

.031 3.2 gallons/hour 13% of pump output

.024 1.8 gallons/hour 12% of pump output

This .024 restrictor will be the setup for the first actual road testing. I just have to extend the flex hose right back to the tank filler-neck, where a fitting added earlier......and wait for warmer weather.

And oh yes, I did take time to actually test the pump volume output with no restrictor in the line. It came in at 22.5 gallons per hour compared to the rated figure of 24. Of course the rated figure would be for completely unloaded conditions at sea level. My actual testing was done with all those feet of 5/16 inch piping and at an altitude of almost 3000 feet. So it all seems right in the ball park to my untrained mind.

Edited May 5, 2013 by JWPATE (see edit history)

[Guest JWPATE]

OK! It is looking good!

Today the temperatures are high enough for serious testing. In fact, it was warmer than I normally would use that car anyway, and after the usual heat soak period - there was no problem whatsoever on restart - no hesitation and immediate response to throttle movement. All is as it should be.

I am going to stick with the .024 restrictor and just leave everything as it is for the next month or so, just to be sure. But the vapor/fuel return loop seems to have cured the problem, and without having to move on to a modern pump. A final effort will be to replace all the temporary flex lines with original-style plumbing and remove the fuel pressure gauge. For now though, it will be further test runs.

The temperature rises were the same as before, but now the fuel pressure falls first to about .5-.7 PSI immediately after shutdown and remains there for about 15 minutes before going to zero. During those fifteen minutes or so, I can clearly see the bubbles of vapor moving through the clear fuel line toward the tank. After that time, they stop forming, no more bubbles move by, and the pressure falls to zero.

Then upon turning the fuel pump on, it immediately starts to click over, as it carries more of the vapors toward the tank and refills the float chamber.

I know full well that there is still the vaporization taking place inside the carbs, but with the fuel lines now supplying fresh fuel, rather than vapors, the carb issues become minor enough to go unnoticed.

Ah, it's a beautiful day!

[TexasJohn55]

JW, I love to watch an analytical mind work! Have enjoyed your thread. TexasJohn

[Guest JWPATE]

Well, after having plenty of hot weather now for testing, I have become accustomed to operating free from any fuel vapor problems, and do not want to go back. So, it is time to make the return vapor/fuel line a permanent change on this R-type Bentley, and replace the temporary flex lines with copper tubes in keeping with the original layout.

So, I started here it the fuel tank filler neck. From here I ran the line forward, then across from port side to starboard while clamping the lines to the same frame member which supports the fuel tank straps.

Here, on the right side, I ran the return lines generally along side the original feed tube from the fuel tank going forward.

At this point, inside a protected cavity in the frame rail just aft of the fuel pump, I decided to add a ball type shut-off valve to the return line. Don't know whether I will actually shut it off during cooler months, but it is easy to do now, so I put in the valve just in case.

As I passed the location of the fuel pump, the return line can be seen against the frame rail, using an old style tubing clamp.

Also, from the pump going forward both the feed and the return lines are new, because I plan to change the routing of the feed.

Here the two lines are fitted with unions so that later on, should the need arise, I can change either half of either tube without any difficulty. This is a well protected location in the frame, and is normally closed on the underside with an aluminum undersheet.

Seen here are the two forward sections from the previous image, and here they are ready for flex line sections to be attached. This is the upper surface of the right-side main frame rail, just forward of the firewall. So we are now in the engine room, and here, in order to match Rolls-royce practice, the fuel lines must be plated. The originals would have been type one CAD plate, but that process has long ago ended under EPA regulations. I am using soft nickel, which is a reasonable substitute. The right side bracket is original, however I have reversed its role from feed line to return. Originally the hard line from the carbs, attached to a bracket located on the engine side, then by flex line to this right-side bracket. I will use the original brackets, but now they will carry the return fuel/vapors, so it doesn't matter so much being bolted to a hot engine side.

The feed line now will be that bracket on the left, which I made up to match the style of the original. From here the feed line will be flex hose directly up to the rear carb banjo.

Now, working inside the engine bay, the space is really at a premium. I shall try to add the second fuel lines near to the original routing, where possible. Some of the old BSF pipe thread fittings are still available, such as this banjo with two inlet nipples, opening at 45 degrees. They were both threaded in 1/4 inch BSF pipe, but I decided to grind off the threads on the inside nipples, in order to make them smaller and neater than using attaching nuts. Then I drilled them for 5/16 tube size, and soldered the inside run, which will connect and communicate the two banjos for the front and rear carbs. They don't need to be taken apart ever, and I believe I can make a better job of it this way. Likewise, I show an SAE fuel hose fitting with both its threads and the hex flats ground off. I first drilled it to 5/16 then temporarily soldered in a length of straight tube, which could be chucked and aid in obtaining a nice rounded surface. This also done primarily to reduce the size of things, where I know space is limited.

It was fortunate for me, that I did have an extra intake complete, inside the spare parts locker. This is how I decided to lay out the fuel lines connecting the two carbs. That long section extending down on the left, is the vapor return, and will connect with the engine block fixture I mentioned earlier. When in place, these lines are not visible from above the engine.

At the forward carb location, I have threaded the inside of the vent line, there at the end of the pointer. It has in it, the same nylon restrictor plug I have been testing, drilled with the tiny .024 inch hole.

After satisfied with the two carb lines, they too are treated to nickel plating.

The fuel feed, from that new bracket on the frame rail, is next in line. It must be carefully bent, in the hard line section, to provide good positioning, among several cables, brackets and etc. which already pass through the intended route. Here I am using the flex hose fitting discussed above, soldered in place to reduce size.

With the pipes now all in place, it can be seen how necessary it was to plan and reduce the size of things to the extent possible.

It does make a satisfactory path down to the attachment on the frame rail top, and I have tested it carefully for adequate clearance during engine startup and rough running. Actually, with the other intake pipes, bottle for windshield cleaner fluid, and etc. all back in place now, the single loop of that fuel feed hard line, is about the only clue that anything at all has changed.

At the front end of things, there is little to give away the changes. Most everything changed is out of sight below the intake horn and other items.

So, with this happy outcome I shall close out the thread. All the testing and experimenting is over now and the modification is complete. Glad to have been able to find a cure for the fuel vapor problem with so little change to the original design.

Thanks for those suggestions and encouragements fellows.

[DavidMc]

Congratulations on solving one of the most annoying problems in hot weather. I have followed your thread with great interest, before you close off why not summarise the result with a schematic sketch of the final fuel system for future readers researching this topic.

[Guest JWPATE]

Thank you David, for that suggestion.

There were several attempts made to simply insure that the engine was operating at as cool a temperature as it reasonable can be expected to do, and a couple of attempts also to shield the fuel delivery pipes from exhaust heat.

The real breakthrough in understanding the problem, came only after fitting a 0-5 PSI pressure gauge and clear fuel hose for observation. In this particular case, the primary problem was not in the carbs but in the copper fuel lines situated inside the engine bay. Following engine shutdown, the increased heat inside the engine room reached a peak after about 15 minutes. At that point the fuel inside those engine bay lines had vaporized to such a degree that the pressure had actually gone up rather than fallen off as is normal. In fact, I don't know exactly how high the pressure went, for it pegged out the gauge at 5 PSI.

What was happening could be understood after observing the pressure rise. I could not explain, earlier in this exercise, why the fuel pump did not make its usual clicking noise when the switch went on, during these hot starting events. Now it became clear. As the fuel vaporized inside the lines there was no where for the expansion to relieve itself. It couldn't pass into the carbs, for they were still full and the needle valves were closed. It couldn't relieve back through the fuel pump, because there is a one-way valve there at the pump. So the pressure went up, somewhere over five pounds. The SU fuel pumps reach a max output of three pounds, and start to pump again when the pressure falls to 2.5. So here we are with the copper lines full of vapor and the pump unable to run because the pressure is too high.

The result, on engine startup, is just what one might expect. A normal start occurs for the first few seconds. Then, as the float level/jet level falls, the needles open to admit vapors from the pressurized supply lines. The engine runs lean on the vapors for a second or two more then stalls out. The fix needed, is to provide another path out of the copper lines for that fuel vapor, and it must run all the way back to the tank. With a path out, the lines will still form vapors, but will not pressurize and prevent the fuel pump operation. Then, with the fuel pump running, it quickly clears the lines of remaining vapors and supplies fuel again to the carbs.

When the system was designed back in the 1940's not only was the fuel a different mix, but the rubber hose available then was no where near as reliable as we are accustomed to these days. So, for safety, this Bentley layout ran the copper hard lines down alongside the engine case to a fitting approximately level with the frame member, and made the flex line connection there. That was fine for many years. But with copper line bolted up to a hot engine we have one more reason for the modern fuel vapors forming. Copper is an excellent conductor of heat. This is why I also decided to route the fuel supply directly from the frame member, in the shortest length possible to the rear carb banjo. Then I used the old fixtures to run the return vapors.

I attach a couple or rough sketches, for I realize my explanation cannot make things clear enough, and certainly the photo images are far too busy to be of much help.