Dual Fuel Pump set up: DIY 1996 e36 (alternative setup)

[bimerok]

So here is my update on what I'm working towards for solving this annoying (to say the least) fuel starvation issue.

After no luck with the typical in-tank dual-pump set up and playing with different ideas mentioned previously in this thread, such as setting up multiple Walbro pickup points throughout the tank or use of in-tank slush proof foam, i decided that some of it will not be practical and some would be too much of effort for not a lot of certainty that the problem will be fixed as the end result.

So I decided that I will go for a surge tank solution. With implementing a surge tank I have a very high percentage of certainty that my fuel starvation problem (due to fuel slushing around the tank) will finally go away once and for all.

Now I know that there are different approaches to the surge tank with few different designs. Some are meant to work with external pressure producing pumps and some with internal. There are also almost ready kits that use internal pumps such as radium engineering, but at $600+ for the set up, it was just a little more than I wanted to spend on this problem and plus I'm enjoying with fabricating what I can. What else is there to do during off season?

Couple of definite requirements that I had for the surge tank setup were: No additional (3rd) fuel pump to be added. The surge tank must be low pressure container. And the most important one was that I don't want to have an external pump. By not having external fuel pump addressed couple of things for me. 1 - I did not have to buy yet another pump, but instead use one of the two I already have floating in my tank. 2 - most of the external pumps that I saw are very loud. 3 - I just have something against the external pumps . The constraint of "internal pump only" definitely complicated things more as far as designing and building the surge tank. I pretty much could have used one off the ebay for under $100 otherwise, but that's boring

So the first thing I start looking for is what I can use for the surge tank made of steel so it is easier to weld with my MIG welder? The idea comes to use a propane torch tank as the base. So I grab one from HomeDepot and start burning off the gas to empty it and cut open to see what's inside.



Cut it open


Vola! Nothing special inside. What did you expect?


After cutting the cylinder open and letting it sit under open air just for few hours it already started to develop rust pretty quickly. As much as I did not want to admit it and move away from steel and already formed "tank" on my hands, I had to give up on that idea and move to something else. That something else was going to be aluminum.

Now I had to find something that I can use as a cylinder/tank made out of aluminum. After throwing few ideas around including finding a piece of aluminum pipe, I could not come up with anything in the right diameter while having right wall thickness and not something with the wall that can withstand an direct hit from a incendiary round All of the sudden my cousin gives me a great idea to use a regular aluminum air intake tube. Bingo! In my garage, I have just what I needed. A big leftover from my air intake set up. A 4" in diameter aluminum tube. All I had to do is to make a bottom that will be welded shut and the top that can be disassembled to insert/service fuel pump. Well... and everything else that is required to do a surge tank

What I needed is to make 3 round "things". One to cover the bottom, one that would be welted to the top with access hole to the surge tank for the pump, and the third one to cover/seal everything tight. After looking around in my garage, I found an aluminum flat bar that was a leftover from my camber plates spacer "project" I did a year or two prior. It was the right size and right thickness - 0.25 inches if I recall correctly.

Now I just needed to make those 3 round wheels. After measuring the right diameters that I wanted to go with, I used my caliper as a drawing compass to scribe the circles in the aluminum bar. Then using the sharpie I marked everything I could to get as close to the preliminary circle as possible using my band saw. Here is the picture of what that looked like.



As you can see the one in the middle, with all of these lines, it is very close to a circle with just few edges remaining. The final finishing was going to be done on my HarborFreight mini lathe. Here is what it looked like when I made circle out of that. This picture also shows some scribing that I've done for the top 2 circles that will have bolt holes around the perimeter and a grove for the sealing gasket.



Here is how it looked like after drilling the holes around the perimeter for the bolts, cutting the grove for the gasket on the lathe and a quick surface polishing.





Then I needed to cut out "access" hole in one of the top lids for fitting a fuel pump through. Also the holes in the middle are to be fill-welled later on. I needed them for machining purposes.



Below you can see what the top, disassemble-able section would look like with the gasket sandwiched in between and the bolts going around the perimeter.











For the various fittings that need to go around the surge tank I got some aluminum weld-in NPT thread bungs and various screw in barb connectors/nipples to go with it. Also for the wires to the fuel pump a special pass through wire adapter. All of that came from Summit Racing and was pretty cheap stuff. Some other fittings and elbows came from HomeDepot racing isle.

With all of the surge tank parts ready to go, all that needed to be done is to weld all of the pieces together. Since I don't own a TIG welder anymore, this job had to be outsourced. I contacted the same person who was working on my exhaust in the past and he was able to turn this welding job around the same day, which was awesome since I'm inpatient and can't wait for anything When I picked it up, I was very pleased with the job he did. Nice strong welds where I don't have to worry about twisting the bung off or breaking it. Nothing should leak through the welds. I'm satisfied.

Here are few pictures of the internals mocked up and what the whole thing will look like.













Once the warmer weather will finally arrive I will start working on mocking everything inside the car. Preliminary location for the surge tank is inside the spare tire well. The height of the tank was made with that location in mind. I don't remember the exact numbers as it was few weeks ago now, but I think the volume of the tank with the pump in it came out to about ~1.5 liters or ~0.4 gallons. This should be more than enough of buffer. Also the final orientation of the fittings will be determined once the final location is determined in the car. The threads will be covered with the yellow gasoline resistant teflon tape and tightened.

The second fuel pump that is currently on the Driver's side of the tank will come out and the OEM siphon system will be installed back. The fuel tank will have one pump in it as it was from the factory and that pump will be feeding the surge tank. The second pump that used to be on the driver's side will go into the surge tank and will be feeding the engine. Here is the quick rundown on each connector on my tank




More updates to come as I have a chance to install it in my car and to try it out on the race track.

- - - Updated - - -

What are you going to use for hose? I think I am going to use the Gates Barricade F/I hose. Was going to look into ptfe but this should work and be easier to work with. Looked into making one but my skill set is not up to it.

Same thing, I am planning on using the regular FI fuel hose for routing everything. Possibly adding some sort of sleeve over the hoses.

[Inflame]

Very cool, it's odd that I've tracked with a nearly empty tank and have never experienced fuel starvation... Hopefully I didn't just jinx myself!

[Maynor]

Ah, nice job Leo! One question, the 1.5L you calculated for the fuel holding capacity, was that with or without the fuel pump installed? The pump and related hardware will displace a lot of the capacity and reduce your available buffer if so..

One safety suggestion, the spare tire well is designed to crush in a rear impact... That's not where you typically want a decent fuel reservoir and a ready made source of spark installed, if you can help it. You may be want to consider putting it directly behind the rear seat, centered, for a little better protection..

[aeronaut]

Not knowing how the OEM pumps and system really works, will you have to prime the surge tank? Aka, could it stay filled mostly with air rather than fuel in the closed system?

[Cobra1956]

Nice work! Aero, no the oem pump fills the surge. The fuel comes back to the surge and there is the fitting on the top of the tank that goes back to the oem tank. Look at the diagram I included a few posts up.

Leo, estimate $$ for your set up?

As for the propane tank, write idea but wrong application:



Cost was more than I could buy one for and my tool box is only hammer, pliers, checkbook, so that limited my fab capabilities. I track friday but won't have this on until after I get back from vacation. I have to believe this will solve the fuel issue.

[bimerok]

Ah, nice job Leo! One question, the 1.5L you calculated for the fuel holding capacity, was that with or without the fuel pump installed? The pump and related hardware will displace a lot of the capacity and reduce your available buffer if so..

One safety suggestion, the spare tire well is designed to crush in a rear impact... That's not where you typically want a decent fuel reservoir and a ready made source of spark installed, if you can help it. You may be want to consider putting it directly behind the rear seat, centered, for a little better protection..

Just measured the volume of the surge tank with the fuel pump and everything else mounted inside and using a funnel through one of the holes on top to fill it up with water. I used a measuring cup that I "stole" from our kitchen drawer

It came out to 7.5 cups with the fuel pump not primed. So priming it will squeeze few drops more

7.5 Cups = ~ 1.8L or = ~0.47gal. So it actually turned out a little more.

As far as location (thanks for looking out for safety as always!), I am planning to mount it towards the front-most wall in the tire well. I know that the location is not ideal, but I still use my trunk for transporting stuff to the race track and a lot of times I put my splitter inside the car, which barely fits from rear of the trunk to the front seats... Hanging it from the top was my other thought, but I haven't try that yet... Will try to play around with the location once I am able to get into my ice-mobile.

Not knowing how the OEM pumps and system really works, will you have to prime the surge tank? Aka, could it stay filled mostly with air rather than fuel in the closed system?

No, once the surge tank is hooked up in to the system it will always be topped off. The fuel pump from the gas tank will always keep it full and the excess fuel will be returned back to the fuel tank. Also return fuel from the engine will also flow back through the surge tank. So all fuel flow goes through the surge tank.

...
Leo, estimate $$ for your set up?

As for the propane tank, write idea but wrong application:

Cost was more than I could buy one for and my tool box is only hammer, pliers, checkbook, so that limited my fab capabilities. I track friday but won't have this on until after I get back from vacation. I have to believe this will solve the fuel issue.

Since I am doing most of the work myself the cost is pretty low. Right now with fittings and welding job I'm at about $120. I will still need some FI hose and few clamps/fittings, so I estimate it will be around $180-$200 when I'm done. If I didn't have the metal that I already had from the leftover projects, it would be maybe $230-$250 then.

Never thought of the SS tanks, but I'm sure it would be $$


Here are couple more pictures I took after measuring volume today. Shows overall height of the container and started to tighten some of the fitting with teflone tape.

[aeronaut]

Those welds look pretty darn good, but you still might want to pressure test it before you go through the hassle of full install. Any idea how many psi that will see?

[bimerok]

Those welds look pretty darn good, but you still might want to pressure test it before you go through the hassle of full install. Any idea how many psi that will see?

I will definitely do a low pressure test of the welds prior to installing it in the car. As far as working pressure of the surge tank, it will be < 1 PSI.

When it fills up with fuel, any excess will freely flow from one of the ports at the top back into the fuel tank. No pressure build up at all.

[aeronaut]

I will definitely do a low pressure test of the welds prior to installing it in the car. As far as working pressure of the surge tank, it will be < 1 PSI.

When it fills up with fuel, any excess will freely flow from one of the ports at the top back into the fuel tank. No pressure build up at all.

Oh cool, that really opens up options for the tank....FYI motorcycle fuel tanks are steel. I'm not sure what (if anything) the OEM's line the tank with, but there are a number of aftermarket epoxies that are made to line the tank to prevent rust and are fuel proof. So, that cheap-o camp propane tank, or similar, could work. And, my MIG welder could weld it. I get fuel starvation around 1/4 tank, 95 m3.

[Cobra1956]

You have plenty of capacity there. The work looks great. The cylinders are aluminum but I found one for @$55. When I factored in the fittings The tank I have coming was a bargain but won't look as good as what you have there. Since my disaster on wheels is track only I don't care too much about the interior but will try to mount about where a back seat panel would mount. Always get a kick out of people nervous about fuel but I could not get my head around hot wires in a fuel tank. I guess if there is no spark you can't get in trouble. As for the pressure test, if it didn't leak when you filled it, it probably won't. Refresh my memory, is your second pump a high volume or just another oem? 36 hours til track time. Hope they have all of te snow off the course. BRRRRRR.

[bimerok]

My second pump is also OEM BMW pump.

[maj75]

Judging by your photos you have less than half a liter of usable volume in that pot. Not enough to make a difference IMHO.

Your pot is 4" in diameter and while it may be 8.5" tall, the fill and return fittings are about 5.5 inches from the bottom. You won't have fuel above that level. Basic geometry says that works out to a volume of 34.5 cubic inches or .56 liters. You then have to subtract the volume of the pump assembly.

You never addressed the original problem, and have no idea whether either of the old pumps you used were providing the proper fuel pressure and volume. You've apparently done two LS swaps but don't want to fork over the money for a new pump or pumps. I have a '95 and did the Bimmerworld dual fuel pump install. There was a wiring diagram discrepancy, but once sorted, no more fuel starvation. I guess I value my time in dollar terms. It's worth it to me to get a problem fixed right the first time rather than spend endless time screwing around trying to fudge a solution.

Verify your original pumps is working properly or you are just wasting you time with this project.

[Maynor]

Judging by your photos you have less than half a liter of usable volume in that pot. Not enough to make a difference IMHO.

Your pot is 4" in diameter and while it may be 8.5" tall, the fill and return fittings are about 5.5 inches from the bottom. You won't have fuel above that level. Basic geometry says that works out to a volume of 34.5 cubic inches or .56 liters. You then have to subtract the volume of the pump assembly.

There's a skill. Its called reading. Like elementary school skills..

Just measured the volume of the surge tank with the fuel pump and everything else mounted inside and using a funnel through one of the holes on top to fill it up with water. I used a measuring cup that I "stole" from our kitchen drawer

It came out to 7.5 cups with the fuel pump not primed. So priming it will squeeze few drops more

7.5 Cups = ~ 1.8L or = ~0.47gal. So it actually turned out a little more.

Might want to apply that skill to the rest of the already answered questions you posed too.

[Cobra1956]

If the pot is plumbed properly it is full as the return from the pot to the oem tank it at the top of the pot. Correct me if I am wrong but it is the brass colored one on the top of the pot. From the research i gathered, independent of Leo's project a one liter fuel supply should last about 1 minute.

[maj75]

I miss read the picture, see that the return to tank is at top not side. My bad.

However, even at 8.5 inches that cylinder can't hold more than .9 liter max. Assuming the dimensions he gave were correct, I'll take the geometry over "I measured it with a measuring cup in the sink."

- - - Updated - - -

Maynor, please feel free to PM me regarding mistakes in other posts.

[Cobra1956]

Couldn't resist. Went out to the van and grabbed a bottle of oil. Measured in at 7 x 3 x 2.75. That contains 1 ltr of volume and you have the neck of the bottle. Think you forgot pie in your math as a 4 x 9 has just under .5 gallons. You should have over 1 minute at wot. I can't speak for other but I appreciate any feedback to these threads as I always want to make sure I have everything covered. I also enjoy diy when it comes to my car. It might take me longer and even cost a bit more in the end but I learn so much from the project and all the adult supervision. If I had Leo's skill and tools I would have made one versus buying one.

[bimerok]

Judging by your photos you have less than half a liter of usable volume in that pot. Not enough to make a difference IMHO.

I typically divide people in to two categories. One would be theoretic. This is the person who would have all the formulas and possible problems/solutions worked out on paper, while arguing and trying to prove that he has a solution for everything, because his theory and formula said so while that person would never touch a physical thing to verify.

Another would be practic[al] – this would be a person who may or may not use formulas and theories to finding and applying the solution. That person is hands on and unlike what the theory and/or formula says that person knows what does work or doesn’t, because it was tried and tested in the real practice – not on paper and not in that person’s mind.

Then there are “assumers”, who think they know more than they really are. These people make significant assumptions without much basis for it or any effort to find the information. They just fill the missing information with the assumption/guess and pretend that this is the truth. Then the whole discussion and points are being based on something partial at best. This is what I see you as 85% and another 13% as theoretic.

…Your pot is 4" in diameter and while it may be 8.5" tall, the fill and return fittings are about 5.5 inches from the bottom. You won't have fuel above that level. Basic geometry says that works out to a volume of 34.5 cubic inches or .56 liters. You then have to subtract the volume of the pump assembly.

Above (along with the rest of your post) is a perfect example of how you value your time in dollar terms where you cannot even see that both of the fittings mid-height are “feeds” as far as tank’s concern. Nor can you calculate correctly using your basic geometry skills what the correct volume of what is available if the tank was to be only filled to 5.5”. I guess you have failed that introductory geometry class. Let me refresh you on that basic geometry skill. Volume of a cylinder = Radius x Radius X Height X Pi or 2 x 2 x 5.5 x 3.14159, which equates to ~69 cubic inches or 1.1 liters (if we are going by what you see in the pictures and using in your example).

…You never addressed the original problem, and have no idea whether either of the old pumps you used were providing the proper fuel pressure and volume. You've apparently done two LS swaps but don't want to fork over the money for a new pump or pumps. I have a '95 and did the Bimmerworld dual fuel pump install. There was a wiring diagram discrepancy, but once sorted, no more fuel starvation. I guess I value my time in dollar terms. It's worth it to me to get a problem fixed right the first time rather than spend endless time screwing around trying to fudge a solution.

Here is the perfect example of where you pull baseless assumptions out of air without any knowledge of what you are talking about. To enlighten you: The original problem is fuel starvation due to the fuel slush under certain cornering G loads. I attempted to “practically not theoretically” address the problem with a solution that did not work and now I am working on the next one.

You are assuming baselessly that my pumps are old pumps, where in fact the feeding pump is a new pump, that was pressure tested and dyno proven to be adequate all the way to the red line of the engine, but for some reason you assume that you know otherwise for a fact.

Also for the reason only known to you, you think that you have a say in where I should or should not fork my money based on the swaps I made, while again having no idea what parts are new and not in my case.

I’m glad that you did BW dual pump install in your car and it works great for you without any issues. I’m not claiming to be the best electrician or anything, but I don’t need to look at any wiring diagrams for something as simple as connecting another fuel pump in my car, so I wouldn’t know if they had any issues with that to begin with nor was I using their kit.

Verify your original pumps is working properly or you are just wasting you time with this project.

I appreciate the solid advice here. It is invaluable.

However, even at 8.5 inches that cylinder can't hold more than .9 liter max. Assuming the dimensions he gave were correct, I'll take the geometry over "I measured it with a measuring cup in the sink."

This is another good example of the person who just theorizes and try to make it look like it is the only right way, when in fact the very basic of your premises is incorrect. You can’t even use the correct formula to calculate the volume based on the information you have and now trying to prove that someone is wrong and you are correct.

I will take “I measured it with a measuring cup in the sink” over “in theory” and “I’m hoping that my formula is right”. You can use everything that Pythagoras wrote, but if I see that I can fit 1.8L in practice, then I will take that over "theory says otherwise".

Posts like yours fill interwebs with tons of mis-information because of "hearsay" and "my formula says".

[Cobra1956]

Pre coffee everyone is a little sensitive. Keep the updates coming!

[bimerok]

Pre coffee everyone is a little sensitive. Keep the updates coming!

LOL. Just re-read my last response, came out a bit more harsher than intended... Having only 3.5 hours of sleep that night did not help

[Cobra1956]

LOL, think you made your point. I couldn't help but think, I wonder if he knows he will never get the time spent on that back? From reading this thread form top to bottom it is obvious that all here are throwing out ideas to make sure we solve the issue. I have a new engine coming. Stupid money but I am not going to deal with the fuel issue so I am very interested in what you get. On vacation this week but my pot is at home waiting and I have to order out my pump. Fuel lines are a question but other than that this seams a no brainer. My car will never see the street again and never race but I want to have something at the track that is well sorted as well as I am not an educated man but I would have been the one with the measuring cup so I can identify to what and how you are going about this.

[bimerok]

It was some very nice weather over this past weekend so I had a chance to work on my surge tank completion and installation. Here is what I did.

Since I already had my dual pump setup all I needed to do is to put the fuel pump that was on the driver's side acting as a transfer pump over to passenger side to now feed the surge tank and take the pump from the passenger side that used to feed the engine and place it into the surge tank. I needed to do this because my transfer pump was older where as the other pump was fairly new and only couple of years old or so... This task was pretty straight forward and in about 30 minutes I had my older pump on the passenger side and newer pump removed from the housing and attached to the surge tank.










I also reinstalled the original syphon transfer system that came with the car back into the driver's side. Next I needed to fish the power wires for the Fuel pump that is inside the surge tank, so just to make sure the wires I used are fuel rated I sniped some from the old Chevy or Ford fuel pump that my cousin had conveniently laying in the trash bin at their shop. Perfect timing that they had a fuel pump job. The wires were plenty long, so I was happy.

After putting everything and closing the surge tank, there was the most important part that I had to do. Check for any possible leaks. So I connected short lengths of hose on every barb connector and plugged them with something (bolts in my case). For the last hose on top I cut up and used a wheel valve stem so I could pump the surge tank with air, drop it in the bucket with water and check for any leaks.



So the moment of truth. This was make it or break it for me . After pumping the surge tank with ~8psi of air there were no hissing sounds or anything. Then filled up bucket of water and submerged the tank. I was very relieved when everything checked out OK without any bubbles. YES!!!

Since I need my trunk space to transport splitter when driving to the track I decided to install the tank in the spare tire well. I know it is not the ideal location, but it will be there for now. My ideal hose routing was to drill holes in the spare tire well and route hoses on the outside, but after looking for any possible solution under the car, I came to realize that there is no way for me to do that. So the hoses got routed from trunk to the access covers beneath the rear seats. To give some extra protection I also fished the hoses inside the protective rubber layer with spiral reinforcement. They look a bit more bulkier, but I don't mind it.

To secure the tank in the spare tire well I used some brackets that go around the tank and screw to the tire well. I used and placed 6 nut-serts/nut rivets to the wall to screw into. They are as easy to install as regular rivets and are amazingly convenient.



Here is the tank mocked up in the well


To wire the surge tank pump I reused the same wiring that I had for my secondary pump by just fishing that wire back and shortening it. As mentioned in the beginning of this thread I used a separate circuit that I made, which gets triggered by the original pump wiring.

Here is the tank all plumed up and wired up.


After everything was installed and plumbed, I needed to fill the tank, turning ignition On and Off would have taken forever, so I just directly powered the pump until I heard the fuel started flowing from the top of the surge tank and into the return side of the BMW tank. At which point I proceeded with firing the car. Without any surprises everything started up and was working perfectly fine. I idled the car for about 10 minutes or so and then drove to gas station as I pretty much had it running on vapors. Used up as much gas as I could before working with fuel pumps

The trip was uneventful and again everything function as expected. Although one of my main goals was not to hear the fuel pump running, I could still hear it a bit louder than I wanted. I guess taking it out from the tank and tons of insulation that it used have and putting it in the little bottle in the trunk didn't really mute it very well, but it is still way quitter than any of the external pumps I have heard and really not too bad. I only hear it when the car is stopped at the light and as soon as I start driving with RPMs just above 1K, the pump is no longer heard.

So far everything seems good. Now all I need is to test it out in the real environment, at the race track... As soon as I am able to test it out, I will report back. I am hoping for a 100% success this time around.

P.S.
Working with the fuel pump area while having the roll bar was not a lot of fun to say the least. Good thing that I still had my seats still sliding and folding backs. I just can't imagine the amount of "FUN" you people with full cages and fixed back seats have when you need to access that area... I fill sorry for you

[douglee25]

Nice work! Crossing fingers.

[aeronaut]

Awesome! Looking forward to the test results!!

[bimerok]

Well, I guess it's official. The real test should be performed on April 19th. Just signed up for a track day at NJMP Thunderbolt...

[douglee25]

Sweet!