Pulse width modulated ECU fuel pump New fuel pump option in the works

I am getting close to completing my "massive overkill" fueling system. I had a bad Walbro and decided that it did not make sense to spend $4k on a shortblock and $69 on a fuel pump... what strange priorities we have, right?

Anyway, Fuelab has an in-tank pump that will make a Bosch 044, dual Walbros, even Aeromotives look like trash... we are talking about Magnafuel and Weldon quality product... the shop I'm using will be offering a drop-in solution for the LGT... more details to follow. I went this route because I don't trust Walbros, one Bosch 044 can not flow enough E85 at 400WHP and surge tanks just seem dumb for a daily driver.

The Fuelab pump incorporates a PWM controller that continuously varies the pump speed based on dwell... Fuelab is working on a controller, but surely the Subaru ECU must use PWM out of the ECU before it sends the signal to the fuel pump controleer which then adjust voltage accordingly to the pump. Th3 Franz seems to believe this is the case also.

The Fuelab pump is fairly choosey- it need a 500-1500Hz signal... so anybody here happen to know where I can get a nice clean PWM signal for the fuel pump (NSFW where are you). I might be the first person with a legitimate claim to need to tweak the hidden dwell cycle (33/66/100). This pump could loaf along at 20% and be just fine for daily driving.

Properly tuned you could approach a "near-returnless" fuel setup. You of course push a bit more fuel than you need, but if pump dynamically follows the load, you end up recirculating a lot less fuel.

I fail to see how this product would fill a need. There's nothing wrong with the factory fuel pump controller. Walbros aren't the best but they get the job done as long as they are "real" Walbros. You're solving a problem that doesn't exist.

Then again, people will buy almost anything on the basis of internet forum hype.

The single Walbro pump can't keep up for E85 400whp applications. I know we don't have that many applications for the LGT, but lots of STI guys are in that range. I'd like to have enough flow with a single in-tank option

I'll be receiving my DAQ card soon and will be taking direct measurements for the frequency and duty ratio of the stock PWM signal. Fuelab is good stuff. My buddy runs in-line pumps and I just got my Fuelab regulator and filter today in the mail. I'd like to stick with the single pump if possible.

The Fuelab pump is fairly choosey- it need a 500-1500Hz signal... so anybody here happen to know where I can get a nice clean PWM signal for the fuel pump (NSFW where are you). I might be the first person with a legitimate claim to need to tweak the hidden dwell cycle (33/66/100). This pump could loaf along at 20% and be just fine for daily driving.

 

Properly tuned you could approach a "near-returnless" fuel setup. You of course push a bit more fuel than you need, but if pump dynamically follows the load, you end up recirculating a lot less fuel.

I saw your PM before I saw this thread... Like I said in the PM, I believe the ECU uses a PWM signal to control the pump, but I don't know which wires you could tap to access that signal. But it sounds like Th3Franz is about to answer that question.

As luck would have it, I have done a little bit of exploration of the ECU code that governs fuel pump duty. Not a lot, but I just kinda stumbled into it while poking around...

The ECU chooses between four possible duty levels: 0%, 33%, 66%, 100%. I have not figured out exactly what conditions cause it to switch between levels, but in watching fuel pump duty in the logger I mostly saw it at 33%, jumping to 100% without much provocation.

There was a guy on RomRaider who had a pump that was very noisy at 33 or 66, and I made him a custom EcuFlash definition that allows you to tailor those to whatever you want. He set them both to 100, so his "four" levels are now 0, 100, 100, 100. His pump is quiet now and he is happier. So if your pump works best as an on/off switch, I can help with that.

Or if you'd like to change the 33% value to 20%, then that's easy too. I think the 100% value can be adjusted as well, for example if you know that 80% is plenty... but I don't recall for certain. I can look into it if you want though.

But it sounds like you are hoping to control fuel pump duty in continuous proportion to fuel demand, and unfortunately I can't help with that. There's just 4 stages to choose from. And unfortunately the logic that governs the switching between the 4 stages is a mystery to me. There's a whole lot of comparisons between different values in memory, and checking flags, and I haven't got much of an idea what they mean.

But the stock fuel pump control scheme seems simple enough that I think there's a good chance that your aftermarket pump will work just fine with it, assuming you can find the right place to tap in. And I think there's a good chance that the fuel pump connector is the right place. But we'll have to wait for Th3Franz to find out for sure.

If you wanted to get really clever, you could try to find a hobbyist EE to make something that looks at the fuel injector duty cycle and converts it to a 1khz PWM signal with the same duty cycle, with some bias around the low end to keep the pump running strong enough when the engine is idling. Plus some logic to run the pump at high duty (the stock tune uses 100%) to prime the pump when you turn the key, and to give good pressure during cranking. Probably not worth the trouble, but conceptually it doesn't seem to hard.

This sounds like a lot of effort for almost no return, the Subaru ECU is not going to be able to control the fuel pump in the way being considered. Even if you could there is no real reason to do it.

You really don't even want to try a true "returneless" system since you need to maintain presure in the fuel rail at all times.

Just put the magic pump in place and be done with it; the stock ECU will still dial it back at idle and cruise based on voltage - just like a Boost a Pump in reverse

If you wanted to get really clever, you could try to find a hobbyist EE to make something that looks at the fuel injector duty cycle and converts it to a 1khz PWM signal with the same duty cycle, with some bias around the low end to keep the pump running strong enough when the engine is idling. Plus some logic to run the pump at high duty (the stock tune uses 100%) to prime the pump when you turn the key, and to give good pressure during cranking. Probably not worth the trouble, but conceptually it doesn't seem to hard.

mpiotro, PM me if the stock signal isnt good enough and you need this. I'm (fairly) local and can build this for you. I did something almost just like that for a class a few years back.

Boostin'- In terms of mass appeal you are exactly right. However on E85 at 400+WHP you would need dual Walbros. Surprisingly (to me) even a single Bosch 044 will not support enough E85 flow at 25+ psi boost. So now you're looking at dual pump options- replacing 1 component with 2 components is not a solution in my mind.

I am not certain the PWM from the factory fuel controller will work with the Fuelab required input freq. After some more discussions this AM, it is HIGHLY unlikley the factory fuel controller outputs a high enough freq range for the Fuelab pump to acknowledge. Factory fuel controller output is almost certainly BELOW the 500Hz lower threshold required by the Fuelab pump input.

RAO- I'm not really talking about a true returnless system. My only other option is to run the pump continuous duty at 60% speed. Even at 60% it will be pushing enough for E85 at ~425WHP. This means it is pushing ludicrous amounts of 93 pump during my 45 minute commute to work on the interstate. The stock 33%/66%/100% duty cycle is probably fine with regard to pushing plenty of fuel to maintain system pressure without being overly redundant in volume.

Apparently I am the only person interested in a big HP single pump in-tank solution. Even if I'm the ONLY one, I still want to do it right and use the best technology that is available. I don't think anyone here will seriously argue that ANY fuel pump running at 30% duty cycle will have the same lifespan as if it runs at 60% duty cycle. So my goal remains to utilize the PWM controller built into the pump.

who came up wit this magic 400whp is too much for walbros number? We have cars in the 500whp range with single walbros and no loss in fuel pressure. I don't think those pumps will last more than a year or two, but at $69 a pop, so what if you have to replace it every year. You pay to play.

Secondly, what makes you think this crazy fuel pump with an overly sophisticated control system is going to be more reliable than a walbro? If you want safety, install a fuel pressure gauge with an alarm. That's the only thing that's going to save your engine... if you catch it in time.

I don't think those pumps will last more than a year or two, but at $69 a pop, so what if you have to replace it every year. You pay to play.

You can pay to play that game, or pay a little more to not play that game... That's way too much hassle just to save a couple hundred bucks.

Surely there are solutions that don't require replacing your fuel pump annually and/or risking a blown motor if your pump decides to do its annual failure a little earlier than expected.

Well in full disclosure my car had a very good tune and I still ended up with 2 cracked ringlands. I ran meth and initially though that could be the culprit. It was detonation plain and simple. We have evert reason to believe the fuel pump is crapping out. It is very intermittent. It only takes once at high load on stock pistons to break a ringland. Walbros may be great bang for the buck, but if you honestly believe a $69 pump is "good enough" when you have $7-8K in your motor I'm calling you a stooge.

RAO- please let me know these E85 single pump in tank solutions you speak of.

Did I say single pump?

Seriously, if you can't figure out how to install dual conventional pumps iin a Legacy which will supply more than enough to reliable power whatever set up you could possible have in a Legacy, you really shouldn't be touching the car. 1000+ whp cars re not all that unusual these days and none of them have the magical mystery pump.

I run e85. Am past 400whp and use a single Walbro. I have a fuel pressure gauge at the FPR and one on the dash. I am with mweinwer2 and rao. Walbro works fine, monitor your pressure, if you need more double pumpers work fine too.

That is the only 400 whp 68hta, or do you have another turbo?

If you are talking about the 68hta, then:rolleyes:. And I own, and like, my 68hta.

That is the only 400 whp 68hta, or do you have another turbo?

 

If you are talking about the 68hta, then:rolleyes:. And I own, and like, my 68hta.

No new turbo, yet...

But the 68 now has an 8cm hotside and some porting. As well as an EWG. Shooting for 450whp now. Added a lot of parts and coatings in this build up. Still not complete. Need brakes. Plus time to put it all back together. Been working on build for 4 months now. Need to take a "staycation" to finish my own car. Too much time spent on customers cars, and new product development. LOL.

mweiner knows how it is, I am sure.

+1

or you have a very optimistic dyno with a high CF

I'm being optimistic. Hoping for crap loads of boost (25+psi) and to finally get the compressor to put out rated flow.

Even if it does not happen, it will be fun and I will be ready for a bigger turbo.

Shop I'm using posted up some photos of the completed setup. Plastic housing is eliminated completely, but all electronics bolt up to the new frame, so factory functionality is retained.

http://wrxatlanta.com/forums/showthread.php?t=29347

Double pump = double the odds of losing an engine when one pump fails.

I tend to side with this, there is something to be said about K.I.S.S .

WTF is the flow rating of that beast?