E85 Discussion - HPFP lubrication and Flow issues Well, E85 is something we all hear about, but is something a bit unproven in the MS3/MS6 platform. I ran a 50% ethanol mix for a while, and then tried straight E85. The 50% mix ran very well, with almost no degradation of mileage. Once the fueling was optimized and the required load curve updated (due to MAF scaling), the car pulled just a little bit harder than what it has before, but at ~3lbs less boost, and 1* extra timing compared to the 93octane tune. So, after a little bit of tweaking of my 50/50 tune, I decided to try out some straight E85. I bumped up the MAF curve just a little bit (~10%), and ran a MAF cal. MAF calibration run is here (2nd gear): http://i64.photobucket.com/albums/h1...al_2ndGear.jpg As you can see, most of the curve is pretty good at about -3%, and the top end is about +11%. Not bad for a guess. I could see that fueling wasn't too far off, and the ECU was able to adjust with the trims after ~30 miles of driving. I decided to brake boost just a touch to throw it into open loop (at 1.10 load), in 2nd gear. The car seemed to jump into open loop with ~11:1 AFR. Again, not bad for a first attempt. ____________________ So I log a quick 2nd gear run and an attempted 4th gear log, to get more resolution of what the car is actually doing. 2nd gear log seemed fine, AFR's in check. The 4th gear run, about 20 seconds later, is where the problem seems to start. As you can see in the following graph, my fuel pressure starts to dip like crazy in decel after the 4th gear run (which i ended early due to the AFR's climbing into the 13's): http://i64.photobucket.com/albums/h1..._HPFP_Fail.jpg ____________________ I could only view the AFR via my AP, so I stopped the 4th gear pulls when I noticed ~13:1 AFR. I figured this was just my MAF curve needing some tweaking. So I turn off the data log, and drive ~1/2 of a mile. The car started to act like it didn't want to accelerate, so I flipped on the data logger on the ~9 minute drive home. At this point, I flipped the AP over to fuel pressure and noticed that it was wwaaaayyy low. Around 60psi cruising, with two points of higher pressure. Unfortunately, I think this was my fuel pump dying :( As you can see in the next graph of the drive home, my fuel pressure is pretty screwed: http://i64.photobucket.com/albums/h1...o_Pressure.jpg Soooo, my fuel pump is an aftermarket rebuild. I'm going to check tomorrow about this whole issue. Hopefully we can work something out. My plan is to tear down the HPFP tomorrow night. Maybe we'll see some tell-tale signs. I have read about ethanol being suspect in many HPFP failures in other platforms (primarily BMW's), but nothing is really proven. I have had my car for just over two years and have run ethanol enriched fuels the entire time. I rebuilt the fuel pump about 6 months ago, and it has run very well since. I am running stock turbo and it was completely maxed before running e85, which somehow lowered my boost after adjusting my MAF curve. I think it would be wise to run some kind of extra fuel lubricant if you plan to run high ethanol contents - Marvel Mystery Oil seems to be mentioned in many places. ____________________ ____________________ Of importance to the big turbo guys is the injector duty cycle. I am running a stock turbo, and this quick blip in 2nd and 4th gear that I logged shows some pretty high duty cycle. I am running the CPe CAI, and the MAF voltage is showing 4.5+. I'm sure this would be a bit higher in higher gears, which would probably drive the injector duty cycle up even higher (for reference, i was logging 285-290g/s on 93 octane at this voltage). It's already hitting 90%+ in second gear with e85. http://i64.photobucket.com/albums/h1...Duty_Boost.jpg ____________________ ____________________ I'll post the actual logs if anyone is interested in analyzing them. ____________________ ____________________ Cliff notes: -straight e85 might kill HPFP's -straight e85 requires very high injector duty cycle with a maxed out stock turbo. BT guys should be aware of this. |
Scratch that, I just found an article from Hitachi that alludes to the current design being non-resistant to E85. Page 357, next to last paragraph: http://www.hitachi.com/ICSFiles/afie...009_07_111.pdf |
I had a chance to pull the fuel pump this evening and take some pictures. [Went to a Joe Pug concert last night with the girlfriend, so the car has been parked in my garage.] My first thought after pulling it apart was "holy shit this thing is dirty": First, the entire album is here: FuelPump_20110428 pictures by phate1985 - Photobucket It is split into many sub albums based on the individual parts when they were immediately removed from the housing and with me touching them as little as possible, and a corresponding 'clean' album. Here are the highlights, so I'm not linking 80 something pictures: Shaft and barrel inside the piece that unscrews (dirty): http://i64.photobucket.com/albums/h1...y/DSCN3000.jpg Barrel pulled out of piece that unscrews. It looks really dirty, but I realized after I wiped a black spot that some of it was the supplied moly lube still on there: http://i64.photobucket.com/albums/h1...y/DSCN3006.jpg Dirty pump shaft: http://i64.photobucket.com/albums/h1...y/DSCN3021.jpg Piece that unscrews, holds the shaft and shaft barrel (dirty inside): http://i64.photobucket.com/albums/h1...y/DSCN3029.jpg Here's the pump housing, itself. It has some gunk buildup, but it's not terrible. There was what seemed to be a carbon deposit in the bottom of this picture. After the pic, I scraped it with a pick and it turned out to be very powdery and dry. Not sure what that is: http://i64.photobucket.com/albums/h1...g/DSCN3038.jpg And after I scraped it: http://i64.photobucket.com/albums/h1...g/DSCN3039.jpg So I cleaned everything with brake parts cleaner (non-chlorinated, of course) so I could look at any damage to the bare metal. The parts cleaned up very nicely. Don't mind the dust, I have no intention of putting this back together as-is. There is definitely some light scuffing on the shaft, but it is so shallow, I would expect it to be only cosmetic. I'll see what the company has to say about all of this. http://i64.photobucket.com/albums/h1...n/DSCN3042.jpg The barrel housing looks decent. The smaller O-Ring (for the shaft) seems to be intact without any signs of wear: http://i64.photobucket.com/albums/h1...n/DSCN3061.jpg Although it does show signs of wear on the spring seat (although I doubt this is anything to be concerned about): http://i64.photobucket.com/albums/h1...n/DSCN3066.jpg And you can see that the spring has started to wear a groove into the centering ring on the spring seat: http://i64.photobucket.com/albums/h1...n/DSCN3070.jpg The piece (barrel) that holds the shaft seems fine on the inside: http://i64.photobucket.com/albums/h1...n/DSCN3059.jpg But the outside seems to have incurred some wear. Running my fingernail across those spots, I can just barely feel them: http://i64.photobucket.com/albums/h1...n/DSCN3051.jpg But it is not even across the entire circumference: http://i64.photobucket.com/albums/h1...n/DSCN3049.jpg ____________________ ____________________ After all of that, I don't know what to think. The tolerances on these pieces are so tight, the wear seen here could make a difference. BUT, with all of the nasty gunk seen in the pics, I'm very tempted to (re)clean the parts, soak them in fresh oil (this time) and try it out. At this point in the day, I'm just going to talk to the company in the morning and probably send them the pics. |
You know that stoich for e85 is around 9.7:1 right? You are running very very lean if you are tuning for gasoline AFRs. You should be tuning with lambda not AFR if you are running straight e85. Lambda will be the same with both fuels. |
Lambda for these cars is 14.7:1, no matter what fuel you use. That is how wideband O2 sensors work: they just convert lambda to whatever you tell it to. .78 Lamda (which is about right for ethanol), works out to ~11.45:1 on our readout, but it is really ~7.5 actual ethanol AFR. There is some good discussion about this in Driver's E85 thread, starting on the third page: http://www.mazdaspeedforum.org/forum/foru...tml#post689173 After looking at my logs again, I'm a little suspicious of the high pressure relief valve. In the second graph of my first post, you can see the pressure hitting 2k psi, which may have caused the failure as I have been running slightly higher pressure since ATR had the ability to control it. I'm going to pull the valve in the next couple of hours. Edit: the .78 lambda is just a starting point for me. I'm sure with the direct injection, the 'ideal' AFR with E85 could be different - possibly at a leaner ratio than the .78 lambda. It's just that no one seems to know, yet. Edit 2: I mentioned in my first post about 'optimizing' the MAF curve for ethanol. If you want the AFR tables to read correctly, the entire MAF curve needs scaled ~35% to compensate for the extra fuel. This throws off just about everything, haha. I'm sure that if I showed the MAF g/s, rather than the voltage, it would show something like 350g/s at the 4.5V. This is absolutely not a true reading of course, but we are indirectly controlling injector pulsewidth with a calculation that includes the MAF V: MAF g/s conversion. My calc'd load values were over 2.5 with a 50% blend, it was ridiculous. All targets need updated with ethanol. |
I just realized that I need to update my timeline and what I've done through today. I thought I had written some of it, but I guess not... After talking to KMD, they think the light scuffing seen in the pics is just normal beak-in wear. I am inclined to agree, as the scuffing seems only cosmetic except for the couple of tiny pits on the outside of the 'barrel' piece. So I cleaned them up (again), soaked them in oil, and reinstalled. After reinstallation, I drove the car for ~15 miles but never saw pressure go above 90psi. [This is still with the e85 in the car.] So cleaning them definitely didn't help anything. Now, I'm unconvinced that the HPFP is actually bad. If you look at my second graph in the first post, you can see fuel pressure hitting just over 2k psi. I'm thinking it could just be the relief valve. After getting back from my 15 mile drive, my AP was still displaying fuel pressure. I get back, shut the car off but flip the key back to the 'run' position. I immediately notice the fuel pressure gradually starting to build. Of course I start to datalog this, and it looks like this: http://i64.photobucket.com/albums/h1...g_20110429.jpg After starting the car once the pressure built to just over 1,000 psi, the pressure immediately dropped to the ~60psi I was seeing before. So I shut it down, again, only to watch the process start over. So right this second, I have the relief valve sitting in front of me, but I don't have an air compressor in my apartment's garage. It has to wait until tomorrow to get tested. ____________________ Edit: Another thought that I have had is that the ethanol is breaking down all of the nasty shit that has been in my gas tank and gas lines for the last 50k miles. Ethanol is more of a solvent, so I'm thinking the extra dirt in the fuel could be messing something up :/ When I pulled the relief valve, the little bit of residual fuel on the valve face was pretty dark. Not a good sign. I'm seriously considering a supplemental in-line fuel filter before the HPFP. |
perhaps the e85 clogged the relief valve or is causing it to stick |
i think this man deserves some free elevated access for this thread. good read for sure. |
Some people did have problems with the relief valve when seeing pressures close to 2k psi. Perhaps the fault is not entirely on the E85 fuel but the higher pressure. Buy new relief valve and try pressure at 1800psi? |
Quote:
Nice work and you have been awarded a VIP status... Marvel Mystery oil is good shit. I have been using it for 20 years in any type of hybrid race fuels i have mixed. In the late 80s i would frequent a local shop where they sold race fuels and race tires. I picked up a lot of info there, esp when mixing race fuel and the requirement needed to add lubrication to fuel additives for increasing octane or the mixing of alcohol fuels with gasoline. |
Quote:
Quote:
|
I called the hydraulics shop and I'm having the valve tested tomorrow morning. This place has been in business for a really long time and they know their stuff. We've had hydraulic lines and fittings made there, from scratch, so I'm hopeful they can figure it out. While I'm there, I figured I would see if they knew of any relief valves that might be a direct replacement for us... |
I went through the hydraulic shops in my area for the RV, I had no luck. All I could find locally was a cartridge type valve...wayyy tooo big for the application. If you ask for an RV, ask for an in line relief valve. You are correct I am not in production of the valves, since I have yet to quantify the results of my setup. Info should be coming soon though, as I am getting ready to tune the car with the new setup. Thanks for the great information and testing!! |
Man, this is interesting stuff - I thought for sure our injectors wouldn't be able to handle the extra fuel volume required by e85, even with the factory k04. I'm interested in seeing what you find out when it's all said and done. |
I just talked to @jwilkins88 on the MSF chat about the HPRV and my fuel pressure issues. He had the great idea that my fuel pressure could just be the HPFP spill valve not getting any/enough voltage (that's what the two prong connector on top of the HPFP goes to). That definitely makes sense, and that will be the next step. I think the voltage is probably the most logical step right now, but getting the relief valve tested will just take a few minutes tomorrow morning. |
High pressure relief valve is good. It held 1,900psi, but opened at 2,000psi. On to the voltage... |
That's your stock HPRV? Time to raise the requested fueling again in ATR. |
Quote:
____________________ So I just reinstalled the HPRV and got everything buttoned up. I started looking into the voltage issue. The spill valve, located on top of the CDFP, is getting voltage but not enough amperage, it seems. I can jumper the spill valve with +12V and the fuel pressure immediately goes up. I didn't want to do it for more than a split second, because I was idling and saw over 1,900psi on my AP readout. Of course I'm datalogging all of this: The spikes are where I applied power directly to the spill valve: http://i64.photobucket.com/albums/h1...pill_Valve.jpg I could hear the HPRV opening and letting off pressure when I did this. I got to hear it earlier today when it was getting tested - kind of a high pitched sound like a very small tea kettle going off. On the spill valve's connector from the PCM when it is unhooked, I get voltage readings of anywhere from 12-17V. I checked for short circuits and everything seemed fine. I cleaned my battery terminals (which were a bit nasty) and checked the two main harness connections to the PCM. No corrosion in the harness connections. I checked for continuity between the spill valve connector and the PCM connector, and both check out fine. I'm unsure, at this point, of what is causing this - anyone have an idea? I've gone through the shop manual checks and it leads to the fuel pump being replaced, so I really don't want to stock out just to have them fuck around and replace the pump (especially since I have aftermarket internals). If the fuel pump were to check out good, the PCM needs replaced, according to the shop manual... |
Slightly modified spill valve theory: it's too dirty to work properly with the supplied voltage and amperage. Reasoning: The internals were found to be gunked up with nasty shit. Why would only the internals be subject to that? I just drove my car for ~25 miles and didn't get a P0091 MIL - that's a "fuel pressure regulator control circuit low" code. I can pull the connector off of the spill valve and it instantly throws a code, so I think it is getting the same amount of power as prior. Sorry if I'm rambling. Any ideas welcome. |
wasn't there a tsb for the fuel pump in which they put a new solenoid on? |
i would try and find another complete CDFP from someone that u can swap in. this would be the only way u absolutely can rule that out. |
Quote:
Once I have the new assembly, I'll take my current one apart to see what all is going on. Hopefully we can rule out that the ethanol caused the problems, directly... Quote:
|
yeah the tsb replaced the pump because the new pump had an upgraded solenoid on it. they did that when they performed the tsb on mine. |
Quote:
you and i been around way too long lol |
Functional pump should be here Monday. If it works I'll swap pump internals and continue tuning. If it turns out to be the PCM, i'm going to stock out and head to the dealership... |
We have PRESSURE! http://i64.photobucket.com/albums/h1...85_101_DL1.jpg I just took it for a quick spin. This datalog is with the completely stock pump, so I didn't want to push it. I am letting the car cool off while I do some studying for an 8am final, tomorrow. I'll swap the internals this evening and dial in the MAF curve tonight if I have time. Hopefully I don't run into any more problems like this. Of course, if it occurs again, then we may have a better idea of what e85 is doing to the pumps. I plan on taking the old spill valve apart to do some diag, as well. This may not be for a couple of days, but I'll post up pictures when I do. Many thanks to @stokes620 for getting the pump to me in such a timely manner. |
so are you gonna keep running straight e85 to see if something like this happens again? |
My rejoicing was a bit premature, I suppose. The spill valve started doing the EXACT same thing, with almost the same mileage, on the same road as the last one. That ended my endeavor a bit early, so I pulled the spill valve out of my original pump (the first one that failed). This thing is fucking NASTY. The parts are sticky, everything was sticking together when I pull it apart. Spill valve housing: http://i64.photobucket.com/albums/h1...9/DSCN3080.jpg After cleaning it up, it looks amazing: http://i64.photobucket.com/albums/h1...9/DSCN3094.jpg Solenoid: http://i64.photobucket.com/albums/h1...9/DSCN3081.jpg http://i64.photobucket.com/albums/h1...9/DSCN3090.jpg Moving pieces: The center "button" has a tiny spring under it that presses against the solenoid. It would barely move when I pulled it out: http://i64.photobucket.com/albums/h1...9/DSCN3087.jpg After cleaning, with the button and spring removed: http://i64.photobucket.com/albums/h1...9/DSCN3093.jpg It looks like the solenoid comes apart, but I'm not willing to venture that far, just yet. I'm letting this dry out overnight, since I don't have an air compressor at my apartment. I'll re-clean the small parts tomorrow night, then reinstall. ____________________ Quote:
|
The ethanol is probably kicking shit loose in your fuel system as well. Should probably expect to have to clean some more stuff after a while. |
show us your face |
Excellent and thorough investigation. The carbon buildup doesn't seem to be from "gunk in the tank" as it seems to be finely deposited on the valve internals. It appears that you are observing a reaction between fluids in the pump solenoid that don't agree with each other and coat the metal components. It could be the E85 - perhaps it mixes with some oil and in the right quantities and correct pressure and temperature results in this carbon film. How long did it take for the pump to become non-functional? I would try normal gasoline for the same time interval and re-inspect the pump to determine if it is indeed the E85 causing the deposits. I have taken pumps aparts several times and I have not seen any deposits on them using up to E10 fuel. |
Thanks, Lex. You are right, it may be some sort of reaction with the E85 and something that was already in the tank/lines. The fact that flex fuel vehicles can switch between gasoline and e85 makes me question this, though. If I were to do this again, I think I would have the gas tank and lines boiled at a radiator shop before switching to e85. I put together my original pump with the cleaned spill valve parts and reassembled with the KMD internals. Reinstalled this morning but no dice; the pressure never got above ~60psi. So it is almost definitely internal to the solenoid. The solenoid is probably not meant to be taken apart, but I'll try to disassemble it tonight or tomorrow night. That split ring lock looks intimidating, right now. In my mind, though, it makes sense that the plunger would be harder to move if the same buildup is seen inside of the solenoids frame. The ECU can only supply a finite amount of amperage without connecting to a relay. |
does spill valve=relief valve? or is it something totally different? |
Quote:
The relief valve is a mechanical ball and spring check valve, attached to the fuel rail right beside the fuel inlet. It only opens when pressure is above ~1900psi. The fuel is then routed back to the CDFP. Edit: I just had a funny thought. The relief valve is like a bypass valve for your fuel system. |
Quote:
Dark deposits that are carbon based are usually found in places that are very hot (ie combustion chamber) |
That's a damn good point. Might be stripping coatings out of your fuel lines.... :ouch: |
Quote:
Quote:
BUT, we shall leave no stone unturned. I called Mazda Motorsports about the possibility of ethanol damaging the lines. The rep wasn't sure, so he's emailing one of the Mazda engineers. Edit: I did find a couple of articles describing a reaction between MTBE and ethanol, but I thought MTBE had been phased out of all gasoline for a few years. I'm going to look into this some more. |
If you think it's a reaction between the standard gasoline and the ethanol based fuel, wouldn't running 100% E-85 resolve that? |
Quote:
Quote:
Quote:
|
In the Evo community, there have been multiple reports of gunk on the injectors from switching to E-85. The prevailing theory is that it is indeed E-85 cleaning the fuel system from regular gunk in 91. In the case of the Evo X, E-85 would eat the stock fuel pump intake filter and deposit it in the fuel filter, clogging it up. |
All times are GMT -7. The time now is 01:50 PM. |
Powered by vBulletin® Version 3.8.6
Copyright ©2000 - 2020, Jelsoft Enterprises Ltd.
SEO by vBSEO ©2011, Crawlability, Inc.
vB.Sponsors