For many reasons (namely a Procharger), I want to be able to run E85 on my 1990. To make things easier, I'm going to add a flex fuel sensor like (I will be running a standalone ECU.)

From what I read, you want to mount this as close to the fuel rail as possible so there's little lag between the sensor input and what's actually going into the engine. It's not obvious the best place or method to do this though. I looked around but couldn't find anyone here mentioning how they added a flex fuel sensor, and I'm a bit wary of doing uninformed modifications to fuel lines myself (I carry a fire extinguisher, but an engine fire would still be no fun).

I'll probably need to do some modifications when I upgrade my fuel filter, so it might be easiest to just add the flex fuel sensor there. That would give me ~5 feet of fuel line between that and the rail, but I could always just run the fuel pump for a few seconds whenever I switch fuels. What do you guys think?

 

If it's a return system (I believe you said you'll be running that still) I wouldn't worry about reporting lag between the reading and what the fuel rail sees. I think you can program a time delay in the tune anyway before it ramps to the higher ethanol content maps. I haven't messed around with it that much but... I wouldn't be too worried about being as close as possible though. You have to figure even with a stock pump the fuel in the line is mostly purged just with the key on prime so realistically what's the time difference at say 5 feet away... I can't see it being more than a second?

 

I hooked mine up within inches of the rail but I have a returnless rail on an LS. Like 84 4+3, if you stay with a return setup, positioning is way more lax and you'll purge the lines with just the pump prime. But alot can happen in a second, so the closer the better. Your greatest variances will come immediately after filling up as gas mixes in your tank, Ive read a couple of white papers that have shown that you can have an inconsistent alcohol% for up to 30 miles after a fillup as the fuel mixes.

On a side note, how are you planning on mounting it near the filter? There's not much room in that area and that was one of the reasons I pushed the flex sensor up top.

PTFE lines aren't easy to use but they're not hard, got to be patient and diligent and check the ferrule is on there right with no strands between it and the PTFE.

 

The pump I'm running should deliver 87gph, which would clear 5' of 3/8" line in just over a second. However, I'm going to be throttling the pump, so that would be more like 10s at idle. I was thinking the prime would probably clear it, but if there were an easier way to mount the sensor close to the rails then that would be a win-win.

That's great info, good to know! I never would have thought it could take that long to mix. I'm going to be running boost, so inconsistent lambda from an inaccurate sensor reading could be damaging. Maybe I could program my standalone to run the pump full blast for the first 30mins after fueling up or something. If you have returnless rails, I'm assuming you're not running the stock lines?

It's generous to call my idea of mounting at the fuel filter a "plan". I was under the car the other day and thought "huh, wonder if I could mount it there?" I'll crawl back under there and think it through a little more.

Thanks for the tip on PTFE lines. Those are the braided steel ones, right?

 

Okay, so with varying the output via the ECM, definitely get it as close as reasonable then. I wasn't sure what you had settled on.

And I totally understand the varying content after fueling. The mixing that occurs in a fuel tank isn't exactly great to begin with... I'd love to see what the CV ends up being after set intervals post fueling. Basically a really bad jet mixer.

 

Running the pump is the common solution, I have a switch wired into the fuel pump relay that will let me run the pump manually, I could probably run the pump while filling as that will get the best results, all depends on how comfortable you are with it.
Stockish lines till the filter, the flex parts of the line by the tank were cut off and converted to -6 PTFE, and everything filter forward is also -6 PTFE. My fuel system is a bit weird and completely still up in the air, I'm struggling to make up my mind about how I want to route it, perks of building your own car lol.
Let me clarify what I meant about by the fuel filter, there is room but you'd have to come up with a interesting fuel jumpern between filter and flex sensor, there's no room to fit both on the same plane.
And yeah, steel braided with a nylon abrasion shield, should get me a really clean result once the motors back in.

The going thought is run the tank dry before filling if you run fuels with elevated ethanol content, no top ups. Otherwise it's a waiting game for ethanol content to settle down, there are tanks that better lend themselves to in tank mixing, like ones where the return is semi-submerged, that'll mix fuel much better than just dropping fuel into the top if the tank.

 

Ahh, yes that will promote mixing, for some reason I thought the return was just a 90* into the tank with no tube. You learn something new everyday.

 

Very helpful, thanks. That's a smart idea to run it as you're filling up - I plan to have a dash-mounted tablet interfaced with my ECU that I could use to toggle the pump when I fuel up. I will also try to get the flex fuel sensor closer to the fuel lines - better safe than sorry. I find myself topping off a good amount to avoid fuel pump starvation on acceleration as described below.

Good info! Lately I've been starving the pump pretty much any time I accelerate in 1st with less than a quarter tank or so. Is it possible that baffle came out of place or something?

 

Fle

 

Note i dont have it ‘input’d’ into my stock 1993 ecm to ‘adjust fuel or spark adv’ automatically.
id love to figure a way to do that.

i have the gage and sensor (plumbed into the return of my 93’s fuel line) for my own knowledge.

i can then plug my laptop into my moates
adapter and manully adjust the tune - lowering the inj lb/hr constant (to give more fuel) and increasing spark adv.

I believe ‘innovAte’ was the brand name of the gage and sensor package. I bought it from summit racing in 2019.

 

The sensor can be thought of as a flow restriction.
you want to mount on the return - right before it dumps into the tank.

Not sure about a 90, but standalone ecm is prob un necessary.

my 1993 idles fine (with stock $da2 batchfire ecm, speed density), even with bigger 224/236 cam and large 80 lb/hr injectors (on both e85 and 93 octane)

 

The continental sensor is good to 600lph with less than a 2PSI drop across the sensor, the manufacture's documentation says max flow with no pressure drop is 400lph. It is restrictive, but not enough to impact performance; At 600lph, they've been shown to only decrease flow by 10-15lph. Even in cases where the sensor is highly restrictive, the sensor should still be placed before the rail and have a bypass loop, that way the restriction of the sensor is taken out of the equation but fuel still flows through the sensor. For a flex-fuel system where the ECM actively adjusts fuel trims and the ignition timing based on alcohol content, the sensor has to be placed as close to the rail as possible and putting it in the return should be avoided. By putting it in the return, the ECM will not be able to compensate for sudden drops or increases in alcohol content until after the fuel has been burned which can prove to be detrimental in a boosted applications, especially if different fuels were used to fill up and the tank had not had enough time to thoroughly mix. Gas stations of the same brand across the street from each other most likely will not have the same alcohol content, let alone gas stations of different brands. If you really want to safely push the limits of what an engine can do, then the best way is to go with an active flex-fuel system, otherwise, you're leaving performance on the table in order to make sure there is headroom for alcohol content fluctuations, or you're retuning every time you top up when you could have just set it and forget it.

 

??? Im finding 78% ethanol seems to be the norm across different gas stations

Theres no way ‘the alcohol ‘at the bottom of the tank’ differs in % content vs. the alcohol at the top of the tank?

Its constantly being mixed with return style systems like ours.

when i go from a tank of 93 to a tank of e85, the gage shows the % change within about 30 seconds MAX of starting up.

Then the % stays consistent throughout the whole tank.

perhaps maybe the OEM has to account for every imbecile who may not wait 30 seconds after starting the car and gunning it onto the highway on ramp.

 

We are talking about mixing flex fuels not going from one tank of fuel to another. The greatest variances in ethanol content come when you top up, not fill up, and once the fuel has mixed you can still see variations; especially with humid days and cool nights, water vapor from the tank sucking in air through the vent condenses and causes the fuel to phase separate, the result is an alcohol slug that will cause the reading to jump at startup and the overall alcohol content in the fuel to decrease. Running the tank low or dry and filling up will result in the least alcohol content variation, which is exactly what you said you do, so no surprises that you get consistent readings quickly.
Secondly, the variation in a flex fuels alcohol content is immense, the ASTM allows anything as low as 51% alcohol content to be labeled E85. That's why most tuners will tell you to find an E station that you like and stick with it, don't go from station to station. If you are seeing 78% consistently then consider yourself lucky, the tanks at the gas station in and of themselves average alcohol content between shipments of fuel, so it's extremely difficult to skew the mix in one way or another and it will most likely stay that way for a very long time.
And lastly, fuels with higher ethanol content are MORE dense, the agitation of filling up a tank is not sufficient to thoroughly mix two fuels of different alcohol percentage, so yes, you can have a higher percentage fuel at the bottom and a lower percentage at the top. If you sit and wait it can take a long time for fuel to diffuse and mix.

I personally have seen a 10% variance after topping up on a road trip, I added 12 gallons to a 26 gallon tank, and watched readings vary for good 20 miles on the highway after letting the truck idle for a few minutes to get the cab nice and cool.

 

Thanks for the info.

That seems like a bad place to put it. As far from the fuel rail as possible, and on the return side so flow through it is lower (although I see that this is an advantage as far as flow restriction goes). Why not put it on the return side right next to the rail? Although I suppose response time isn't very important if it's not feeding into your tune.

I'm surprised and impressed that yours idles well with that cam and those injectors. What's your idle speed set to?

Plenty of people have made power on the stock ECM, but there are a lot of sacrifices. For one, I don't want to have to carry a laptop around with me and flash a new tune every time I refuel. You would need to patch the BIN to correct for E content automatically on a stock ECM, and the time you'd put into learning assembly language and reverse-engineering the BIN would eclipse the cost of a standalone. Second, running a 2- or 3-bar MAP on a stock ECM will mess up things like baro correction. Lastly, GM used a lot of clever tricks in the ECM's software to make it work well with the limited computing power they had at the time. Unfortunately, these solutions came with the cost of needing to be calibrated (with much effort and cost) to a specific engine setup. One good example is their intake tract heating model. The second you make changes, you need to recalibrate these.

If your engine has stock driveability with all your mods, you are the exception. It takes a lot of work and skill or a lot of luck to get that to happen. I've put a decent amount of effort into correcting these, and still, off the top of my head: 1) my car runs a little rough until it hits closed loop 2) my car runs very lean on a hot start 3) every now and then after a warm start my engine will buck and die below a certain RPM. Standalones have so much more computing power that they don't need such a specific solution and can brute-force their way to making things run well on various engine setups. Not to mention the fact that their datalogging is so much better to help you fix it yourself if you need to. Tuning transients (MAP/TPS accel/decel enrichment) is pretty rough with a 4Hz datalog.

While these points are sufficient for me, I have many other reasons for wanting a standalone. These include objective benefits (like being able to individually trim banks based on wideband O2 sensors) and just wanting something to tinker with.

I'm looking at 10psi of pressure drop just through the stock lines, so 2psi would be pretty minor. Is that this one? That's clever to run a bypass loop, although I think you'd still lose some flow to the injectors. The bypass would pull more flow, which would drop the pressure the pump could provide as well as causing a little more frictional loss through the lines. I have no idea how significant this would be, and I'm guessing it would still be worlds better than having the flex fuel sensor inline.

Edit: This is wrong, as ThatOneKid points out below you can program in a delay that counteracts the sensor reacting before fuel gets to the rail.
I don't see a big difference in putting it before or after the fuel rail, though. Say it's after and you switch from gas to E85. E85 will hit the rail while the ECM thinks you're still running gas, and you go lean. On the other hand, say it's before the rail and you switch from E85 to gas. E85 will reach the sensor while you still have gas in the rail, and you go lean. Of course flow through the supply line will be higher than flow in the return, but the difference shouldn't be much unless you're at WOT.

Agreed on the benefits of a flex fuel sensor.

Additionally, it's very well documented that ethanol blending can differ from station to station and especially between summer and winter. I'm surprised you've consistently seen the same value, but that's the exception rather than the rule.

A few posts ago someone referred to some papers showing the tank can still be poorly mixed 30 miles after a fillup. This may not be the rule, but it can happen. It's good to know that you've seen it level out after about 30s. What pump(s) are you running, and is the fuel system constantly at full blast? A large pump running constantly would obvious mix quicker than a small pump, a large pump that's throttled at idle, or a staged pump setup where the second kicks on at X rpm or whatever.

That's an interesting point on separation in the tank! Also an interesting anecdote on your readings varying for that long.

 

Continental makes 3 flex sensors, the first one you linked is a continental sensor made for GM, just one without mounting holes and short tubes.

Putting the sensor before allows you to use an ethanol correction delay, most standalones support this feature. It's a short delay that can be tuned to delay the corrections until the fuel is estimated to have reached the injectors, if you set this to 0 then the ECM will instantly correct based on sensor inputs, which is also equally as bad as having it in the return. Retroactively correcting fuel trims, which is all you can do if you put it in the return after a negative content spike puts your motor on borrowed time if you're pushing ignition timing to the limit because by then the fuel will already have burnt or be burning in the cylinders.

 

My Continental sensor is on the return, on one side of a Y, right before it returns into the fuel cell.
Zero issues. Ever.

 

The more you know! It didn't occur to me that you could input a delay, but I stand corrected.

 

You can very much put it in the return and have great results, I'm saying if you do and are pushing the limits, it's a riskier move especially if you are not using a conservative timing curve, which most street tunes do not. Most street tunes or DIY tuners push timing till knock retard kicks in and back off, which is nowhere close to conservative.

 

No knock sensors, aggressive tune with boost (25 degrees of timing at 9 PSI), flex sensor on the return next to the fuel cell. 250k mile junkyard engine is still alive and kicking since being treated this way since late 2016. The little stock 4.8 made 570 whp on 9 PSI through a TH400 and a 9" converter and ran 9s. Would have went 8s with where I had it in the fall before I decided to sell the engine.

Most of the risk assessment done in this thread with putting the sensor closer to the fuel rail is unfounded.