Fuel level sensors interfere with each other

[FFRWRX]

I'm (hopefully) sorting out the last electrical issue with my build. The fuel gauge does not seem to be working. I took the wiring off the tank and did some measuring. When I put a jumper wire across the output from the 2 fuel level sensors the gauge read full. The ohm reading across these, when in the tank with it almost empty, was 124 ohms which I don't think is correct. But then I took the one plate off the tank (the smaller one) and may have found an issue. Some reading on the forum also pointed to this.

It seems that the 2 fuel level sensor floats tangle with each other when installed. I think that may have been pushing at least one of them down below the fuel level. Anyways, is this normal and how to you get around it? I tried bending the wire supporting one of the floats and broke a piece of the plastic holder. Before I order a new float assembly it would be nice to know how to get them to not tangle with each other. I'm wondering if I can cut off the one float assembly and just use one, but not sure how to get the gauge to read properly.

Rick

[Dave 53]

I recall that each sensor can be mounted in various positions based on the bolt pattern on the tank. You just have to find the combination where they aren't bumping into each other.

While you're on this step of mounting the level sensors, it would be a good time to figure out your vent strategy. There are some good threads in the 818 fuel forums.


Note: I guess what I meant is I just remember having to fiddle around with it until it worked. A common theme (especially mounting the doors).

[FFRWRX]

No, the sensor plates can only be mounted one way. The F5 manual states this and it seems to be true from checking the bolt pattern on the round plate. The way they mount pretty much puts the floats interfering with each other.

[FFRWRX]

I'm thinking the best way to eliminate any possible issue is to remove the 2nd float unit; keep the plate and just cut the stuff off below it. Then wire the one float sensor (on the fuel pump assembly) to a unit like this:

Screenshot 2022-11-20 at 4.49.47 PM.jpg

https://www.tanksinc.com/index.cfm/p...rod/prd748.htm

Cost pretty much the same as getting a replacement float sensor assembly and it can be calibrated to any system.

Anyone use something like this?

Rick

[blomb11]

I just bent the rods so they would not interfere with each other. You also need to make sure and adjust them because the FFR tank is shorter than a WRX tank. So you have to bend the rods so they sit flat on the bottom of the tank and then do the same but at max height.

[FFRWRX]

I just bent the rods so they would not interfere with each other. You also need to make sure and adjust them because the FFR tank is shorter than a WRX tank. So you have to bend the rods so they sit flat on the bottom of the tank and then do the same but at max height.

At least I know that it is a possible issue and not something I have assembled wrong. How could Factory Five not point this out in the instructions?! They just say to bolt them in. In my defence, this stuff was already assembled when I bought the car so I didn't know about any possible issue with it. Anyways, I did break the plastic on the one float assembly trying to bend the wire; it is pretty tough stuff. So I have to buy another one and then bend it and maybe bend the main one as well. That's why I'm thinking of buying the Metermatch gizmo and eliminate the dual sensors and bending....and possibly breaking another one.

[idf]

At least I know that it is a possible issue and not something I have assembled wrong. How could Factory Five not point this out in the instructions?! They just say to bolt them in. In my defence, this stuff was already assembled when I bought the car so I didn't know about any possible issue with it. Anyways, I did break the plastic on the one float assembly trying to bend the wire; it is pretty tough stuff. So I have to buy another one and then bend it and maybe bend the main one as well. That's why I'm thinking of buying the Metermatch gizmo and eliminate the dual sensors and bending....and possibly breaking another one.

I likely have the OEM sensor and would be happy to send to you if you decide to replace it.

[RPGs818SNA]

Unfortunately, Rick, as the sensors hang before bending, they do interfere with each other as they rise with the fuel level. They can be adjusted so as not to, but it’s hard to see them in the tank.

I took a couple of large paint stirrers, laid them over the two sensor while they were installed in the tank, and marked precisely where they fit on the sensor covers. I then removed the sensors and clamped them to the paint stirrers per my markings and hung them so I could see the floats move with the fuel level. Sure enough, they bumped into each other as they rose with the fuel level, but it was easy to see how to adjust them by bending the wires a little so they didn’t bump as they rose.

The dual sensors are used to measure and sum the fuel level in the two lobes of the saddle shaped gas tank in the donor. If only one is used, the gauge will read between half full and full as it assumes the other lobe is full. Using the MeterMatch device you mentioned should let you make the gauge read correctly with a single sensor. That may be the easier route to get your gauge reading correctly.

RPG

[FFRWRX]

I appreciate the offer "idf". I was almost going to go this way by getting a replacement sensor unit.

"RPGs818SNA", don't know why I didn't think of taking both out and simulating them out of the tank to do the bending. But, the fact that I would have to do some bending, and had already broken one sensor by bending it, I saw bad things happening with this approach.

I ordered the MeterMatch and it was delivered this morning. So the plan is to remove the sensor stuff from the secondary unit, and hook up the MeterMatch to the primary one and the gauge. I'll reach in the secondary sensor port and push and pull the primary float up and down to calibrate the gauge. In my head it all works great.

I'll report on how this matching unit works.

Rick

[Hojo]

In a thread last year, I asked for help with my weird fuel gauge. It would read full and go down normally until it got to a little more than 1/2 a tank. It would then go to totally empty in 10 seconds. It would stay at empty for 20 miles or so then pick up where it left off. I pulled the floats out and the ohm reading varied between the values in the WRX manual. I tried bending the float arms but apparently didn’t get them untangled. I will now unhook the one closest to the passenger door and get the MetrrMatch. I never heard of the MM until this thread. Thanks.

[FFRWRX]

I hooked it up temporarily yesterday and did a quick calibration with the one float/sensor. It seemed to work very nicely. I'll post more details when I tidy it up and make sure everything is good.

Only one issue. When you take the fuel pump out of the tank with the lines removed but all the electrical connectors attached, aim the outlet more or less at yourself, and turn the ignition on, the result is pretty much what you would expect.

[FFRWRX]

I think it's beaten me.

I can't get the MeterMatch to work. I can calibrate it with the float both up and down per the instructions, but when I power it off and then back on again, it isn't holding the calibration.

I think it may be something to do with the fact that the fuel sensor and gauge ties into the computer. So I isolated it as much as I could. At one point I took the instrument cluster apart, removed the gauge, and hooked it up to 12V and the sensor. It worked, but the little resistor on the back of the gauge motor got hot....really hot. Checking out the circuit on the instrument cluster, it looks like the "ignition" voltage goes through some resistors; to step the voltage down I assume. So I put the gauge back where it should be so it was getting the proper voltage. Then I cut the leads from the one sensor I'm using (since one leads seems to go to the ECM and I thought that may be messing things up). I grounded one sensor lead and connected the other one directly to the input on the gauge cluster to the fuel gauge (connection A2). Again it worked while doing the calibration, but when powered off and back on the gauge went to full when it should have gone to empty.

I'm not sure what else to try. There are some resistors and connections on the back of the gauge cluster that I have no idea what they are for. The "ground" connection on the back of the gauge seems to go through a resistor before actually going to ground. Whether something is still tied into a computer chip or something else that is messing this up I don't know. My knowledge of all this stuff is pretty much at an end.

Some pictures:

Fuel gauge removed from cluster:


Screenshot 2022-11-27 at 2.54.42 PM.jpg

The back of the gauge motor. The bottom pin is not used and the others are ground, ignition, and "unit" which is the sensor connection:


Screenshot 2022-11-27 at 2.55.21 PM.jpg

This is where things get confusing to me:

Screenshot 2022-11-27 at 2.54.14 PM.jpg

You can see the 3 connections to the gauge motor: The ground runs through a resistor before going to what I assume is a ground. The ignition runs through 2 large resistors (on the other side of the board but you can see the solder connections) as well as through a couple more small ones. The unit (sensor) connector runs up to pin #2 on connector A, but also runs down to some resistors and then maybe to ground?

So that's why I don't want to start cutting traces on the board to eliminate more stuff from the basic gauge and sensor connection...I don't know what they do. Other than maybe step the gauge voltage down. But why the sensor connection also has resistors on another path I don't know.

If anyone has any thoughts on this, feel free to contribute. The simplest thing at this point would be to get the other stock sensor and hook it all up as "normal". As long as I don't break something bending the float wires!

Rick

Edit: I don't give up easily! One more thing I can try. Remove the gauge unit from the cluster again, run leads from the cluster to the gauge for power and ground (so the gauge gets the correct voltage) then a separate wire from the float/sensor to the gauge. That will bypass any of the electronics on the circuit board for the sensor wire. After that.......then give up.

[RPGs818SNA]

I like your spirit Rick.

Before giving up on the MeterMatch, you might try measuring the voltage on the fuel sensor when it is connected between pin A2 and ground. Measure it with the float in both the full and empty positions. These readings should be half the voltage when both the sensors are connected in series.

Next wire in the MeterMatch, program it as before, set the rotary switch back to 0 (the LED should go out), and measure the voltage at pin A2 for full and empty. It should be twice the readings with the single sensor, and the gauge should read correctly.

Next, power the MeterMatch down and back up. If the voltages at pin A2 are different than before, the MeterMatch isn’t storing the values correctly and may be defective. If they are the same, try powering the meter down and then up and check the voltages again. If they are the same, the meter should read correctly. If not, and you’re game, I’ll suggest some more steps.

Good luck,
RPG

[FFRWRX]

Thank you very much for the help RPG. I tried what I wrote under my edit above, and it didn't work. I have to think the unit is faulty and not holding calibration settings. I emailed the company to let them know what I've tried.

My latest setup had the gauge powered by the instrument cluster, but the float/sensor (one only) going directly to the sensor pin on the gauge. Like before, the calibration settings work fine, both empty and full. But when I turn the ignition off and back on the gauge does not go to empty where it was for the "empty" calibration. I'll double check today, but I'm pretty sure every time I go back to the calibration setting, it is lost. What I mean is, if I go back to calibration setting #4 (tank empty) the gauge will not go down to empty as I think it should. So it is not remembering the previous calibration setting. I'll see what the company says about it.

I may try what you suggested just to further confirm things.

Rick

Edit: Got a reply from the company right away; I am very impressed with that! They suggested a couple basic things to check, like that my float/sensor is working correctly. Pretty sure I measured the resistance at one point but I'll check it again through the full sweep at the gauge sensor input.

[J R Jones]

Rick, I did Fighter/Interceptor electronics in The USAF which does not qualify me for Subaru or fuel indication; it does make me dangerous.
Working with the original parts is a deviation possible? Cheating?
1) cut off one float and fix it full up. I assume you may never get below 1/2 sweeping the other float up and down.
2) cut off one float and fix it full down. I assume you may never get above 1/2 sweeping the other float up and down.
3) cut off one float and fix it to the mid-point. I assume you may never get above or below 1/4 / 3/4 sweeping the other float up and down.
4) can you fix both rheostats to one float so that they sweep together?
jim

[FFRWRX]

An update while I'm taking a break.

JRJones: One issue is that I physically broke the nylon part of the second float assembly, so the float arm is separate from the resistor thing that the float arm sweeps on. If it wasn't submerged in gas I would have tried to fix it, but not sure what adhesive would work. Didn't want to take a chance. I could maybe put the 2 broken parts together at a fixed position, but that shouldn't be any different, operation wise, than getting rid of it and just using the main sensor.

Speaking of the main sensor.....that might be the issue. I put an ohm meter across the 2 wires and didn't get any reading when moving the float up and down. Then suddenly did get a reading. The numbers were way off what they should be (I was getting 222 ohms empty and 173 full) but at least it was working; sort of. Then the readings went up by 100 ohms at each end. A bit more playing around and I found that when I pushed and pulled the white connector that goes into the tank, the readings went all over the place. This is the part of the connector that goes into the tank, not where the wires plug in:

Screenshot 2022-11-28 at 1.53.37 PM.jpg

So out came that assembly and onto the bench for more testing:

Screenshot 2022-11-28 at 1.54.26 PM.jpg

I pulled the connector off the bottom of that main white connector, found the wires leading to the float assembly, and tested them:

Screenshot 2022-11-28 at 1.53.57 PM.jpg

Perfect readings! 51 ohms empty and 2.1 full; just what they should be. And it was consistent.

So I'm getting somewhere. Just have to make sure the connections are good and why me pushing on the main connector caused the readings to change.


Rick

[FFRWRX]

Finally got back to working on the car and sorted out the fuel gauge/sensor issue. It was the electrical connector inside the tank that was causing the issue. Once I tightened it up and made sure the connection was good, the sensor worked consistently.

The MeterMatch has 4 calibration points. Normally only full and empty are used, but you can use the other 2. So I calibrated it with the float full up and full down, then did a 1/4 full and 3/4 full so the gauge should be fairly accurate. I had the pump assembly out of the tank for this, and propped the float up at the appropriate positions. I won't know how well it all works until I put some mileage on it with various amounts of gas in the tank. But when put back together it is at between 1/4 full and empty which looks about right.

And in case it isn't clear, I cut off the sensor from the round cover so there is no connection on that one; no physical float assembly on it now. I am just using the main float on the cover with the fuel pump.

I left the MeterMatch accessible behind the passenger seat under a 3D printed cover, so fine tuning of the calibration later is easy. Notice the sunk-in part of the 3D printed cover. That's what you get when using PLA in a car; not even at high temperatures. I'll heat/reform it and reinforce it on the inside, or print another one using a better material (PETG maybe).

Rick

Screenshot 2023-01-06 at 9.37.53 AM.jpg

Screenshot 2023-01-06 at 9.38.07 AM.jpg

[Ajzride]

I had been planning to swap to ASA for all my 3D printed car parts, and then TeachingTech tested and recommended ApolloX (modified ASA) for car applications. He even printed an AirBox in the engine bay out of it:

https://www.formfutura.com/shop/product/apollox-2779