Gas gauge not working? I think I fixed it.

-
KitCarlson said:
I can make a circuit with a 50 cent micro-controller and a few other components that will enable calibration of the empty and full. It works by separating the sender measurement, from driving the gauge. There will be a special calibration procedure performed at empty and full. It will correct for off specification senders and gauges, if they function in a monotonic manner.

If more than a few have interest, I will start a separate thread, develop the schematic, parts list and simple software for the system. I do not have interest in selling units, but I can see helping someone get setup up.
Click to expand...

No offense intended, but talk is cheap. No one has ever gotten as far as I have fixing this problem. If you can do this, then why haven't you done it long ago. There has been a bunch of threads dealing with this problem.
Why have you waited until someone else found a solution to claim you can fix it.
My fix cost $1.15 for the regulator and a few feet of scrap wire I have here and some solder.
The one great thing about these cars is there are in NO COMPUTERS.
My fix is cheap and requires no computer or software.

So far I do not know if this will calibrate properly to full. It very well can. but I did not make a claim until I can prove it. When I finish with the calibration for empty, I will fill it up and then I will tell what it does.
All anyone wants, that I have heard is, to know when the gas tank is close to empty. So they do not have to drive around with a gas can in the trunk. My modification does just that.
Maybe a little offense was intended, but I tried to hold it down
 
Pawned,

No offense taken. I do more than talk. Have more than 30 years of professional product design experience. My designs always work.

What I suggest has be done before by me, for other similar measurement systems, and likely by others. It may cost a few dollars more, but it will also calibrate the full end of scale. It will not fix wiring problems, or mechanical problems like float travel limitation, a broken sender or gauge. If the sender changes resistance with travel, and the gauge is able to swing from empty to full, with a variable voltage below gauge reference voltage, it should work.

Once the controller is programmed, it is just a part, so others need not worry about that. The calibration means, will use a couple switches, so the unit can be re-calibrated again, if moved to a different car, or sender or gauge changed. No pot to drift or adjust.

I will start a new thread called "Gas gauge calibration unit". I am tied up by grandchildren at the moment, so it might take a day or two get things started.
 
Pawned, I've conversed with Dave on other issues. I can assure you he knows of what he speaks. Dave dammit don't be so modest. Remind us? of your web URL on the homebuilt EFI.

Just remember, there's always another way to skin a cat, although I don't know why anyone would want to. Skin a cat, that is

surprised-cat1.jpg


508044849_683.jpg
 
I probably have a good 40 hours or more working on this modification. I restarted from scratch 3 or more times because my approach was not working.
I never hid the fact that I was working on this modification, I was asked more than a few times if I would post my mod I said I would when I found the answer and when it worked.
So as a courtesy to the other members on the board, I post the modification for everyone to see and copy if they wish. Only to have people claim my mod can not work, that they found the fix first, or they have a better way. They never have anything to back it up.
Kit you may be a designer, I was/am an electrical engineer. I was the one who always was called to the customers site to make the equipment work the way the designers promised it would.
As I said before, when I start out to fix a problem, I either fix it or die. whichever comes first.
Contrary to popular belief, electronics is not a precise art. Tolerances are usually very liberal.

So if I get a little testy when challenged, now you know why
 
A final follow up. This circuit is working perfectly. For how long, I have no way of knowing.
Many years ago, I was working at a very top secrete government installation. There was some higher up guy from either the Department of Defense or dept of energy grilling me on the cause of the failure of the piece of equipment.
I finally got fed up with this guy when he demanded to know why an SCR (the size of a hockey puck) failed exactly when it did.
I looked him square in the eye and told him, "God in his wisdom, looked down at the SCR and said 'You are going to die'".

I truly do not remember the ramifications of my explanation, except I was briskly escorted off that military base.
 
Looks like this thread has run its course. There should be a way for the originator of a thread to close it. Like there is in the parts wanted and for sale forum.

Thanks for coming. I will probably unsubscribe to this thread, but PM me if you have questions
 
FYI. The other day I filled the tank and the needle is pegged so far right of full, I can not see it.
Will have to see if it comes down.
Put the gas can back in the trunk
 
Pawned, I have been having similar issues as you but with both temp and fuel gauges in my '66 Fury. I made a 5v regulator using a 7805 chip but the gauges still don't work when hooked up to their respective senders. I borrowed a spare gauge from my brother to do some bench testing. His gauge reads 21 ohms resistance, which sounds like the correct range from what I read on a thread here.

I hooked the test gauge up to a 5v computer power supply and the needle pegged high as it should. I then used a variable resistor to simulate the fuel or temp sender resistance. Problem is, the test gauge reads half scale with 7 ohms resistance and full scale with 2 ohms resistance. These resistance values are quite a bit less than what the senders supply, which are on the order of 10 ohms for 'full' or 'hot' and 75 ohms for 'empty' or 'cold'.

I've been thinking of making a simple variable voltage regulator using an LM317 chip and a couple resistors (one of them being variable to adjust the voltage). I could try using a 7808 or 7806 but thought the LM317 would let me fine-tune my voltage to what the gauges need.

Thanks for your post. It has given me hope for getting the gas can out of my trunk.
 
mcmlvii said:
Pawned, I have been having similar issues as you but with both temp and fuel gauges in my '66 Fury. I made a 5v regulator using a 7805 chip but the gauges still don't work when hooked up to their respective senders. I borrowed a spare gauge from my brother to do some bench testing. His gauge reads 21 ohms resistance, which sounds like the correct range from what I read on a thread here.

I hooked the test gauge up to a 5v computer power supply and the needle pegged high as it should. I then used a variable resistor to simulate the fuel or temp sender resistance. Problem is, the test gauge reads half scale with 7 ohms resistance and full scale with 2 ohms resistance. These resistance values are quite a bit less than what the senders supply, which are on the order of 10 ohms for 'full' or 'hot' and 75 ohms for 'empty' or 'cold'.

I've been thinking of making a simple variable voltage regulator using an LM317 chip and a couple resistors (one of them being variable to adjust the voltage). I could try using a 7808 or 7806 but thought the LM317 would let me fine-tune my voltage to what the gauges need.

Thanks for your post. It has given me hope for getting the gas can out of my trunk.
Click to expand...

Have you thought about the variable voltage regulator?? I ask because today I received the 5 that I ordered from China. $1.15 each.
The problem is it took almost a month and a half to receive the shipment from China.
Good luck.

Ed
 
Pawned said:
Have you thought about the variable voltage regulator?
Click to expand...

I happened to see a variable regulator at Radio Shack when I was picking up the rheostat, so I'll probably get one of them. Sure will be nice to get my gauges working.
 
I got the variable regulator built tonight using an LM317 chip. I'm feeding about 9 volts to the gauges and they appear to be responding well. I'll need to go through more gas (not hard to do with the barge) to see how accurate things are when the tank is nearing empty.

IMG_20140215_195553_252.jpg
 
Currently working on my 64 Valiant instrument cluster on the bench and dealing w/ same issues. I'll post more details in a new thread, but a few quick comments:

Pawned stated a few pages back that you can't do anything about the sensor's response, but not quite true. I added 27 ohm in parallel w/ the fuel sender on the circuit board (just ring terminals on studs) to get it to read correct at empty (where sender is ~90 ohm). I adjusted my new electronic Vreg ($30 ebay "voltage limiter) to get "F" to read correctly. For a 2-pt calibration you need 2 "knobs" you can turn. If too much signal (too low resistance at sender), you could add series resistance. Not sure why Pawned stated 9 ohm empty, since that should be when full.

Now fooling with the coolant temperature gage. Initially, the temp gage read open (infinite ohms), so I popped the top (drilled 1 rivet) and found the wire burned. I wrapped a very thin wire around the stub wire left off the arm, soldered and soldered the new wire to the output stud. The factory wire is likely Ni-Cr so I doubt solder sticks, but it does hold the Cu wire tight around it, but might not prove a long-term repair. The factory used pink adhesive at the stud, which I picked to bare Cu. With a 20 ohm resistor to gnd, simulating the sensor, the gage now reads at "ideal" mark (I think, 1st tic, just below half scale).

I am now searching for info on the sensor specs. Nothing in my 65 Dart shop manual, nor Haynes. Rockauto and others give conflicting PN's, even listing on/off switch sensors. I suspect most gage-type sensors have identical response and just vary in pipe thread and output connector. I put a TS-18 in my Dart (slide-stud connector), whereas my 64 Valiant has a threaded-stud type (can't read stamp). They read 318 & 363 ohm, resp. at 61 F. I'll post if I find manufacturer specs.
 
Pawned said:
FYI. The other day I filled the tank and the needle is pegged so far right of full, I can not see it.
Will have to see if it comes down.
Put the gas can back in the trunk
Click to expand...

After I shut off the car today, I stopped to see if the needle was pegged or stuck at F. I saw it work its way to Empty, thus the needle was not stuck or broken off
 
Bill,

Most temperature sensors I am familiar with are NTC, or negative temperature coefficient. I have found that after sitting a long time (years) without use, they read open, however when heated they work again. It is as if the internal connection is made using a metal with a low melting point for connection.

Typically NTC have a standard nominal resistance at 25C. Then at a second temperature, check the NTC table to confirm. I typically use a second or third order polynomial to approximate the Steinthart equation, for use with EFI.

I found what I thought was a bad sensor, then found another, it was bad too. After heating to 100C both were working again.
 
Have driven the car around. I now can see the needle on the right side of the full. I put the gas can back in the trunk, just in case
 
mcmlvii said:
I got the variable regulator built tonight using an LM317 chip. I'm feeding about 9 volts to the gauges and they appear to be responding well. I'll need to go through more gas (not hard to do with the barge) to see how accurate things are when the tank is nearing empty.
Click to expand...
My disclaimer as I do not know the voltage I am sending to my meter. I am guesstimating it as 5.7-6.7 volts

I do not know that I would want to send that much voltage (9 volts) to the the fuel gauge. As it is almost twice what the factory was.
It is sending a live current through the fuel tank.

Full using 5 volts the current is .53 amps
Full using 9 volts the current is .9375 amps

Empty using 5 volts the current is .065 amps
Empty using 9 volts the current is .118 amps

Worst case are sending almost twice as much current thru the sending units wire wrapped resistor, which is in direct contact with Gasoline.

It may not make a difference in practice, but it is something to consider.
 
Pawned said:
I do not know that I would want to send that much voltage (9 volts) to the the fuel gauge. As it is almost twice what the factory was.
Click to expand...
I was thinking 9v was just asking for a burned out gauge but that was what the fuel gauge needed to work properly on the bench. After installing everything and driving the car a couple times, I had to turn down the voltage regulator because the temp gauge was reading full scale with a hot engine and 9 volts. I haven't put a meter on the regulator to read voltage yet but I'm guessing it's in the 6 volt range based on where the pot is adjusted to on the regulator board. As a bonus, the fuel gauge started working better at some point after the adjustment. Not sure what that's about but I'll take it.
 
mcmlvii said:
I was thinking 9v was just asking for a burned out gauge but that was what the fuel gauge needed to work properly on the bench. After installing everything and driving the car a couple times, I had to turn down the voltage regulator because the temp gauge was reading full scale with a hot engine and 9 volts. I haven't put a meter on the regulator to read voltage yet but I'm guessing it's in the 6 volt range based on where the pot is adjusted to on the regulator board. As a bonus, the fuel gauge started working better at some point after the adjustment. Not sure what that's about but I'll take it.
Click to expand...

If you go to post #44 and picture 12. You will see that I am feeding each gauge with different voltage. The Temp gauge is still fed by the 5vdc regulator. My variable regulator ONLY feeds the fuel gauge. I cut the trace on the PCB and then fed them by the two separate voltages
 
Looks like I will never know if my fix worked.

I was putting the car up on the jacks so I can start working on cleaning the underside.Well, it done fell and the jack severely dented the gas tank.
It spit about a cup of gas out of the gas cap and I was waiting to see the gas pouring from the bottom of the tank.
I think God intervened and did not allow the tank to split. Thanks God

Until I drop the tank and pop out the dent I will have no idea what is happening.


I am confident my fix was working. Today I saw the needle at the full sign. Up until today it was pegged so far to the right, I could not see it.

I may need a new tank and it seems for sure I will need a new sending unit.
 

Attachments

  • IMG_20140224_192142_289.jpg
    32.9 KB · Views: 218
Well that's an "aw $$$$" moment, now, ain't it. Man that sucks!!
 
67Dart273 said:
Well that's an "aw $$$$" moment, now, ain't it. Man that sucks!!
Click to expand...

At least it did not split open and spill all the gas out.
I was thinking of filling it with water, drilling into it and use a dent puller.
But that sounded dumb.
I may do that yet, but keep that one for a spare.
 
Actually it does not take that much pressure. You would have to make some wood forms for the top and bottom or the thing will look like a rather boxy egg, LOL

If you try it in the car, don't forget to bend the pickup back into shape
 
You must log in or register to reply here.
-

Back
Top