Fuel Gauge/Temp Gauge Regulator Replacement Experiment

[Dave The Sparky]

My fuel and temperature gauges have never been right or in any way accurate and seem to read what they want so i suspected it was probably something to do with the regulator mounted in one of the gauges to blame.
So here is what i bought to try and give me a stable and slightly more accurate set of gauges a cheap but rather bulky voltage regulator 12V in and 5V out the beauty of it being it has an adjustable output via the little white dial next to the terminal,
Everything i read says the regulator that is fitted (which actually works in an entirely different mechanical kind of way which i wont go in to here) kicks out 5 Volts
My ultimate plan is if i could work out what voltage the gauges actually need is to buy a far smaller regulator and set it at the voltage required but i need to know if this will actually work in the first place and if it does what voltage is the correct one.



I left the regulator as it came at slightly more than the 5 volts it should be and the fuel gauge read 1/2 full which is wrong as its not that long ago i filled up and suspect its more like 3/4 full.

 

[Dave The Sparky]

So i cranked the regulator up to 6 Volts and what do we have but just under 3/4 full tank.

 

[Dave The Sparky]

Took the 8 out for a good run to warm it up and fuel level has dropped as you would expect but i actually have a temperature reading, the needle barely moved before.

 

[Chamba]

Bloody legend. I have the same issue ( but my fuel gauge drops fast enough to stir up dust with the 351W). I'll be keen to see your solution.

 

[Dave The Sparky]

I wont know until i fill the tank up to the top (or run out of fuel first due to a false reading 😁 ) but its looking pretty good at the moment so will report back when i do.
I would expect the proper components to do this to come in for very little cost so long as that voltage is constant on every Jeeps gauges

 

[FLCJ8]

I was going to ask if you had seen this:

Fuel gauge Voltage Regulator BYPASS

Well, not exactly a bypass. But close. I don't like the temperature based VR. It works, it is simple, and I give full marks for clever on its design. The problem I have with it, is that nobody makes a new one that lasts more than 3 months anymore. And I was tired of contact surface...

www.cj-8.com

But then I saw you replied to the post
Had you tried the setup in the original post, or are you like me and finally worked your way down your to-do list to this project?

 

[Dave The Sparky]

Guilty as charged its just taken me a while to get to the stage where it annoyed me enough to sort it out.
So long i dont even remember replying to that post 😁 .

 

[Dave The Sparky]

Well in the interests of science i thought i would test my little regulator on another set of gauges and using a variable resistor and a little 12V battery mimic the resistance settings that the workshop manual says should work the gauges..

 

[Dave The Sparky]

I wont bore you to tears with a photo of every setting but this is what i did.
I simply set a resistance near enough as to what the manual asks for and then watched what the gauge did.

 

[Chamba]

Yes, but that's an American speedo where your other one is British. Therefore yours is likely corrupted by the metric system where the American speedo is pure imperial. I put forward that this is an unknown variable and therefore skews your results.

 

[Dave The Sparky]

Well a US gallon is different to a UK gallon so you could be on to something

 

[Dave The Sparky]

Life is too short for this sort of sruff but anyway here is what i found out so you dont have to!

First the temperature gauge with the numbers from the manual as to what it should be and in brackets what my resistance setting had to be.
This is with a standard multimeter and the test leads are 0.2 ohms which i didn't factor in so all readings will be 0.2 ohms out i guess.

Cold 76 ohms (not a lot i can do about cold)
Beginning of the band 36 Ohms (Needle actually started moving at 24 Ohms to get it to the C mark)
End of band 13 Ohms (12.8 Ohms see photo above)
Hot 9 Ohms ( 9 Ohms)

So temperature although not perfect at the colder end is completely fine as it warms up and thats really the important part and where it needs to work.

Second the fuel gauge
Empty 73 Ohms (Needle started moving at a lowly 40.5 ohms)
1/2 full 23 Ohms (19.2 Ohms)
Full 10 Ohms (10 Ohms)

So i wont run out of gas as the empty mark is way out of spec but in a good way where-as the 1/2 full isn't actually miles out and full is spot on.
I would be interested to know where 40 ohms is on a fuel tank sender as it may be that the 73 ohm reading is bone dry in which case 40.5 ohms could mean 2 or 3 gallons are left in the tank which is fine with me.

My voltage regulator tops out at 6.1 Volts or i would of tweaked it up a bit and give it a little more voltage and see if it improves these numbers.
The post @FLCJ8 re-posted had a 5 volt regulator and there is no way 5 Volts would work accurately on ether this setup or the one thats currently in my 8 but that could just be something to do with my setup and gauges.
As also mentioned in that post i have no idea if temperature will affect any of this or the age of these gauges either as this one in particular hasn't been used for many years but seemed okay.
Just for information its been 21 Deg C today which according to my converter is 69.8F so warmish nut not too hot and not too cold either.

Tried to use the mechanical original regulator to compare but its either stuck on or there simply isn't enough load with just the one gauge so it just puts out 12 volts permanently which is no use to anyone.

Not sure if this post helps or confuses but i am quite pleased with the readings in my 8 just need a bit of time to find out how accurate all this is.

 

[FLCJ8]

today which according to my converter is 69.8F so warmish

Dave,
I was going to say great job on your testing, but...


in what world is 69.8F considered warmish?
This puts all your calculations into question.
Now I need to locate the link I was trying to find in the original post from 2016 so you can test it too.

 

[Dave The Sparky]

Average Temperature in Halifax UK​

The warm season lasts for 3.0 months, from 11 June to 9 September, with an average daily high temperature above 16°C. The hottest day of the year is 22 July, with an average high of 19°C and low of 12°C.
The cool season lasts for 4.1 months, from 17 November to 19 March, with an average daily high temperature below 9°C. The coldest day of the year is 6 February, with an average low of 1°C and high of 6°C.

Yep as stated red hot yesterday, ice cream weather almost... 🍦🍦🍦

 

[FLCJ8]

The warm season lasts for 3.0 months, from 11 June to 9 September, with an average daily high temperature above 16°C (60.8F). The hottest day of the year is 22 July, with an average high of 19°C (66.2F) and low of 12°C (53.6F).
The cool season lasts for 4.1 months, from 17 November to 19 March, with an average daily high temperature below 9°C (48.2). The coldest day of the year is 6 February, with an average low of 1°C (33.8F) and high of 6°C (42.8F).

Now you're just showing off.

Yep, definitely a heat wave.

 

[cbford]

Hadn't been on in a while, surprised to see this thread. I can offer this comparison of the 2 controllers used, not sure if it is helpful or not...

The LM voltage regulator I wrote about is an integrated linear regulator and operates on a single ground. This means all energy that flows into the regulator also flows through the gauge/sender circuit all the time.
Whereas the SD controller Dave used is an isolated, switching controller. This means energy that flows into the controller does not all flow out constantly. This device monitors the demand side of the circuit and only sends what is required up to the voltage and amperage setting targets.

The LM7805 is designed as "5 volts" output, but that is at 10 volts input. At 12-14 volts input range, it has output potential of 9-11 volts but no matter voltage, it is capped at 1.5 Amps (good up to 35 Watts total).
Because it is integrated to the load circuit, as the load varies the LM drops excess voltage (energy) into the heat sink it is attached to. When your tank is full (73 ohms) your heat sink is very hot. When tank is empty (10 ohms) your heat sink is ambient temp.
If anyone made one of these 3 pin regulators that was 12 volts, 1.5 amps with no low voltage drop diode- it would essentially be the same thing as the stock voltage regulator. Inefficient, but just fine in a car with metal gauges housings and metal dashboard, etc... that don't melt.

The SD voltage controller only ever sends the demanded energy into the circuit up to it's max setting. The Jeep gauge needs nearly 12 volts to read full, but only about 3-4 to read empty (amperage varies from .16 to 1.2 across swing at 12 volts). With 5 as a max voltage, that is why gauge will not read up to full but is accurate towards empty. If you had an SD that could reach up to 12V potential and keep amps low (so as to not burn up small winding wires in gauges) it would make the gauges work perfect. And, nothing would get hot - which is why these are the preferred voltage controls today in plastic and composite housings.

The LM regulator is still ticking after 6 years. Gas gauge works perfect off it, but temp gauge, though consistent, is always on low end of scale. This might be amperage potential of the LM7805, but I never bothered to figure it out.

 

[Dave The Sparky]

Ah okay that makes sense as 5 Volts wouldn't cut it on my test setup as there simply wasn't enough power to move the gauges to anywhere near where they should be.
But as i found out 6.1 Volts and 9 to 10 ohms ((tank full or engine hot) shifts the pointers to the end of the scales it doesnt take anywhere near 12 volts to do it.
It appears like i have effectively made a more efficient (less heat) but less accurate at he lower ends of the scale regulator for mine than you did?.

Anyway i have put the regulator back in my 8 and i would say that the temperature gauge works fine and i have every faith that the hot reading at the top end of the gauge is bang on now after playing with it on the bench.
The gauge reads about a 1/3 at running temperature driving around but thats a 1/3 more than it ever has and i never want to see it climb any higher that that anyway.

I am going to see where the empty is on the tank (which could take a while as when you want the Jeep to use fuel it suddenly turns in to some sort of Prius 😁 ) and report back on how much it takes to fill it from that point .
As stated before I can live with it saying empty if its still got a few gallons in there, i was far more worried about it saying 1/4 full and there was nothing in there at all...

 

[cbford]

You are exactly right, that is what the SD does, just a more efficient control solution. The formula to use to figure out what you would need to reach full sweep on the gauges is the combo of Watts and Ohms laws.
Your resistance ranges are known so you could play with the 0-12 volt range and the 0-"burned up" amp range (this seems to be 2-ish amps at 12 volts based on my acciden... uhm, experiments) to find the max settings that would support your resistance load ranges properly.

 

[Chamba]

Now you're just showing off.

Yep, definitely a heat wave.

This is why none of the houses in the UK have air con and why most of the bathrooms are carpeted.......

 

[Dave The Sparky]

This is why none of the houses in the UK have air con and why most of the bathrooms are carpeted.......

Yuk carpet in a bathroom no thanks! underfloor heating is the way