Gas gauge calibration unit

[KitCarlson]

I have read many posts on this forum about fuel gauge problems after replacing the tank fuel sender in our A-Bodies. Some problems are related to mechanical issues of float travel, the instrument voltage regulator, and sender ground. However many have found even after fixing those, the gauge readings for empty and full still persist as problems.

The original system is simple electrical circuit, consisting of a primitive instrument voltage regulator, a gauge and the sender unit in series. If all components are in specification it works. Like most circuits, it is a good example of Ohms's law. If the instrument supply voltage is increased, both the gauge indications for "empty" and "full" will increase. Adjusting the voltage may work in some situations, but not others.

I think it is possible to make a circuit to separate the reading the sender value from driving the gauge. The separate drive of the gauge, enables a means for adjustment. So the basic fix will involve supplying the sender with a constant current. The sender voltage will vary, directly related to the sender resistance. The sender voltage is then measured by a micro-controller on a continuous basis. The gauge is driven by a transistor, not the sender, using PWM (pulse width modulation). The PWM is a convenient means to vary the duty-cycle, to vary the gauge reading. There will be a minimum PWM that results in "empty" reading, and a maximum that results in "full". The program in the micro-controller, will align the sender minimum reading with "empty" PWM and the maximum sender value with "full" PWM. Switch 1 and 2 provide the means to indicate when at empty and full. This is a bit more involved, but will save that for later.

The installation will not be difficult, the sender wire in the OEM circuit will be removed from gauge terminal and placed on measurement board terminal. The gauge drive from board will be connected to the gauge terminal where the OEM sender wire once connected.

The other feature is a solid-state regulator replacing the primitive and now expensive OEM IVR. The same regulator will also supply the micro controller. Many have already done some sort iof IVR conversion, and know about that.

My car does not have gauge problems at this time, so I need to find a fuel gauge and sender for development. I am doing this to help others. I plan to build and test a prototype, but I do not plan to make and sell units. I will work with others and perhaps receive a small royalty for each unit to recoup my development costs. I am thinking RedFish or others may have interest?

Attached is a first draft of the circuit. There are ways to shrink and cost reduce, but the suggested circuit will be easy and robust for development work.

 

[KitCarlson]

Rev. 2

1. added LED for Empty warning
2. added input protection
3. hopefully larger schematic view

 

[straightlinespeed]

If by chance you need a sending unit and a gauge I have both that Im not using, although Im not sure if you would need the entire cluster or what. If your interested send me a PM and we can discuss further.

 

[67Dart273]

I havta say, that diagram is hard on my old eyes. Contrast, mostly, I think.

Do you really need a sender? If it's adjustable, why can't you assume the so called "standard" sender range, plus/ minus a factor?

 

[red67gts]

watching this with interest, thanks for sharing, Joe

 

[zhandfull]

Subscribed

 

[KitCarlson]

Good thought on sender. Heck I have an old decade box and other resistors, so no need for sender, just gauge. I guess taking care of two 4 year olds has corrupted my mind. Been doing it most of a week. Been way too cold to get them outside.

I will try to get a better schematic view. I used standard export to png format at 150, and 300 dip with no improvement. I typically use .pdf, not sure how to post that.

I only have about an hour in the work so far. I may find time Saturday.

 

[TrailBeast]

A little better.
Nevermind, the site resized it.

I uploaded a 150% image here if it helps www.letsgocomputers.com/images/GuageCal2.jpg

 

[KitCarlson]

Trying monochrome. Using the right click to open image in new tab, then using control+ key to zoom works some better.

More about the calibration. I am still thinking of a better way. What is required is a means to determine then the tank is empty, and PWM value for the gauge to be at empty. The user knows when the tank is empty, and sees the gauge, so buttons are used to indicate to micro, "store values". Same needs for full. The procedure corrects for both off spec senders and gauges and stacking tolerance errors.

Right now the plan is to do the following: When tank is "Empty", press "E" button, the gauge will read somewhere near 1/2 and decrease at a slow rate when at "Empty" or slightly below press button again. Then when "Full" press "F" button, gauge reading will increase slowly, when "Full" is reached press button again. The micro will automatically save settings. The button pad could be made detachable using header pins and cable.

I am open to other ideas.

 

[68gtxman]

Kit, great way to fix this! Thanks in advance for your ingenuity!

 

[KitCarlson]

You are all welcome for the kind words. I like to solve problems, design things and work with my hands. I have slowed down some the past few years, keeping at it keeps me going. I have many things to do on my own stuff, but multi-tasking is good too. I often learn things thing that can be used on other projects.

Back to the project, thinking about a few details:
I am thinking I can get by without an actual gauge. It would be helpful to know approximate gauge resistance. I am thinking it will be fairly low. If anyone has spare gauges, please measure resistance and report.

Del reported the sender resistance values as:
L = 73.7 Ohms (empty)
M = 23.0 Ohms (1/2)
H = 10.2 Ohms (full)
I can estimate gauge current for full scale with sender of 10.2 ohms. It would be nice to have the unit setup with default calibration values. Since the drive circuit eliminates the 10 Ohms resistance, it has the capability to over drive the gauge, however the PWM duty-cycle is used to control the average current. The over drive will be helpful for tired gauges that require more current to reach full scale.

The sender level will be measured using a current source to supply the sender. I plan to use a current source of 20mA. The 20mA is considered a safe value, and is likely less than the OEM circuit. If I know the gauge resistance, I can determine that. The OEM current varies with sender value, worst case at full tank.

With 20mA the sender will put out about 0.2V full and 1.5V at empty. The measurement circuit range is 0 to 2.56V at 10bits, or 1023 for maximum. The resolution at 0.2V is about 1 part in 80, so that should be good. The empty resolution will be excellent at about 1 part per 600.

 

[Pawned]

If memory serves, the factory service manual has empty at 73 ohms +/- 12 ohms
Full is 9.6 ohms +/- 1 ohms.

This seems to be a lot more expensive than my fix that cost exactly $1.15 and works perfectly.

And BTW my modification is free to whoever wants it and the only ROYALTY I want is a "Thank you"

 

[67Dart273]

If memory serves, the factory service manual has empty at 73 ohms +/- 12 ohms
Full is 9.6 ohms +/- 1 ohms..

Can you find a definitiive source for this? I have been unable to do so.

 

[Pawned]

Can you find a definitiive source for this? I have been unable to do so.

I just copied it out of the 1970 factory service manual. I would cut and paste it but the pdf file does not allow it. It is on page 8-84 in the right column.
It goes on to say that you need an accurate ohm meter

 

[67Dart273]

Thanks. The way I post these is to take a screenshot, then crop, scale, etc. Somehow, I've missed this. But this does not say that's what the gauge says, this says that's what the sender is "at the stops" So maybe the info I've got for what we might call "operationally empty" that is the actual reading the gauge "wants" might be in error when the float is actually indicating "tank empty".

I always assumed the figures I dug up was because somebody took apart the gauge test box. Boy would I love to get my hands on one of those, which is supposed to actually cause an empty / 1/2 and / full reading

 

[KitCarlson]

Pawned,

Thank you for that information.

It is good your solution worked for your car. While you price was very good, it came from China, and just the chip for the switch mode regulator costs more here. It makes me wonder if parts that do not meet spec, or other issues, end up being sold, in a gray market.

The solution I propose should work for many, it is not a single point calibration. It costs a few dollars more, but much less than some IVR being sold now. It will be possible to improve gauge linearity, something that is not very good in OEM system. It might be possible to set the 1/2 tank calibration, by pressing both buttons at the same time. It is a matter of code, and the way buttons are interpreted. I need more experience with that. Most systems I work with, have communication features and get the job done without buttons.

 

[Daves69]

Fluke 114 on a 3/8 stainless a-body sender, new, not installed.

E = 79.4 (on stop)
Mid = 37.6 (measured distance between stops)
F = 10.5 (on stop)

68 barracuda gauge = 20.7

 

[Rocky_JS]

The gauges fuel, oil, and temp all are the same resistance @ 20 ohm +/- 1 ohm outside that value assume the nichrome resistance wire or its insulation is fried.

It would be helpful to know approximate gauge resistance. I am thinking it will be fairly low. If anyone has spare gauges, please measure resistance and report.

 

[67Dart273]

I'm not sure resistance is all that helpful. These are a "hot wire" meter, and surely change resistance between hot and cold

Here's what 3V gives you: Two gauge units in parallel, and in series across 3V Since both units read substantially the same, I assume either they are both fairly accurate, or both "off" by the same amount, LOL

 

[67Dart273]

The other factor here is "just what is" the design supply voltage of these gauges? I'm not sure anybody really knows. Seems to me the service manuals describe the original IVR as being a "50 percent duty cycle" so that would NOT be 5 or 6 volts on a running automobile, it would, rather, be closer to 7 volts

"Everybody" seems to be using 5V regulators. Frankly, I bought an RTE, and never bothered to check the output voltage!!!! when I installed it.

 

[pishta]

How about a rheostat in series? Put it between the gauge and sender and just dial additional resistance in to read E when you run out of gas. Let F end up wherever as E is more important to me. Feasible?

 

[cudajim]

The other factor here is "just what is" the design supply voltage of these gauges? I'm not sure anybody really knows. Seems to me the service manuals describe the original IVR as being a "50 percent duty cycle" so that would NOT be 5 or 6 volts on a running automobile, it would, rather, be closer to 7 volts

"Everybody" seems to be using 5V regulators. Frankly, I bought an RTE, and never bothered to check the output voltage!!!! when I installed it.

I've tried an adjustable regulator for my IVR, adjusted it to 6.0 v and it was not right. I then readjusted it to 5.0 volts and it seemed to be much more accurate.

 

[65 Dartman]

Subscribed.

 

[Pawned]

How about a rheostat in series? Put it between the gauge and sender and just dial additional resistance in to read E when you run out of gas. Let F end up wherever as E is more important to me. Feasible?

This will not work. I tried 3 or 4 times, in series, in parallel a combination of both.

The problem is the gauge is a bi-metal resistor. as it heats up it uncoils. The problem is the resistance characteristics of the resistors have changed over time and use.
You need to work with the resistance that is given, it is like a worn out car, in order to get it up and running you have to give it more gas.
In the meters case the gas is equivalent to the voltage. give it a little more gas and it will still run like before.

Increase the voltage and you meet the meter on its terms

 

[Pawned]

I've tried an adjustable regulator for my IVR, adjusted it to 6.0 v and it was not right. I then readjusted it to 5.0 volts and it seemed to be much more accurate.

If you ever took apart a OEM voltage regulator, it is nothing more than a fast acting flasher. It turns on and off very fast and lets the capacitor smooth out the voltage. a very antiquated PWM type system.

On my modification. I have tuned the voltage regulator and it is running like new. At this point I am not going to disassemble the cluster to see what voltage that is.