Fix for adding a second Gauge to measure resistance, without interfering with existing gauge? is it possible?

Thread Starter

immusicman

Joined Jul 23, 2021
30
Hi. New to forum. I have never tried this before, but I am attempting to add a 2nd fuel gauge to an existing system that already has one fuel gauge, and there is one sender in the gas tank. The gauges both operate off 33 to 233k resistance. The sender is perfectly matched for both. The problem seems to be when both are connected to same sending unit, both gauges drop on the readings and it is incorrect. I have figured a work around by using a switch to choose which gauge to use, however, when I use this method, the original gauge has a built in warning system when fuel level is low, so it is triggered by the switch when using the added fuel gauge. The original gauge shows full, and when I add the other gauge, operating both at same time, the gauge drops to about 3/4 of a tank, and the new gauge reads about the same. I didn't realize when simply adding a measuring device to an existing system, it would change the resistance going to the original gauge, apparently. Is there any suggestions or fixes for this. These fuel senders are using a ground circuit return, so there is no power to them. Typical fuel senders use a rheostat, and ground circuit goes to gauge which measures the amount of resistance and gauge correlates amount of fuel based on that resistance reading. Any help greatly appreciated.
 

MrChips

Joined Oct 2, 2009
30,701
Assuming that you have a good reason for having a second readout here is the problem and my solution.

The fuel sensor (sender) is a variable resistor. If you want to tap off a signal from the sensor you have to do so without disturbing the resistor value. All gauges have resistance. Hence placing a second gauge across the first gauge will disturb the reading.

The solution I can offer is to tap off the sensor with a circuit with very high resistance. In general, we want the resistance of the receiving circuit to be at least 10 times that of the sensor. We can do this with a JFET or CMOS op-amp.

Since the sensor and the gauge are both non-linear, you would have to modify the op-amp circuit in order to get the correct reading on whatever type of gauge you choose. One remedy for this is to use a microcontroller. An MCU can be programmed to convert the reading from the sensor to the correct readout. Using this method you can have as many gauges as you wish including analog, digital, graphical, data logger, alarms, etc.
 

Ian0

Joined Aug 7, 2020
9,667
Ask Mr. Heisenberg!

How about using an op-amp as a unity-gain buffer. Connect the input to the junction of the sender and the first gauge, and connect the output to the second gauge. The first gauge should read correctly, because there is no additional load on it. The second gauge should see the same voltage as the first, and should therefore read the same as the first.

I suggest you check that resistance reading - most of these gauges are "hot-wire" ammeters and take rather a lot of current, so 233kΩ seems a lot, but 33Ω seems realistic. You might need something more heavy than an op-amp to drive it.
 
I've only worked on a couple old gauges, but they were "old school". The regulator was like 2.5V and it pulsed. Pulsing can be a problem.

In one case, there was a pulse of voltage and a damped current meter read the fuel value.
The other case may have been pulsed too.

You can measure resistance with a constant voltage or a constant current source. That's problem 1.

It probably is constant current.

You need the resistance of the sensor at both extremes and the open circuit voltage for starters. Measuring the current would help too, because that would tell you if it's constant current. You should also be able to tell if it does anything wierd. When you insert a meter, it will change the reading.

You could find the linearity with a potentiometer.

For low fuel the car used a thermister.

if it's constant current source, you can measure the voltage across the sensor.

Probably my suggestion is to build a current to voltage converter. That device will have a negligible voltage drop. On the order < 1 mV if done right.

Any linearization would have to be done in a processor or a re-drawn scale of an analog meter.
A damped current meter would be ideal. These are fluid filled unless the second gauge is used for filling, If the guage is not needed when in motion, then you can place a keeper on the meter.

A voltmeter could be used, but the sense could be wrong (fixable) or the linearity can be wong. It might need an offset.
 

Ian0

Joined Aug 7, 2020
9,667
Measure the resistance of the gauges.
There's another type of gauge which works a bit like a stepper motor. You can easily spot that type as the fuel gauge doesn't return to "empty" when the ignition is switched off. They have three terminals, positive, sender and ground. The winding between positive and negative sets up a fixed magnetic field, the winding between positive and sender sets up a variable magnetic field, and the needle takes up the position determined by the two opposing fields. The needle is heavily damped so that it stays put when neither winding is powered.
 

SamR

Joined Mar 19, 2019
5,031
The gauges work off voltage measuring the voltage drop across the variable pot sensor if it is a standard float sensor. Place the gauges in parallel.
 

Irving

Joined Jan 30, 2016
3,843
@immusicman The sender resistance is 33 to 233ohm, not 233kohm I hope.

You are all making it far too complicated!

If these are common automotive gauges they are typically 80 - 150ohm and are current driven through two opposing coils. The dual coils take care of errors due to battery voltage. Here is a typical arrangement:

1627061197146.png

And below is a partial wiring diagram from my 1965 Triumph TR4A. Item 32 is the gauge and 33 the sender, Items 30 & 31 are the water temperature circuit which works in exactly the same way. Note the ground connection on both gauges. On my instruments the ground is the casing and the connection is made through the retaining clip rather than an explicit wire - corrosion leads to odd readings. Despite the 'balanced coils' approach item 27 is a crude bimetallic voltage regulator just for those two instruments; from memory it regulates at approximately 9 - 10v across the 10 - 14v range of the battery.

1627060827011.png



This simple circuit will do the job. Despite the gauges being current driven this measures the voltage at the sender and replicates it at gauge 2. This works if the gauges are identical. Q1 is running on the edge of the gate threshold - there's only about a 50mV variation of Vgs to move Vd from 3.5 - 8.5v (or 35 - 110mA) for 80ohm gauges, but it works fine for any gauge from 50 to 350 ohms. I didn't search for a MOSFET with a softer knee, most modern ones aren't very soft and it has to dissipate up to 750mW in an automotive setting without any significant heatsinking so I didn't mess with an SMD part. R2/C1/R3 is just a bit of LP filtering and spike suppression since the DC input <1Hz. Its not essential, just R3 on its own would probably be OK. R4/C2/C3 just tidy up the supply a bit though U1 is automotive spec and OK to 40v.

1627061989879.png
 

Ian0

Joined Aug 7, 2020
9,667
Based on my suggestion in Post #3
This simple circuit will do the job. Despite the gauges being current driven this measures the voltage at the sender and replicates it at gauge 2.
Considering how accurate these gauges normally are, I was considering these two possibilities:
Screenshot at 2021-07-23 21-01-48.pngR1, R4, R5 and R8 are the gauges
R2 and R6 are the senders.
But until I have some values for the current and voltage ranges, I can't calculate a suitable value for R6
 

Thread Starter

immusicman

Joined Jul 23, 2021
30
Thanks so much for all the replies. I should have mentioned the 2nd gauge is an intellitronix digital bar graph gauge. I will take a pic of documentation that came with it when I get a chance. I’m working sunup to sundown next few days. I really appreciate all the information and it’s obvious to me I’m in a forum with absolutely brilliant minds.
 

Ian0

Joined Aug 7, 2020
9,667
https://www.intellitronix.com/product-page/complete-bargraph-gauge-set-w-chrome-bezel
It says: "User programmable to work with Ford, GM, Chrysler, VDO, and aftermarket 240-33 ohm industry standard fuel senders."
I bet you have a 240-33Ω sender.
Mine and @Irving 's fixes are dependent on the gauge having the same resistance as a normal gauge. If it measures the sender in a different way, it's not guaranteed to work. To guarantee it will work, you need two identical gauges.
 

Ian0

Joined Aug 7, 2020
9,667
I built a digital fuel gauge when I was about 19. It used an 8-bit A/D on the sender which was connected to 8 address lines on a 512x8 bipolar "fuse" PROM. The 8 outputs drove a pair of 7-segment displays directly. The 9th address line connected to a 555 which multiplexed between the two digits.
It was damned accurate, but only if the car was parked on a flat road, because I forgot to include any form of filtering. As soon as the car was moving, on a hill or going round a bend, the reading was all over the place.
 

Irving

Joined Jan 30, 2016
3,843
https://www.intellitronix.com/product-page/complete-bargraph-gauge-set-w-chrome-bezel
It says: "User programmable to work with Ford, GM, Chrysler, VDO, and aftermarket 240-33 ohm industry standard fuel senders."
I bet you have a 240-33Ω sender.
Mine and @Irving 's fixes are dependent on the gauge having the same resistance as a normal gauge. If it measures the sender in a different way, it's not guaranteed to work. To guarantee it will work, you need two identical gauges.
Possibly true, but I would expect an after-market gauge to use a similar current to the standard gauge to overcome wiper and connector resistance. Its easy to modify range and offset on mine to accomodate any differential.
 

Ian0

Joined Aug 7, 2020
9,667
Possibly true, but I would expect an after-market gauge to use a similar current to the standard gauge to overcome wiper and connector resistance. Its easy to modify range and offset on mine to accomodate any differential.
If I were designing a digital gauge, I'd be using 3.3V or 5V logic; so, instead of making an awkward level-shifting and supply voltage compensation problem for myself, I'd connect the sender to my regulated 5V supply with either a pull-up resistor or a constant current source. I'd not connect it to 12V.
 

Irving

Joined Jan 30, 2016
3,843
I think it all depends. In a classic car you want to retain the original dials, but possibly add a modern secondary but easily removable display (though if I was doing it I'd make it a head-up display!) I'm sure the TS has a reason for doing this, though $450 is, to me, a good reason not to!

As to changing voltages etc. I don't see the need. It works well enough as is, if it ain't broke, don't fix it. The sender is pretty non-linear to volume. Most fuel tanks are not neat cuboids that level accurately reflects volume remaining. 'Full' usually means full, 'empty' in all cars I've owned usually means 'you've got about another 20 to 40 miles in the tank, fill up soon' and anything in between is merely a comfort factor.

Of course, if you put an MCU in the loop and calibrate the tank by topping it up 1/4 gallon at a time... But is there any real benefit?
 

Ian0

Joined Aug 7, 2020
9,667
I nearly fell over when I saw that price tag, but it is for a set of 6. A single one is $50, which compares with about £40 for a good quality (Durite) analogue guage. I remember when all cars had Smiths or Jaeger instruments but now you need a second mortgage for a new Smiths gauge for a classic car.
I know that the sender is horribly non-linear (and so is the gauge) from calibrating one all those years ago: I used to think that the spacing between turns on the wire-wound sender was adjusted to linearise it against the shape of the tank, but soon realised it wasn't - I did once have a car where a reading of about 2 gallons meant "empty".
I'd still recommend that the TS checks before connecting the buffer circuit to his new gauge, by connecting the gauge to the sender with the tank empty and measuring the input voltage.
 

Thread Starter

immusicman

Joined Jul 23, 2021
30
https://www.intellitronix.com/product-page/complete-bargraph-gauge-set-w-chrome-bezel
It says: "User programmable to work with Ford, GM, Chrysler, VDO, and aftermarket 240-33 ohm industry standard fuel senders."
I bet you have a 240-33Ω sender.
Mine and @Irving 's fixes are dependent on the gauge having the same resistance as a normal gauge. If it measures the sender in a different way, it's not guaranteed to work. To guarantee it will work, you need two identical gauges.
It took a while, but from what I could find on gauge one, it is digital, it is 233-11. It’s an oem by Kawasaki. The gauges both seem to measure the same when operating one or the other. As I said before, they both drop off full a bit when trying to run both
 

Thread Starter

immusicman

Joined Jul 23, 2021
30
I think it all depends. In a classic car you want to retain the original dials, but possibly add a modern secondary but easily removable display (though if I was doing it I'd make it a head-up display!) I'm sure the TS has a reason for doing this, though $450 is, to me, a good reason not to!

As to changing voltages etc. I don't see the need. It works well enough as is, if it ain't broke, don't fix it. The sender is pretty non-linear to volume. Most fuel tanks are not neat cuboids that level accurately reflects volume remaining. 'Full' usually means full, 'empty' in all cars I've owned usually means 'you've got about another 20 to 40 miles in the tank, fill up soon' and anything in between is merely a comfort factor.

Of course, if you put an MCU in the loop and calibrate the tank by topping it up 1/4 gallon at a time... But is there any real benefit?
I’m adding a really nice custom late model dash fairing to a motorcycle that has only a tank console gauge setup. I don’t want to lose any oem functionality. If I can get pic attached here, I’ll upload it87229E76-AFEE-4F0A-B808-7327EDE3FEE3.jpeg
 

Thread Starter

immusicman

Joined Jul 23, 2021
30
If I were designing a digital gauge, I'd be using 3.3V or 5V logic; so, instead of making an awkward level-shifting and supply voltage compensation problem for myself, I'd connect the sender to my regulated 5V supply with either a pull-up resistor or a constant current source. I'd not connect it to 12V.
I actually checked the voltage on the gauge input wire for sender just to see what it reads. It read 5v when the gauge was powered on in 12v system
 
Top