Matching fuel guage to sender

[macdonald.dm]

Re: Matching fuel guage to sender

Hi Keith.

I see. Any chance you can take a measurement of the resistance of your actual SW gauge?

I can't see that adding a resistor in series would do the job, because if you placed a 100 ohm resistor in there to bring the 'empty' resistance up from the Sierra's current value of 140ish ohms to the 240 ohms required by your gauge, that would have the effect of bringing the 'full' value up from 22 ohms to 122 ohms. This is half the empty value, so the gauge would only read up to 1/2 full max.

So, I'm wondering about using a series resistor along with one placed in parallel with the gauge. I'll need the existing gauge resistance (which should be constant) to arrive at values - if I manage at all...

I'm a tad concerned about messing with parallel resistors, tho', as this could cause a greater current to flow through the sender. I wonder what the chances are of making the sender warm up - gulp!

(Surely there is some protection there however - I mean, if the wires to the gauge became shorted together, this would cause max current through the sender, and it must be designed to cope with this without going 'boom'?!)

[keith_h]

Re: Matching fuel guage to sender

Hi Donnie,
What the chap wrote down was as in the drawing but I don't know where the 30 Ohm nor the 7.2v come from

[macdonald.dm]

Re: Matching fuel guage to sender

Ah! Now that's quite clever! And cunningly simple!

What he's done is to use a potentiometer to provide 7.2 volts to the sender instead of the usual 12V.

The 7.2 Volts is obtained by turning the pot until, well, you get it. If you can imagine the arrowhead on the pot in your diagram moving upwards towards the 12V rail, then the voltage at the arrowhead will increase until it BECOMES 12V.

Turning it down has the opposite effect - ie: it will ultimately become 0V. If you imagine setting the pot so that - of the 30 ohms TOTAL resistance - there are 12 ohms resistance ‘ABOVE’ the arrowhead, and 18 ohms below, then the voltage coming out will be 7.2V (it's all about ratios).

The actual voltage of 7.2V isn't critical - you just turn the pot until your gauge works as you want; ie: on an 'empty' tank, adjust the pot until the gauge needle only JUST touches empty, and this should be ok.

He arrived at 7.2V because this provides the same current (0.05A, or 50mA) as the gauge WANTED when supplied by 12 V, but didn't actually get (tho’ I actually get 7.35V if I’m doing it right).

Ie: the gauge WANTED a sender resistance of 240 ohms when empty, on a voltage of 12V. This means it would have received a current of 0.05A (I=V/R, I=12/240, I=0.05A).

Since, however, you couldn't INCREASE the resistance (from the actual 147 to 240), you friend came up with the idea of REDUCING the voltage instead. With a current of 0.05A (established above), and an actual resistance of 147 ohms, this meant the voltage would have to be dropped to V=I x R, V=0.05 x 147, V= 7.35V.

The only slight snag is that when FULL, the gauge really wanted a current of 0.36A to give the ‘full’ reading (obtained by V=12V, and WANTED sender resistance of 33 ohms. I=V/R, I=12/33, I= 0.364A). With the new, lower, voltage of 7.2V, it will actually only get I=V/R, I = 7.2/22, I= 0.327A. In theory, this will mean it won’t – quite – go to ‘full’, but I think it’ll get pretty close!

There is another complication too (!): If you set the potentiometer to give around 7.2V with the SENDER at the ‘full’ 147ohm resistance, when the sender then travels to ‘empty’, and its resistance drops to 22 ohms, this will actually affect the potentiometer and cause the voltage coming out of it to drop right down – to around 5.5V.

Confused? Clear as mud?

It’s a nice idea, and sounds worth trying to me.

[keith_h]

Re: Matching fuel guage to sender

Hi Donnie, thanks for that and clearer than mud but still back at the same question - where do I get a pot of around that value without raiding the bank?

[macdonald.dm]

Re: Matching fuel guage to sender

Simple, Keith - make your own!

This wire on Ebay (130154199963) has a resistance of 60 ohms per metre, so you only need 1/2 a metre. Coil it around a former such as a piece of wood dowelling, tape it up at each end, and then take 'taps' off of it at various places until you get the reading you want!

If you do decide to go this route, the actual amount of wire you use isn't critical - just make sure it's at LEAST 1/2 metre (30 ohms) or else the overall resistance will be too low, and the current flowing through it will be greater than you want (it'll get warm!).

Enjoy...

[keith_h]

Re: Matching fuel guage to sender

Thanks Donnie. It's amazing how in this time of chips and microprocessors etc. how we can forget the basics

[macdonald.dm]

Re: Matching fuel guage to sender

Hi Keith.

Oops, fly in t'ointment time...

Your friend's diagram assumes the sender is an electrically isolated unit that you can get two wires to, but it's actually likely that your sender is earthed through the tank?

No chance of this being too easy...

[keith_h]

Re: Matching fuel guage to sender

Hi Donnie, I don't think it is as the tank straps are lined with rubber, the fuel lines end in 18" of flexible hose (special rubber) and the filler has 9" of rubberish tube connecting it to the tank. I shall go chasing wires at the weekend. (I contacted the supplier of the resistance wire and he is supposed to be coming back to me with its watts/current rating). You're quite right - nothing is ever easy. I once worked with a bloke who multiplied every job time estimate by the Stevens factor of 2.4 (his name was Tony Stevens) and it was amazing how often he was right

[donsayers]

Re: Matching fuel guage to sender

I have the same problem, Triumph gauge ford sender. the two don't match fuel runs out before gauge reaches zero. I made all the measurements, tank empty, tank full etc etc and designed and built a circuit using operational amplifiers that would sit between sender and gauge and compensate.

However in the meantime I totally rewired the car, better contacts newer wire, and the behaviour of the gauge changed completely, and I never did get round to fitting the circuit to compensate.

One of the problems is that ideally you want a box with two knobs; one adjusts the full reading without affecting the empty one and the other knob adjusts the empty reading without affecting the full one. Not possible what you get is one knob that changes the whole range up and down and a second the changes the spread, almost impossible to calibrate.

I hear that Europa do senders that fit ford tanks that work with BL Lucas dials. Much the easiest solution I think.

[macdonald.dm]

Re: Matching fuel guage to sender

Hi Don.

You are absolutely right, of course. True linear calibration won't be achieved by the method suggested by Keith's friend.

However, due to the fortunate resistance ranges involved here (22 - 147 ohms on the sender as opposed to 33 - 240 ohms required by gauge), it looks as though it is possible to get a pretty good correlation between them, with the gauge just failing to indicate 'Full' with a full tank.

Having said that, it's all theory at the moment! It's a simple - hence very attractive! - solution, and probably worth trying.

The main problem I can see is that the tank sender is earthed, whereas it would need to be completely isolated electrically for the circuit to work without modification.

The main problem

[h_m_cumming]

Re: Matching fuel guage to sender

Are the gauges your using the normal bi-metalic strip with a heating coil type? If so then you'll find two slotted adjusters on the back. One adjusts the range of movement and the other the zero.With a little bit of trial and eror you can get the reading pretty close, especially at the zero which is the important end.

[macdonald.dm]

Re: Matching fuel guage to sender

Blimey, Hugh, could it really be that simple! How wonderful if it is!

IF you still need to go the 'pot' route, Keith, the pot is easy to wire up - much easier than I thought!

Your gauge should have two wires going to it: one is the switched 'live' from the fusebox, which only goes 'live'(+12V) when the ignition is turned on, and t'other is the line to the fuel sender. Should be easy to identify with a meter. Label them, or note colours if they are different, to keep ID clear. (I'm assuming here that the sender IS 'earthed', by the way, which it appears to be from my Sierra diagram).

Looking at my diag (sorry, no CAD here...), your Pot would be wired as follows:

1) The switched live to the gauge goes INSTEAD to 'A' on the Pot.
2) The other end of your Pot goes to an earth point.
3) The 'tap' you're taking off of your Pot then goes to the terminal on your gauge which USED to have the switched live on it.
4) That's it! The gauge will now be receiving whatever voltage you chose by adjusting the take-off point on your pot - presumably around 7V will give you what you want.

(A word of caution: We need to ensure the Pot wire isn't being overloaded which would make it get pretty warm! A 30 ohm Pot will mean that around 400mA will be passing through it all. In addition, when the sender is at 'empty' and has a resistance of 22 ohms, I'm seeing a further 300+mA going through the 'top' part of your Pot. So, that part of the Pot could be carrying anything up to 800mA... This looks like around 6 watts of power that part of the Pot needs to handle. Check that your wire will handle that!

I'd be tempted to start with a Pot resistance of more than 30 ohms (40-50?) and see how that works. This will reduce the gauge reading further, so you may have to move the take-off point higher to get a greater voltage going to the gauge, but it will mean that the whole Pot will be taking less power. Trial and error...)

This is all theory, of course! Just how well it works in practice hopefully you'll be able to tell us soon!

Another advantage of this system is that the gauge should be less affected by voltage fluctuations since the normal changes in battery voltage (<12V after cranking to >14 when charging with no load, etc) will be lessened in proportion to the lower voltage you are now using.

To set this up, I think the way to go would be to get the tank to the 'empty' stage you want to consider being 'empty' - ie: you probably want a small reserve? Then tweak your Pot to get the gauge indicating 'E'.

Please make sure you are happy with the above before proceeding. It isn't a complicated circuit by any means, but that doesn't mean I haven't made an ar*e of it at some point.

[macdonald.dm]

Re: Matching fuel guage to sender

Keith, I've just re-read your second post on this thread at the top of the page where you state the sender values.

I'm confused about all the values you list. Can you confirm that the sender you have in your car really is 22 ohms on Full, and 147 on Empty, and not the other A-B, B-C numbers you've posted?

[keith_h]

Re: Matching fuel guage to sender

Hi Donnie,

The 22 full / 147 empty came from a Ford technical bulletin that their Customer Services folk very kindly sent me. The ABC figures are what I actually measured but, thinking back, probably in circuit so I suppose I have to ignore them as there is no telling what else I was measuring. I've got visitors this weekend so I will take a fresh look later. Resistance wire is on its way.

[donsayers]

Re: Matching fuel guage to sender

Keith, the "Guy from Dunstable" was probably me Sorry. I had very little luck (see above) mainly because of the difficulty of calibrating the unit once it was installed. I have now bitten the bullet and I am trying to calibrate the gauge and unit out of the car prior to refitting it. I will keep you posted.

Donnie, I am intrigued by your circuit, however a potentiometer only gives a constant voltage if no current is drawn. My first attempt at a solution was like yours but using a large power transistor biased to provide the correct voltage. It didn't work. The whole problem with smiths derived dials is that they are current driven not voltage. If they were voltage driven the task would be much easier. As it is the usual problem with the plating on the Lucas spade connectors corroding and going open circuit (or nearly so) is adding to my difficulty.