Is the Series 1/2 temperature gauge adjustable?

[Tom H]

Jerry, The dwg you show is not how the gauges work. It is not a galvanometer located within a resistive bridge. Rather, two of legs of the bridge are actually the coils that pull the magnetic pendulum / needle left and right (toward full scale and zero scale). The third leg is a fixed wire wound resistance and the fourth leg is, of course, the sender. I've got a pretty good handle on this. And I just received two more Temp gauges and one more Fuel gauge to get a better idea of what the design and specs are or were. What I am seeing with the senders is that the 52700 is probably an exact or near exact replacement for the original 4800/00 sender. But it seems neither the original 48u00/00 nor the good substitute 52700 is presently available from suppliers. The 3800 is apparently the closest match of any available senders and that is why suppliers have been suggesting it as a replacement. It seems it has a higher cold resistance but a steeper curve, such that at around the critical temps between 170 and 230 it comes close to matching the gauge. Once I finish testing the two additional gauges, I will refine the data on the senders and correlate it to approximate Temp error (rather than ohms error) when used with a good gauge. Mike Hartman and I will then publish a document showing these results. These gauges have the ability to be adjusted by loosening a pair of nuts on the back to reposition the magnetic coils to calibrate the gauge. But I am pretty sure my recommendation will be to NOT attempt that. Unless the gauge was misadjusted at the factory, or was somehow damaged, or some owner or shop misadjusted it, I don't think the calibration would have changed over time. Coils don't lose turns, and the laws of magnetism do not age. Furthermore we do not know exactly how the gauges were calibrated, although you can see various cal point indicators on the dial.

We hope to complete this in about 2 weeks or so.

Tom

 

[loose_electron]

Rather, two of legs of the bridge are actually the coils that pull the magnetic pendulum / needle left and right (toward full scale and zero scale). The third leg is a fixed wire wound resistance and the fourth leg is, of course, the sender.

One coil winding is in series with the fixed resistor and the other one is in series with the sender resistance, right?

Ah, then it is not a Wheatstone bridge, but rather a magnetic field balancing meter. (actually a current balancing device) truly old school, circa WW2 methods.

That said, that fixed resistance should be adjustable to get a very accurate temperature reading at one particular temperature. Similar method of calibration and substitution would work.

Just tossing ideas out here, this EE gutted the electrical system and used Speed Hut gauges instead.

 

[Tom H]

One coil winding is in series with the fixed resistor and the other one is in series with the sender resistance, right?

Ah, then it is not a Wheatstone bridge, but rather a magnetic field balancing meter. (actually a current balancing device) truly old school, circa WW2 methods.

That said, that fixed resistance should be adjustable to get a very accurate temperature reading at one particular temperature. Similar method of calibration and substitution would work.

Just tossing ideas out here, this EE gutted the electrical system and used Speed Hut gauges instead.

Yes, not really a Wheatstone bridge, but somewhat analogous and it also explains why it works well even without a voltage stabilizer.
And the fixed resistor and variable resistor are each in parallel with one of the coils. Here's a link to a description of a similar (Fuel) gauge used on an MGA. But the MGA gauge is upside down compared to the Alpine and the Temp gauges of both are reversed from the Fuel Gauge, such that on the Temp gauge a low resistance causes a high reading:

http://mgaguru.com/mgtech/electric/fg_03.htm

Nothing wrong with switching to modern gauges! But this old EE is fascinated by how these early engineers solved problems with clever electrical/ mechanical solutions.

Tom

 

[loose_electron]

But this old EE is fascinated by how these early engineers solved problems with clever electrical/ mechanical solutions.

It was a differencing circuit/machine, and the common mode rejection of the differencing device mostly gets rid of the power supply variance. The method of differencing is still in very common use in modern electronics. Signal processing in most semiconductors is pretty much all differential, and not ground referenced. In an IC, that helps get rid of power variance, ground bounce, and common mode noise.

 

[BobTR]

Tom, as you suggested in our PM, yesterday I went ahead and tested the Intermotor N0. 5270 sender that I got off eBay. I also tested the Smith's 3800/00 sender. Both of these were sender tests only with no gauge connected. I again used my pan of water on a hotplate with the sender setting in the metal holder. For this testing I made sure the water was up over the metal holder and a little onto the base of the sender just under the top electric tab. I did not put the tab under water. This time I also kept the water moving in the pan with a flat putty knife (always in the water) as it heated up and then as it cooled down to ensure consistent temperature on the sender and thermometer. The numbers on the left are from the Intermotor No. 5270 and the ones on the right are from the Smiths 3800/00. As you can see, even though the room temperature and cold water temperature resistance readings on each sender were a lot different to start with, they pretty much came close together as the water heated up. I thought that was interesting. Hopefully you can make use of these numbers.

Intermotor 5270 (no number on unit) Smith 3800/00
Thermometer temp ohm reading ohm reading ohm reading ohm reading
heat up cool down heat up cool down
room temp 226 na 298 na
cold water 235/236 na 340/344 na
100 F 167 na 210 na
105 F 145 na 191 na
110 F 121 na 154 na
120 F 100 na 118 na
135 F 71 na 85 na
150 F 48 52 58 61
160 F 38 41 46 48
165 F 34 40 41 42
170 F 30 34 36 38
175 F 28 29 33 33
180 F 24 26 30 29
185 F 22 23 26 26
190 F 20 20 23 23
195 F 18 18 21 21
200 F 16 16 19 19
205 F 15 15 17 16
208 F 14 14 16 16
210 F 13 13 15 15
Fast boiling –211/212 F 12/13 12/13 14 14

Bob

 

[BobTR]

Ummm, the numbers did not stay in columns as they looked before I hit the send button. I tried to attach the Microsoft Word doc but it must be too large. Tom, let me know if you want me to email the Word doc to you. Thanks!

I think you may be able to interpret the numbers from this "key"
1) First row is thermometer water temperature in F
2) second set of numbers is No. 5270 ohm reading with water heating up
3) third set of numbers is No. 5270 ohm reading as the water cooled down
4) fourth set of numbers is the 3800/00 ohm reading with water heating up
5) fifth set of numbers is the 3800/00 ohm reading as the water cooled down.

Bob

 

[loose_electron]

Ummm, the numbers did not stay in columns as they looked before I hit the send button. I tried to attach the Microsoft Word doc but it must be too large. Tom, let me know if you want me to email the Word doc to you. Thanks!

I think you may be able to interpret the numbers from this "key"
1) First row is thermometer water temperature in F
2) second set of numbers is No. 5270 ohm reading with water heating up
3) third set of numbers is No. 5270 ohm reading as the water cooled down
4) fourth set of numbers is the 3800/00 ohm reading with water heating up
50 fifth set of numbers is the 3800/00 ohm reading as the water cooled down.

Bob

A screen shot picture of a table works well to not mess up formatting.

 

[Tom H]

Nice work! I had the same problem posting columns of numbers, so I am able to read it correctly. You can see that in the critical area between 190 and 2112 ( and my results in oil up to 230) the 3800 ( and the unmarked one) may be useable for Series I,II gauges. I'll include your data in our write up.

And BTW, your data on the 3800/00 matches my test results within 0.5 ohms from 150F to 210F. I think the right Ohms value at each point is the number halfway between the heating and cooling readings. The actual temperature of the thermistor inside the sender is almost surely lagging the water temperature in both directions. Measuring while heating and cooling is a good approach. In my case I tried to minimize that error while only measuring while cooling, by cooling as slowly as possible, putting the hot oil bath in a Styrofoam container to insulate it. With your data and mine confirming the results on the 3800, it adds confidence to my results on the 52700, done simultaneously with the 3800. Now to get some more gauge data and hopefully we'll have some useful conclusions.

OH! And one more thing! The unmarked one looks to be closer to the new 52700 I tested than your 3800 in the 170-210F area. It also looks to be the best match to your gauge of any sender we have tested. So maybe it really is a 52700, like the box says!

Tom

 

[BobTR]

Good! I'm glad you could interpret that latest testing data and can use it for your analysis. When I get a chance I'll give the unmarked "No. 5270" sender a try in my car. May be a while before I can get around to installing it.
Bob

 

[Ken Ellis]

Trial image posting

 

[loose_electron]

So, if you adjust the gauge to 210 degrees F with a 14 ohm reference load you should be in pretty good shape.

My immediate thought would be to take the gauge, replace the sender connection with a 14 ohm resistor, replace the resistor on the other side with a 10 turn potentiometer, and adjust to 210 degree gauge reading. Pull the resistor and plug in the sender and you are good to go.

 

[Tom H]

Ken, Thanks.

Jerry, Sounds reasonable, but I would advise against that. Adjusting the gauge involves loosening two nuts on the rear of the gauge that fix the position of the two coils that move the needle moving those magnetic coils to achieve the proper indication at several cal points. I'm guessing that one could move just the coil that affects the pull toward the upper end, but it appears pretty finicky to me. Furthermore the gauge scale is quite compressed above 200 deg or so, with no markings to really tell you where to set the needle indication for 210F. You could easily mess up the calibration such that the needle might peg high while still actually below max temp. A better solution I will be proposing in our write up, will be to first test the gauge with some fixed specific resistors to verify that the gauge is close to original condition (based on my tests of 4 gauges). Then install the sender and run the car bringing it up to normal operating temperature. Note the needle position on the gauge. Then measure the actual temperature of the coolant as quickly as possible, and then simply observe the gauge noting any time it goes significantly above the "norm". This is how just about all of us use the Temp gauge on our regular cars, which rarely have any numbers, but maybe Green, Yellow and Red areas. If you find the car is actually significantly hotter than the gauge indicates, especially running near boiling, you almost certainly need a new sender. All of the 3 new senders I have tested are actually about 13 to 15 ohms at boiling point, pretty close to the original design resistance. The problem is that the new recommended replacement senders, although OK near the critical boiling point, are not so OK at other points. The unmarked "52700" sender that Bob has already looks pretty close to a good match to an original gauge.

Tom

 

[loose_electron]

Ken, Thanks.

Jerry, Sounds reasonable, but I would advise against that. Adjusting the gauge involves loosening two nuts on the rear of the gauge that fix the position of the two coils that move the needle moving those magnetic coils to achieve the proper indication at several cal points. I'm guessing that one could move just the coil that affects the pull toward the upper end, but it appears pretty finicky to me. Furthermore the gauge scale is quite compressed above 200 deg or so, with no markings to really tell you where to set the needle indication for 210F. You could easily mess up the calibration such that the needle might peg high while still actually below max temp.

My original temp gauge is buried in a storage box since I use modern gauges. I would never advocate taking the actual gauge apart because some people don't have the tools or skills to deal with such things.

The calibration resistance is buried inside the gauge and is not an external add on then?
OK, well you can still balance/adjust the gauge with external resistors.

You should still be able to get a single point accuracy by adjusting the resistance by one of two methods:

Adding resistance on the battery side in series with the connection
OR
Adding resistance on the sender side in series with the sender connection.

In one of those two positions you should be able to use a trim pot to get the gauge where you want it to be for a specific temperature.