MG-Cars.info

Welcome to our Site for MG, Triumph and Austin-Healey Car Information.

Recommendations

Parts

MG parts spares and accessories are available for MG T Series (TA, MG TB, MG TC, MG TD, MG TF), Magnette, MGA, Twin cam, MGB, MGBGT, MGC, MGC GT, MG Midget, Sprite and other MG models from British car spares company LBCarCo.

MG MGB Technical - Temperature Gauge / Sensor Problem

Car is a 72 BGT that was “restored” by the PO. The temperature gauge problem I have is that at about 195 deg F the needle is at the "one width from the H block" position.

I have 2 new senders (ordered for the 72) and have calibrated both with nearly identical temperature- resistance curves. They give the same results when installed in the car.

Checked the voltage at the sensor and it is fluctuating at less than 10 VDC, as would be expected with the 10 VDC supply through the gauge resistance.

Temperature sensor calibration gave 45 ohms at 212 ̊F, 82 ohms at 180 ̊F, and 550 ohms at 72 ̊F.

In the thread “Cooling Question” Max Fulton asked if possibly a temperature gauge out of a 75 or 76 might have been used since that, combined with the earlier sensor, would give the higher readings that I am seeing. The Moss catalog does show different part numbers for the 68-71 gauges and the 72-76 gauges, but the same sensor (760-180) is used for both. I think he was probably thinking of the later 77-80 gauge and sensor.

Questions are:

Does anyone have any data on the resistance of the 68-76 temperature sender at any specific temperature?

Does the 77-80 temperature gauge look the same as the 68-76 such that it could be substituted in the 72 dash and look correct? If the answer is yes, how do I determine one from the other?

Assuming that I do have the correct gauge is there a way to calibrate the gauge?

Thanks in advance

Larry

72 BGT
58 A

Larry Hallanger

I would have expected a lot lower than 10v with the gauge near its maximum deflection, that is where the sender is at its lowest resistance and the greatest voltage dropped across the gauge, hence the least across the sender, also much slower fluctuations than that.

I've checked my fuel gauge and on 3/4 tank it shows just over 3v wrt ground, and pulses on briefly about once per second. That is with the engine off. With the engine running the system voltage will be higher, which will mean the 'on' periods are shorter, but the off periods should be about the same. Whilst this will mean it will pulse faster with the engine running, it will only be about 20% faster and not 4-6 times faster. I know this isn't a temp gauge, but it is about the same era as my 75, and AFAIK the characteristics of fuel and temp systems are about the same. It would be interesting to swap over the temp and fuel senders and see just what happens. The characteristics of your temp sender are in the same ball-park as the fuel gauge sender for max and min.

There isn't a 10vDC supply for the gauges. The factory stabiliser outputs system voltage which can vary between 12v and 15v, however it switches it on and off in a varying duty cycle so that it *averages* about 10v whilst the system voltage varies as above. If you had an aftermarket stabiliser which regulates to 10v, then it would be a *constant* voltage you would see, not fluctuating. With the stabiliser switching on more frequently the gauge will read higher. In theory this should be compensated for by switching off more frequently as well, but as your stabiliser seems to depart so much from what I have observed with both mine, i.e. it is probably operating outside its design spec, then there is every reason to expect that the duty-cycle is out as well and it is averaging higher than 10v.

The pre-77 and 77-on gauges might be the same size, but the earlier gauges have the needles pointing downwards and the later pointing up, so quite different in appearance. Max may well have got the year wrong, but surely the principle still applies and you could have a 77-on sender. I note you have ordered 2 earlier senders and they give the same results, but they could be labelled/stored incorrectly, I'm sure some time ago someone else went through this and they were being sent the wrong senders.

I've no idea whether temp gauges are the same as both mine are capiliary but *fuel* gauges can be calibrated by twisting two little slotted plates on the back of the gauge, possibly covered by cork plugs. They each adjust the reading when the pointer is on that side of the the gauge. Use a good-fitting screwdriver, as the plates are only thin, can be stiff, and a small screwdriver with twist a hole in them. Otherwise you could insert a 'calibration resistor' in series with the circuit to lower the reading to more closely represent 205, which I reckon should be less than mid-way between N (straight down) and H.
Paul Hunt 2

Larry, I don't have a calibrated gauge but I have checked the 180 degree thermostat in a pan of hot water. The stat begins opening at 180* and isn't fully open until around 200*. That is fairly typical for all 180 stats I have ever checked. Of course my 74 MGB heat gauge is marked C N H. In normal driving the gauge reads almost to the N with a stat I have tested. I have to assume the temperature is around 200 degrees when leaving the stat so N must be around 200 degrees. A few years ago I did some average voltage and resistance readings ranging from cold to normal plus a heat soak. I couldn't check at H because my engine temperature doesn't ever get to H. My cold engine readings were at around 72 deegrees air temperature. I failed to get a voltage for the heat soak condition The readings were measured with a good quality digital meter. These readings are from my 74. At the time I had a 68 and checked to same readings on it, the figures were not exactly the same but were very close through the range. All voltages and resistance readings were taken at the sender terminal.

Cold engine 720 ohms 10V
Needle left of N 78 ohms 6.8 V
After 30 min. needle at N or just above N 72 ohms 6.4V
Engine off, several min. heat soak, voltage not read. 65 ohms

Clifton
Clifton Gordon

Larry,

I think I have a couple of spare temp gauges in my stash, including a pre- and post- 77. They are very different in appearance, but you are welcome to either or both if they'll help.

Paul K

"I couldn't check at H because my engine temperature doesn't ever get to H."

It will do if you put a piece of cardboard across the front of the radiator.

Clifton - what voltage stabiliser did you have, and what digital meter i.e. does it have an averaging function?
Paul Hunt 2

Paul: I realize cardboard in front will increase engine heat but I did those measurements to have a reference for normal running conditions.

I have the OE stabiliser. My meter is a Radio Shack 22-183 Digital meter about 15 years old and was their best meter at the time. I know Fluke makes much better meters. It does not have an averaging function. It does vary some when reading stabiliser voltage. I used the higher of the varing readings for for each measurement. Perhaps not very scientific but it gives me some reference data for my car.

Clifton
Clifton Gordon

Clifton - that should be fine, as long as you aren't varying the engine speed or electrical load during the tests. Did you note the rate the stabiliser was switching? I.e. was to closer to the 4-6 times per second of Larry or the 1 per second I get on both my cars?
Paul Hunt 2

Paul, Best I could tell it's about 1 time per second. I looked at my sender again yesterday with the engine cold around 70 degrees, engine not running, ignition on. At the sender I read a varying voltage of around 9.8 volts to around 8.3 volts. I wonder if the swings in voltage would be greater with a analog meter or is the limited swing due to damping of the heat gauge? I should try to fins a decent analog meter. My old Simpson 260 died several years ago.

Clifton
Clifton Gordon

My analogue meters respond in significantly less than 1 second so give an easily-read instantaneous reading, swinging between 0v and what ever voltage is derived from the potential divider of the sender and the gauge. A digital should really give much the same, dependant on its sampling rate, although it probably has more chance of showing zero than a voltage as the on period (voltage) is shorter than the off period (no voltage). The gauge itself has no effect on how *fast or slow* a meter responds, either analogue or digital, that is purely down to the construction of the meter. The only effect the meter does have is on the actual voltage reading, and that is from the effect of its resistance in the potential divider with the sender.

The thermal construction of the temp and fuel gauges gives a huge amount of damping on *their* readings, it can take getting on for a half a minute to reach a 'steady' reading (actually cycling up and down very slightly once it gets there if you you look closely, which shows the 'stabiliser' switching on and off). And when you first switch on the ignition the factory stabiliser itself is 'cold' and so connects full power to the gauges for the first few seconds, which actually results in the gauges reaching a steady reading faster than they would with a solid-state regulator.
Paul Hunt 2

Larry,

here are some readings of the temp gauge sender unit:

°C 0, °F 32 -> 2.307 Ohms, °C 20, °F68 -> 885, °C 40, °F 104 -> 384, °C 45, °F 113 -> 317 (Gauge Min.)
°C 60, °F 184 -> 155, °C 75 °F 167 -> 112 (Gauge Mid.)
°C 85, °F 185 -> 82, °C 90, °F 194 -> 71, °C 95, °F 203 -> 62, °C 100, °F 212 -> 54 (Gauge Max.), °C 105, °F 221 -> 47, °C 110, °F230 -> 41 Ohm

Hope this helps

Ralph
Ralph

Very useful Ralph. What sender does this relate to, there were two senders and gauges and getting the wrong combination results in incorrect readings, so one assumes the senders are different.

Also I don't think 213F/100C relates to max on the gauge i.e. H as temperature calibrated gauges were graduated up to 230F, with 90F minimum and 182F equating to the position of the 'N' on later gauges and the temperature of the standard stat on most models.
Paul Hunt 2

Paul,

the sender units only differ from the lay out of their housings. The inside mounted thermistors should read the same resistance within normal tolerances of the production (+/- 10%).
There are two different electrical temperature gauges for the B.
They differ from the styling of the chrome ring but are internally the same.

Ralph
Ralph

I have plotted the temperature - resistance data from my tests, Ralph's data from his email, and a couple of data points from Bob M. They are basically the same curve at the higher temperatures which are the ones that matter. Looks like I have a voltage stabilizer problem which I will address this weekend. Will post the results when available.

Thanks to all.

Larry

Larry Hallanger

My experience indicates a significant difference between 'early' and 'late' senders. A late sender used with an early gauge will yield readings significantly higher than normal. Likewise fitting an early sender with a late gauge will give lower readings.

Years ago (pre-internet) I had a 71 GT that ran abnormally hot. It never overheated, but the gauge always read very high (almost pegged on 'H' IIRC). Even went to the expense of having the rad redone with a heavy duty core - it made no difference. Finally swapped out the sender and all was well. Put the old sender in a late car and it read correctly there.

Regardless, the gauge can be recalibrated as Paul described.

If only the temp gauge is reading incorrectly I'd doubt that it's the voltage stabilizer at fault. If both gauges are off, then check the stabilizer first.

Tom
Tom Sotomayor

Tom

Your experience sounds like an exact match to mine. Am I correct in assuming that by "early" and "late" your are referring to 68-76 and 77-80? Moss has the same part number for 68-71 and 72-76, and a different one for 77-80. My car is titled as a 72 but was built in late 71. If Ralph is correct that only the housing is different between the sensors. I am assuming that my fuel gauge is reading correctly. It shows full after a fill up and low when I hit 200+ miles (have done 240+ on a tank but it was very close to the empty mark).


Ralph

What "model" sender is your data from? Can you describe the physical differences between the early and late senders?


Paul H2

I still suspect that the voltage stabilizer may be at fault since the cycle rate I measured was so much faster than yours. Just checked the wiring diagram and the back of the fuel gauge, which is also fed by the stabilizer, is much easier to access than the back of the temperature gauge, and probably the voltage stabilizer itself. Will check that as soon as I can get to it, probably this weekend (was out of town last weekend).

Paul K

Thanks for the offer. Will hold off until I track down the other possibilities first.



Larry
Larry Hallanger

As I recall, the sender change happened c1971-72. Around that time (I remember where I was), we ran into goofy gauge readings with replacement units - sorry I do not recall which way they went. The early OE ones had black plastic insulators, the later OE one was red. Otherwise physically interchangeable. Current replacements all seem to be black, and I suspect that they might be a sort of averaging of the two OE versions.

Thanks to Paul for the nice writing on the elegant ambient compensation in the instrument design - I hope all the people who think modern is "better" read it carefully! Note that it is just a refined implementation of the rather strange Magnette instrument.

FRM
FR Millmore

Larry,
Yep, 68-76 and 77-80 are the break points I was thinking of.

It sounds like your fuel gauge is working correctly to me.

To adjust the gauge, pull it out of the dash. If you need to, make up some jumper wires with a male spade connector on one end and a female on the other. Make the leads fairly long - you'll find lots of uses for them! This makes the adjustments a snap. Just at a guess if you have a 180* thermostat, adjust the needle to "N" just as the thermostat opens (upper rad hose heats up). Adjust it using the 'hot' side adjuster on the back of the gauge.

Tom
Tom Sotomayor

There are *four* altogether in the Parts Catalogue although two of them are alternatives from different manufacturers:

BMK 1644 from 18GF 101 through to 18V 672/673 up to 74 or 74 1/2.
88G580 from 1975 to 76, possibly 18V 797 on.
13H 5602 or 13H9715 from 77 on.

Haven't got time to check whether there were head or gauge changes at that time.
Paul Hunt 2

I bought both early and late senders from Moss to add to my Frankensteined '69 GT. Both were smaller than the original that always read low. The early one read pretty hot and the later one read way, way too hot. I then decided to investigate the t-stat. Sure enough 165*. Replaced with 180* t-stat (which was verified in hot water before installation) and original sender and the gauge now reads normally. Verified that at N posiition the t-stat housing was at 180-182* with IR thermal gun.

My theory is that the early sender that Moss is selling is not the correct part.

Darrell
darrell

I never understood what Larry's complaint was anyway - "...at about 195 deg F the needle is at the "one width from the H block" position." is within the normal range of imprecision.

From a 77 I did recently, after all bad grounds and low voltages were corrected; temps measured at the rad fan switch with a good digital thermocouple instrument.
Temp gauge reads mid point @180,
low edge of red @ 210,
high side of red @ 250

FRM
FR Millmore

Larry,

i only mentioned the early version up to 1976, as my cars are all built before this date. Sorry that i did not quote this.
For this reason my answer to Paul Hunt's question might be not 100% right.
The physical difference is the housing material made of brass, while there is another version that has a galvanized body with a copper tubing that projects into the cooling system.
As far as i can realize, the brass version was fitted during the production run until at least 1975.
If you have a look upon the 'genuin spare part senders' they were galvanized (the one i bought in 1976 was).
I also tried one from a A-Series engine that was matching the readings of the B-Series sender.

Ralph
Ralph

"Sure enough 165*. Replaced with 180* t-stat (which was verified in hot water before installation) and original sender and the gauge now reads normally."

This doesn't make sense. A 165 stat will open earlier than a 180, and so the gauge should read *higher* with the 180 than the 165. Until, that is, the engine starts generating more heat than the radiator can get rid of, at which point both stats should be fully open and irrelevant anyway. The only way a 165 stat could cause high gauge readings is if it were faulty and opening too late or not enough, in which case replacing it with another 165 would have made more sense, particularly in a warm climate. And replacing the stat when you have 'too hot' problems is one of the first things you do anyway.

Looking at the early numerically calibrated gauges 195 is barely 1/3rd the way from the N position to the H position, and is the normal switch-on point for electric fans (90C/194F). This seems to happen at about mid-way between the N and H positions of the later 'C-N-H' capilliary gauges. It's going to be harder to judge on the electric gauges with their narrower angle of operation, but I certainly wouldn't be very happy if the needle got to a needles width of H before the fans came on.

Paul Hunt 2

Paul-
I went through the same thought process re Darrell's story. The key is "... the original that always read low". In other words, it WAS running at 165, and now it is at 180. But, the replacement units had it too high, even with the low stat.

I'm not very happy with gauges that read that high either, when the engine is in fact not in an overheat condition - but they are crap gauges, and that's why they took the numbers off them.

Don't even get me started on the damned - RANDOM - fan switches!

FRM
FR Millmore

Yes Millmore has it right. My gauge always read low which was due to a PO putting in a 165 t-stat for some reason. I did not know this at the time. Replacing the sender was easier than replacing the t-stat, so I tried that 1st only to find that now it always read hot (early) or really hot (late). My gut was telling me that the engine was not running hot, if anything it seemed cooler than it should be.

Replacing the t-stat and going back to my original large bodied sender worked properly.

Darrell
Darrell

I've been through the Parts catalogues and they at least show the following changes:

Sender:
Nov 61, start of MkII 18GF to 18V 672/673, 138401-368081, BMK 1644
Dec 74, start of 75 model year 18v 797/798,368082-410000, 88G 580
Sep 76, start of 77 model year, 410001 on, 13H 5602 or 13H9715

Gauge:
Nov 61, start of MkII, 138410-258000, BHA 4686
Aug 71, start of 72 model year, 258001-410000, BHA 5090
Sep 76, start of 77 model year, 410001 on, AAU 3030 (BT 2231/01)


So when the gauge changed in 71, it must have had the same characteristics as the sender didn't change until 75. And when the sender changed in 75, the gauge remained unchanged until 77, so again the characteristics must have remained the same. It was only for the 77 model year that both gauge and sender changed. E&OE in the Parts Catalogue, of course.
Paul Hunt 2

By the way, the 'BT 2231/01' number is the Smiths number. On a fuel gauge at least this can be found on the rear part of the face above the F mark, tucked up behind the front part of the face (see Image).

Paul Hunt 2

>So when the gauge changed in 71, it must have had the same characteristics as the sender didn't change until 75.

Not necessarily. I don't have a parts book in front of me so I'm speaking purely from conjecture here, but if the gauge change coincided with a thermostat specification change, they could have for example gone to a hotter thermostat and by changing the response curve of the gauge, kept the needle centered. Again, I don't know that they DID do this, only that they could have....

Of course, given the precision of the gauge, why would they have bothered? ;-)
Rob Edwards

It was always impossible to get a straight answer to this, and there were clearly errors in listings at the time. However, I distinctly recall that it was mid 72 through 73 that I first ran into the problem. That would mean that the cars affected were newly out of warranty, probably 71 or early 72 cars. The "old" sensor was black, and the "new" one was red. "Old" sensors gave wrong readings in "new" cars & VV.

From The Roadster Factory book - they are usually pretty good on these things - 68-71 = GTR104; 72-76 GTR 101
Those would be Unipart numbers, superceding the BL ones.

I believe that the GTR101 item is the same as was used on Austin America (1300 on the other side) and such. On those the "new" sensor came in earlier, or maybe from the beginning of production. I have two AA engines here known to be 69 & 70, with red sensors.

FRM
FR Millmore

According to Clausager standard thermostats were originally 180F. 'Soon after' Sept 64 this was changed to 165F. In March 69 the standard stat changed back to 180F, and remained so until the end of production. None of which tie in with either gauge or sender changes.

From Sept 64 there were always hotter stats for cold countries, and cooler stats for hotter countries except while the standard stat was 165F.

Moss US quotes only two senders - one from 68 to 76 and another after that, likewise Vicky Brits. The Roadster Factory online catalogue has what you say, FRM, but gives the dates as 68 to 74 and 74 to 80, which sort-of agrees with the first two in the Leyland catalogue. It would be interesting to get temperature/resistance curves for those two. It's really only relevant with original senders and gauges, because as we know too well what comes out of parts houses can be way off, despite the numbering.
Paul Hunt 2

I use the old TRF print MGB catalogue, which has not been available for a long time. It is a very nice redo/summary of factory books of the time, though it is missing a few things. My favorite thing about TRF is that he uses factory & OE supplier numbers, and they are all cross referenced to each other. I despise the renumber practices of other suppliers.

As I said, the TRF print listing agrees with my memory of the time, and I was running my own shop, exclusively working on British cars and Volvo - about 20 cars a day throughput. It was clear then that the Leyland/Unipart listings were not to be trusted; since aftermarket listings were derived from them, you can't trust those either.

But as you say, with what we get now, does it matter? Which is why I questioned the original problem Larry had. The only reasonable solution is to live with your known gauge behavior, calibrate the gauge, or buy real instruments.

FRM
FR Millmore

This thread was discussed between 06/09/2007 and 15/09/2007

MG MGB Technical index

This thread is from the archives. Join the live MG MGB Technical BBS now