It is one of the most common questions asked by TA, TB and TC owners, and it usually arrives in the same shape: most electronic gauges are made for negative earth, and my car is positive earth, so what do I do? The answer is that the question contains a false assumption. The gauge that suits these cars best has no electrical connection at all, so the polarity of the car is simply not a factor. There is more to say about where the bulb should go and what the reading means, but the hard part of the problem dissolves as soon as you stop looking at electric gauges.
Which cars had one, and which did not
The starting point is worth stating, because it explains why the question comes up so often. Most T-Types never had a temperature gauge.
| Model | Water temperature gauge | What the dashboard carried instead |
|---|---|---|
| TA | None | Speedometer, rev counter with eight day clock, ammeter, oil pressure gauge |
| TB | None recorded | Grouped with the TA and TC in every catalogue consulted, none of which lists a temperature gauge. Probable rather than proven. |
| TC | None | Jaeger S.461 speedometer, Jaeger K30 rev counter with integral eight day clock, Lucas ammeter, Jaeger oil pressure gauge |
| TD before chassis 13914 | None | A separate oil pressure gauge |
| TD from chassis 13914 | Yes | A combined Jaeger oil pressure and water temperature gauge replaced the separate oil gauge. Mechanical, with a capillary. |
| TF 1250 and TF 1500 | Yes | A separate temperature gauge alongside a separate oil pressure gauge, ammeter and clock. No fuel gauge. |
| Y-type | None | Oil pressure and petrol contents only. The Y Register lists the temperature gauge among the things the car does not have. |
Two things follow. The first is that fitting a gauge to a TA, TB or TC is an addition rather than a restoration, so nobody should feel they are spoiling an original car by wanting one. The second is that the factory itself answered this question in 1952, when the late TD gained a capillary gauge reading from the radiator top tank. That is the pattern to copy.
Owner photograph to follow
The gauge that does not care about polarity
A capillary gauge, also called a mechanical or Bourdon gauge, is a sealed system with no electrical connection whatsoever. A metal bulb, usually brass, sits in the coolant. It is partly filled with a volatile liquid, and a fine bore tube runs from it to the instrument. As the bulb warms, the fluid’s vapour pressure rises, the tube carries that pressure to a Bourdon tube in the gauge head, and the Bourdon tube straightens slightly and moves the needle. It is a pressure gauge that happens to be calibrated in degrees.
Because nothing about it is electrical, it has no supply, no earth and no polarity. It does not know whether the car is positive earth, negative earth, six volt or twelve. One specialist notes the giveaway: a capillary gauge does not fall back to zero when the ignition is switched off, because it never had power to lose. The only electrical part is the panel illumination bulb, and a filament lamp does not care which way current passes through it either.
That is the whole answer to the original question. For a positive-earth T-Type, a capillary gauge removes the problem rather than working around it.
One source describes the fill as diethyl ether and works from vapour pressure; another describes the system simply as liquid filled. Both are current specialist sources and both constructions genuinely exist. The archive has not established which a given Smiths or Jaeger automotive gauge uses. It makes no difference to how you handle one.
The capillary tube, and the one mistake that ruins a gauge
The tube is not a cable. It is the pressure path of a sealed system charged at manufacture, and it cannot be shortened, cut or separated from either end. Cut it and the charge escapes, the needle drops to the bottom of the scale, and it stays there. Repair is possible in principle, since the gauge can be re-charged, but it is a specialist operation and the fluid is not easy to obtain. Treat a cut capillary as a gauge lost.
- Do not attempt to separate the gauge or the bulb from the tube.
- No bend tighter than a one inch radius anywhere along its length.
- Clip the tube every four to six inches where it runs along the engine.
- Leave three or four coils of not less than two inches diameter between the engine and the bulkhead, so engine movement and vibration are absorbed by the coils rather than by the tube.
- Coil any surplus length neatly behind the dashboard rather than cutting it. Surplus is normal and is not a fault.
- Do not test the gauge by dipping the bulb in boiling water. It can damage the instrument.
Whether a mechanical gauge will reach
It is often said that a mechanical gauge cannot reach from the radiator to the dashboard. On the evidence this is not so, and it is worth correcting because it is the reason people talk themselves into an electric gauge they then have to fight.
Seventy eight inches, which is six feet six, is the standard Smiths capillary length. A hundred and twenty inch version is catalogued. Aftermarket gauges are sold with tubes of 1800, 1850 and 2000 millimetres. And the factory itself managed it: the late TD’s combined gauge read from the radiator top tank with a capillary, which on a T-Type is one of the shorter runs on the car rather than one of the longer ones. Surplus tube is the normal outcome, not a shortage of it.
| Gauge | Capillary length |
|---|---|
| Smiths TG1310-01C078, water temperature 30 to 110 C | 78 inches |
| Smiths TG1301-30B120, water temperature 30 to 120 C | 120 inches |
| Smiths GD1301-64, dual oil and water | 78 inches |
| Vintage Car Parts CA721, water temperature, 3/8 inch BSP bulb | 2000 mm |
| Vintage Car Parts CA1310 duplex oil and water | 1800 mm |
| BHA4737 dual oil and water, the Austin-Healey and MGB gauge | 6 feet 6 inches |
That last one is the gauge sometimes described as the big Healey combination gauge. It is a 52 mm instrument with a five eighths UNF bulb and six feet six of capillary, listed for the Austin-Healey 3000, MGA, MGB, Sprite and Midget. Whether it drops into a T-Type dashboard aperture without modification is not something the archive has established.
If you would rather have an electric gauge
It can be done, and the polarity question then becomes real rather than imaginary. Three separate things have to be right, and they are commonly confused with one another.
The sender does not care. Temperature senders are thermistors, which are simply resistors whose resistance varies with temperature. A resistor has no polarity and does not know which way current is passing through it. What a sender does care about is its earth. Most are earth return types that complete their circuit through the sender body into the metal it is screwed into, which means PTFE tape or thread sealant on that thread will break the measuring circuit. The manufacturer’s own instruction sheet says so plainly. Note the trap for a T-Type: the club’s published radiator adaptor article recommends PTFE tape, which is entirely correct for a capillary bulb that carries no current and entirely wrong for an earth return electrical sender.
The gauge may or may not care, and the discriminator is what is inside it rather than what is written on the dial. Bimetal gauges work by resistive heating, which does not depend on current direction, and the maker states that it does not matter which way round the power and sender wires are connected. Modern 52 mm electrical gauges are a different matter: the current instruction sheet says they are to be fitted to negative earth vehicles only. So a positive-earth car is not shut out of electric gauges, but you must match the gauge and not merely the stabiliser.
Where a gauge does take its earth through its case, and most classic Smiths instruments do, it is earthed by being clamped against the metal instrument panel. Isolating the body means mounting it on nylon or fibre bushes and washers so that neither case, bezel nor bracket touches the panel, proving with a meter that there is no continuity to the panel, running a dedicated earth wire of the correct polarity instead, and remembering that the illumination bulb holder usually earths through the case too and needs the same treatment. That is the general principle for panel-earthed instruments. The archive has not found a published procedure specific to a T-Type instrument panel.
The voltage stabiliser
An electric temperature gauge of the bimetal type does need a stabilised supply, and this is not optional. Without one the gauge reads by battery voltage as much as by temperature, so it will read differently at a fast idle from how it reads at rest. The manufacturer is blunt about the consequence: unstabilised, these gauges will massively over read and may be damaged internally.
The Smiths stabiliser is not a linear regulator. It is a thermal switch, a bimetal strip with a heater coil that chops the supply on and off about once a second, so the time averaged voltage is around ten volts whether the battery is at twelve or the dynamo is pushing fourteen. The instrument sheet gives the working window as 9.6 to 10.3 volts. The output is never actually a steady ten volts, which is why the gauges themselves are slow acting bimetal instruments that respond to the average rather than the instant.
Which gauges need one depends on type. Bimetal fuel and temperature gauges require it. Moving iron gauges must not have one. Capillary gauges need nothing at all, having no supply to stabilise. The stabiliser itself is polarity specific, unlike the sender and unlike the gauge terminals, and positive earth versions are current stock items rather than rarities. One is listed as part 13H3554 and another simply as VSP. Ratings differ between the maker’s own documents, one saying a maximum of two gauges per stabiliser and another saying one to three, so two is the conservative figure to work to.
No T-Type was fitted with a voltage stabiliser at the factory, which follows from the fact that its temperature instruments were mechanical. That is established by consistent absence from the catalogues rather than by any positive statement, and is stated at that strength.
Where to put the bulb, and why it is argued about
The factory location is the radiator top tank, and it is the only place on a T-Type with a catalogued adaptor: Abingdon Spares 65-052A, cross referenced as Moss 361-050, described as threading into the radiator top tank to take the sender, listed for TD and TF. For a TC there is no catalogued part, and the club has published a home made answer: a three eighths inch tapered BSP brass plumbing nipple, bought for under a pound, bored out slightly on a lathe because its internal diameter will not otherwise pass the bulb, and sealed into the radiator. The author of that piece used a Jaeger gauge.
The cylinder head is the alternative, and here two named authorities disagree, which the archive reports rather than resolves.
| Source | Position taken |
|---|---|
| Neil Cairns, MG Car Club Y Register | The sensor is far better fitted to the cylinder head. There you get a more accurate reading and see any faults sooner. |
| Eric Worpe, MG T Society | The header tank gives more information on the cooling system, with the possible added bonus of indicating whether the coolant level in the header tank has fallen. |
Both are defensible and they are optimising for different things. The head gives the earlier and more direct warning that the engine is hot. The header tank shows how much cooling headroom is left, because as temperature rises less flow goes through the bypass, and it will also betray a falling coolant level in a way a head sender never will. The head is also where the trouble starts: because of the way an in line engine is laid out, and because the coolant enters the block at the rear and feeds up into the back of the head, number four cylinder runs hottest.
The two positions do not read the same. At the header tank some of the hot coolant leaving the thermostat mixes with what is already in the tank, so a published figure gives the head outlet at about 74 degrees and the header tank typically 65 to 70. That is a mixing effect rather than a calibration error, and it means a gauge reading is only meaningful once you know where its bulb is.
There is no catalogued cylinder head adaptor for a T-Type. Anyone taking that route is fabricating rather than buying, and should say so when describing what they have done.
The radiator cap thermometer, and its name
The period solution, before dashboard gauges were common, was a thermometer mounted in the radiator filler cap and read by the driver looking forward through the windscreen. The correct name for it is a motometer, usually a Boyce MotoMeter after the dominant maker, patented in 1912 and largely obsolete by about 1930. It is sometimes called a calorimeter, which is a different instrument altogether: a calorimeter measures a quantity of heat in joules, not a temperature. The word is simply a misnomer, and worth correcting because the two measure different physical quantities.
Moss list it as a period accessory for cars with an exposed radiator cap and give the fitment as MG TC and TD, 1945 to 1953. For the TA the archive found no fitment listing, and given that the motometer was already going out of fashion by 1930 a TA motometer is at best a late survival rather than period practice. Not established either way.
It should be understood for what it measures. A motometer reads the temperature of the vapour and air in the radiator neck above the coolant, not the coolant itself and certainly not the head. It was designed for unpressurised thermosyphon systems, it is sensitive to coolant level and to airflow past the cap, and the standing historical complaint against it is that it failed to warn drivers in time. It sits at the coolest end of the circuit and is further behind events than anything else discussed here. As an ornament with a function it is charming. As an instrument it is the weakest of the options.
Do you need to convert to negative earth
For this purpose, no. That is the practical answer and it is worth putting first, because the suggestion comes up every time the question is asked. A capillary gauge has no electrical connection, and positive earth stabilisers are current stock items, so neither sensible route to a temperature reading requires a polarity change. Converting turns fitting a gauge into a job involving the dynamo, the control box, the ammeter, the coil, the clock and the radio.
If you are converting anyway, for other reasons, these are the parts that care. The battery is reversed, usually needing to be turned round for the leads to reach. The dynamo must be repolarised, because it retains residual magnetism and will otherwise fight the system; three published guides give three slightly different techniques for the momentary connection that does it, all achieving the same thing. The ammeter connections are reversed. The coil primary connections are reversed, and on a positive earth car with a modern coil the printed markings are the wrong way round, which if ignored puts the full battery output through the primary winding to earth and overheats the coil. A period radio is the most vulnerable single item and can be destroyed outright.
The starter, the lighting, the switches and the heater motor do not care. A series wound starter develops torque in the same direction whichever way the current flows, and filament lamps are indifferent.
There is a T-Type specific sting worth knowing. The eight day clock built into the TA and TC rev counter head is polarity sensitive, and the club’s own guidance notes that it may need an electronic replacement mechanism which is not cheap, and that the ammeter will need attention. The club’s framing of whether to convert at all is a use case judgement rather than a matter of principle: if you tour, and want to run a fan, a sat nav, a phone charger or a tyre inflator, the answer may be yes.
The originality objection is real and widely held, since the car left Abingdon positive earth, but the archive could not find any club or concours document stating that a conversion is penalised, and will not invent one.
What the needle should show
Here the archive has to report a disagreement rather than a figure. Two named club authorities give different running temperatures for the XPAG.
- Eric Worpe, MG T Society: the output water temperature is held at a nominal 74 degrees, with the thermostat almost fully open at around 74 degrees under hard driving.
- Neil Cairns, MG Car Club Y Register: the thermostat will control the engine’s running temperature at about 82 degrees.
The likeliest explanation is that they are describing engines with differently rated thermostats, but neither states the rating, so the gap stands unresolved. Add the mixing effect described earlier and the sensible conclusion is that no single number is the right answer. What matters on a gauge is the temperature your own car settles at, in traffic and on the open road, when it is known to be well. A gauge earns its place by showing a change from that, not by matching a figure from a book.
For the MPJG engine in the TA the archive found no published running temperature at all.
What to fit
For a positive earth TA, TB or TC, a capillary gauge reading from the radiator top tank, using either the catalogued TD and TF adaptor where it fits or the club’s home made nipple where it does not. It ignores the polarity of the car, needs no stabiliser, no earth and no supply, and it is what the factory itself chose when it added a temperature gauge to the TD in 1952. If you want oil pressure and temperature on one dial, the dual gauges exist with the same six feet six of capillary.
An electric gauge is a legitimate second choice if you prefer the look of one, provided you buy a bimetal gauge rather than a modern electronic one, fit a positive earth stabiliser, and keep sealant off an earth return sender’s thread.
Common questions
- Did the MG TC come with a temperature gauge?
- No. The TC dashboard carried a Jaeger speedometer, a Jaeger rev counter with an integral eight day clock, a Lucas ammeter and a Jaeger oil pressure gauge. There was no water temperature gauge. The same is true of the TA, and of the TD before chassis 13914. The late TD gained a combined oil and water gauge, and the TF had a separate temperature gauge.
- Does a positive earth car stop me fitting a temperature gauge?
- Not if you fit a capillary gauge, which has no electrical connection at all and is therefore unaffected by polarity. If you want an electric gauge, buy a bimetal type rather than a modern electronic one, since current 52 mm electrical gauges are specified for negative earth vehicles only, and fit a positive earth voltage stabiliser.
- Will a mechanical gauge reach from the radiator to the dashboard?
- Yes. Seventy eight inches, or six feet six, is the standard Smiths capillary length and longer versions are catalogued. The factory's own late TD gauge read from the radiator top tank. Surplus tube is normal and should be coiled, never cut.
- Can I shorten the capillary tube?
- No. It is a sealed pressure system charged at manufacture. Cutting it releases the charge and the gauge reads dead low until it is re-charged by a specialist. Coil the surplus instead, with no bend tighter than a one inch radius.
- Should the bulb go in the radiator or the cylinder head?
- Published club authorities disagree. The head gives the earlier and more direct warning that the engine is hot. The header tank shows the cooling system's remaining margin and will reveal a falling coolant level. The header tank is the factory location and the only one with a catalogued T-Type adaptor. Expect the header tank to read several degrees lower than the head.
- Do I need a voltage stabiliser?
- For a capillary gauge, no, there is nothing to stabilise. For a bimetal electric gauge, yes. The manufacturer states that without one the gauges will massively over read and may be damaged internally. Moving iron gauges must not have one.
Sources
This article draws on the MG T Society for TC instrumentation, the TC radiator adaptor article and Eric Worpe’s cooling articles; the MG Car Club Y Register and Neil Cairns for XPAG cooling and the Y-type instruments; the MG Owners’ Club marque guides for the TA and TF; Caerbont Automotive Instruments, the current makers of Smiths instruments, for the voltage stabiliser and gauge instruction sheets; the Triumph Owners’ Club Christchurch technical papers on Smiths gauges for sender and gauge types; Speedy Cables for capillary handling; Grace’s Guide for British Jaeger; and the Moss, Abingdon Spares, SC Parts, Vintage Car Parts and Complete Automobilist catalogues for parts and lengths.
The archive could not establish: whether the TB or the TD Mark II differed from their contemporaries in instrumentation; whether the TF’s factory gauge was capillary or electric; which fill a given Smiths or Jaeger gauge uses; whether classic moving iron gauges are intrinsically polarity sensitive as opposed to earth sensitive; any T-Type specific procedure for isolating a gauge body; any catalogued cylinder head adaptor for a T-Type; a running temperature for the MPJG engine; and whether any radiator is sold ready tapped for a temperature bulb.
Read next
This guide is a general introduction and is not a substitute for the official Workshop Manual for your car.