Which thermostat for the XK120 140 150: a study contd

Part 2: Thermostats that you can fit to an XK120/140/150 and early Series 1 E Type that use the bellows-sleeve type thermostat. (All must use only the 4lb max radiator cap). The Jaguar Factory Service Manual provides specs: starts to open 60-63C, full open at 80C and 3/8inch (9.5mm lift). The 9.5mm lift is the important part as that is how far the sleeve moves up to close the bypass port. You can use any of the listed bellows sleeve ‘stats with the various temperatures.
Smiths The Smiths coding system changed in early 1958 whereby the X.43xxx numbers were replaced by X.85xxx numbers, some with and others without the X prefix; Note: there were more Smiths thermostat-temperature versions manufactured than the ones Jaguar had chosen. They are sometimes close to the nominal value Jaguar opted for. Examples: 85025/70 instead of 85025/74. With the only difference bein the temperature spec, they can be used, but must be bellows-sleeve type.
Do not use: any non sleeve type: Do not use: Smiths X85032 X, 85035, X.85024/74, X.85025/74).series, (Service Bulletin 235/239 refers to Mark 7, not the XKs, and revises only late Mark 7 use of the above # s in place of the std C3731/1, again, ONLY applies to later Mark 7. Do Not use any “British Thermostat” fitted to early Mark7 only. Do NOT use the AC TC series which have no sleeve,
Although Jaguar fitted various temp thermostats thru the models, that really does not matter. Any of the NOS different numbers that are applicable to the XKs “fit”. So far all exterior dimensions have measured to be the same across all manufacturers and numbers. All work the same, except for the opening temp and full open temperatures. Once open, all else being equal, the car will operate at / near the thermostat’s open temp spec.
Jaguar part codes: Jaguar used part # C.3731 and at the factory fitted only a Smith’s thermostat bellows type. Later part number C3731/1 is just a different opening temperature. It is the same exact fitment part, no exterior difference. Current vendors that use the Jaguar C3731/1 numbers do NOT supply new old stock Smiths. They supply whatever current, and different type thermostat that they select that they feel fits and works, and just use Jaguar’s part # to identify that it is “a thermostat”. Jaguar revised the thermostat part number thru the XK140/150, (such as Jaguar code C7105 (83° C, or C.12867) but the only difference is the temperature spec. The part fitment in the housing is the same and all are interchangeable within the 120/140/150, and early E.
Use any of these:
Smiths:earlypart #, laterpart # but essentially the same Jaguar part#
X.43655 opens 60 to 63°C , open 80°C. 85026/60 C3731
X.43570/5 opening 73°C (163° F). X85025/72 (1 de C diff) C3731/1
X43605/3 83C same as X85025/86 3 de diff C7105
X.43570/28 73C same as X85025/74 1 de diff C12867
X.43570/16 79 same as X85025/80 C12867
X85025/74 C13944
X85025/82 C13944/1

AC: use any of the TF: TF1, TF2, TF3, F4…, But NOT the TC series which has bellows but no sleeve. Each TF also has a manufacturing code (next to the temperature stamped on the top flange to refer to as an additional check.
• TF-1 72° C 2235 mfg # code
• TF-2 80° C 2233, 2664 and later 2310
• TF-3 68° C 2202
• TF-4 86° C 2307, there exist TF4 incorrectly stamped 68° C , which should have been 86° C, The production code 2307 will still prove it is the correct TF4 86C.
Lucas (Motostock) In the 1980s some of Smiths had been sold to Lucas, which started selling Smiths thermostats, although they were most likely made by AC, Lucas LF1 72° C (160 F) correct bellows/sleeve, and Lucas LF2 80° C (176 F) winter, correct bellows / sleeve
Quinton Brivec / Quinton Hazell (QT on the part #) Quinton Hazell QT 100/200 series, QT 100/70, QT 100/74, and the QT 200/80 and QT 200/86: both winter temps
Remax much like the Quinton Hazell, the part # was similar: QH refers to QT100, Remax: one example is the NT100. 68C P154
SNG Barrett: C3731/1, One top cross bar and, top mounted sleeve and a visible spring. The flange has a small hole. (the top mounted sleeve moves downward, not up as does the Smiths, AC, QH, Remax. Possibly the same as the current XKs C3731/1 090160.
XKs current C3731/1 (the XKs 090160,) is a 160C: not marked as such on the one I have, 160C is on the cold side, but as said, once open—thermostats are open. The car will likely run warmer with the stat open all the time, except in very cold air. It has one top cross bar and, top mounted sleeve and a visible spring. The flange has a small hole. Note that an older XKs version,( likely a sleeve added RobertShaw), was sold as 09-0160 Superstat that has 3 cross top bars over the flange, a brass enclosed bottom, but does not have a piddle valve or tiny air hole in the flange: It fits, it works, has the sleeve, however, if it is to be used, a small hole should be drilled where possible and functional; Some of this older type are still around, so see the current one, one top bar, as an example. The hole is an air bleed hole to allow air to escape to radiator when filling the cooling system after a drain. Smiths type have the piddle valve.
Moss Europe bellows/sleeve thermostat, part 434-156: this is the exception to the “all the same”, as this is not old stock, but a new manufacture, listed for Triumph, Austin Healy, but not Jaguar. What we have now is an original “Smiths type bellows/sleeve thermostat” being manufactured currently. It fits the housing, but remains to be seen if it closes the bypass when open to radiator: the bottom positioned sleeve moves upward. It travels less to full open at 8.382mm than did the Smiths at 9.5mm, but the sleeve itself has 1.256 more height to compensate and maybe thus close off the bypass. I have no knowledge of an actual test in Jaguar housings as of yet: Does it close the flow to bypass sufficiently? I hope so. It may be that the sleeve stops can be filed to be shorter if necessary. A $123.99 part that is new, is IMHO better than a hard to find $70 NOS that is 60 years old.

The thermostat installs in the XK120 housing, bellows/sleeve into the housing. You can opt to a test of the entire assembly in hot water to prove if the bypass fully closes, or you may test only the thermostat to be sure the sleeve moves at least the spec 9.5 to 10mm upward. With 60 year old parts, do the test before install.
Check points: The thermostat body inserts into the housing; be sure that the side flats that connect bottom to top are positioned so as to not block off the bypass port. The piddle valve or small hole in the top flange should be at 12 O-clock (positioned so that air can escape). Check upon fitting that the sleeve does not bind on the housing; there should be very tiny clearance so that the sleeve can freely move. It is a good idea to test any new (or old when servicing) thermostat in a pot of water to be sure it operates, and opens fully, 9.5mm travel,( except for the Moss at approx. 8.3mm travel which remains to be tested in a housing.)
Use only a 4PSI radiator cap which has a long reach so that the cap internal will contact the radiator neck. Be sure the cap contacts the radiator top and leaves a “witness” mark on the rubber seal. An original cap does not have a rubber seal ring within the cap so the detached seal is necessary in the radiator neck. A later cap will have the rubber seal ring within the cap itself, so if used, no rubber seal in the neck. 50/50 anti-freeze has a boil point of approx. 223F, a 4lb cap is 3lb more, at sea level, The boil point decreases about 3 de F for each 1000 ft. of elevation. At 50-50, the freeze point is -34F. Most anti-freeze specs found on the container have a boil spec if 265 but this is (in fine * print) with a 15PSI cap. Boil with 50-50 is 226 F with our 4lb cap at sea level. A 70-30 provides -84F and 276F (spec is with 15lb cap which we do not use) but may cool less efficiently. Most have 5 year coolant life, some are now more. Anti-freeze left too long creates some gooey junk…nuff said. The 4lb cap is a safeguard and will release too much pressure which will destroy the thermostat and likely a head gasket or core plug: even a block can crack with excessive hot spot heat and / or steam pressure. Most thermostats start to open 10C to 25C lower than the full open temp spec. If you test one in a pot of hot water, the sleeve should open fully at/near the spec temp, fully means the sleeve has moved 9.5mm -10.1mm. There is a “stop” which prevents it from moving much further. My tests on a Smiths showed the sleeve at the low-cold point is 16.6mm below the inside of the top flange. When fully open the sleeve is 6.5 mm below the inside of the top flange, thus the sleeve moved up 10.1mm: it did not reach the “stops”. The “stop” is 4.6mm below the inside of the top flange. The thermostat reacts very quickly to cool and close to radiator: as soon as out of the hot water the sleeve starts to drop and the poppet starts to close.
For the Moss, where the stop does stop the sleeve at 8.38mm yet untested is whether modifying the stop would allow the sleeve travel to move any further without internal damage and possible failure.
Comments/errata to Nick Email:mountainwolf8034@aol.com
See separate post" Thermostats for the XK120".

Nick, I have not physically checked the thermostat in my 1950 XK120, but I do know it is working. I have a spare NOS AC TF2 - is that suitable for normal UK conditions? I don’t quite understand why you insist on a 4lb radiator cap? I’m using a 7lb AC one at present - is this safe, in your opinion? 7lb doesn’t seem a very high pressure, especially if you’re used to American V8s… I also have a very good aftermarket alloy radiator fitted.


I am assuming that your 1950 car does not have a heater. In the interests of raising the boiling point, given the problem of the somewhat marginal nature of the XK120s cooling system, I opted for a higher pressure rad cap. My car, a 1953, does have a heater and I managed to eventually rupture the core with the higher pressure cap. After replacing the core I reverted to a 4lb cap.


Good point, Chris. My car had a heater added at the factory in 1952, though that was in turn replaced with a new one in 1998. I certainly don’t want rusty water all over my new carpet, so I will be reverting to the correct 4lb cap. Does anyone know where to get the authentic-looking cap? UK, preferably.

Hi…4 lb is the correct cap, all the way thru the XK series. Use of a 7lb, with the extra 3 PSI really only comes into play IF and when the boil point is reached. If the car ran below boil all the time the pressure remains under say 1-3 psi. Little if any addtl pressure is created just by hot water. (but still…,don’t open the cap when hot) (there is some water expansion, in the XK120 system the radiator fill mark is well below the top, and it is not a closed/catch tank system) So 7lb…not needed…it would only come into play in a boil situation…and that is exactly when you don’t want the additional steam pressure in the system, even tho it is not much more. The sleeve bellows thermostats could be damaged, heater elements, core plug as well. Radiator cap: XK120/140 : note that the inner neck surface that contacts the inner cap is “deep” on the XK radiator. A correct cap must contact the inner neck seat,.
Roger Payne and Rob Reilly have posted: Two types, The correct original XK120 and early XK140 cap is a MADE IN ENGLAND AC Model RB5 , a Circular Top, with a closed (shrouded pressure release spring) and no sealing washer within the cap itself, with a 1-inch depth valve seat, in conjunction with a (red) seal/gasket mounted in the radiator neck, and also rated at 4psi.
The later replacement, not factory original is a MADE IN USA AC Model RC5 (USA AC Part No 850799). This is a Circular Top, but with an open/visible larger spring loaded seat with integral rubber washer, with a 7/8in depth to valve seat, and rated at 4psi.
In either case, check that a “witness” mark is on the seal, proof that it is in firm contact. Do not use 7lb or more: if a bellows sleeve thermostat is installed more than 4psi can destroy its function. Even if a wax thermostat is installed, the rest of the system, from water pumps to engine core plugs are not made to handle higher pressure. The “need” to raise the boil point is a false need:…something is not correct/optimal in the cooling system. Note that 50 50 modern antifreeze mix (freeze 034F) raises the boil point to about 223F. at sea level. A 4lb cap raises about 3F more. If you use the search forum and type in radiator cap you will find photos and more info. Caps are available from the usual XK parts sources.

See the recent post: Which thermostats for the XK120. Yes, the TF2 can be used. It it is an 80C stat…means full open at 80C so your cooling system will run at, or just above that temp, and not any colder because the thermostat’s job is to keep the temp at 80C. Once at 80C this does not make the coolant run any hotter. On a cold day, maybe with a 73C stat your car would run at say 75C…with a TF2 it will run at 80C. The TF2 was a standard winter thermostat. I doubt many owners switched thermostats, summer to winter…but maybe did select one based on average year round temps. The lower rated thermostat provides a bit more “situation” cushion, say in traffic on a hot day,.,.but only 7-10D.
Just do check any NOS 60 yr old parts…be sure the sleeve is where it is supposed to be cold, not tilted to one side (a damage), and test in hot water to watch it open fully,the sleeve to move upward at least 9.5mm, and the top poppet to open. Our vintage gauge ends at 100C, boil point with 50 50 antifreeze and 4 lb cap, is about 107C (226F) at sea level,. but less by 3 o for each 1000 ft elevation.
For a cap…try BroadSport…Guy Broad. Do be sure it is a long reach cap,see my recent post on "radiator caps,. "

Thanks for such a full answer, Nick! I’ve managed to find a NOS AC RB5 cap with the shrouded pressure release spring, complete with red seal. It has the rather lurid red & blue decoration, but no matter. I hope the radiator neck on my aluminium radiator is for the long reach cap! The 7lb cap that’s on now is an AC RC8 one.

Thanks also for the thermostat info, Nick. I’ll check for any possible damage on the TF2, still in its box. I might keep an eye out for a TF1 72 deg one, too.

the Smiths , QH, and AC bellows sleeve thermostats come up on Ebay from time to time, I posted some item #s recently on the XKs forum. An ebay search…for Bellows thermostat, bellows sleeve thermostat…or Thermostat TF1.(a 72C) or TF3. (a 68C) a QH is a QT100.72. Lucas aLF1. …it takes some different searches to find the listings, fair prices seem to be $40 rto $50 USD, I just obtained several. If you wait…we may know if the Moss Europe will work,still to be tested, it is new…and about $123.99 USD.

Thanks for the tip on Ebay, Nick.
I now have two irig Smith 80 deg thermostats, BUT, I still dont get it.
The stat determines at what coolant temp the water is directed to the radiator for cooling. A high, say 86 deg stat for winter is good in order to get the engine hot faster. But once hot, and with te stat open, I still have a very high 95 deg on two engines, in both cool autumn temperatures, an very hot summer days.
So it must be other issues, such as clogged rads or damaged water pumps. Those original rads were sent off for pressure testing etc before being installed along with rebuilt engines. I know that on one of them there is a NOS 86 deg stat, perhaps on both…?!
Very confusing.
No other engines reach 95 deg. whether original or modern alu rads.

Hi Peder…yes, the thermostat determines the coolest the coolant can be, but not the warmest since once fully open to the radiator which happens at the thermostat spec…so say an 80C stat, then it is the rest of the cooling system (clean, effective water pump, the design, the coolant itself) , the ambient air temp, the amount of air thru the radiator that cool the coolant that determine the coolant temperature. The thermostat is on vacation…fully open to radiator, and should be fully closed to the direct path to engine (the bypass closed). Various engine coolant passage internals will play a role. The XK120 was known from the start to overheat in traffic etc, I have articles from back then, new cars at the time so no age factors, that describe overheat issues on a hot day in traffic. Due likely to radiator design and air intake, fan, and water pump, as well as internal engine design of the coolant passages. If all were efficient, then with an 80C stat the car should run at about 80C…not lower tho…the stat would close the path to radiator. SO… we can: clean the internals, be sure the radiator is clean, try to close off air escape to the sides of the radiator, be sure the tune is correct as a poor tune creates excess heat, maybe install an electric fan in front of the radiator for slow times/ traffic stops, maybe use only water with Water Wetter product in summer as it cools better than 50 50 anti-freeze (drain refill with water,…then winterize in the fall when temps cool). and maybe the trick with the windshield washer pump hose directed to the front of the radiator. If an 86C stat is used, …then 86C to 92 C would be expected. But if a 73C stat…it may still be at 86 to 92C on a hot day. Use of just water and Water Wetter, .it may cool a few degrees better…but one must be careful of the boil point at lets say 100C. Even with a 4lb cap, that is only 3o…and if at 1000 ft above sea level…then minus 30…so 100C. With 50 50 anti freeze it is 107C. If you were to use a 73C thermostat it would open to radiator sooner…but the same conditions that cause overheat with an 86C will come into play…once the stat is fully open. It may provide a little room for short traffic stops. It would be interesting to make the change to a 73C on same car…and note how it works out. The 73C, whether Smiths, or AC TF1 or 3, or the QH QT100/72 can be found. The newer type SNG Barrett and XKs is a 71C (160F). My car has that one…71C. The ONLY time I run at 71-75 C is on a very cool day…otherwise about 75-80C and quickly to 90C in any traffic on a warm day. It will continue to climb…if I don’t find some speed and nice air soon. That is when I switch on the spray pump and auxiliary front fan. I have a WX128 water pump and the radiator was fully flushed. Nick