XJ-S Coolant Recovery Reservoir Retrofit

Edward F. Sowell

The Problem

Although later XJ-S cars are fitted with coolant recovery reservoirs, earlier models (such as my 1976) are not. I donít know which models have them, but I have been told by a Jaguar parts person that they have come and gone a few times between 1976 and the present. I donít know if this is true, but I know for a fact that my 1976 (actually December 1975) car does not have one, and that Ď87- í88 models do. The easiest way to find out if you car has one is to call a Jaguar parts store and ask them to look in the parts manual. Or, you can look to see where the hose connected to the filler cap neck goes. On my car it pokes right through the engine compartment wall and dribbles the overflow out above, and in front of, the left front wheel. If you have a recovery system, it will instead go into a left front wheel well cavity. Apparently, on some cars it is in front of the left front wheel, while on others it is behind it.

One consequence of not having a recovery system is that the coolant header tank cap has to be removed to check or add coolant. This is an inconvenience in that the engine must be cold just to even check the level. Another disadvantage of this design is the difficult in judging if more coolant is needed. The problem here is that if you fill to the top of the neck, as instructed in the Repair Operations Manual (ROM), the first time the engine heats up the expansion volume of coolant, perhaps a liter or so, will be released onto the street. Then, upon cooling down again, the level in the header tank will drop. Therefore if you check your coolant the day after you topped it up you will find itís down again! Of course, if you have a good eye you may be able to judge where you should fill it to in order to avoid loss, but I, for one, have never been very good at that. As a consequence, I probably have run low on coolant more often than Iíd like to admit to.

One other problem should also be noted. If you look at the coolant temperature sensors (the one for the dash gauge as well as the one providing the ECU signal) seem to be at levels somewhere near the top of the header tank. Consequently, if the coolant level is below the neck, as it surely will be without a recovery reservoir, these sensors may well not be immersed when the engine is cold and at rest. I observed this to be the case when I changed my sensors and no coolant gushed out. Of course this will probably change very quickly once the water pump spins up, but in the meanwhile this could give a misleading signal to the ECU at start-up.

I also want to mention that even the cars that have a factory installed recovery system have a disadvantage that is very irritating to someó its location under the fender. This means you canít see how much is in it, to start with. I, for one, have gotten used to looking under the hood of my "ordinary" car and quickly seeing the coolant level. Without being able to see it, you are stuck with opening the cooling system proper to check the level. I address this issue in my retrofit. Thus this retrofit may be of interest to even to those of you with factory recovery systems.


There are many after-market coolant recovery reservoirs available. The problem, however, is finding a place to put it in our crowded engine compartments. One solution, applied by at least one XJ-S Lovers list member, is to move the windscreen washer bottle a little one way or another to allow a small coolant recovery reservoir, perhaps fashioned form a plastic bottle, to be placed beside it. However, the volume you can accommodate in that location appears to be about one liter, not quite enough according to Mike Morrin who took this approach.

As noted above, the space problem is resolved in the factory installed systems by use of the cavities in front of, or behind, the left front wheel. You get to these spaces by removal of the wheel and undoing 5-6 sheet metal screws or plastic body fasteners holding the closure panel in place. The space is large enough to easily accommodate a 2-3 liter, properly shaped reservoir.

To do a retrofit on a car without one, you can use a reservoir from a later model XJ-S, or get an after-market unit. My first thought was to use a Jaguar system, since I assumed this would be pretty much of a drop-in installation, and if purchased at a junkyard would not be too expensive. Well, this turned out not to be the case, at least for the unit my Jaguar parts house located for me among their used parts sources. They wanted $50 for it, and it turned out to be incomplete. I would have had to order several new parts, such as the hold-down strap and the cap with hose and feed tube that drops down into the reservoir. Moreover, it happened to be off of a model for which it was installed in front of the left front wheel. On my car, thatís where the charcoal canister is, so I could not put the coolant recovery reservoir there. While I might have been able to mount it behind the wheel, it clearly was not a straightforward task. Or, said another way, certainly no easier than installing the $15 after-market unit from Parts Plus!


So letís cut to the chase. What does it look like when installed? This is shown in Figure 1, which shows the view looking into the cavity behind the left front wheel, panel removed.

Figure 1. Reservoir in cavity behind left front wheel.


You can see that I chose to hang the reservoir from the underside of the body structural member at the top of the cavity. This is the obvious thing to do, given that the mounting rack has two tabs for mounting screws at the top and none at the bottom. The instructions pointed out that you can bend these tabs around as needed to attach to either a vertical or horizontal body surface, but I chose to leave the mounting tabs vertical, attaching to two angle brackets screwed to the body structural member. The reason was to drop the reservoir down a bit, as mounting it too high can cause filling problems (see below). In retrospect, it might have been a little better to hang it even a bit lower in order get the filling tube going all down hill (more on that later.), but that would have required mounting hardware more complicated than the hardware store angle brackets I used.

It is also imperative to mount the unit as far back as possible, so as to clear the cavity closure panel when it is reinstalled. The limit on how far back you can get it is probably determined by your hand drill and your dexterity with a screwdriver. The first set of mounting holes I drilled was too close, and I couldnít get the panel back on. The second set was as far back as I could reach with my drill, and fortunately it worked. By the way, assembly is easier if you attach the angle brackets to the reservoir-mounting frame, then screw the angle brackets to the car body. Hint: make a cardboard template to match up with the mounting brackets and us it to locate the holes to drill in the body.


I spent most of the time in this project worrying with the connections to the tank. As received, the tank has two ports, each a 5/16" diameter nipple molded into the tank for attaching vinyl tubing. The one that is supposed to come from the radiator (the feed port) is at the bottom and front with respect to the way I have it mounted. As you can see, I have blocked off this port. It is just too close to the cavity closure panel. Any attempt to attach a tube to this port will result in a kink when you reinstall the panel. In lieu of the provided entrance port I use an elbow hose barb fitting that passes through the lid of the reservoir, with a tight fitting grommet. Note that if you donít use an elbow here the tube will kink when it gets hot coolant. More on hose barb fittings later.

One of the clear vinyl tubes you can see in the picture is the feed line from the header tank neck. It penetrates the engine compartment wall and enters the cavity just forward of where the wiring harness comes through. I drilled the hole for clearance around a 90-degree hose barb elbow. It is held in place by the hose connecting it to the reservoir. Originally, I tried this without an elbow, using a continuous tube from the header tank neck and passing through the harness hole. Kinking of the tube, a major frustration in the project, nixed that idea. In summary, the feed line has a tube from the header tank neck to the leg of the elbow inside the engine compartment, the other leg of which fits into a hose connecting to the elbow at the lid to the reservoir, and the final segment to reach to the bottom of the reservoir inside. All are single-barb joints secured with clamps because these joints must be absolutely air/water tight or the recovery will not work.

The other provided port on the reservoir is for over flow. It is at the top and back so canít be seen in the picture. I do use this port. In the photo, you can barely see the overflow tube dropping down from behind the reservoir a little to the left of the bottom fitting. It passes under the reservoir and goes to the closure panel. I drilled a hole in the bottom, stepped edge of the closure panel so a short length of tube could stick out and drain to the street. It is clamped down to the panel so it canít be accidentally yanked out or pushed back up into the cavity.

Of all the connections, the one that caused me the most grief was the least essential one, that is, the one for viewing coolant level. Obviously, you canít see the reservoir due to its mounting location. What I did to address this was to devise a "sight glass," such as was (or is?) used on old-fashioned boilers. This is a section of clear vinyl tubing attached to the outside of the closure panel, Figure 2. (The white you see in the photo is plumberís putty, sealing the closure panel to the body.)

Figure 2. Sight glass

Its length and vertical position correspond as close as possible to the reservoir, as the coolant level in this tube will be exactly the level in the reservoir. It is secured to the panel by a hose tie-down clamp. At each end is a hose barb elbow that penetrates the panel. As you can see in Figure 1, there is a fitting attached to the bottom of the reservoir, which goes to the elbow at the bottom of the sight glass. The top elbow is open to atmospheric pressure. Obviously, it canít be sealed or the pressure above the liquid would compromise the accuracy. Not evident in the photos is a loop in the hose connected to the top elbow before it terminates in the cavity; it is merely to discourage any debris that might be in the cavity from getting into the sight glass. The important thing is that by looking under the wheel well (an action surely to raise curiosity at the filling station!) you can check your coolant level.

The main difficulty with the sight glass idea was first trying to use the bottom front reservoir port, originally intended for the feed tube, to connect to the bottom of it. I donít know why I thought it would work after I had already rejected it as a feed port, but I did. Perhaps it was simply that I could not find fittings necessary to do it any other way. I assembled it this way, but after driving a week or so I realized it was not registering properly, which could only be caused by a kink induced by the closure panel. I then ripped it all apart and renewed my search for fittings to do it properly.

Note in Figure 1 that the original feed port is blocked with a brass screw. The hole happened to be about the right tap drill size for an #8-32 thread, so I tapped it and inserted the brass screw, coated with silicone glue. Hope it holds!

The final bit of plumbing is a filling system. It is a vinyl tube passing from the engine compartment to the fender cavity through the harness hole. It goes to a fitting I installed in the top of the reservoir. To fill it, I insert a small funnel into the end of the tube, Figure 3.

Figure 3. Filling

This design has one problem I should mention. As you can see, due to the height of the reservoir and fitting the filling tube is not "downhill" all the way. Consequently, to fill the reservoir the tube and funnel have to be held high. After filling, I also blow on it a bit (after cleaning the end!) to send all of the coolant on into the reservoir. Another mistake to report here: originally, I tried to do the filling with a tube to the top end of the sight glass. This was just too slow. Moreover, it meant that the end of the filling tube in the engine compartment had to be left open in order to get an accurate fill reading, which did not appeal to me. The separate filling tube is a much better idea.


I chose to go with the after-market reservoir from 4-Plus Auto Parts. This unit is called Radi-Acc, model LBI and is made by ASV Plastics in Elkhart, IN. It is a 2.5-quart (2.37-liter) tank made from high-density polyethylene. Although I wound up using nothing but the tank and the mounting rack, it came with hoses, clamps, and even a new radiator cap (13 psi, so useless for me!).

The only problem with the tank is the location of the ports for hose connections, which were not just right for the installation I settled on (see below). Consequently, I needed some special fittings. Although my local hardware had some plastic hose barb fittings, the selection was extremely limited (Ace ainít the place!). In particular, they had ľ" and 3/8" elbows, but no 5/16". I visited several others, including a major industrial hardware store in my area with essentially the same discouraging results. And fittings to suitable for attaching a hose barb to a thin-walled plastic tank are simply not available at any of these sources.

After much searching, I found some good Web sources of such fittings. First, I found Eldon James Corp. http://www.eldonjames.com/. They have very nice hose barb elbows of various plastics, using the style that has a single barb and land for a hose clamp. Their literature promotes this style as far superior to multi-barb fittings normally used without clamps. They are a manufacturer, but were happy to send me anything I asked for as samples without charge. I used their elbows throughout my project.

With the elbow problem resolved, I needed something to attach the fill tube and the sight glass tube to the tank wall. In this regard, I should point out that the tank wall is too thin for simply screwing in a pipe-threaded fitting. Moreover, you cannot reliably glue to the high-density polyethylene, a fact confirmed both by the online "GlueGuru" and ASV Plastics. At this point I realized I needed a "bulkhead" fitting that I could pass through the wall and secure with rubber seal and lock nuts. Unfortunately, the Eldon James line did not have fittings of this type that I could adapt to hose barbs. More searching led me to McMaster-Carr (http://www.mcmaster.com/). Their online catalog is immense, so I sent them an e-mail describing my needs functionally rather in terms of specific fittings. They replied in a day or so with exact page and part numbers, and 3 days later I had my fittings. Excellent service! They recommended their 51055K4 bulkhead connector for connecting to the tank wall. This fitting has a 5/16" female self locking port at both ends. To attach to the hose, they recommended their 51055K246 barbed elbow which has one smooth leg that fits into the bulkhead connector--- just pushes in and locks, water tight! The other leg is designed to insert into the 5/16" vinyl tubing. Unfortunately, these are multi-barb style fittings, but since there is no pressure to speak of I decided it was good enough. No leaks were apparent.

How does it Work?

At the time of this writing I have been driving the car for two weeks. Everything seems to work perfectly. I filled the tank so that when cold the level in the sight glass is at the top of the clamp that holds it in place, a little more than half of the reservoir capacity, and an easy mark. When it gets to normal operating conditions it is up to about ĺ of the capacity. Since I have not driven it on a really hot day yet. I donít know what it will be under the most extreme conditions. But if it overflows a bit, no problem.

There is one minor, possible problem. I notice that after cooldown there is an air pocket in the feed line. Apparently, it is drawing a bit of air in when it is sucking the coolant back into the radiator. Perhaps I did not get a perfect seal at the various clamped joints in the feed line. Iíll keep an eye on it.