Congratulations on your solution. Of course a lighter spring would be the easier solution.

While you were working on it I saved up some green and installed Gus's valve. It was simple, didn't spill a drop of fluid, and took less than an hour. It was one of the less expensive fixes on our XK8.

 

This is a draft of what will become a web page describing this mod. I'll then link to it from my sig. line. I think many guys will have enough with what's here now to do the mod. Comments appreciated.

This mod involves replacing the spring in an internal control valve on the convertible top hydraulic pump so as to reduce the peak pressure in the latch circuit from 1600 to 1100 PSI.


Procedures for installing replacement spring

1) Open the petcock on the pump.

2) Carefully remove the pump from its mounting (as with any operation where the pump is to be disturbed, be very careful with the solenoid connectors!).

It will be helpful to use a box or similar as a platform to support the pump for what follows.

3) Remove and save the fluid from the pump reservoir. Siphon it, pump it out, or flip the pump on its back to drain it; whatever you find easiest. With a little care, there is no need to lose any significant amount of fluid.

4) Remove 4 fasteners (these may be hex or torx, depending on pump vintage) securing the reservoir to the pump body. Carefully pry the reservoir off of the pump, working it back and forth. Tilt the reservoir down as it separates from the pump so as to retain any fluid remaining in it.

5) Unscrew the control valve found on the lower left of the pump body (see photo). The valve body may be hex or cylindrical in cross-section. Use a vice-grip to grab hold of it if it's cylindrical.

Working with the removed valve ...

6) To free the adjuster screw, drill out the locking dimple on the side of the valve body, just far enough to free the screw. You can feel when it frees up by gently applying torque to it with a screwdriver. Remove the screw and the other valve components (see photo).

7) Substitute the replacement spring described below for the stock spring found in the valve.

8) Reassemble the valve. Find the adjuster screw position where the valve just closes, then turn the adjuster screw clockwise an additional 4 flats. (The easiest way to find the "just closes" point is to attach a snug-fitting section of plastic tubing to the valve's external threads, and submerge the adjuster-screw end of the valve in a fluid. Watch for bubbles in the fluid as you try to blow air into the tubing. Where the bubbles just stop, that's the point where the valve just barely closes.)

9) If desired, you can lock the adjuster screw by using a punch to scar a spot on the edge of the screw where it meets the valve body threads (it's soft metal). I don't think the screw will try to move, and so don't believe this step is really necessary.

10) Re-install everything in reverse order, steps 1) to 5) above.

11) Check fluid level. Test convertible top operation before returning the pump to its mounting.


Source for replacement spring

This mod requires a replacement spring for the pump's latch-circuit control valve. A suitable spring was located at Century Spring Corp. Their part number is S-3205. These springs sell for $1.36, probably about $1.00 in quantity. (OK, that's not zero cost, but pretty close, right?)

Century Spring Corp. is not set up for "quantity one" orders, so I'll be buying a supply of springs to keep on hand for anyone who might want one. Send a PM please if interested.


FYI ... Selection of replacement spring and pre-load setting

For simplicity, everything here is expressed in round numbers.

The first of the two graphs below shows the behavior of the stock control valve. Two cases: 85 Bar (1200 PSI) is used in the ram circuit, and 110 Bar (1600 PSI) is used in the latch circuit. In both cases the valve begins to open ("cracks") at a pressure set by the adjuster screw, and is fully open at a pressure about 1000 PSI higher. This 1000 PSI spread is governed by the stiffness of the spring, and is the same for both the 85 and 110 Bar settings.

The cracking pressures were measured through a calculation that uses: the adjuster screw settings (0.5 flats for 85 Bar, 2.0 flats for 100 Bar), adjuster screw thread (24 TPI), spring constant (425 lbs./inch), and the size of the hole behind the valve ball (.01 sq inch). Crunch, crunch, crunch ... 200 PSI for rams and 600 PSI for latch. If this method seems like a stretch, consider that a mechanical pressure gauge in effect uses exactly the same sort of pressure-to-force-to-movement calculation to work correctly ... to position its indicator at the correct point on the face of the gauge.

We want to reduce the peak pressure from 1600 to, say, 1100 PSI. By experiment it was found that the convertible top will not work if we just change the adjuster screw on a stock valve to do this. The cracking pressure is then too low. So we need a valve that retains 600 PSI cracking pressure but also gives 1100 PSI peak pressure, as represented by the red line in the 2nd graph below. This requires fitting a new spring.

Selecting replacement spring: For the valve to crack at 600 and be fully open at 1100 PSI the slope of the red line in the graph must be 1/2 the stock value (600-to-1100 PSI v. stock 600-to-1600 PSI). This in turn calls for a spring with a spring constant of 1/2 the stock value. (The chosen spring is actually 225 v. stock 425 lbs./in.) The new spring also has to meet size constraints to fit inside the valve body, and other specs for closed/ground ends, load handling, max deflection, etc. Springs can get complicated!

Replacement pre-load setting: Since the cracking pressure is to stay at 600 psi, the pre-load force on the spring must also stay the same. But since the new spring is 1/2 as stiff as stock, the number of turns on the adjuster screw must be twice that used with the stock spring. (4 flats v. 2 flats)

That is how the replacement spring and pre-load setting were chosen.

 

This mod of course calls for changing out the spring in the pump's internal control (relief) valve.

A shipment of springs is on the way.

If you expressed an interest earlier by PM (or even if you didn't) and you want a spring, please send me a new PM with your mailing info.
$2.00 each will keep me whole. I'll provide PayPal info by PM, but please don't use it until you have the spring in hand. (If you don't have PayPal, don't worry about it; just give the $2.00 to a cause of your choosing.)

I'll post an expanded set of installation instructions shortly.

 

Just to sort of wrap things up ...

If you have a spring coming, please take a look at the final(?) installation instructions here (also linked in my sig. line).

Error correction: In post #83, Jandreu suggested that fully compressing a spring might change its characteristics, and I replied that I thought it would not. Wrong. It turns out many springs (including the one we're using) come with a "maximum compression" spec, beyond which we should not go. A typical value is about 1/2 of the difference between the free length and the fully-compressed length. Springs get complicated!

All done, just in time for winter!

 

Although my '02 XKR has not had a hydraulic leak as yet, as insurance I obtained an alternative spring for the 110 bar pressure valve from Dennis 07 and installed it as instructed. Initially the top would stall about half way up. I had my wife operate the top while I observed the hydraulic reservoir and could clearly see a stream of fluid coming from the 110 bar (modified) valve. Another "half flat" clockwise turn resolved the problem with the top operation returning to normal. As an additional layer of insurance, a 100W resistor was installed on the power line. The top has continued to function with the addition of the resistor although slightly slower.

 

I have been lucky so far to not have any problems with my convertible top. I hope that I didn't just jinx myself,
Anyway, I've been reading this and based on decades of working on hydraulic systems of commercial aircraft. we have two options.
One is what everyone is talking about, controlling the pump output pressure.
Another option would be to control, or lower, the system pressure. Boeing would have two hydraulic pumps feeding a pressure module that incorporated a relief valve to control system pressure as a backup to the pump's internal control. We could do the same thing with our pumps. The pump would do what it was designed to do, produce 1600 psi. But when the pressure gets out of the pump, some would go to the system and some would be directed back to the pump/reservoir. The hoses to the top would never see 1600 psi because the relief valve would only allow a lower pressure to the system.
This would not involve doing anything to the pump. Instead, install a tee on the pump pressure line (output), and another tee on the return line. Connect the tees together with a relief valve in the line. It should be easier to set up the hardware because we only have to match the threads where we tap into the original hoses. The rest is whatever it needs to be.
Thoughts? Is there any reason why this wouldn't physically work?

 

Would work, something like this already,
JagRepair.com - Jaguar Repair Information Resource

The right accumulator would also 'work' to reduce peak pressure, but it's not clear that spreading 'pressure impulse' (the overall pressure x time, or integrated pressure over time) would be helpful compared with the relief valve.

The genus of the spring solution and the relief valve solution are that the allows the system to still operate (with lower margin than before) while demonstrably reducing both the peak and the impulse. Should be helpful, at least until the sheathing under the crimps turns from the current crackly bits into dust.

 

Gents,

Seems to me what Stu46h proposed would be doable. As crbass said, not so different from external relief valve designs used to date. (Tee fittings are not required because there are ports in the pump body available for connecting a relief valve).

Just a comment or two:

This pump is of the positive-displacement type, and such pumps ideally deliver a fixed volume of liquid with each revolution, independent of pressure. Without internal control valve(s), such pumps can't be said to have been designed to produce a particular peak pressure (there is some amount of pressure at which the motor will stall or things will break, but that's another topic entirely). In this application, the designers incorporated internal control valves set at 1600 (latch circuit) and 1200 (ram circuit) psi. Early on, nobody here knew of these valves and so -- max pressure having been measured at 1600 psi -- we came to think of it as a 1600 psi pump, requiring external components to reduce the pressure. But it ain't so.

Later I discovered that the observed peak of 1600 psi derives from one simple valve setting inside the pump. That being so I came to see that just re-setting that valve to the desired pressure is an attractive alternative to doing something "downstream" with other components, which can cost hundreds of dollars.


Concerning pulse "shape" (on a pressure v. time graph) ... I believe that the shape of the pressure pulses does matter as one factor contributing to hose failures. Without getting all down in the weeds ... there is much of interest here. Test equipment for hoses such as ours provides for programming the details of the pulse waveform (shape), not just setting peak pressure, temperature, etc. I spoke about this to a rep at Haidar, Inc. years ago and was quickly in over my head, but came away convinced that pulse shape was part of our hose failure story.

 

A few years ago, I benefited enormously from the dogged pursuit of the solution of The Green Shower Problem by Dennis07 and the other contributors. I installed Dennis' spring to the latch PR valve, long after this thread was started, and it has been working fine.
May I again say: many thanks to all, especially to Dennis for his heroic efforts.
The goal was solely to reduce the latch system pressure to the minimum reliable level to avoid hose failures at the latch piston. None of the work would affect the operation of the top itself, as far as I can see.
The problem I have now is that the top is reluctant to disengage from the windshield frame, and will not open far enough to cause a operation of the conv/top 'fully open' microswitch. As a result, the pump motor will keep running for as long as the button is held, and the windshield latch claw will not retract. A message "Convertible top not latched" message is then permanently displayed.
The fluid level in the reservoir is above the minimum mark, and there are no indications of a leak, nor evidence of air bubbles in the reservoir.
My wife pressing down on the R-side of the (almost) folded down top, as I hold the 'open top' switch, causes the ram microswitch to operate, stops the pump, and retracts the latch claw.
In the closing direction, the top has to be helped (slight forward pressure via the finger slot to make reach / contact with the claw, after which it latches properly. This all suggests that there is not enough force being generated to obtain full ram travel.
The question is then what to do about it. Should the ram pressure be slightly raised, by tweaking the ram pressure screw slightly CW, or am I overlooking something more obvious? What are the concerns about slightly increasing the ram pressure, if any.
I would much appreciate any suggestions.

 

I had the same issue when the spring was changed. A slight (<1/8 turn CW) turn of the screw solved the problem.

 

Thanks for the reply: Just to be clear, only the latch hydraulic pressure relief had its spring changed, per Dennis07's recommendations. Nothing was done to the hydraulics that operate the rams that raise / lower the soft top. Did you experience a raise /lower problem after changing the relief spring in the latch relief valve?

 

Thanks for the kind words in post #109 below. It turns out that there are times during the raise / lower top cycles when both the ram and latch control valves are open. The upshot of this is that changing the pressure setting of one control valve can affect the operation of the other circuit. For example, lowering the pressure setting of the latch valve might cause the ram circuit to be unable to raise the top.
Bottom line: I'd recommend not making any changes to the ram valve. In my experience, operating only on the latch valve can resolve any troubles like the one you're experiencing.
(I'll respond to your PM as soon as I can; probably this evening.)