Hello folks, I've upgraded my TPS to a more modern hall-effect sensor and despite me still chasing other issues I thought I'd document and share what I did since it seems to function well.

You'll have to excuse the lack of pictures I was excited to get it in the car and totally forgot!

Why go with a hall-effect sensor over the OEM sensor or the adapted mustang TPS many people use? The track on the OEM module is well documented as wearing out in spots if it is held in place for longer periods of time (e.x. highway cruising), while the mustang TPS is slightly newer and much cheaper it still has a similar design flaw, outputs the wrong voltage at idle, and decently made ones are becoming harder and harder to find. The solution to this is the hall-effect sensor, it senses the magnetic field of a magnet attached to the shaft to determine the angle of the shaft very accurately and with no contact whatsoever meaning a much longer lifespan.. it also allows idle and WOT voltages to be set correctly without any hassle.

Here is a link to the one I used and can guarantee actually fits, there are lots of identical ones around that will likely work as well (google 0-5v 90 degree hall effect sensor), .step and .sldprt files have also been included so you can modify the mounting hardware for use with another sensor if need be
https://www.ebay.com/itm/225990385526?var=525092764483



Files to make the adapter can be found here or in the attachment, do NOT use standard PLA for these they will melt in the engine bay, I used PA6 nylon for mine as it's appropriate for the application and also affordable
If you get everything together and it outputs odd/negative values flip the sensor 180 degrees, I also recommend gluing the sensor shaft to the adapter so the adapter cannot pop out of the little slot on the shaft.. DON'T glue the throttle side it'll ruin your day
https://grabcad.com/library/jaguar-xjs-tps-adapter-1


If you do not own a 3d printer there are services online that will print and ship for you.. this option may also be cheaper than buying an entire roll of nylon, I'm not sure

 

Thanks! This looks like a reasonable solution. How did you set the idle voltage? Just rotate it a bit like the stock one?

I wonder how hot the engine bay gets. I don't have a filament printer, only a resin one. My current resin is good to 180C apparently. Not sure that's good enough.

 

Yea the flange that comes with the sensor allows for ~10 degrees either direction which is more than enough, just make sure you're not sitting below the lower limit of the sensor at idle

180c should be plenty high enough, thermostats on the v12 open around 80c, under-hood temperatures probably sit a little higher than that.. you can also anneal parts to supposedly make them tougher and more temperature resistant, something I've been meaning to try

 

Very cool.
I'm not too familiar with TPS's & how exactly the XJS ECU uses the signal, but is this sensor effectively an exact match then - maps across the same voltage for the same throttle opening angle of OEM sensor? Is this an industry standard thing?
Did you notice any immediate driveability changes once you had replaced it?

 

From what I understand the TPS is used to adjust the mixture immediately when you hit the gas, then the o2 sensors catch up and trim fuel more accurately a couple seconds later.

The voltage range across the original sensor appears to be linear, ~0.05v/degree, also no I have not noticed any difference aside from an issue I was having before being fixed (erratic behavior at certain throttle angles)

 

Do you have a pic of this installed?

 

Sorry for the delay getting pictures! This is after several long mountain drives so I can confidently say pa6-gf will survive the underhood temps.. after a couple heat cycles I would recommend you double check the idle voltage though and readjust if needed

 

Slight revision:

The TPS is used to adjust the mixture immediately when you hit the gas, then the baro sensor in the ECU catches up a couple seconds later.

In essence the TPS acts like an accelerator pump found in (most) old carburetors, covering the momentary fuel delivery gap when the throttle is opened.

The baro sensor is the primary mixture control device. The O2 sensors trim the mixture slightly afterwards.

Cheers
DD

 

Agreed Doug, but I cannot believe that is the whole and sole function of the TPS. If it was, why bother with the rising-rate voltage as the TPS opens? It must be telling the ECU to do something, mustn't it?

I recall this discussion has been had on the Series II saloon forum, but I seem to remember that did not come to a firm conclusion.

 

It isn't the sole function.

The TPS tells the ECU when the throttle is opening, or closing. And if the throttle is fully closed, or fully open. It's easy to understand what the ECU would do with such information.

What comes up in discussion is what does the TPS do at steady throttle? Throttle not moving, not at idle, not wide open. As though you were driving on cruise control.

My conclusion, after many discussions over the years, with input from people a lot smarter than me?

Nothing. It does nothing.

I'm not saying I can't be wrong

But what would the ECU do with a steady throttle reading that isn't already done by the baro sensor?

Cheers
DD

 

I finished this installation on Friday. There was one thing I had trouble with that is worth mentioning. While the sensor only has two possible positions on the adaptor, the adaptor has three possible positions on the capstan. I was getting frustrated with either 0V or 5V and no where in between, so I left it for a bit. When I returned to it I ignored the mounting holes and found where it wanted to be, then repositioned the adaptor to suit. I got it set at 0.34V at idle. Part of my frustration was needing three hands, one to hold each multimeter lead and one to adjust the sensor. I now have some leads with aligator clips and they're so much easier!

But anyway, this made a huge improvement to the car! The lurch when I step on the gas is gone, and the overall throttle response feels faster and more predictable. Took the car for an hour drive, about 45 miles, and it was one of the nicest drives I've been on with the car. Everyone should consider doing this.

 

Haven't had a chance to look into it but I wonder if there may be one with dual traces. I have facelift car with the 6.0L and it uses one for the transmission in addition to the engine. It has 6 wires and is no longer available like so many parts nowadays. Haven't pulled it apart to see if it might be cleaned but I know that would only be a temporary solution at best.

 

I can tell you that on the 6CU Super Enhanced ECU from AJ6 Engineering (a mid/late 90's iteration), the throttle input was connected to an op-amp hot-glued onto the board which suggest that the throttle delta was simply being amplified to improve throttle response.

~PK

 

nuts! I definitely should have included markings for mounting direction, sorry for the frustration.. glad to hear it turned out well in the end though!

 

Kitte
Could you please explain how you got the capstan spindle to engage with the TPS spindle? I am thinking of one of these as a spare.

 

There are two 3d printed parts. The first one is the adaptor bracket visible in the photos. The second is an adaptor for the shaft. It has a D slot on one end and a tab on the other. The tab fits into the slot on the sensor, the D slot fits onto the capstan. The adaptor needs to be glued to the sensor.

A couple of screenshots to illustrate:

 

I'm going to have to work on my installation today, it kinda failed yesterday. The car made it home, but it wasn't happy. I'm not yet sure if the screws worked loose or if something started melting, but I can easily move the sensor by hand. The first 60 miles were great, the second 60 it was back to lurching as I stepped on the accelerator, and dying at red lights, which was a new experience. I'll post an update once I figure out what went wrong.

 

How do those of us without a 3DP get one made?

 

There are a lot of 3D printing service companies. They all work similarly; you upload the model and choose the material, they give you a price, you place the order, they send it to you.

Shapeways is the best known, they're in the Netherlands. The cost varies depending on material and process, for the larger part the lowest cost is $12.93 for nylon(sls). I would probably go with the nylon (MJF) at $14.25.
If you really want to be fancy, they have a service where they will print in wax and then do a lost wax casting in metal. Brass would be $565, gold is $43000.

https://www.shapeways.com/materials/...et-fusion-pa12

 

Well this was entirely my fault. The cause is that the screws backed out and allowed the sensor to turn. They did this because I didn't tighten them beyond snug. The other problem is with how I 3D printed the part. OP has a filament printer, mine uses liquid resin. Mine is better for fine detail, but the parts need to be supported. I supported the larger piece in a way that resulted in one of the surfaces being quite rough. I thought it would be ok, but I was trying to tighten the sensor against a couple of high points instead of a smooth surface.

I modified the file to have threaded holes. Currently printing a model train part, when that is done I will reprint the adaptor.

You can see here the gap I was dealing with.