Replacing the Fuel Gauge Sender Float

I've refitted the sender. it's the thing in the tank that tells you how much fuel you have left in the tank. the cork that makes up my float was cracked and I believed would become saturated and heavy if fitted as it was. So I fitted a new float.

The first thing to say....is that you can buy replacement modern fuel gauges with a modern, sealed float.

Modern reproduction fuel level sender for a DS

However I've read very mixed reviews of these. Some of which point to poor build quality and others that point to accuracy - though that could be as much to do with how they are set up when fitted. No. that wasn't for me. I intended to re-use my original jaeger-branded sender. I just had to find a suitable float.

I've seen some fuel gauges that were badly corroded - outside and in. I'd taken a look inside my Jaeger to make sure it didn't look like this.....

No. it looked to be in good condition. I gave the wound wire and moving contact arms a gentle clean. They are both quite delicate.

The float on the original jaegers is a rectangular box with a flat bottom. The arm hangs down in the tank and the cork floats on the fuel. As the fuel level lowers, the float lowers, altering the electrical signal back to the gauge on your dashboard.

When I removed my fuel tank I stuck an endoscope camera inside and did some tests to see how the arm and float of the sender behaved when fitted. Videos like this....

VIDEO: fuel gauge sender arm inside fuel tank

As you can see, when the arm is hanging at its lowest position ('empty' on the gauge) it REALLY is at the bottom of the tank. So 'empty' really does mean 'empty'. And I though the flat base of the float was probably significant to giving an accurate fuel level reading too.

My float was cracked and some of its shellac coating was coming off. I thought abut repainting with shellac but had read that it's not resistant to the ethanol in modern fuels. I thought about coating it with epoxy - but thought that would add weight to the float and alter its buoyancy. So I was on the lookout for a float solution.....

I'd seen plain, uncoated cork used.

(Credit: Jarno Muona)

I'd seen people replace the float with bits offcuts of closed cell foam - and that seemed to work.

Float made from closed cell foam

The Jaeger float is held on the arm only because of a wobbly section at the end and so a piece of foam is perhaps a good substitute if you're keeping the original arm.

And I'd seen some carefully made metal floats.

Homemade metal float (credit: Buttercup Bob)

Several years ago someone on Facebook found a cheap float on eBay and used the float off that. 

Modern plastic float (credit: Simon Broadhurst)

I bought one as well and squirrelled it away. More recently, I thought it wouldn't hurt to see if I could find a different sender float of about the same shape and size as the old Jaeger - with a flat bottom. And I found one! The float was about the right length.

And it was about the right depth. And used a slot in the end to stabilise the float - just like the cork original.

I thought about trying to bend and re-shape the long arm that came with it for use on my Jaeger, but found it just too thick to pass through the hole in the sender.

So I also bought some more sprung steel wire - the same size as my Jaeger arm - to make a new arm. Removing the Jaeger arm and cork was just a matter of opening up a couple of crimps. They were all that held it on.

I was careful to engineer the new arm so that - when fitted to the Jaeger, it hung down the exact same amount as the old cork float. 

Setting the arm length and height

I was really pleased with the result. But I'd forgotten one tiny detail..... the float was too big to fit through the hole in the fuel tank! It was too wide! So, ultimately, I've had to resort to using the 'barrel' float that I'd bought previously. 

Again I was careful to set the arm length and height so that 'empty' really does mean 'empty'.

I've refitted it in the tank now. New seal and a good earth contact.

The sender will end up sealed in the 'time capsule' under the seat, but I suspect I will still keep my eye out for a suitable modern sender unit - but with a smaller float.

Painting and Refitting The Fuel Tank

The fuel tank is back in and looks a treat.

Fuel tank refitted

Lifting the covering lid to reveal the tank had been like opening up a time capsule. 

Opening the lid over the tank was like opening a time capsule

And if you read my post HERE, you'll know that by some miracle I found my tank seems to be in remarkable condition internally.

My petrol tank is (currently) shiny and rust free inside

Because of this, I sold the new replacement tank I'd already bought and was thinking of just  shoving the old one back in and closing it away again.

Closer examination showed a thick layer of dust on the flat surface by the filler neck. It was REALLY thick...I assume that since my car had sat in a barn for 14+ years without rear wings, it had blown in? it really was VERY thick....Anyway, dampness had lifted the paint under the dust and the surface was just starting to rust. As was the top face of the seam around of the tank at that point. 

THICK dust

These things being so, I took advantage of a hot spell to get the stripped and painted. And besides, I'd already painted the area the tank sat in.

Painting The Fuel Tank

My tank has been empty and off the car for a number of years. 

I blasted it through with compressed air to remove any lingering vapour - though I doubted there was any. I removed and kept the non-matching red and brown vinyl strips that had sat under the tank straps. They looked so 'of their time'. I was definitely putting those back on.

Protective strip under the tank strap

Removing the paint with a wire wheel on a drill was easy - but very noisy. After a de-grease with panel wipe, I applied rust killer all over - but particularly to the area over the filler neck and around the seam, and I primed in the seam between top and bottom halves with a rust curing product.

Rust-killer applied

The tank came up very well.

Prep-ed and ready for painting

The tank is big and awkward. And heavy. I rigged up some hooks on the patio to hang the tank from the pergola. I masked off the openings and screw threads and applied a primer coat of two part epoxy. This comes in rattle can form and I've used it a lot as a base coat for black parts.

It's not at all cheap, but adheres exceptionally well and gives great rust protection. Since i was bothering, I also painted the dust shields over the rear suspension pipework. I had *just* enough for one good coat on everything. 

I used an epoxy primer on the tank

I allowed a day for the epoxy to cure and then broke out the compressor and spray gun. For top coats I used Eastwoods 'Underwood Black'. No - I'm not sponsored by Eastwood! As far as I'm concerned, it gives a lovely mid sheen finish - more than satin, but less than gloss. Just like old black parts you find under the hood....

I already had a plan for how and where I'd leave the tank to dry.

Note how I had to prop the engine crane up with a stick! See my post HERE.

I'd already cleaned and glued back-in the rubber strips that sit under the tank to stop it rattling around.

Before: rubber strips under the tank

Fitting the Fuel Tank

Fitting really is the reverse of removal. I covered that HERE. Now it came to putting it back, it took me a while to work out where the little foam rubber spacers on the tank seam would best sit. It partly depends where the little metal securing tabs are around the tank hole Once I'd decided what needed to go where, I glued them on - to stop them being knocked off when I was fitting the tank. I carefully located the filler neck in its hole and then lowered the tank in place.

Perhaps the trickiest job at that point was  squeezing the tank short sides past the bent metal tabs without catching them. The side-to-side fit of the tank seemed to be particularly tight. I used a wide piece of thin metal to cover the tabs and then slid the tank over that - then removed the metal.

Not a lot of room down the tank sides

Now that I had sight of where the metal fixing straps would go, I slapped some contact adhesive around on the top of the tank - with the same care-free abandon evident when I removed the tank - and refitted the red and brown vinyl strips that would sit under the tank straps. And of course then the straps themselves were fitted. That tank wasn't going anywhere.

Fitting The Pick Up and Breather Pipes

It's easier to refit the fuel breather from the rear wheel arch with the grommets removed from the chassis. Mine had been removed for painting anyway. I first slotted one of the grommets onto the breather pipe and fed it  through the two holes in the right hand wheel arch chassis near the fuel filler tunnel. 

I fitted the grommet on the wheel arch side. In the engine bay I slipped on the other grommet and fitted that. Done!

The nylon pick up hose running under the outer sill is stiff and prone to fold in on itself if forced to curve too tightly.  There isn't a lot of 'bending room' between where the hose passes through the sill and the eye through the bottom of the 'B' pillar.

Pick up hose passes through an eye in the B pillar

For that reason, in the sill I removed most of the clips that held the nylon fuel hose in place on both sides of the B pillar and gave myself extra scope for feeding and bending the hose for the next task - fitting the sealing sleeve over the hose....

The fuel pick up hose is a snug fit in a curved tube in the tank - but it has an air gap around it. That gap needs to be closed to prevent fuel vapour escaping and I think to also aid the suction from the fuel pump. Citroen fitted a black rubber sealing sleeve over the pick up hose. 

The sealing sleeve

At the tank end, the sealing sleeve goes over the end of the curved metal tube that the hose passes through. The other end of the sleeve ends at a point in the outer sill. The original Citroen sleeve was so tight on the hose, that there are no sealing clips at either end. The tight fit is sufficient to seal any air gaps. When you do as I did and replace your fuel hose, it's near impossible to refit that original sleeve. For this reason, simple kits are sold that are a far easier fit over the fuel hose. These are a looser fit over the fuel line and so rely on jubilee clips on both ends to provide an airtight seal.

Sealing kit

I had one of these kits but first tried - and succeeded - to refit the Citroen original. (I did say 'near impossible'....). Why I bothered though, I'm not really sure! Bloddy-mindedness I suspect.....

Unlike the fuel breather hose - where the two holes in the chassis it passes through are in line - the two holes through the sill for the pick up hose are offset through the sill sides......It was VERY fiddly trying to the fuel hose and sleeve through the two grommets in the sill.

Pick up fuel hose passing through the sill

But it was still marginally easier than trying to fit the grommets afterwards. Yes. I tried it both ways. Once the fuel pick up was in the fuel tank area I fed it through the curved metal tube until it came out of the filter plug hole in the floor of the car. 

Fuel pick up emerging from where the filter plug will be fitted

But I then had to re-adjust (several times) the position of the sealing sleeve on the hose/ through the sill to both take up that hose slack in the sill and to still fit over the metal spigot on the tank. Just to ensure their sleeve was air tight I added a Ligarex clip at both ends of the sleeve and refitted the hose clips on the sill. 

I fitted a new short, curved hose that fits the breather to the tank. That hose has Ligarex clips on it, and there is a little brass bush inside the breather hose so that tightening the Ligarex band around the curved hose doesn't crush the breather hose. And that was that.

Before: curved breather hose and pick up hose on the tank

After: note the angle of the pick up hose through the sill

I want to investigate changing the cracked cork float on the fuel sender for something more likely to float, so am not fitting that yet.

Setting Ignition Timing On Cars Manufactured After June 1971

On an internet forum there has been a very long and protracted discussion about helping a fellow DS owner trouble shoot problems setting up the timing on his car. He happened to be fitting 123 electronic ignition, but the big debate wasn't really 123 specific. It concerned the underlying instructions for 'pre-setting the static timing' on Ds manufactured after July 1971.

In Citroen's 'manual 814' and it's French equivalent 583, the instructions for setting the timing are essentially these:

• Use the notch in the flywheel to find the static timing position
• Mark this as 'position zero' on your timing gauge.
• Set the distributor up to cause an ignition spark at that point
• Run the engine
• Set/ adjust ignition advance (move the ignition point) for the particular engine.

In essence, Citroen provide these same instructions for cars built before July 1971 and for cars built after July 1971. What's the significance of that date? 

• For Ds built before July 1971, the notch in the flywheel  - used to pre-set the static timing - gave 12 degrees before top dead centre for number one cylinder - 12BTDC. This is established and proven to be an adequate static timing point for completing the rest of the timing set-up process
• After July 1971 the piston of the notch was moved. it now gave top dead centre for number one cylinder - TDC. 

Notch in the flywheel

Regardless, the Citroen instructions for ignition timing are essentially the same. 

Citroen timing instructions for pre-July 1971 engines and.......

......Citroen timing instructions for post-July 1971 engines.

In both cases the instruction is to use the point given by the flywheel notch, as the point at which to set up the initial distributor ignition - the static timing.

The upshot of this is that for post July 1971 cars, because you start at TDC, when you set the static timing, initial spark ignition would take place slightly later - when the engine was at TDC. And combustion would occur when the piston was well into its down stroke. And when setting the advance, you would expect to need to move the distributor and ignition point more (degrees) to reach the same ultimate timing position than you would for pre July 1971 cars.

Well that's theory.....even though the intended process of setting the advance would ultimately leave the engine with the correct degree of timing, the problem and debate on the forum concerned practice: whether or not beginning the process at TDC is a viable start point. The recent debate on the forum concerned the wisdom of running an engine with ignition happening this late, whether it might cause engine damage and whether the engine would run at all. If you can't get the engine going, you can't set the advance......

The general consensus was that - for Ds built after July 1971 - following the Citroen workshop manual and starting the timing setting process from a 'static timing position' of TDC was simply not viable. Citroen were lambasted, and  - for the DS owner being helped - the 123 manufacturers were pilloried for not correcting Citroen's mistake.

Worried and confused? So how can you tell what flywheel is in your car? How do you set the timing on a DS manufactured after July 1971?

Worried

Don't be.

Confused

Read on.

Does Your Flywheel Notch Give 12 degrees TDC?

If you are not sure what flywheel your car has - or simply want to check - you can go about it this way:

• Remove the spark plugs and slowly crank the engine (or if a five speed, jack up a front wheel and turn that).
• Identify the beginning of the compression stroke of number 1 piston using a balloon or similar.

Balloon deflated - engine not at compression stroke

Party time! Compression stroke found!

• Stop just as the balloon begins to inflate. Remove the balloon and slip a long dowel or screwdriver into cylinder number 1 as a 'tell'.

Using a marked dowel as a 'tell' to find TDC
(The white cap is only there to centre the dowel)

• Insert a 6mm pin (a drill bit is ideal) into the timing hole under the alternator.
• Continue to slowly turn the engine until the drill drops into the notch in the flywheel and 'locks' the engine. (You should also have noticed the dowel/ screwdriver rising as the piston continues up on its compression stroke - until the point the flywheel locked.)
• REMOVE THE DRILL.
• This is the crucial bit: continue to turn the engine slowly and, at the same time, observe the dowel/ screwdriver sticking out of the top of cylinder one.

Case 1: If the dowel/ screwdriver continues to rise for a while before beginning to descend, then the point at which the drill engaged with the flywheel was BEFORE top dead centre (aka 'TDC'). For carburettor DS cars built before July/ August 1971, the expectation is that you will find 'case 1'. On these cars the point at which the drill drops into the hole beneath the alternator (the timing point) gives 12 degrees before top dead centre (the firing point). The timing point and the firing point are the same.

For this 'case 1' scenario, I covered the process for setting up the static timing with a mechanical distributor HERE and HERE, and for fitting a 123 electronic distributor - including setting the advance - HERE.

Case 2: If the dowel/ screwdriver immediately begins to descend, then the piston is also descending and the point at which the drill engaged with the flywheel was already AT top dead centre. 

From July/ August 1971 onwards, the expectation is that you will find your car/ engine meets 'case 2' above. On these cars the point at which the drill drops into the hole beneath the alternator (the timing point) gives zero degrees before top dead centre - TDC. So how do you set the timing??

How To Set The Ignition Timing Where the Static Timing Position Gives TDC

The general consensus was that a static timing position of TDC cannot be used to set the static timing. The position of the crank and piston needed to be altered before the static timing could be set. In essence a slightly earlier static timing point needed to be found and used. Some people say the engine needs to be "wound back" but in reality this means "winding it forward" again.

This would seem to be a good way to go about things......

• As above, find the intended statistic timing position using a 6mm pin. This needs to be using number one cylinder and when the piston is on its compression stroke.
• Slowly rotate the engine to cause the 6mm pin to drop into the timing slot of the flywheel. (For these engines/ flywheels this is of course TDC).
• Put a thin white mark on the flywheel opposite the 'zero' of your timing gauge.

White mark opposite the zero

• REMOVE THE 6MM PIN
• Rotate the engine again until you again find the start of the compression stroke for number one cylinder again. 
• Now slowly continue to rotate the engine until the temporary mark you made is opposite the '6'  mark on the gauge.

You might find the engine over-rotates and your mark goes past the '6'. if so, repeat the exercise  until you successfully stop the mark opposite the '6'. 

• Without disturbing the engine any further, set up your distributor to cause ignition to happen at this point.

For an  old mechanical dizzy, this is the process of using a test lamp to indicate and so set the precise point at which the contact breakers open.

And if you are fitting 123 ignition, this is the point at which the little green LED light comes on.

• having found this point, clamp the distributor up just so that it can only be turned stiffly.

What have you just done? Even though the notch in your flywheel finds the TDC point, and your white mark on the pulley indicates TDC, you have set your 'static timing position' when the engine was 12 (crankshaft) degrees before TDC. As such, you have mimicked the process for cars manufactured before July 1971 and should be confident that the engine will run sufficiently for you to complete the timing process - setting the advance.

Setting The Engine Advance

If you have followed the steps above, the key point to bear in mind here is that while you have set up your distributor to trigger ignition at 12 degrees before TDC, the mark you have made on your pulley equates to zero degrees of advance. So when setting the advance, you need to see the full amount of advance on your timing gauge. You don't need to do any column 'C' maths to account for the 12 degrees as would otherwise have to do for a pre July 1971 car. Follow column 'B' in Citroen's timing values table. But do remember to halve the value....

Here is an example. This is an extract from Citroen's table of timing values.

Illustrative extract from Citroen's timing table - manual 814

Using a 2347cc DS23 of the 1973 year as our example, (second row here) the table says that the advance timing should be carried out at 2000rpm and the total advance should be 23 crankshaft degrees. That's a total of 23 degrees. Note that column 'C' is empty. The assumption for your DS 23 is that, when setting the advance, you need to set all 23 degrees, because you have been instructed to start with TDC as your static timing point - your 'static setting' as column 'C' calls it is zero.

If you had a DS23 covered by that timing curve and those values, and if you had followed the process I've described here, that is what you would need to do when setting the advance: you would be looking to see 23 degrees. Since you would be using the pulley to set this, you would be wanting to see the mark you made move by half that value '11' 30"

Note also for the first row above (a particular 1985cc 'DV' vehicle) that column 'C' is populated. Why is this? Well that row covers certain vehicles/ engines used both sides of the fateful July 1971 change in flywheels.

• For vehicles in this row but before that date, the static timing position and the white mark made on the pulley already gives 12 degrees. So of the total 24 degree specified, only another 12 remain to be set in the 'advance' process. 12 degrees of crankshaft movement are measured on the pulley as 6 degrees of movement against the gauge: from zero to 6. That's what column 'C' is telling you.
• For vehicles made after July 1971, the assumption again is that you are starting from a zero degree position and will need to set all 24 degrees - just the column 'B' figure.

That's just about the key difference when setting up the timing on post July 1971 cars. Any error in the process is not the fault of 123 ignition per se though the instructions that came with my car were pretty cryptic. It's the Citroen factory workshop manuals that say to start the process from TDC, and fail to suggest that the engine should be/ needs to be 'wound back'. 

FOOTNOTE

Someone has recently sent me PDF (so presumably downloadable) instructions for fitting 123 that are specific to the DS and which do mention the need to 'wind the engine back' 12 degrees before you fit the 123 distributor and rotate it to light the little green LED.

Extract from 123 electronic ignition PDF instructions for Ds

I've also re-read the cryptic generic instructions that came with my 123 and I think that the process outlined would be sufficient for setting the static timing on a DS made after July 1971. The instructions would work.

Generic and cryptic - 123 ignition instructions

Here is what they say in essence:

• Using the distributor cap as the guide, put a mark on the perimeter of your old distributor body that aligns to the 'cylinder 1' HT contact of the distributor cap when it's fitted.
• With the distributor cap removed, crank the engine so that the rotor arm points to the mark you have just made - the point where the rotor arm would cause a spark to number one HT lead if the distributor cap was fired. 
• Leave the engine in this position and replace your old distributor with the 123 distributor
• Twist the 123 distributor anti-clockwise until the little LED light comes on.

What the above process essentially does, is it use the position of the distributor about to be replaced, to find the current timing position of the engine and use that as the static timing position. And because it uses the distributor cap HT lead position for cylinder 1 as the reference point for the mark, when the rotor point to that mark, cylinder 1 must be on it it's compression stroke. So using the position of the rotor against the mark made on the distributor perimeter, is a rough equivalent of rotating the engine so that any mark on the pulley aligns with '6' on the gauge - about 12 degrees before TDC. As such, it should be enough for static timing.

The caveat to this is that the instructions assume the engine hasn't been cranked or otherwise rotated with the distributor out. They rely on the distributor being in the same rotational position as when the engine was last running. That isn't mentioned, but the generic 123 instructions are written from the point of view of swapping out an old distributor - so perhaps that omission is understandable.

Refitting the Rubber Door Seals

This is one of the jobs that - for me - transforms the hard metal chassis into a car. For that reason I was  looking forward to it, but at the same time was also dreading it. it's a tricky little job that doesn't easily forgive mistakes. And I only wanted to do it once.

Door seals refitted

The rubber door seals go around the door frames. The doors - and their glass - close against these seals. The seals fit into metal channels around the door frame edges. 

Metal channels are parts 11 and 14. Door seal rubber is part 15

The channels are used to crimp the seals in place. Fitting the seals is all about the channels. The channels, the channels, the channels......

The original channels were made of quite thin metal. That made them easy to close up when fitting the rubber seals. Unfortunately being thin also made them prone to rust - and to them needing replacing. 

Water wicks up and rusts the channels

And that meant you need to take the rubber seals out. Removing the seals when carrying out restoration is easy and satisfying: tugging on the rubber at an angle opens up the crimped edge claws and out they came.

The rubber seals survived this treatment well. The trouble was that the channel crimping claws  - being only thin metal - opened-up and became badly mis-shapen and wobbly along their length. You can see that in both the photos below.

As such, putting the seals back on then became a lot harder......Some restorers replace the whole crimping channel with modern reproductions. 

Reproduction channels are available. They are 'handed' so be careful what you buy

This might be because the old, thin channels have become very rusted but might also be because - having removed the rubber seals - they can see the damage that's been caused to the claws of the old channels. We should be grateful that reproductions are available, but I'm told they are not a good fit around the curves of the door frame and require a lot of fettling.  And being pressed into a 'W' profile - they are not easy to reshape!

Reproduction channels

Curve is not quite right.....

The other problems with the reproductions are that:

• They`re made of thicker metal. They are less easy to close up once the door seals are fitted and there is the risk of damaging new chassis paintwork when you attempt to do so.
• The crimp arm and claw is a bit too big. It doesn't fit the shape of the inner door trim.

That's why I disturbed the seals as little as I felt I needed to......

Seal only partly detached and tied up out of the way

.....and used as little replacement channel as possible.

I used as little new channel as I could get away with

Old channel being removed

Patch section of new channel welded in

It's also why - when I thought about painting the chassis - I was keen to avoid removing the seals. Originally, after fitting the new outer sills, I seriously thought about re-seating the rubber seals in their channels and then - somehow - masking off those areas for when it came to painting the chassis. 

More recently, as I stripped old paint from the chassis, I realised how ridiculous that was. Also -  when I was working on the outer sills - I'd caught the rubber seals with my angle grinder leaving some ugly nicks and scars......

Telltale damage from an angle grinder

......and my ego made me keen to replace those door seals with unmarked ones. So.....even knowing the damage it would do to the crimp claws, I had carefully pulled all four seals completely off. 

Now I've gone and done it......

It isn't really viable to try and open up the crimps to remove the seals as you still end up with wobbly claw edges.

Because of the damage this inevitably caused to the channel claws, as part of the preparation for chassis painting, I invested a disproportionate amount of time in re-shaping the claws on the channels. More specifically, I spent the time on restoring the 90 degree angle to the crimping arms and claws and making the channel arms straight and consistent as I could over their lengths. To help me, I made a couple of tools to fit inside the crimp claw.

Home made dolly to re-shape the channel claw edges

Smaller tool for tight work

The idea being that these acted as 'dollies' on the back of the claw and I used a hammer to reshape the metal and restore the 90 degree angle to the claw as best I could. 

I used the tool to re-shape the claw angles

These tools worked surprisingly well but it still took me a good couple of days to work all around the four doors - inside and out - straightening out the claws of the channels.

Fitting the seals is a bit of an art. The metal channels have what I and going to call a 'W' profile. You need to use your imagination a bit.

'W' profile of seal holding channel

The big black rubber seals need to fit into both of the valleys of the 'W'. Note that the two arms of the 'W' have small 90 degree 'claws' folded inwards. It's these claws that crimp the seals in place. 

On the door seal, the edges of the seal fringes that get crimped in have fattened lips to help hold the seal behind the crimp claws.

Fattened lip on the edge of the seal fringe

When crimping, the idea is to close up the arms of the 'W' so that the claws fold over the top of the seal lip.

Arms and claws closed

If the crimp claw is mis-shaped - and especially if it is squashed flat or closed - the seal edge will not/ cannot locate behind it. The seal will sit proud (not straight) and/ or will start to lift out of the channel at that point.

To make it more interesting, on the insides of the door frame verticals is additional trim. In my case it's white ribbed piping trim. I think later cars had black?

Door piping trim inside the cabin

The trim has an 'L' shaped groove along it and hooked edge.

Groove and hook of the inner trim

This piping fits over the inner claw of the channel 

Now. Whereas on the outside of the car, the claw crimps directly to the black rubber seal.....

.....on the inside of the car you crimp the white trim against the door seal to hold both pieces of trim in place. 

Again, if the claw edge of the channel is mis-shaped it affects the fit of the white trim and that - in turn - can affect the fit of the door seal. You can see in the photo above that the fattened edge of the door seal is designed to fit in behind the round profile of the white trim. If the claw of the 'W' channel is mis-shaped, that will be difficult.

By the way, notice how - in my demo photos using reproduction channel - the white ribbed trim doesn't properly fit on the channel. It's not the groove in the trim that's wrong - it's the reproduction channel that's a bit too big. The trim can't properly roll around the 90 degree claw, meaning the white trim doesn't sit as intended. It hangs fairly loose and doesn't leave the intended space behind it for the fattened lip of the seal. It can make it harder to locate the seal behind the trim - because it's free to moves and takes up the seal space - and since the white trim is meant to be holding the door seal in place, there is a risk they both work loose. As I had retained the original channeling, then unlike in the photo above, my white trim fitted well to the lip and arm of the 'W' which made assembly less fraught.

You are also perhaps realising that, on the inside of the doors, the seal and internal trim need to be fitted at the same time - before you crimp. And it's the same for the black finishing strip over the tops of the doors and down the front 'A' posts

Internal black trim over the 'B' pillar

That needs to be fitted at the same time as the rubber door seals. And as a bonus, that black horizontal trim has a foam bead tucked inside it making it a little more awkward to manipulate. All in all it makes fitting the door seals a bit fiddly to say the least. Even harder if your roof is still on the car!

This is how I went about it

My black rubber seals had silver paint on them from some poor quality car re-spray in the past, and some kind of mastic along the crimping edges. Cellulose thinner helped to remove the silver paint and picking away until I was close too tears and my fingers were raw was sufficient for removing that f**king mastic.....What I was left with was REALLY loved quality rubber: very, very supple. Good as new. 

With door seals ready, the channel edge claws re-shaped and the chassis now painted, I put a thin bead of 'CT-1' sealant in the outer valley of the 'W' channel. The idea was that this would fill-out the channel after the seal was crimped and prevent moisture wicking up and causing unseen rust. That's the idea anyway....

Fitting the seals is tricky because each rubber seal is surprisingly heavy. And flexible. And long. The job is made harder by gravity - with the door seals sagging and twisting like a snake. Even worse if it drooped after fitting in the groove filled with 'CT-1' as the sealant went everywhere.

My roof is off my car, so I could access the roof rail. Readying some lengths of masking tape, I roughly positioned the seal along the top edge of the door frame - and then loosely taped it in place. I then took time to align the top fringe edge of the seal with the top of the doorframe and particularly its corners. they were my reference points for the rest of the refit. I carefully positioned the fattened lip of the rubber seal....

Fattened lip on the outer fringe of the seal

.....under the channel claws in the two top corners of the frame and 'tacked' the seal in place by lightly closing the outside claw at strategic points. This was the point at which I'd wish I could have grown another pair of arms and hands. Then - without stretching the seal, I worked my way around the outside perimeter and down, just lightly closing the claw edge as I went. Just enough to hold the rubber seal in place.  I reviewed the fit as i went - just to make sure the seal remained fitted as it should. 

The 'S' bends at the top line of the door were particularly hard to do. Even though the rubber is moulded to shape, if the claw of the channel is not adequately re-shaped here so that it grips the seal, the seal will pull itself out and will not follow the intended line.It's also hard to evenly crimp the lip in this area. 

Once I was happy that the seal was sufficiently held in the outer claw, I moved to the inside of the door. I 'rolled' the white ribbed trim onto the claw of the channel. 

You need to 'roll' the trim onto the claw edge

This is why that claw edge needs to be straight or straightened out. If it's not, the trim simply won't fit. And if it isn't fully engaged on the claw - then the ribbing of the trim will accentuate any bumps and uneven fitting. It will look awful.

The top of the inner door frames similarly have a black finishing strip. But it's one long strip on each side of the car that goes from the back of the car, right across the tops of both doors, along to the front edge of the from door and down to the sill. That also needs to be rolled on to the claws of the door frames before you locate the seal.

When the vertical and lateral trims were both located on the claw edges, I then pushed the loose fattened lip of the rubber seal.....

Fattened lip on the inside fringe of the seal

.....in behind the trim. 

The seal should fit behind the white trim

I then made the crimp, closing that half of the 'W' channel together using a rubber coated 'dead hammer' applying light, even, taps along the length of the trim. I took my time to close the crimp gradually.

The inner fringe of the seal  was now properly fitted to the inner face of the door frame. However this has the effect of stretching the seal over the middle hump of the 'W' and pulling the outer lip out of place again.......

So, turning my attention back to the outside of the door frame, I checked that the lip of the outer edge of the seal was still engaged behind the claw and worked my way around the perimeter, re-seating the seal as necessary and now fully closing the crimp edge. To do this, I used a hammer and another special tool.....

I used a wide, thin piece of wood as a drift

Forcing a fatter tool in the gap between the chassis and the seal would have added to the forces already pulling seal lip out of the claw  - at the exact point and moment that I was trying to closing the crimp. Whereas I could locate my thin piece of wood in the gap without disturbing the seal. 

Using a thin piece of wood as a drift to close the crimp

Wood was also softer and kinder on the paintwork and being a wide piece of wood, the load from any hammer taps was spread over a longer section of the channel arm - meaning it created a neater, more even crimp. I used a narrower tool to work around the 'S' bend in the channel at the top door edges.

Similar tool for the 'S' bends

When I was happy with the fit - and that the seals were not going to pull themselves out, I wiped off any excess CT-1 and left things to settle. While i was fitting, the CT-1 worked as a lubricant - sometimes helping me locate the seal but at others helping it pop out. Once the CT-1 cured, it helped 'cement' the seals in place.

I did that for all four doors of course... It took a lot of time but it was a very satisfying to see the results. It's not perfect. But it never was going to be once I'd tugged the seals off. It did leave me wondering though, how Citroen did it in the first place?