Citroën BX do-it-yourself

Repairing the window lifter

root — Tue, 10 Apr 2018 12:44:15 +0000

A sharp popping sound and the door window slides (or falls) down. Nice surprise but it does happen as our cars age. The window lifter mechanism uses galvanized steel Bowden cables and these tend to rust over time, especially if the climate is rainy and wet, or the door gets damp inside for any reason.

What to do now? First of all, don't press the switches in this case. Allowing the motor to rotate will eventually jam the cables now loose, allowing them to jump from their proper guiding channels. This can lead to more extensive damage than just the snapped cable.

Then, the very first step is to make the window close until it can be repaired. The best solution is to find a piece of wood of the appropriate size (or more than one stacked) and wedge up the window (you'll need 50 cm of wood). Make sure you saw a notch in the top of the block to make the window seat properly, without the risk of it jumping out accidentally if you plan to use the car during the repair. You can secure the window temporarily with duct or gaffer tape but the wood block is better, especially as it's completely concealed when you park the car anywhere.

To start the repair, remove the internal trim panel the usual way (a large screw behind the window lifter switch, another one at the bottom face of the upper end of the grab handle, the front doors with the speakers also have three smaller screws behind the speaker, so remove that first). The door lock knob is a bit tricky, especially as it doesn't tolerate repeated removals if you're not cautious. Use a small screwdriver to pull on the little tab in the exposed slot while you remove it and be careful not to break it.

You have to remove the complete lifter mechanism from the door. The motor is secured by three bolts, the guide rail by two:

To release the plastic lifter block that attaches the window to the actual mechanism, reach behind this block to find a plastic hook at the bottom. Pull the hook down, this will disengage it and you'll be able to lift off the window from the large plastic pin going through a hole in the bottom of the glass. Duct or gaffer tape the window to keep it above halfway up. You have to disconnect the motor now: its connector is a weatherproof one, use a small screwdriver to prise the red cover on the connector, as you pull it, it will disengage the connector itself.

Remove the rail and then the motor through the lower aperture. Finally, you can lift the window manually and place the wooden block to secure it. I found it much easier to use two pieces of wood than to try to wiggle a longer one into the door through the apertures. For instance, a block of 15 cm width across the bottom, a plank of 35 cm length on top of it, vertically. As the bottom of the door is a bit angled, you can use the bottom block to wedge the plank into place.

You need new cables. The original one was made of galvanized steel, probably 7×7 strands, relatively rigid even when new. For a replacement, you can do better. Shop around for the same outer diameter (1.5 mm) but thinner strands like 7×19, stainless steel. This is a very common size, you should be able to find it without any trouble. You need twice 1.3 m for one window (this length already includes all the slack necessary for the assembly), so, especially if you plan to do all the other windows preventively, or just to store up the cable for the inevitable future, being as cheap as it is, you might just want to buy 15 m or so, with plenty to spare.

You'll be amazed to feel how much more flexible this thinner stranded, stainless steel cable is. So much so that we can forget the problem of how to apply stop sleeves onto the cable. Pressing sleeves isn't exactly easy to do at home, they would require crimping pliers of a very considerable size (or a hydraulic press), and even then who knows how strong you manage to make it. But with this very flexible cable, you can simply tie a knot – yes, seriously. Don't worry, there will be room for the slightly larger size of the knot in our mechanism.

So, when you have your cables ready, let's start with the drum. Have two cables cut to the exact size of 130 cm, and tie a knot to one end of each. Grab the cable in a vise, pull on the loose end with pliers to make the knot as tight as possible. If you have a gas torch or a beefier soldering iron, you can add a bit of solder to the outgoing knot end to make it absolutely rock solid, without the possibility of the strands fraying. But this really only works if you have something more substantial than a regular electronic soldering iron (and even then, don't forget to use flux). Failing that, just wrap a bit of regular insulating tape around the wire and cut through that, this will also help against fraying.

Thread both cables into their anchor slots in the drum and wind them around in the grooves, in equal lengths. Both cables will have two full turns and will meet nicely in the middle, just between the two anchored ends.

If you have any doubt about any of your grooves, now is the time to clean, cut and re-form the walls using an Exacto knife, a small file, a small screwdriver or similar. You want the cables to lie perfectly in the grooves in order to avoid any possible jamming later when in operation. With the drum refitted on the spindle and the cover in place, this cover basically keeps the cables in position (the cables should meet in the middle). Grease the top of the drum before you fully tighten the screws.

Now, place the components on your table like this: rail to the right, with the pulley at the top and the convex curvature pointing upwards (the fastening bolts facing down). Motor to the left, with the drum visible on the left side (again, fastening bolts facing down). Thread the upper cable into the sleeve, across the pulley and down the middle of the rail. Make sure the springed end of the sleeve is housed properly in the plastic guide (it helps if you keep it temporarily in place with tape, it has a tendency to pop out from the holder while you manipulate the cable). The cable will reach to the middle of the rail.

Do the same with the lower cable. Slide the lifter block to the side, out of the way (or simply slide it off the rail temporarily). Put two marks 10-12 mm apart in the center of the rail with a marker. The task is to tie a knot on each cable, making the internal edge (the side coming from the motor, not the loose end) of the knot to align with the respective outer mark, basically leaving 10-12 mm distance between the two knots when pulled completely tight. Forgive the picture, it doesn't look completely precise here but that's the best I could come up with, with only two hands:

If you're interested how this distance can be determined: the tensioning strings are 37 mmm uncompressed, 16 compressed, giving 21 mm of tensioning capacity for each, 42 mm for the two strings combined. The lifter block has an overlap of 27 mm between the two cables fully seated in their slots. This gives an extra distance of 42 - 27 = 15 mm needed to achieve the fully tensioned state but nothing is that precise, hence the 10-12 mm recommendation. It's better to err on the upper end of this interval, though, going closer to 15 mm.

Now that you have all the knots ready, the mechanism can be assembled fully. Seat the lower cable firmly in its slot and force the lifter block upwards, completely tensioning and compressing the spring of the lower sleeve. Either use clamps to keep it in place (but be careful not to damage the rail) or ask a helper to grab both sleeves and to press them together into ther guides, compressing both spring as much as possible. With this tension in the setup, you can grab the upper cable and seat it into its recess in the lifter block, using pliers.

In the end, you want both springs compressed to near maximum. When you fit it back into the door, add the weight of the glass, and run everything to bed in, the tension will inevitably decrease a bit, so it's important to start with as much as you can.

Back at the car, reconnect the motor temporarily and grease the exposed wires. Run the motor in both directions to bring the grease into the previously hidden places and grease all cable runs becoming exposed now. Without the window attached, there is no end stop and you can run off the rail very easily, watch out. Remember not to try the one-touch switches yet, just the internal arrow ones.

Don't grease the rail itself yet, it's better to do that once already in place.

Now comes the tricky part, putting the mechanism back. It came out easier with the snapped cables and without tension but now its overall structure is much more rigid. Here comes the only really viable procedure I could come up with: start with the rail. Keeping the rail horizontally, feed it into the lower aperture from the center towards the outer edge of the door (the end with the pulley has to go in first). You have to keep the motor close to the rail as the cables allow it but don't even try to fit the motor yet, just the rail. As you go deeper into the door, you'll have some room, not much, to start to turn the guide vertically. If you end up with the guide almost vertical, pulley on top, pushed far back to the edge of the door, this clears up some space for the motor to follow. The only orientation the motor will enter the aperture in, also considering the cables attached, is with its metal body pointing towards the car.

Keep the motor turned as visible on the picture, and sink it into the aperture with the bottom first. Nudge the rail top make room if necessary and squeeze the top part of the motor into the aperature. No easy but, as illustrated, it's possible. Once inside, you can start rotating the motor to its real orientation, to line up its bolts. Only place the upper two nuts, without tightening. With the motor flush with the door and tightened up, you wouldn't be able to reconnect it, so leave it as loose as possible. Route the cable so that the red fixing tab (which you have to open fully, until you hear a click) faces upwards, and while fitting the connector, push on the tab firmly to snap it closed. There is an opening in the middle of the door you can peek through to see what you're doing, and in the end, this opening is where you can route the cable harness.

Leave the glass lifter block positioned behind one of the apertures where you can reach it easily and sink the window manually to the point when it sits on the locking pin. Reach behind the lifter and lower the window just a little more to let the pin go through the hole in the bottom of the glass. Squeeze the lifter so that you can push up the locking hook to snap it securely.

Now, with the window firmly attached, you can try the full operation of the system.

When refitting the trim panel, the only tricky part is the lock tab. Make sure the small plastic tab of the internal sliding mechanism is in the lower slot. Place the tab onto the pin, making sure it sits properly and while you press it down to lock into place, use a small screwdriver to push the locking tab in. The locking tab has to jump to the upper slot as the whole plastic part slides into place but it won't do that without you helping on that locking tab by pressing it in.

LEDs everywhere

root — Wed, 14 Jun 2017 22:14:38 +0000

dashboard

bulbs

LED

A project I've been contemplating for quite some time now, and with the constant advances in LED technology, I thought the time had finally come: LEDs all around the car. Not in the external traffic lights, for sure (with one exception, I'll return to that) but inside, with the large collection (nearly a hundred, really!) of bulbs everywhere. Let's count: instrument panel, warning lights, LH and RH display, switches, A/C panel, ceiling and rear lights, gloveboox, boot, door sill courtesy, radio surround, ashtray, cigar lighters, sunvisor. Quite a lot.

The LEDs were sourced directly from China using

Instrument panel and friends

Let's start with the small W1.2W T5 bulbs. My original intention was to use 5050 SMD LEDs all around because of their larger light output (5050 simply means the size in tenths of a millimeter, so a 5050 will be 5 × 5 mm). But I couldn't find them in socketed, colored variants. You have to pay attention: T5 is the actual bulb in the socket (whether it's removable as with the tungtsen bulbs or functionally equivalent in the socketed but inseparable T5 LED bulbs). The socket itself has another standard and there are several variants: B8.3D, B8.4D and B8.5D, denoting the relative size of the small connectors. We need 3D but the 5050 LEDs on the market (probably due to size constraints) only come in the larger, 5D size.

So basically, we have two options.

we can use socketed bulbs with a potentially less potent LED, then we have the nice touch of having colored sockets according to the color of the LED; this helps with the replacement and later handling;
we can reuse our existing sockets and put standalone T5 LED bulbs inside. We lose the possibility of colored sockets and those bulbs are somewhat longer (the socketed variant can do away with the connectors of the bulb and build the two halves together).

Finally, I decided for the second option. I really liked the idea of colored sockets but the actual performance is more important.

In any case, we need the LED bulbs in the color of their function. As I experienced it when I did the internal lighting of some switches, using colored LEDs for colored display positions gives a nicer look than using white LEDs. There is less waste of brightness and the final color will be more vivid and pleasing. Using white LEDs will result in washed out colors.

And there are places where we need the socketless, standalone bulbs anyway: around the radio, in the suspension and ignition switches.

We have two options: either these colored bulbs, more reminiscent of traditional bulbs, with their connectors formed of thin wires, or the PCB-based counterparts. These second ones look a bit strange (and come without color markings, but you can help yourself by marking them on the edge with colored permanent marker pens) but have two distinct advantages. First, the connection in our sockets in much more reliable, due to the much larger contact area. Second, they are a bit shorter and we have a few places where this counts.

In both cases, it pays to prepare everything first on the table, with a voltage source of 9–12 V. With the wire end bulbs, you often need to bend the wires to the outside to make them work in our sockets. When it's lit, place a little mark on the bulb housing on the positive side with a permanent marker (preferably in the color of the LED in the case of the PCB-based version). Later, when you install them in the car, use the following diagrams. Note that they show the position of the positive contact when you insert the socket into the hole, not after twisting it. In case of the PCB-type bulbs, if you feel it's not that easy to fit them into the holes, you might need to chamfer the corners of the PCB very slighty with a small file or sanding paper.

It's more complicated than to replace everything with LEDs, though. First, the airbag warning light circuit doesn't tolerate LEDs, so V1 should be left as a traditional tungsten bulb (with LED, it keeps blinking constanly, warning of an airbag problem). Second, the nice glowing orange gauge pointers depend on a specific kind of backlighting. V20 and V24 are most reponsible for the pointer illumination but V17 and V22 also take part. T5 white bulbs are of the cooler white variant, lending a bluish tint to the pointers: better than nothing but still not ideal.

Additionally, V18 and V19 are there for the turbo gauge, so if you don't have one, you don't need to put green bulbs there (it would create a much brighter background at the bottom of the tacho). Still, you can use those positions with less bright LEDs to help the orange pointers. This is something you really have to experiment with to find the best solution for your own needs but my current setup is:

a yellow LED in V18; this helps the tacho pointer without adding a visible hot spot in the tacho left corner,
a traditional tungsten bulb in V19; this helps the redline marking without adding a visible hot stop in the tacho right corner, and provides good illumination to the pointers,
a traditional tungsten bulb in V24; this feeds mostly the pointers and doesn't add a very visible discoloration to the green backround there.

There are other things to consider, too. Due to the way these colored LEDs are manufactured, there are brightness differences in color, even if the bulbs are externally the same. White and green are always the brightest, blue is OK (we only need it in a single position, anyway) but red and yellow are dimmer, no matter what. Yellow is also a bit orangeish to start with. So, there might be another workaround for the yellows: a less potent LED but in white:

It's less bright than the 5050 chips described earlier but it's white, so the brightest color of the game. If we forget the original idea of same color as the warning light position in this single case, it will be less orangeish, more yellowish in shade and just on par with the other warning lights. But this might be a question of personal preference, too.

Warm white would be even better but I've never seen these T5 bulbs in warm white.

The rest are green everywhere. The suspension switch has a really small confined area inside, so the illumination effect of the LED is very strong and nice, even in daylight. The two buttons beneath the radio are acceptable but the coloring is not absolutely ideal. This is the place where the shorter, PCB-based bulbs perform better:

Display background

The displays (LH, RH, A/C) use the same bulb but I found an even better solution. There are T5 LEDs sporting two large 5730 SMD LEDs on both sides. They might look strange at first being based on a PCB but they fit into a T5 socket nicely. They are excellent in places where there is a need for a uniform, large display background rather than a spot-like, local illumination. All three displays mentioned above are perfect locations. Instead of local light, they flood the entire display compartment with enough light to give a nice, uniform background effect. The left on the picture is perfect in layout but, unfortunately, only comes in cool and warm white.

This one is a bit more interesting because it has a third LED on the front that we don't need (it would cause a very strong hotspot). I was lucky with this bulb because its chips are on separate circuits. I simply desoldered the front chip and the remaining two went on working. With a different bulb you might need to replace the chip with an appropriate resistor, though. So, using the second variant requires a bit of work but the end result is better: it's slightly less bright which is an advantage here (we only want the actual characters to be lit, not the background) and it can be bought in green. And if you find them still too bright, you can even desolder one more side chip. Leaving only those facing the center of the display unit (but try experimenting with tape covering the LEDs first before you grab the soldering iron).

The RH display has two or three light bulbs depending on the variant, the trip computer version has one more than the simple clock. But with three such side illumination LED bulbs, we already overdo it. I found two (center socket simply left empty) to be better. Thanks to the side lighting capabilities of these bulbs, there will be no ugly darker center spot.

The only temporary problem I faced was in the A/C panel: the bulbs are a bit long and don't fit there. But then came the Aha Moment: cutting pliers! The connector end of the bulb is quite long, as you can see on the picture. Cutting about the third is the solution in the A/C panel. Still long enough to be housed in the socket properly and there will be no problem with fitting. And the display is just beautiful, just like in the LH display:

Apart from these displays, they are also my bulb of choice in the ashtray and cigar lighter positions where there is room and the principal direction is to the sides, not to the front. However, we'll have a little trouble with the cigar lighters. All other T5 sockets have connectors arranged on the opposite sides (left and right). The cigar lighters, unfortunately, have it on top and bottom. This means an immediate and sure short circuit if you insert a PCB-type LED bulb here (the manufacturer could get away with this stupid solution in the case of wire ended tungstens because those are folded only to the opposite sides). You need a bit of insulating tape to cover the diagonally opposite connectors (say, right connector on both sides) of the small bulb, there is no other solution.

And finally, this is the perfect LED bulb for the ignition switch lighting as well. Don't try to force it open in the middle, the brown connector part and the white cylindrical cover is glued together. Prise off the black end cap instead and use needle-nosed pliers to remove the original bulb. A white two-sided LED bulb will be perfect here, too.

Internal lighting

Most of the internal lighting uses T10 W5W bulbs. Two in the ceiling spots, four for the rear seats. There is one in the glovebox, one in the boot and four in the door sills. I found several possible LED bulbs. The most logical one has several LEDs all around to give a 360-deg light pattern, while the other is a flat one with chips on both sides. Also, there is a clear trend on the market towards COB LED bulbs, these no longer have separate chips but a large, homogenous yellow area. The larger this area, the brighter the LED becomes:

The second proved to be more pratical in the glovebox, the boot light and the door sills. In these lamps the principal direction of the light is to the sides, so the flat bulbs perform excellently (note that all these pictures were taken in broad daylight, just in the shade of a door or similar, to show the real results. It's too easy to make very bright looking pictures with today's overautomated phone cameras).

The map reading lamps in the ceiling are fine with the 360-deg bulbs (again, note the daylight, there is a usual shade behind the windscreen, that's all). I know it's a question of taste but I really like this pure white instead of the old tungstenish yellowish tint:

But the same bulbs would be to bulky in the rather shallow lamps of the rear seats, we need a different format there.

And it wasn't my intention to replace the halogen bulb in the ceiling lamp but I was clumsy enough to drop it while hot and this is not something bulbs like. This BA9s socketed bulb being a one-trick pony in the XM, I didn't even have a spare, so I looked it up in the store finally. It will match the map reading lights now:

The result is also very good here. The T10 bulbs for the rear seats could be just a bit more uniform but with the limitations of this shallow design, they perform still rather well:

The internal lights are driven from a timer circuit that dims them gradually rather than switching them off abruptly. I'm happy to report that this function works well with LED bulbs, too. This is important because there were some speculation earlier whether the different characteristics of LEDs would allow dimming or not, whether they would be switched off rather abruptly or not. In reality, although the timer interval is very slightly shorter than with the original tungsten bulbs, the dimming effect remained.

Sun visor

This is less important in brightness, the benefit of chips would be more in the cooler operation. Almost all festoon LEDs, in order to achieve a brighter ouput, have the chips on a protruding platform in the middle, which is not acceptable for us because of lack of space. So, I chose the newer, COB LED variants that actually look much like the old filament inside the glass housing (we need the size of 36 mm):

Let's see everything compiled into a table for easier overview. Note that the quantities show more than actually required. Stores sell these bulbs in lots, so I had to round up to the next possible amount. Also, I wanted to have plenty to spare, who knows, in spite of the usual life expectancy of LEDs, I might still need replacement one day.

Bulb	Quantity	Use
T5 standalone	20 green 20 red 20 yellow (white) 5 blue	instrument panel (9 G, 6 Y, 1 B, 3 W), warning lights (10 R, 8 Y), radio flap (G), ashtray flap (G), suspension switch (G)
T5 double-sided	10 white 10 green	ashtrays (2 W), cigar lighters (2 W), LH display (2 G), RH display (2 G), A/C panel (3 G cut), ignitions switch (1 W)
T10 360-deg	5 white	map reading lamps (2)
T10 stubby	10 white	rear seats (4)
T10 double-sided	10 white	glovebox, boot, door sills (4)
BA9s 360-deg	2 white	ceiling lamp
Festoon 36 mm	4 white	sun visor (2)

How much really? Well, that's 69. And if you add the following section, the switches: 79 in total.

Windows switches

They have already been discussed in Switches with LEDs. However, handling all these ready-made bulbs made me thinking. Why not do the same and prepare a small PCB, just the right size to fit the switches, keeping the LEDs and their resistors neatly packed?

Apart from the neatness of the solution, it has real advantages over the original: twice as much LEDs, meaning brighter display.

This takes care of all the internal lights. But what about the outside? Well, not much. H1-H3-H7 are definitely out of question, not only the light output is still very much beyond expectations, their geometry is so much different from the old filament bulbs and the headlight projector geometry dependent on it, that the light pattern is completely wrong and, of course, it's therefore illegal.

But large manufacturers like OSRAM have already jumped the bandwagon with replacement P21W and similar bulbs, so I think the legal situation might change one day, with the bulbs gaining the necessary EU and similar approval logos and accepted for road use.

Reversing lights

There is one location where I still thought a bit of experimenting will be in order: the reversing lights. We all know that our reversing lights fell prey to the beautiful rear design of the XM. Housed in the plastic trim between the indicators, the coloring reduces the light output to a large extent. So, with the risk of needing to go back to old bulbs for the MOT temporarily, I chose these:

They look very serious, don't they? Not the usual dozens of LEDs all around. But those present are from Cree (the Ali store is

circuit

diagram

circuit_diagram.pdf — Downloaded 16564 times

The XM is a complicated beast when it comes to electrics and electronics: there is no denying that. It entered the executive car market with practically all the toys of the era. There are billions of components, thick harnesses, a plethora of connectors, everywhere. Finding anything without the proper diagrams is nigh impossible.

The factory manuals have all the diagrams for sure, as well as after-market manuals like Haynes in the UK or RTA in France. But they all reproduce the factory schematics that, still before the fully computerized era, were drawn traditionally, and for obvious cost reasons, in black and white (although the layout diagrams made use of the red spot print color already available because of the Citroën logo).

The diagram booklet we offer promises a little bit more. Although based on the original manuals in layout and many details, it uses color to help understanding. Parts, although they also indicate the factory numbering scheme, document themselves directly on the pages, without the need for constant cross referencing. Wires and connectors are color coded instead of cryptic literal abbreviations. Usual automotive icons help identify every important part. Harnesses are displayed in an unobtrusive light type that's available when needed but doesn't really get in the way when not.

Layout diagrams not only show the actual location of the component but have added clarifications whenever possible about precise whereabouts (LHD and RHD, where applicable) and various other pieces of information like bulb types, diagnostic codes and similar. Components use the usual icons to quickly convey their purpose, often without the need to cross reference to the numerical identifiers of the schematic diagrams.

The diagrams show a fully loaded car, with all the usual engines (diesels, petrol 2.0 and various V6 versions), Mark II, as a good starting point for most cars in use. And the diagrams underwent numerous small modifications during their years of use in the XM community: inevitable errors and omissions of the original factory diagrams were corrected, based on the actual cars.

Apart from the PDF, we also offer the diagrams in the form of a mobile app. Not only they can be always with you, the app provides easy search for parts by name and number and a plethora of other information about the car and its diagnostics.

The Citroën Technical Guide

root — Wed, 14 Jun 2017 20:10:30 +0000

guide

citroen_guide.pdf — Downloaded 11585 times

Željko Nastasić and I are proud to present the Citroën Technical Guide. This book, available in PDF format (a Electronic Fuel Injection — Gasoline EFI and EMS systems

The HPi Engine — Newest gasoline EMS systems

Diesel engines — Mechanical diesel pumps

Electronic Diesel Control — Electronically controlled EDC systems

Diesel Direct Injection — HDI diesel technology

Suspension

A Suspension Primer — Hydropneumatics and its advantages
Hydropneumatic Suspension — The basic HP suspension of the DS, ID, GS, GSA, SM, CX and BX
Hydractive I — The first Hydractive suspension of the XM
Hydractive II — The second Hydractive suspension of the XM and Xantia
Anti-sink system — The anti-sink system fitted to some XMs and Xantiae
Activa Suspension — The Xantia Activa suspension
Hydractive III and III+ — The new Hydractive systems of the C5
Hydractive Summary — Functional description of the suspension behavior

Steering

Power Assisted Steering — The standard DIRASS steering
DIRAVI Steering — The steering of the SM, CX and some XMs

Brakes

Standard braking system — High pressure brakes integrated with the suspension on all models before the C5
Anti-lock Braking System — The additional elements of the ABS

Electrical systems

Multiplex wiring — of some XMs and contemporary Xantia, C5 and Xsara

Air Conditioning

Air Conditioning — every form of aircon and climate control

Appendix

ORGA number — Calculating the ORGA (Replacement Part) number

Index

Index — Full hypertext index of all important components mentioned in the book

As already mentioned, not all chapters are complete and finished. We used a very large selection of materials and information sources to compile the Guide but we remain very interested in your opinion and some additional material you might have to help us further. Please, contact us but don't send large files at once, let's talk it over first, we might already have them...

Welcome!

root — Tue, 13 Jun 2017 09:30:06 +0000

This site was originally dedicated to Citroën's upper middle category cars, the BX series of automobiles, from the point of view of those owners who were not afraid of tackling some maintenance and repair work on their car themselves. As years have passed, more and more pages of its larger brother, the Citroën XM were added. Although the name of the site hasn't been changed for the sake of history, many of the ideas presented in these pages, especially all the pages originally published in the Citroën Technical Guide apply to many other Citroën models, not only BXs.

Although all DIYers should have a workshop manual like the #908 published by Haynes Publishing, many questions are not answered by such books. They describe the various components, their removal, dismantling and refitting in great detail, but they are often incomplete when it comes to troubleshooting procedures and ideas. Often they suggest component renewal as the only possible cure, however, many of these parts can actually be repaired with a little patience, skill and determination.

We tried to assemble a useful collection of tips and ideas to help our fellow Citroënthusiasts to get the most out of their cars while keeping the dreaded repair costs at bay. We also provide detailed descriptions of how to retrofit some nice features available only on bigger models (eg. coolant temperature meter in the instrument panel, one-touch driver's side electric windows, front foglights, warning beeper for headlights without ignition) and even such gadgets that were never fitted to BXs (eg. washer fluid level warning). In addition, we describe some modifications to standard BX solutions (eg. modifying the dashboard digital clock illumination level, swapping the rear fog and stop lights on Break/Estate models, killing the self-cancelling feature of the turn indicator stalk).

No such collection can ever be complete: thus, please feel free to contribute to it with your ideas.

If you have just purchased a used BX or think about buying one, check out the Maintenance page.

Switches with LEDs

root — Thu, 16 Apr 2015 12:47:58 +0000

switches

LED

The XM Series 1 was born in the era of automotive bulbs. When it came to the Series 2, LEDs were already entering the scene. New switches like those on the dashboard started to come with LED illumination but earlier switches retained (and shared with other Citroëns of the time) remained with plain old bulbs in many switches of the door windows, electric seats or some center console switches. Retrofitting LEDs to these is a simple process with huge benefits: longevity (practically, eternity) of the LED, cool operation temperatures and lower power consumption are all welcome changes.

Window and electric seat switches

The bulb holders inside the switches are not huge, they are only large enough to house a small socketless bulb:

Using traditional (through hole) electronic components, although possible, is a very tight fit. The LED itself can take the place of the bulb but the compulsory series resistor has to be housed outside the holder, with care taken not to obstruct the operation of the switch. And even then, the positioning of the LED is not optimal: bulbs are not directional but LEDs are and placing them horizontally, as the holder recess allows it, will not direct its light at the switch surface. Note that you need green light LEDs, not white. White will illuminate, too, but the colors will be washed out, green is much better.

Smaller, SMD components are better here, but at the expense of more skill required to handle them. It really helps if you have a helping hand stand with a magnifying glass.

The SMD components should be of size 1206 imperial (3216 metric, this actually denotes the physical size, 3.2 by 1.6 mm). Window switches have a separating plate in the middle of the cap, shadowing the illumination of the opposite arrow, so it's best to place an even number of LEDs into the holder. The 1206 components fit easily into the holder, with some room to spare but a resistor of this size can only handle 0.25 Watts and we would need slightly more. The first thought would be to use a 1210 format resistor but there is a better solution: two resistors in parallel. The main advantage is that resistors come with resistance in specific, standarized steps and using two makes it easier to fine tune the necessary resistance still using common, run-of-the-mill values.

The original version I suggested earlier had the components soldered together, with wires attached to the ends:

Not bad but we can do better than that. With all the experience gained in LEDs Everywhere, I came up with a better idea. Instead of attaching these components end to end with tedious work, a small PCB can hold them much better, is easier to solder and with careful packing, has room for twice as many LEDs:

But first, you have to know the electronic details of the LED you use in order to calculate the resistor values. Two pieces of information are important from the supplier catalog or the component datasheet: forward voltage (U_f) and forward current (I_f). The resistor value is easy to calculate:

R = (U_s – 2U_f) / I_f

where U_s is the board voltage, 14 V to be safe (the nominal voltage of 12 V lead batteries is actually 13.8 V, anyway). If we take an example of U_f= 3.1 V and I_f= 20 mA (the Foryard FYLS-1206PGC LED I used) the resistor will be (14 V – 6.2 V) / 20 mA = 390 Ω.

This is where the parallel resistors come in handy. Although you could buy 390 Ω SMD resistors for sure, such uncommon values might not be that readily available. These SMD components leave the factory on reels of 4,000 components. Just as an example, in the store I buy my components from, bog standard values can be bought in any quantity, they always have reels available and cut off as many components as you wish. With uncommon values, you might be forced to buy in larger quantities or maybe per reel because they probably won't be able to sell the rest that easily.

Just a reminder, you calculate the common resistance as:

R = (R₁ · R₂) / (R₁ + R₂)

Two very common usual standard resistor values, 1 kΩ and 620 Ω give us just what we need (differences of a few Ohms aren't important): (1000 · 620) / (1000 + 620) = 382.7 Ω.

Board

A schematic diagram is hardly necessary, all we need is two serial LEDs fed from two parallel resistors.

The printed circuit board layout (Gerber files are available if you want to have them produced):

So, if you use the same (or equivalent) LEDs as I did, the components needed will be as follows. With different chips, modify the resistors accordingly:

Position number	Window switch	Seat switch
R1	1K0	1K0
R2	620R	620R
R3	1K0	—
R4	620R	—
D1	FYLS-1602PCG	FYLS-1602PCG
D2	FYLS-1602PCG	FYLS-1602PCG
D3	FYLS-1602PCG	—
D4	FYLS-1602PCG	—

Soldering

Soldering SMD components is a challenge, especially if you only have a regular soldering iron and eyesight. Using a helping hand stand with magnifying glass or head magnifier glasses helps a lot. The usual way would be to use hot air rather than a conventional iron and a soldering paste that is tacky enough to hold the small components in place until you solder them. But you can still do it if you only have the regular equipment. Have a pointy iron tip, and buy two extra items: a flux pen and a desoldering wick. Flux pens are made under many brands, they look like a thick marker pen, have an alcohol or water soluble flux fluid coming through the tip. Don't think you can skimp on this; with through hole components, it was OK to use the flux (usually resin) in the solder core. But you won't solder SMD parts without a flux, believe me. This is what makes it work.

The desoldering wick is a simple braided copper wire that, due to its capillary action, sucks up unused solder like a sponge. Because you will add more solder than needed in many spots, the wick makes it possible to remove the unnecessary extra. Find a thin one (1 mm or less in diameter) because when you use it, you need to heat the wick and a thicker version will require too much heat.

The best way to solder SMD components that I could find on the Net and can verify by my own experience is as follows:

1. Tin one pad of the component you want to solder. If the PCB has been tinned in manufacture, there is no need for flux here. Just make sure there is a small bump of tin on that pad (practice will teach you how much but really not much).

2. Apply flux onto that tin now.

3. Place the component on its pads. It won't lie completely flat because of the tin on one pad. Clean the tip of you iron: there's no need for extra solder at all. Carefully pushing the component down in place with a forceps or needle-nose pliers (as you prefer) heat the end lying on the tinned pad. As soon as the pad gets hot enough, the solder will melt, the flux will do its job, the component will sink into its flat position and it will be soldered in place all right.

4. Now turn over, apply flux to the other pad and component end.

5. With a very small amount of solder on the tip of your iron (practice will teach you how much, really small), heat the other end. The flux will do its job once more, this end will be soldered in place, too.

With enough practice, that's all there is to it. If the blob of solder is too large, remove it with the wick, apply flux and heat again with a smaller amount of solder (if any) on your tip than before. Use a strong magnifying glass (25x to 30x) to look around closely.

The PCB itself can be held still with Scotch or applying double-sided adhesive to your work area and pushing it into the adhesive (you can also use a stretch of Scotch tape upside down, fixed in the end with normally positioned tape). A piece of glass can serve as a perfect work area (I always use an old BX mirror plate). Solder doesn't stick to it and you can apply and removes adhesive tapes with ease.

Always think about the order of components before you start. In our case, the R2 resistor needs to be soldered first because once its neighbors are in place, it would be much harder to reach its terminals with the iron. The rest isn't important with this particular PCB. Follow the markings on the board with polarized components like LEDs (the dotted end of the contour lines denote the orientation of the terminal marked with a small dot on the LED).

You'll also need wire connections (use a multi-stranded wire, remove the insulation and use the individual strands):

1. Tin the pad you want to connect to. Tin the end of the wire.

2. Apply flux.

3. Solder the wire end to the pad in one go.

Soldering, act two

Having said all that, using the "correct" method of soldering SMD parts is with a hot air gun. It's much easier, much faster, much more comfortable and much more fun, so really consider it if possible. Basically, instead of traditional solder, you use soldering paste, a mix of very tiny solder droplets in a tacky flux glue.

1. You just apply a small amount of solder paste onto the pad. I found the paste-in-a-syringe to be the easiest to use, you can very easily put small amounts of paste to the PCB surface. You don't need flux, you don't need tinning, just paste straight on the pad (a little blob in the middle, about covering just about half of the pad of these PCBs, there's no need for more). If the tip of the syringe starts to clog up (the paste tends to dry out when left exposed for longer periods), you can gently, very gently warm up the tip to make it flow again easily.

2. When both pads have the paste, simply place the component, press it into the paste with tweezers. The paste is a glue, the component will stay there. Again, you don't have to apply flux, just push the component into the paste.

3. Heat the component, both ends, using the hot air gun (250 °C is just fine with the usual soldering paste). You just move the tip of the gun here and there to heat both sides. When the solder paste melts on both sides, it will turn into regularly looking, shiny metal solder and now the magic happens. If the paste blobs at both terminals of the component are hot enough to flow, surface tension will literally move the component, it will align itself automagically with the pads, jumping into the perfect position–it's really fun to watch.

For the soldering itself, you don't even need flux but it helps to have it handy if you're not completely satisfied with the positioning or the amount of solder applied. Reheating and resoldering is easier and results in better finish if you add a bit of flux for each reheating operation.

However, soldering the connecting wires at the end is still easier with the regular iron.

Assembly

As soon as you have all components and the two connecting wires (length: 38 mm, diameter: 0.3 mm or similar) soldered, you can start assembling. Hold the board with the black resistors facing upwards, the wires to the left:

Have your green bulb holder with the opening facing up, flat part to the right and thread the two wires through the holes in the left side wall of the holder. Yes, I know, it's not that easy; still, it can be done. Use tweezers and a magnifying glass. If you're desperate, enlarge the slots leading to the holes (I don't recommend forcing the wires down the original slots, the wires are thin and the soldering might also let go) with an Exacto knife or similar:

At any rate, you need the board inside the holder, LEDs facing down (into the closed part of the holder), wires sticking out from the left wall holes. Route the wires to the sides and wrap them around the pegs twice. Cut the excess wire. Although the key in the green holder and the notch in the PCB, as well as its wire will probably keep the board in place, you might want to add just a blob of hotmelt glue in the end to make sure the board never wants to peek outside any more.

With the switch (be it either window or electric seat switch) in your hand, hold it in the way shown on the picture, the separated wall on the bottom connector upwards:

While you do so, note which side of the switch is to your left:

On the seat switch it will be obvious because this is where the symbol is but the window switch is two-sided. In either case, make sure you put the green holder with its non-flat, illuminated side rotated to the left side of the switch as you determined:

If you fail to do so, there will be no light: the LEDs are sensitive to polarity by their very nature.

One-touch driver's window

This switch requires a different angle of attack. Remove the caps (the inner part is just like the regular window switches, the outer parts can be prised off with a small screwdriver; keep it parallel to the switch surface and try to prise the side closer to the center of the switch). The center operating flaps just slide out and remove the rubber seal carefully. The bulbs are wedged into the slots of the connector plate. Cut out the bulb, leaving as much of the wires as possible. If you happen to remove the wires, there is no way to push new ones into their places (in this case, the only alternative is to solder new wires onto the connector plates. Not impossible but requires a relative hot iron (around 450 degrees Celsius) and, consequently, very skilled, very quick and very careful soldering so that nothing plastic will be damaged). Leave the wires, it's better that way.

Just like with the main switches, I now started to think about using the same PCB-based route instead of components soldered directly in situ.

This small PCB will be held in place by the bulb wires (or replacement wires you soldered there), preferably use up to 0.6 mm to make it more rigid). Again, as with the other switches, the board hosts more LEDs than the older approach could handle.

This is actually a double board, not a single one. After having soldered the components, it has to be snapped in the middle and the halves soldered to the left and right side of the switch, as marked on the PCBs. The lab I have my boards produced by couldn't put a so-called V-cut onto such small a board, hence the row of small holes to make it still possible to break the board into two halves. A V-cut is a small groove scored into both sides of the board to, similarly, make it easy to snap and this is the usual approach with larger boards that need to be separated.

Now hold the switch so that the plastic frame around the pins has the missing side at the bottom:

If you turn the switch around, still the missing side at the bottom, connect the two boards as indicated by the L and R markings. The LEDs will, obviously, face upwards.

Careful with the iron, try not to melt the plastic of the switch.

Center console switches

The switches in the center console (rear windows, rear windows disable, electric armrest) have a different bulb inside, not that green holder. The bulb is cylindrical, with a diameter of 4.2 mm and its terminals are soldered into a PCB at the bottom of the switch. The rear windows disable switch already has a LED for the yellow light position.

You can either buy component LEDs and the necessary series resistors or use the T5 bulbs described in LEDs Everyhere. In the second case, you need to disassemble them (straighten the wire terminals and pull out the internals from the colored plastic holder). These are LED chips with usually two small resistors attached to both ends and fit very nicely in the place of the original bulb. Having two resistors instead of one (obviously, with both half the value of the required resistance) has a real advantage: as the voltage drop is only half on each resistor, they can be of ¹/₈ W instead of ¹/₄ W, giving a smaller resistor. Even if there are two of them, they fit very nicely.

You'll probably find these 5050 chips to be much too bright. Simply replace the resistors with larger values (remembering that our brightness perception is logarithmic, so double the value isn't half the brightness). In my case, and that's about typical with 5050 chips, I needed to go up to 2 × 3.3 kΩ (that's approximately ten times larger than the usual ≈600 Ω of 20 mA LED chips). Or, if you use component LEDs, you have to experiment with your required brigtness level, anyway.

Note the polarity before you solder. The easiest way to keep track of polarity is to place a little mark on the PCB with permanent marker. Positive feed comes in at connector A1 of the window disable switch and A5 for all the other center console switches. With component LEDs, make sure the dome of the LED protrudes somewhat from the switch body, otherwise the illumination will be too centered in a hotspot. You might need to make some tests but as a rule of thumb, at least the domed part of the plastic case should protrude, or maybe even a bit more. The 5050 chips with their PCB holder sit very snugly on top of the switch body, in a small recess there, giving a relatively large angle of distribution, so I finally voted for them.

In the case of the yellow LEDs of the disable switches, I simply replaced them with current, brighter yellow LEDs without any further modification.

Dashboard switches

The switches in the dashboard row (foglights, hazard, heated window, etc plus the rear wiper and washer) have LEDs from the factory, however, technology back then only allowed much dimmer LED chips than today. If all your other switches are brought to current illumination levels, you might want to replace those LEDs as well. You need to buy 3 mm component LEDs, around 150-300 mcd (green is always much brighter, test your sortiment before you decide). All switch caps have a green symbol and a yellow filter for the operation telltale, although the hazard switch has a red LED behind it. If you prefer, you can replace some or all of the other yellow LEDs with red ones (in the end, I changed the rear window heater and rear foglight switch telltale to red instead of yellow. I always had to think about momentarily which foglight is which; with yellow telltale to the front and red to the back, it comes intuitively now).

Dismantling is easy, you don't need to take the whole switch apart. Prise free the cap and you'll find a plastic part serving as a reflector for the two LEDs. You can prise its lugs free (one at the top, two at the bottom). The small PCB with the LEDs can be removed carefully, use a needle nosed pilers to grab the LED terminals. Snug fit but it can be removed: it sits on prongs protruding from the internals below, serving as connectors. Once you have the PCB out, there's no need for more disassembly. Note the polarity of the LEDs (there are switches with the same PCB but don't rely on it, check every single one yourself).

The SMD series resistor is on the back of the PCB, 562 Ω, approximately correct for the typical 2.2 V and 20 mA LEDs. Desolder, bend the new LED terminals outwards (mimicking the original), make sure the domed part protrudes from the reflector holder (I found it better when it protrudes more than the original but that depends on your actual LEDs). Refit carefully, take care of not bending the connector prongs it sits on.

If the brightness is too much (it probably will be), consider replacing the resistor mentioned above. Because my LEDs were quite bright, I had to go to as high as 6.8 kΩ to get the result I wanted.

Rearview mirror switch

This switch has a single LED soldered in the middle. If you're satisfied with the brigtness, there's nothing to do. If you need somewhat more, just replace it with a contemporary green LEDs (probably greener, the original is somewhat yellowish). Just bend the terminals like the original LED was (quite rudimentary for a solution, all the other switches are better designed, but it works).

Anti-theft keypad

Again, only start to work on it if you're not satisfied with the brigtness. The keypad background is done with tungstens in a very special holder, so unless you have a complete replacement keypad to play with, it's unlikely to be worth trying to replace those. The LEDs, however, are very nicely designed, in a push fit socket, no need for soldering. Just remove the LED, observe the dimples on the terminals and use a needle nose pliers to make similar ones on the new LEDs. Both color and polarity are correctly marked on the PCB, there's no need to worry about that. The original LEDs are 5 mm but I found that the 3 mm ones I used all around are just fine and illuminate their front panel hole perfectly. The color is nicer for my new green LED and the red one, also serving as an alarm warning, is considerably brighter now.

End results

These pictures were taken an hour before sunset, the switches jut shaded by hand. They start to become visible to the naked eye around this time of day, the window switches being close to the window are not yet really glowing but the seat switches are nice green already:

In pitch darkness:

As you can see from the switch photo above, the green holder isn't symmetrical and the lighting will be offset to one side, hence the difference in the two arrows. A solution could be to modify the holder a bit: namely, to file or cut half of the green material covering the side farther from its arrow to illuminate. This might result in a slightly different tint for the two arrows but our eyes are more sensitive to the brightness here than the colour difference. But if you don't mind the extra light, you can remove the green material completely, leaving all LEDs exposed. I tried both and I like my switches not only to faintly glow in the dark but to give a nice green light, so I decided on complete removal but you have to experiment for yourself. The holder material is much tougher than one would think but with a good Stanley knife or a coarse file and with a little perseverence it can be done.