[cfnutt]

Hi Phil,

You are certainly giving us a master class in electronics. When you have finally perfected your design, will it be available in kit form through the Club with detailed instructions?

Martin

[Phil Bradshaw]

Dunno about making kits for DiY installation; the current version (16.11-3.8) takes a week or so to build from scratch and several days to install.

[Phil Bradshaw]

Assembled control board -



Board assembly to retainer -





Alteration of repeater LED wiring to thinwall 10/0.1 stranded cable to reduce wiring bulk -



The lamp lens can be parted from the base using a knife and then re-attached using polystyrene cement. A blob of 703 silicone on the back of the circuit board around the wires retains the board to the base and seals against water. Terminals for a KF2510 connector are crimped to the wire ends.

Polystyrene cement has to be allowed time to cure properly before repeaters can be attached to the mirror shells to avoid separation of the lens when the securing screws are tightened.

Addition of remote potentiometer (10/0.1 cables) and optic fibre indicators to mirror shell to avoid having to remove mirror glasses for setup and in service adjustment -



0.3mm diameter optic fibres used are cut from a £3 Lidl optic fibre lamp bundle as possibly the cheapest source compared to buying optic fibre by the metre. Larger diameter fibres and ferrules could be used but I've gone for small and unobtrusive.

Inset shows exterior view of optic fibres shaped to fit countersunk ends of the 1mm ferrules. Shaping is achieved by carefully melting the end of each optic fibre to form a crude lens; finished fibres are retained in the ferrules using 703 silicone.

The position of the indicators ensures that there is no direct view from the driver seat. Their main functions are to aid setup and service adjustments and provide means of fault finding should the system fail.

The 1x2mm silicone tubes are a tight push fit on the ferrules (outer diameter about 1.8mm) and need no adhesive to retain them.

The 1mm ferrules, the 3mm ferrule for screwdriver access to the potentiometer and the potentimeter board are retained using hot glue. Hot glue also is spread over the wires to the potentiometer board to provide mechanical support to prevent connection failure of surface soldered wire ends.

Short lengths of silicone tube and PVC sleeve are used to keep things tidy with heat shrink tubing covering the ends of the screws retaining the repeater lamp.

Control board installed in each mirror shell, all wires and optic fibres below the centre web of the mirror shell -



The remote potentimeter assembly is handed, the position being dictated by the mirror arm obstructing access to the lower edge of the mirror shell opposite the repeater lamp.

Optic fibres are cut flush with the ends of the 1x2mm tubes which require no adhesive to be retained by the 2x3mm tubes on the indicator LEDs.

Remote potentiometer wiring colours don't match circuit layout colours above: someone didn't check before gluing everything in place. Duh.
Upshot is reversal of the black and purple wires in the connector plug so that clockwise rotation of the adjustment screw increases temperature setting.

Mirror glass assembly connected to control board (right side) -



Trial fit -





The rubber seal has to be softened in hand hot water to ensure best fit all around the mirror assembly then held in place while it cools.

The fit of the glass in the mirror head shell isn't very good when new - hence the hot glue to cut through when dismantling new mirrors - and as a result the rubber seal barely covers the edge of the glass due to the mica sheets adding to the thickness of the heated glass assembly.

The simplest remedy is to resort to adhesive again once I'm happy with the installation wiring and can finally finish each mirror assembly. Clear silicone seems to be the best option because it will stick to the mica sheets surface and can be parted more easily than the original hot glue.

Meanwhile some burn-in is required so that I'm sure everything works as intended: lash up test rig -



The K-type digital thermometers are the older version of what's available on eBay. This type can be calibrated because the circuit board is provided with low and high range adjustment potentiometers (low range being closest to the display). Once a meter is taken apart two of the four holes beneath the label that align with the potentiometers can be punched through from the front and then covered with a date label after each calibration.

Calibration for the range 0-100°C is sufficient for present purposes, i.e. range between temperature of melting ice and boiling water, the latter adjusted to accommodate atmospheric pressure. Calibration of two meters so that they agree to ±0.5C is possible by ensuring that they respond together to small changes in temperature. Actually nice kit for not a lot of pennies; the later versions for much the same cost have no provision for any adjustment and can be adrift by more than 10°C but can be identified by the label being different.

The 14.5V supply is a 12V 30A unit maxed out on the voltage adjustment to simulate maximum voltage of the vehicle charging system. The control unit and switch are a story in themselves ...

[Phil Bradshaw]

Timer Control Unit

The present timer control unit is the result of a number of attempts at achieving reliable timer operation while providing a suitable supply for the mirror heater temperature control boards. Cheap timer PCBs off eBay for the most part don't fit this requirement, some needing surgery to even work at all in the desired range. Momentary trigger operation of some also is a demerit (more bits required) as is slow reset time found with one type: what's needed is a timer that will operate once when powered then reset faster than the system can be switched.

Simpler timers (i.e. minimum of components taking up limited PCB space) such as a Darlington pair or NE555 arrangement reliant on capacitor-resistor input are fine for short time periods but can be very hard to set with any precision to much over 1 minute without running into capacitor leakage and other circuit constraints even when using zener control on the input (about 3 minute tops). For what I'm after (90-120 second, possibly longer) a CMOS programmable timer seems favourite, preferably one that won't fry at 12V.

Oddly enough this is what I found inside an Omron H3Y timer relay when it occurred to me to take one apart. Inside the 12V version is a CD4541B 20V CMOS with ancillary components to go with it all squashed into a compact unit requiring only 3 connections to make it work. Comes with a knob and scale too. No point re-inventing that wheel then: chop the timer part out and discard the rest including the screw terminal base in favour of a PCB mount DPDT (double pole, double throw) G2R-2 relay.

Tearing apart a relay doesn't need any finesse because only the upper part of the shell is needed -



(H3Y-4 shown here, one of several bagged off eBay as being cheaper at the time than the equivalent H3Y-2.)

Once the relay and timer assembly is withdrawn the knob can be pushed from the shell from inside. When reassembling the timer part there is a bit of juggling to seat the knob into the potentiometer, both set together to minimum or maximum to ensure that the knob isn't fitted 180° out of alignment.

Details -



Relay, shell case lower part and screw terminal base are not required.



This is an earlier version with the shell cut off about flush with the edge of the board carrying the header pins. Later version uses more of the shell (about 5mm longer) so that the cut end of the shell can be shaped to make a bridge in order to provide room for other components and wiring on the control board. This also permits hot glue fill of the base over a card insert flush with the edge of the board to retain the board in the shell and provide mechanical support for the the header pins, the latter being essential to prevent failure of the solder connections to the board.

[Phil Bradshaw]

Timer Control Board

Circuit schematic -



Switch wiring is unchanged from earlier versions except for orange replacing yellow to distinguish between vehicle system voltages (orange) and regulated 12V output to the mirrors (yellow). In the process 0.35mm² thinwall cable replaces 0.5mm² cable for the pink and black wires for the switch LED in order to reduce harness bulk.

Output connections.

A: 12V supply positive for temperature control boards.

B and C: heater element supply.

D and E: heater element supply.

Connection E also is common ground for the 12V supply and LED repeater lamps.

The L7812 (12V) linear regulator is required for the HK4100F relay because connection 3 from the switch is at vehicle system voltage, i.e. up to 14.5V. The relay will operate down to about 8V (so will tolerate regulator losses when vehicle system voltage drops to 12V) but will start to cook at anything much over 12V. The two capacitors are required to quieten electrical noise from the regulator which can affect other board components and also interfere with external electronics, VHF radio receiver for one: made me jump when Radio 4 suddenly started whining like a banshee when testing a regulator before soldering it to the board. No heat sink is required for the L7812: with only the HK4100F relay to run the regulator barely warms up (about 30°C maximum) and it has internal overheat protection that will prevent damage to the regulator while the board fuse melts when a short circuit is introduced.

The present rating of 7 Amp for the board fuse is a compromise between protecting the system and limiting voltage drop across the fuse during full power operation on 14.5V to about 1.1V when the system is cold. The few tracks that will be handling full power current of about 5.5 Amp have solder overlay to increase their capacity but damage still could occur in short circuit conditions if too high a fuse rating is used.

LM2596 step-down buck converter PCB is so cheap that it isn't worth building from scratch and can be mounted on header pins soldered into the input and output connections at the corners of the board, thus raising the converter PCB above the main PCB to allow for wiring beneath the converter.

A multi-turn potentiometer on the buck converter board allows suitably fine adjustment of the output which is used for the timer and DPDT relay as well as the mirrors heater control. A 50kΩ single turn potentiometer is provided on the control board for adjustment of the switch LED intensity. Along with the timer setting it is expected that converter output and switch LED intensity should only require occasional adjustment in service.

The 4 indicator LEDs in addition to the switch LED, relay (PWR, UP) and buck converter indicators may not be needed for the final version except perhaps the green LED indicating power getting to the board when the system is switched off. They have been very useful during development though so they can stay for now.

Physical layout -



'See through' showing board tracks and cut-throughs -



The size and shape of the board is dictated by the box that it is to be fitted into -



This type of box comes with 10 posts in the base only 6 of which are used: two larger posts for 3mm self tapping screws for securing the board with the four smaller posts retained as side supports for the board. Unused posts are readily trimmed down using a twist drill of larger diameter than the posts.

Two PG7 plastic glands are used for cable entries. The blanking grommet opposite the two glands in the photograph fills a hole made in a previous arrangement having cable entry at each end of the box.

Assembled timer control board -



This board uses a 10 minute timer initially set to 2 minute; a 3 minute or 5 minute timer will do just as well if set time of 2 minute is enough for all conditions. Red marking of 6 holes near each of the board corners (likewise on the 'see through' image) is to indicate where not to solder anything so that the board is evenly supported all round.

Bridge shape of the timer shell with hot glue used to secure the side opposite the timer connections (upper view) and buck converter board stand-off using single headers (lower view) -



Mirrors harness -



Switch harness -



During installation the switch first must be fitted to the dash after which the harness connections can be made to the control board screw terminals and vehicle wiring. I'm still trying to find sockets to fit the type of switch used so that the switch can be pre-wired to the control unit prior to installation. However, I'm not sure about using sockets because this likely will entail de-rating of the switch which so far has proved reliable even in short circuit conditions sufficient to toast several tracks on one board.

[Phil Bradshaw]

First install of the mirror assemblies took a while with dodging snow and rain. When it came to testing them before installing the timer control unit and dash wiring the left side temperature control unit threw up a fault: the relay started buzzing as the temperature reached the cut-off point. This struck me as sounding like only one transistor was working properly because the relay eventually would switch over and the buzzing would cease.

I was nearly right. The culprit turned out to be cross-track leakage due to unexpectedly rapid corrosion which has affected every bit of metal within the mirror. After a bit of thought this isn't surprising because there's a fair mix of metals connected together one way or another including the ball seating plate that comes with the original mirror. All that is needed to accelerate differential corrosion is a bit of heat ...

The main issue was the accumulation of green residue on all the strip board tracks of the control board and on the back of the mirror glass assembly which had migrated sideways in places and in places had created leakage paths between adjacent tracks. Um. Something Needs To Be Done.

Strip boards for one mirror assembly after a quick clean to assess the damage -



Luckily no permanent damage has been done other than one of the transistors dying, duly replaced along with its mate for good measure and also the relay in case the contacts had been damaged by arcing thus shortening its expected service life.

I can't think of a completely effective way of preventing water from collecting in the mirror assembly, which it will do either directly or due to condensation, because of the way the mirror is fitted to the arm. Poor fit of the glass in the shell and the rubber seal don't help either but a non-setting sealant like Hylomar Blue may fix that.

Remedy that I've come up with to deter board tracks corrosion is to paint the tracks side of the the strip boards using spray lacquer that is its own primer (allegedly - time will tell). Like so after 4 thin coats, each coat low-baked in an electric fan oven set to 55°C -



Metal case thermal switches can be replaced by equivalent plastic case switches but maybe not with the present version of mirror glass assembly because I'm already thinking of redesigning the heater element so that it fits within the rim of the mirror shell rather than on it. More on that if and when I get round to it. Meanwhile a dose of WD40 should help prevent further corrosion of the switch cases and the rest of the metalwork inside the mirror assembly including the original ball seating upper plate which already has lost some of its galvanised finish.

[Phil Bradshaw]

Final installation took a while ...

Each temperature control board has to be connected to incoming wires before the board can be secured to the standoff posts -



Plugs for adjustment potentiometer and LED repeater connected, optic fibre sleeves in place -



Wiring from the arm assembly is passed through a sleeved grommet secured from inside the door skin using a circlip -



After passing forwards of the drop glass front channel the wiring is routed down to the lower hinge where it is cable tied to the wiring from the dash where it passes below the hinge -



Dash wiring enters the A post (hinge post) through the grommet used for the courtesy light switch wire and terminates each side in a 5-way connector for the mirror heaters and a 2-way connector for the lock motor about 150mm from the grommet.

Waterproof connectors used for the junction of the dash and door wiring with wiring length suitable for cable tying below the drop glass bottom buffer -



Highly expensive anti-rattle device -



Lock motor wiring is for one of these in each cab door -



Lock motors are controlled by the existing alarm which has a generic 2-wire output. The motor kit comes with a link rod the end of which is bent to form an S shape to fit the empty lower hole in the lock lever operated by the lock button.

The rest is plug'n'play once the dash wiring from left side to right side is installed (heater duct above engine cover has to be removed for access). All new wiring is wrapped up in convoluted flexible trunking and made long enough to allow the timer control unit to be stuffed into empty space in the corner of the dash above the fuse box or dangle below the dash for adjustment.

The system is working already; photos and initial setup description to follow in due course.

[VDUB384]

I must applaud your interest and determination to get heated mirrors and indicator repeater lights in the mirrors you have put a lot of time and work into it:D so when are you doing the heated seats lol .
Dave

[Phil Bradshaw]

Heated driver's seat sorted already - cheap seat heater with cigarette lighter plug from Lidl.