Original page created on 23/09/2020; updated on 11/05/2023.
I’ve already installed the tail lamps I’ve just described, equipped with a 0603 LED, on a lot of rolling stock, especially goods. There is still something that bothers me: the problem is that equipped coaches can no longer be used except placed at the end of the train. For a while, I’ve been thinking about a lamp that would be removable. Only two small holes would be visible, in the absence of a lamp, for the passage of the electrical connections, at the level of the lamp holder.
To achieve this result, 0603 LEDs are a little too large to fit into the lamp without jamming. The best will be of 0402 size (1 × 0.5 mm). Then, the connection could be done on a 1 mm pitch connector. It does exist, but it’s not easy to find.
Equipement of a Roco Est coach. The lamps are fixed permanently.
I therefore need “board to board” connectors (pin header) with a 1 mm pitch in SMT mounting. Indeed, as already done for incorporated headlights, the printed circuit will be glued to the end wall of the coach. All devices (the connector only actually…) should be on the other side of the board.
Here is what I found:
These connectors are interesting particularly because they can receive the male pins from the rear, through holes drilled into the PCB.
Once I’m pretty sure I can get this type of connector, I can start the project.
I’ll base this on the equipment of a UIC coach. I have many train compositions that end with such a coach. But there are also, jumbled together, DEV, WL MU, U, F, Rapide Nord, USI, etc. So I’ll have to make sure that the circuit can adapt to many different cases.
The original idea is to fit the rear of the lamp with a male connector that will plug into the coach’s drilled end wall. The male pins suitable for the female connectors found have a section of 0.3 × 0.3 mm. But, on second thought, I’m going to use round pins made of ⌀ 0.3 mm nickel silver rod, which will allow for more discreet holes (a 0.3 mm square has a 0.43 mm circumscribed circle, which would require drilling to 0.5. Round pins will be content with 0.4).
Here are the first drawings of the project. On the right, (theoretical) verification of installation on a B5Dd2 UIC already equipped with embedded tail lights. In fact, I am thinking of combining the two possibilities on the same circuit: removable or embedded lamps. In addition, integration of series resistor, and two groups of power pads, in case the circuit needs to be shortened.
Due to the small size of the components, it is necessary to make a template to place and solder the LED on its pins.
The template I originally made turned out to be very complicated to implement, and dangerous for the extraction of the LED after soldering. You can nevertheless find it on the old version of this page.
In a bakelite plate (or any other heat insulating material), I engrave two triangular grooves with a PCB engraving mill bit. Another plate is screwed on top of it to enclose two pins. The LED will be held between these pins (arrow).
Note: most of the following pictures are taken during the manufacturing process, using a digital microscope which is finally very useful to work with such small components. But the quality of the pictures goes from very poor with the oldest one to decent with the new one, bought at the end of 2021.
I cut two sections in ⌀ 0.3 nickel silver rod; one 5 mm long, the other 6 mm. This difference in length has two reasons: to facilitate insertion into the lamp, and to remind the LED’s polarity.
One end is chamfered on sandpaper.
The other end is flattened by crushing it with a pair of flat pliers, and shortened to about 0.5 mm (I did not do this, and it caused problems for insertion into the lamp). Here are a number of prepared pins, marked with a coloured marker to make them easier to identify.
The two pins are inserted into the template, with their flattened sides well parallel. The LED is then placed, noting its polarity (the LEDs used, although tiny, have a fairly visible polarity mark). The technique is similar to that of chopsticks or sugar tongs: with the LED placed upside down on the desk, I approach the template with the pins slightly apart, so that they pinch the LED.
It remains to adjust its position at best. This is finally the most delicate step of the process. By the way, I bought a pair of special SMD tweezers made of non-magnetic steel at the same time as these LEDs. I’m glad I did, because any other steel object, scribe, screwdriver blade, irresistibly attracts the LED.
Then a little soldering paste is applied to each side, taking care not to touch the LED, which is just waiting for a flick to escape…
The next step is soldering, with hot air because attempts with a soldering iron have all ended in the loss of the LED, either by disappearing into the environment, or by destruction.
Air temperature: 250 °C. The hot air gun in itself deserves an article, which I will probably write later, after having experimented with it more. Finally, no: no article!
Close-up view. Once the LED is soldered, a good cleaning is necessary to remove the remaining flux.
After the stress of hot air soldering, a working test is not useless. Note that none of the LEDs soldered with hot air showed any malfunction.
Some finished LEDs.
The first one, above, is crooked. The second one, on the right, has an excess of solder, which does not prevent it from working. From the third one onwards, the result is satisfactory. I am making progress…
Two ⌀ 0.4 holes 1 mm apart are drilled at the bottom of the lamp. Warning: for centring, do not rely on the rear circle, which is not coaxial with the front. So drill from the front.
The equipped LED is inserted into the lamp. Pay attention to the polarity. I systematically place the anode (the +, the longer pin), upwards.
The whole is fixed in the vice, with the interposition of a piece of waxed paper, then a drop of CA glue is deposited, followed immediately by the installation of the “glass”, which will have the effect of chasing down the possible glue excess. The glue leaks slightly through the holes, which will help the pins to hold.
The perpendicularity is rectified by placing a wedge (wooden pick here) under the paper. Although the CA is supposed to cure very quickly, I prefer to wait a couple of hours.
Here is the finished lantern.
Another working test…
Red SMT LED in 0402 case
LiteOn LTST-C281KRKT
£5.2 per 100 — price 2020
at TME
Non magnetic special SMD tweezers
Bernstein 5-078-13
£10.40 — price 2020
at TME
Andonstar ADSM201 Microscope
$138 $ (€124,52) — price 2021
at Shenzhen Andonstar Technology Co., Ltd
Hot air soldering gun
$52.52 — price 2021
at Ebay
No article, as this item has never worked properly:
fancy temperature display, untimely stop of operation,
then total failure after only a few uses. Contacted,
the seller promised a replacement, which I have been
waiting for three months (note of May 2023).