Motor constantly on.

There is nothing obvious. Have you done the static voltage checks? They are a good indication of most of the likely problems. Beyond that it is just a process of elimination.

What happens when you connect the decoder to a DC supply? Does the motor reverse OK then?

If you have a PIC programmer you can try programming the address to something other than 3 and then reading the EEPROM location 01 in the PIC programmer to see if it has changed. If it has take a look at the motor drive circuit, if not take a look at circuitry between the track connections and pins 6 and 7 on the PIC.

I've just built up the 2nd of the kits I bought, exactly the same result. Motor constantly on.

All voltages are correct again, and I programmed the pic using my new Pickit2 programmer. The pic verifies ok as well.

I've programmed the chip using "DIY Motor 629.hex", is this the correct hex file?

It sounds a little like it is not decoding any DCC data. I think that the next thing to try is to see if it works on DC to rule out any incompatability with the Elite.

1. Remove the L272 chip from its socket (or disconnect the motor).

2. Install the function decoder firmware.

3. Connect the red and black wires to a regulated DC supply (anywhere between 7V and 20V) and measure and note the voltage on pins 2, 3, 6 and 7 of the PIC.

4. Reverse the polarity of the red and black wires. The voltages on pins 2 and 3 should have reversed, following the reversal of pins 6 and 7.

This will demonstrate that the PIC is functioning OK in DC mode and that the track signal is getting to the PIC. If this is the case it looks like there may be an incompatability with the Elite.

If you are not seeing the correct voltages at pins 6 and 7 check the values of the resistors connected to those pins and check for shorted tracks.

ok, I've got some voltages. How do these look? The only one that really reversed was pins 6 and 7.

PIC was still in circuit, and 17v was coming from the Hornby controller.

----------Normal------Reverse
Pin 1 ------3.5V------3.5V
Pin 2------0.3V------0.3V
Pin 3------0.2V------0.3V
Pin 6------1.7V------0.3V
Pin 7------0.3V------1.8V

The regulators are 78L05 (came in the kits). I was very careful with those, looked at the datasheet to make sure I had them the correct way around. I'll see if I can get a couple of replacements tomorrow and see if it makes any difference.

The tant is also the correct way around.

I've replaced the regulator and the cap (I used a 1uF electrolytic as I had one to hand).

Voltage into the regulator hovers around 17V. I'm still getting 3.6V on pin 1 of the pic. I've tried two different (brand new) regulators, both give the same result.

Just out of curiosity, I removed the regulator (I put some sockets in so I can swap it out quickly), and I get the same results (motor only going one way). The voltage at pin 1 of the pic is 1.5v if I remove the regulator.

Apart from the cuts being missing under the L272 it looks OK at first glance. You need to investigate why you are not getting 5V out of the 5V regulator though. I think that is fundamental to what is wrong. The output from the regulator should only be connected to R14, C3 and pin 1 of the PIC. Only C3 is required to be connect to the output for reliable operation of the regulator. R14 is 10K so should not provide a big load - have you checked the value of R14?

You should be able to get the 5V power supply working. It is quite straightforward and it is always worth checking the 5V is OK before plugging in the PIC and L272.

I've checked those caps, they are ok (can see the numbers fine outside in natural light, 104 and 474).

Also checked the rest of the components.

D1 - Correct (1N4148)
D2 - Cannot make out the numbers due to position
D3 - Correct (1N971B)

D4, 5, 6 and 7 - all 4003

R1, R2 - Both 10k.
R3, R4 - Both 1k.
R15 - 4.7ohm
R14 - 10k

I'm going to attempt building it on breadboard, but cannot find a 4.7Ohm resistor locally, will a 10ohm do instead?

Thinking about it, would two of the 10ohm in parallel be ok instead of 4.7ohm?

I give up.

I built up just the power supply part on breadboard (using new parts) and got 4.9V out of the regulator. Transferred it to some strip-board and got 4.9v. Added R1, R2, D2, C4, R14, C4, R3 and R4 plus the jumpers, tested the voltage again and now I'm getting 0v. Disconnected R14 and D2 (which should have completely disconnected the new components from what was working previously), and I still get 0v.

I give up. I've wasted hours and hours trying to get both the kits and now a new build to work and just failed miserably.

There is no harm in copying the application notes on the datasheets.

The capacitors are there to stop instability. Generally you can put as much as you like on the input, but it is the little ceramic one (C2) that stops the oscillations due to being much lower impedance at higher frequencies. The tantalum capacitor is quite low impedance so you could probably replace both C1 and C2 with a high value tantalum and put it in the C2 position.

Sometimes putting a big capacitor on the output can cause instability. I would recommend that you always have a bigger one on the input than the output, although there should be no need for a really big one on the input since the current consumption of the PIC is very low. If you are using a big one on the output I would recommend a low impedance type such as a tantalum rather than an aluminium type which have a large effective series resistance. Large value ceramics tend to be very big and very expensive

If you are getting oscillation in the regulator, it will cause your meter to display strange values, and only an oscilloscope will confirm what is actually happening.

There is a possibility that the load on the regulator is too small, so it may be worth trying to reduce the value of R14 (just put a 1K in parallel as a test) to see if that makes a difference.

So should it make any difference if I swap C2 (100nF on the output) and C3 (470nF on the input)? I'm not quite sure if the problems I've got are due to dodgy multimeter readings or something else.



On my latest experiment, R14 is not even in the circuit. Could this cause the dodgy multimeter readings?

Just an update.

Bought a cheap multimeter on ebay. The two newer decoders I have built are now showing a steady 4.9v coming out of the regulator, so hopefully that is ok.

Now I need to put in the rest of the resistors/caps which go next to the PIC. Probably will not be today though.

I feel like I'm getting somewhere at least. The worst thing is I probably had the correct voltages earlier on (Pin 1 ------3.5V------3.5V), my multimeter was giving me dodgy readings!!!