Maintenance

[TerryJC]

I'm not sure if the capacitor actually made any difference or I was just lucky.

I would suggest that if you can reproduce the effect next week with the 100 nF cap and approx 7 cycles, then we aren't far from getting what we want. We might get away with this by trying smaller caps on the FET Gate and / or larger caps to slug the effect overall.

We still have the option of adjusting the pot to increase the threshold if all else fails.

[Penri]

OK, I'll go along with that, and I might add some decoupling on the power lines as well.

[Penri]

Controller reinstalled and wired into circuit today, over-current problem still persists.

[Geoff]

I went into the Model Town on Monday and found the underground train at the far station and did not respond to the proximity sensor.
On closer inspection I found the fuse had blown. On replacing that with a 1 A fuse it started working fine. Investigating why the lights were not working correctly I found that they were connected to terms 11,12,13 and 14 when they should be connected to terms 12,13,14 and 15 Term 15 was left empty!
Which meant one of the LED lights was connected to the Guard Rail Button. I don't know if this caused any overload problems?
On correcting this the lights worked as they should.
When I left everything was working fine.
I don't know when you collected the spares for the underground you also picked up the spare proximity sensors which were in a pink antistatic bag, or where they might be?

[Penri]

Geoff

As I recall I picked up the underground spares box and took it up to Terry, I don't think I removed anything.

Thank you for checking the wiring.

Although the underground was running today, its behaviour was not as expected, I'll post more on this later.

Penri

[TerryJC]

Penri,

I returned the spares box to you when I brought the repaired system back to your house.

[Penri]

Terry

Yes that's correct and I put it back in the railway store as it was, I think the only thing you queried was wether the motor drive board in the box was a good one or the defective one that Geoff swapped out.

Penri

[Penri]

Hello

When I arrived at WMT Mandy told me that the underground train was going but only moving on the far right hand side of the track (as you look at things from the visitor's viewpoint).

I cut the power, moved the train back to the left hand side, it's normal starting position, then reapplied power. The train started, seemingly normally but only got partially into the tunnel before stopping, then continued on, stopped mid tunnel then started it's return journey, stopped once more then came to rest partially out the left hand side of the tunnel. The second run as much like the first but now shifted further right, never regaining it's "normal" starting position. Eventually after the 4th or 5th run the train was "stuck" at right hand end of the track.

I took some timings, very crudely with my phone's stop-watch function (in seconds), for each stage of the journey:

Transit time start to first stop 5.33 Travelling left to right
Time at first stop 6.96 Stopped
Transit time first to second stop 3.52 Travelling left to right
Time at second stop 6.67 Stopped
Transit time second to third spot 3.29 Travelling rIght to left
Time at third stop 7.46 Stopped
Transit time third to fourth stop 2.36 Travelling rIght to left

Do these timings make any sense?


Penri

[Penri]

Geoff

I should have apologies for getting the connections wrong, sorry for both getting them wrong and not apologising.

Can you provide a sketch of the underground circuit please, it may help us understand some of the interactions we are experiencing, plus help me understand what a Guard rail is and what it does. The last time I came across Guard rails was in designing circuits to use them to remove the influence of stray electrical paths when measuring passive component soldered into circuit boards under test.

Penri

[TerryJC]

When I arrived at WMT Mandy told me that the underground train was going but only moving on the far right hand side of the track (as you look at things from the visitor's viewpoint).

I cut the power, moved the train back to the left hand side, it's normal starting position, then reapplied power. The train started, seemingly normally but only got partially into the tunnel before stopping, then continued on, stopped mid tunnel then started it's return journey, stopped once more then came to rest partially out the left hand side of the tunnel. The second run as much like the first but now shifted further right, never regaining it's "normal" starting position. Eventually after the 4th or 5th run the train was "stuck" at right hand end of the track.

I'm thinking that this may be caused by something slowing the train down so it never reaches the sensor. By default, the train is accelerated out of the station to maximum speed and runs until the first sensor is passed. It then decelerates into the tunnel and stops. Each segment is supposed to work the same way.

In order to control this in a sensible fashion and also to prevent the train hitting the buffers in the event that a sensor is missed, there is effectively a timeout. This timeout is set and is used to calculate how many acceleration and deceleration iterations (loops) the software will need to get the train to the designated stops and is based on the maximum and minimum speeds required and the time to reach that speed. The maximum and minimum values for PWM are also defined and will obviously have an effect. It the loop timeout is reached before the sensor, then the train will stop too soon; much as you have observed. If the train is running slowly, compared to when the system was commissioned three years ago, then this will be the result.

There are obviously a number of reasons why the train might run slowly; eg track resistance, failing motor, high resistance current path to the tracks. (I discount the latter because it is difficult to envisage a high resistance and a short circuit in the same path, although a partial short would rob the train of power.)

They may be other causes that I haven't thought of, but running too slowly would seem the most likely to me.