Reed switch speed

[duncan webster 1Participant @duncanwebster1]

I want a device which will react to the passing of a train on a 5″g track. I’ve used track circuit in the past, but that won’t work on welded steel track. I know I could have a short bit of track insulated, but is there any chance a Reed switch would work, magnet on driving truck. Suspension travel is 10mm,and 8 mph is 140 “per second, so it will have to be fast. If anyone has tried this it could save me some experimenting. Same question for Hall effect.

[Michael GilliganParticipant @michaelgilligan61133]

Last September, I bought some  little reed-switches [one of those impromptu ebay purchases ‘for stock’] … very impressive specification, and silly-cheap

Reed Switch Hamlin 59170-510 Sub-miniature Overmolded Reed Switch (PACK OF 10)

ebay Item No. 304455711869

MichaelG.

[Robert Atkinson 2Participant @robertatkinson2]

A reed switch should be fast enough as long as the magnet pole is large (long) enough. Around 20-30mm should be enough.
A hall effect switch will be more than fast enough. They work for ABS wheel speed sensors.
The bigest issue with both will be physical alignment with the magnet for either sensor. The Hall effect could be fitted to a tapered pole piece to concentrate the field making a larger and more robust target.

[SillyOldDufferModerator @sillyoldduffer]

Reed switching speed is of the order of 100uS, but that assumes the actuating magnetic field is to specification, which could be hard to maintain, and field strength decreases rapidly with distance.  The magnet would have pass consistently close to the reed, and I don’t know how practical that is on a bouncy train.

Not tried anything remotely like it, and although my armchair says reeds should easily be fast enough, armchairs don’t do experiments.

Dave

 

 

[Don CoxParticipant @doncox80133]

I used to maintain a couple of TXE2 telephone exchanges in my BT days and they were 100% reed relay switched with many hundreds, if not thousands, of reed elements in the place.  The setting up speed (that’s the time from the customer picking up their ‘phone to the point when they received dial tone) was limited to 120 milliseconds after which time the call was abandoned and a second attempt made with the information relating to the switching paths used for the failed call recorded on a paper record.  This was fast enough for the customer not to even notice, I can’t remember the exact sequence of events now, it was about 40 years ago, but several sequential reed operations would have been involved in that process so I think your speed requirements would easily be met.
Reed reliability was a problem and for that reason the checking sequence, partially described above, was an important feature of these exchanges.
My subsequent experience with burglar alarms at home taught me that the door gap had a critical effect on reliability, but the distance between the actual reed element and the quite puny magnet was probably still 10mm even in the ideal circumstance so I would think that should be achievable in your project, a bit of experimenting with a multimeter will tell you.

[John Doe 2Participant @johndoe2]

Might be simpler and easier to have a standard microswitch mounted between the tracks, operated by some sort of sprung lever fixed to the underside of the train.

The type of switch I have in mind are those in a plastic oblong case about 28mm x 16mm and 10mm thick, made by Honeywell for central heating valves, with a plastic button that operates the switch, (see below). You can get ones with a metal lever on top which pushes down on the button.

Or even better, file a shallow slot out of the top of one rail and fix the microswitch outside of the track with its operating lever over that slot, in such a way that the train wheels will pass over unhindered, but will push down the lever and close the switch when it passes over. Microswitches of this type are available to switch several amps, so should be able to switch signalling without needing extra relays etc. which a reed switch probably would.

This idea will work for any engine, since they won’t need modifying. And be much easier than messing about with magnets and clearances; trying to get the magnet aligned to switch the reed reliably. And having to do that on all the engines.

Or, if you really want to go the electronics route, a photo-diode or transistor between the tracks looking upwards: Switched when the train goes over the top, blocking the ambient light. No modifications to the engines, or magnets or alignment required, but will need compensation for ambient light levels, and electronics. Will work for any train or carriage, (or any runaway).

Switching speed simply requires positioning the switch a suitable distance before the train reaches the point where the switching needs to happen, taking into account the fastest train speed likely.

[An OtherParticipant @another21905]

Hi, Duncan.

I don’t know if you are ‘electronics’ capable, but another alternative would be a photo-sensor/lightbeam arrangement across the track, interrupted by the loco. It could be hidden in trackside furniture. Getting a bit more complex, I know, but I would agree with the comment that reliability of reed switching would be questionable due to the loco jiggling about.

Going a stage further, there is plenty of Arduino and similar related stuff online, and would also give the facility to introduce delays or operate associated switching/indication.

 

[duncan webster 1Participant @duncanwebster1]

Thanks for the replies, I’m sufficiently encouraged to do some experiments with reeds and Hall effect. The magnet will be on the driving trolley, so I don’t need to equip all locos.

I’ve discounted opto beam interrupt as it could get triggered by something other than a train.

I’ve no doubt that on a dry day the microswitch idea could be made to work, but keeping the rain out would not be easy. I only need to switch 0.5 mA, as the signal logic is driven by a processor. I think I’ll put a 555 or similar very close to the sensor to extend the pulse, as a very short pulse fed into a longish twin core might get mopped up by the capacitance?

I’ll report back when I’ve got some results, but the weather isn’t that conducive at the moment, it seems to rain every time I go to the track

[Author]

Posts