LDT RS8 inconsistent block detection

[sotars]

On my layout I have 15 LDT RS8s and about 100 blocks. Just finished the block diagram in Traincontroller Gold 8 with all blocks allocated relevant occupancy sensor and stop and brake markers.  Spent an incredibly frustrating couple of hours profiling some locos, setting brake compensation etc then having the devil of a job getting consistent stopping at the stop markers.

 

I realised that a large number of the block sensors were not turning on until the complete loco was in the block as opposed to switching on immediately the contact spot entered the block.  This happened whether it was a 5" 0-6-0 pannier or an 11" Western. Further on leaving the block the sensor switched off immediately the contact spot left the block.  The inconsistency is that for example on one RS8 (No 78) outputs 1 & 3 operated as normal or as expected outputs 2 & 4 operated as above. 

 

All rail sections have power droppers and all block sections have isolating fishplates on both rails.

 

Has anyone else seen this behaviour (the layouts not mine which was at times quite extreme!!!) or indeed have any suggestions as to where I might try to look for a solution?

 

Thanks

Alex

[two tone green]

Hi,

 

Why do you have both rails isolated?

 

Have you checked the operation of the sensors on the Lenz handset to make sure if its a hardware problem or software problem.

 

Also you need to set the memory settings in the memory tab of the properties dialog box to stop it turning off straightway when the block is unoccupied, stops chatter due to0 diry wheels etc. 

 

But first check the operation of the sensors using the LH100.

[sotars]

Two Tone thanks for the reply I do not have an LH100 but an LH90 and a +Smarthand. Just reading the manuals now to see how to do it.

[sotars]

Cannot see in the manuals of either handset how to view the occupancy sensor.

 

Just done a test where the loco (0-6-0) is completely within a block. Occupancy sensor is lit. Move it forward and as soon the first set of wheels contacts the point entry (so 2 wheels on the point 4 wheels still in block) the sensor switches off. Lift the two wheels in the point off the track and maintain contact in block the sensor comes back on.

 

I isolated both rails in one section to eliminate that as a possible cause. All sections have individual droppers though.

[two tone green]

I take it that some form of convention has been established whereby one side of the track is always the side that is gapped as the detector rail and that the opposite rail is the common return. You don't change side for the detector rail.

 

There is no need to isolate both rails as there is a common rail as such all the way around the layout if you follow the above convention and the return rail is common any way on the RS8 due to the way is is configured.

 

Are point electrofrog or insulfrogs. Sorry but trying to grasp the set up, how the layout is wired. How are the points wired to feed them with DCC power. Have you a spare channel on an RS8 to take the output to one of the problem points and feed it from the RS8.  

 

I know that on the LH100 you can interrogate each feedback sensor to see if it is active or not, occupied or not. That is the clue to all this. Sorry, don't have the manuals with me as I'm away from home and have not used Lenz for a few years now so cannot remember the key pushed to get at it on a LH100. No idea at all about the LH90 never did like it.

[two tone green]

Another thought, are all the problem blocks all on one RS8

 

Are you sure you dont have duplicate settings in TC for the detectors?

[sotars]

There are 2 power districts each fed by its own booster. The outer rail is always connected to the detector and the inside rail is the common return (although as I explained I did add a number of isolated sections in one locality to remove that as a possible cause). All points are electrofrog and provide power to all rails. Frog is switched by point motor. Layout operates normally with no shorts etc. The only problem is the occupancy sensors not switching on until loco is completely in the block.

 

Did you mean to wire a point (turnout) to the RS8?

[sotars]

No it covers all the RS8s all RS8s have different addresses and these are unique on the system.

[two tone green]

Possibly the 0-6-0 loco is not the best to test with as the centre wheels may not have pickups on and also the loco may not be properly sat on the track causing the detector to go off and on.

 

A diesel may be better so it is well and truly spread across the block and point

[sotars]

I get exactly the same results with a Western. One bogie in the block, one bogie out then no sensor activity. Lift the bogie that is out of the block from the track then sensor is activated.

[two tone green]

Ok, but what about the addresses in TC are they all unique.

 

Yes, feed the point off an output of an RS8. It is sometimes possible to see the loco change speed as it moves from a tack fed via a RS* to a track fed directly from the controller. The RS8 causes a voltage drop so by affecting speed but it should not affect the detection in the section connected to the RS8.

 

I wired all my layout, track, points, crossings etc via RS8's to get over the speed difference in areas of complex and long sections of pointwork. Also allowed full locking of points when occupied

 

I can say I never had any of these problems but others have had issues which in one case turned out to be a problem with the serial version of the Lenz PC interface.

 

Have you tried disconnecting the booster one at a time to see if this affects the problem detectors in the other power district.

 

Sorry, without actually seeing, being there it is difficult to grasp the full set up.

[two tone green]

Are you sure the power districts are fully isolated from each other and from each other via the RS8's and don't have some sort of fancy feedback through a RS8.

[RFS]

I would strongly suspect that your problem occurs when the loco crosses from track that is not in a block section (eg turnouts) into a block that is fed by an RS8 port.  Track fed directly from the bus on a Lenz system will be 16 volts, but fed through an RS8 it will only be 14.5 volts, caused by the use of diodes to do the detection. So if you have a loco straddling the join then the RS8 will not detect the presence of the loco until all of the pickup wheels are in its section: any pickup wheels on the directly-fed track will draw the power for the loco.  I had a lot of problems with correct stopping at first due to this problem.

 

There are two solutions: either feed the turnout sections from a spare RS8 port, or connect power to them via a set of 4 BY251 diodes.  There is a PDF file describing how this is done.

 

EDIT: here's the link - http://www.ldt-infocenter.com/dokuwiki/_media/de/anschlussbeispiele/page_497.pdf

 

I purchased my diodes from this Ebay seller in Germany (far cheaper than the UK)  - http://www.ebay.co.uk/itm/BY-251-Diode-25-Stuck-/160593866857?pt=Bauteile&hash=item256423f069

[sotars]

Certain the power districts are isolated from each other. One or two RS8s do feed both power districts but each block of 4 outputs is powered from the relevant power district.

 

I will disable those RS8s tomorrow and see if that works.

 

Powered down one power district and problem still there.

 

It must be the wiring. Somehow the connection of the feed to the isolated RS8 detected sections and the feeds to the points is causing this anomaly. It cannot be the common return as I have isolated a section and the problem is still there.

 

As I said I will disable all RS8s tomorrow that feed both power sections and test again.

 

Many thanks for your time this evening.

[sotars]

Thank you RFS for that information. It does tie in with one of TTGs suggested solutions.

 

Of the 15 RS8s 6 are ones that were used on previous layouts without problem and still do not exhibit this behaviuor. They are Rev 2.2 just like the other 9 bought new this year. They are also Rev2.2 but there is a slight change in the layout of the components. Most if not all display the problems as described.

 

I will order a quantity of the diodes and get the soldering iron out again. It is a much cheaper way than buying more RS8s.

 

Thank you for your input and the links, will let you know how it goes.

[RFS]

All my RS8s are rev 2.2 and exhibit the problem.  I've just connected up the diodes using terminal blocks - I'll put up a picture tomorrow to show you how I've done it.

[RFS]

This is how I've configured the BY251 diodes for a block of turnouts in the station throat area.  The bus wire spur comes in at the bottom, and the connections at the top go off to the point motor accessory switches (Tortoises). All relevant wires are blue, the others being for the other rail. The diodes must be orientated as in the picture.

 

Incidentally, although I did have this problem quite a bit, it was intermittent which is why it took time to find this solution. But you should be able to reproduce it with something like your Western, by placing on the track so that one bogie is in the monitored section and the other on an unmonitored turnout.  The RS8 won't "see" the loco until all its pickups are in the monitored block.

 

[sotars]

Thank you both for your help.

 

I have wired up a group of points through the RS8 and the connected blocks do now indeed work as expected. I have ordered some BY251 diodes from the source in Germany via ebay and will install these as shown in your image RFS.

 

Thanks again for your time and the solution.

 

Alex

[two tone green]

Worked out well. Let us know the outcome once you have all your diodes in place.

[sotars]

The BY251 diodes arrived today. I bought some Veroboard and made up the unit as shown. Surprisingly it actually works. This is my first attempt at building a usable accessory. Thank you both (TTG and RFS) for your time and the solutions to the problem.

 

Alex

 

[Edwin_m]

I built a similar 4-diode arrangement into my non-detected blocks as I thought this problem might occur.  Having found out that it does, I'm quite pleased I went to the trouble! 

[two tone green]

Well done Alex. Have fun.

[maidstone]

The BY251 diodes arrived today. I bought some Veroboard and made up the unit as shown. Surprisingly it actually works. This is my first attempt at building a usable accessory. Thank you both (TTG and RFS) for your time and the solutions to the problem.AlexLDT RS8 Current Limiter.jpg

Just found this post but I am confused by how the wiring / diodes are connected on the back of the Veroboard as it is not clear which the strips run on the back of the board. Can you help sort out my confusion.

Many thanks

[Edwin_m]

Just found this post but I am confused by how the wiring / diodes are connected on the back of the Veroboard as it is not clear which the strips run on the back of the board. Can you help sort out my confusion.

Many thanks

Not sure if the poster in question is still around, but it looks to me as if the strips run left to right and there are no cuts, at least in the central part of the board.  It doesn't need all five terminals so I suspect the ones each end are electrically connected via another uncut strip underneath and the multiple ones are used simply because those type don't come in singles.  You then just connect two wires, one to any of the terminals at each end.  My equivalent was based on choc-bloc connectors! 

[maidstone]

Not sure if the poster in question is still around, but it looks to me as if the strips run left to right and there are no cuts, at least in the central part of the board.  It doesn't need all five terminals so I suspect the ones each end are electrically connected via another uncut strip underneath and the multiple ones are used simply because those type don't come in singles.  You then just connect two wires, one to any of the terminals at each end.  My equivalent was based on choc-bloc connectors!

 

Many thanks for your reply and help - using your thoughts I will have a play and see if I can make it so work. I am just going to use the double connectors at end and try a copy you Chocolate Block diagram albeit on Veroboard.