Problem with using two eLinks

[LuLuJo]

I have a problem with my system. It has 2 eLinks each powering separate buses for 1. Track. 2 points And signals. Both have 5 amp supplies. Both were working ok. Then I extended the track and had a short which stopped eLink 1 working. But eLink2 still worked okay. Found the short due to a mistake in wiring and now the track is okay but the eLink controlling

the points seems to have stopped working as nothing will move but it seems to be ok on the system. Tried repeated restarts but always the same. Both controllers seem to be okay but nothing actually operates on the 2nd bus. Checked the ini file attached which seems okay but it’s not quite the same as an older backup. Windows device manager shows both eeLinks on the right ports. Tried a spare eLink as controller 2 but that made no difference. A couple of screenshots attached. Any ideas?

[Chrissaf]

What version is your RailMaster installation? The current version is 1.74.5

Your "railmaster.ini" file image only shows the lower half of the file.

If there is a .INI line "reset eLink on start=n", that is 'off screen', then make sure that n=0

Other than that suggestion, I have no other suggestions to make at this time, there is nothing obvious out of place with your screenshots or text comments.

[LuLuJo]

Thanks. I had checked the FAQ and made sure that the reset elink line was set to 0. My only thought now is that the elink may be defective. I am pretty sure the spare one is.

I suppose I could set up a test with a spare decoder and one solenoid and see if i can get that to work.

[96RAF]

Simply swap the eLinks over and see if the problem follows the unit or if it stays with the bus it is powering. No need to alter any settings.

[St1ngr4y]

Hi,

What information is displayed when you "hover" your mouse cursor over the Controllers icon at the top right of the main window ? ......

Are both controllers shown as active with version 1.07 firmware?

Are you able to measure any voltage on the track output terminals of eLink 2 ?

Ray

[LuLuJo]

Thanks for the various suggestions as I seem to have solved the problem. The fault is with the controller 2 eLink which is not outputting any discernible voltage. It is shown as working correctly everywhere but clearly nothing is coming out. Fortunately the spare eLink I had is working today.

My efforts were not helped by a disturbed wire to the Adfx8 accessory decoder controlling the points immediately in front of the control point.

So I can now get on with adding decoders to the 15 new points at Carnforth.

Interestingly I could not find a new elink for sale anywhere in the UK. Its not listed by hornby so it may be on its way out?

[96RAF]

It never was sold on its own, only as part of an RM package. Since most RM licences are now bought from within the app the eLink will only can be found cheap on the second hand market or on the odd retailers very dusty shelf. S/H ones can be very cheap as the original licence is not transferable to a new owner once registered thus making the licence worthless and the eLink package virtually unsaleable second hand.

Edit - eLink c/w RM is still for sale by Hornby and at the time of writing still in stock.

[LuLuJo]

Hornby presumably wanting to sell an elite at £250 rather than an elink at £100. That's sad as its a nail in the coffin of railmaster. It used to be a cost effective way to get into DCC

[Fishmanoz]

First there was Elite, then RM, then eLink.

So if they wanted only to sell Elite, why develop eLink?

Then when it first came out, it was part of two excellent DCC sets (Majestic and Western Master if I remember rightly) that convinced quite a few to get into DCC.

What changed?

PS. You can find Majestic on eBay but you’ll need very deep pockets.

[LTSR_NSE]

Surely what has changed is RM (i.e Hornby no longer owning/supporting it).

Since Elite works as a stand-alone unit, Hornby are perfectly happy selling & supporting it (from hardware perspective). If anyone wants to use it with a third party piece of software they can seek support from the software ‘developers’ HRMS.

However eLink is solely designed to work with RM - how can (in today’s litigious world) Hornby sell a product that they then provide no support for?

[96RAF]

Odd then if eLink is unsupported that it is available c/w RM and is still in stock

https://uk.hornby.com/search?encoded=lVbNb9sgFP9r6guatOYwbQdrykelHTopWqrtUEUIw6uDQowHuIn

My original post earlier amended to indicate eLink can be bought new.

[LTSR_NSE]

Whilst the checkout appears to indicate 50 are available, the actual product page goes nowhere:

https://uk.hornby.com/products/elink-railmaster-software-r8312

However with this new(ish) website, broken links/info prove nothing, so i’ll defer to your greater insider knowledge @96RAF.

Mod note: broken links are reported to Hornby for rectification as and when they are found, including this one.

[Fishmanoz]

And HRMS has always owned RM, consequently was the case when eLink developed.

When eLink first released, hardware and firmware issues abounded and were handled by Hornby, usually with few questions asked replacement eLink. That doesn't sound like no support to me.

[96RAF]

Going back to the OP I don’t suppose by mistake you have the points bus plugged into the Prog terminals of eLink B instead of Track.

Note that only controller A is capable of programming in RM.

Another thing to check is in Design Mode are all your points designated to controller B, as I have known them to default to A during a reshuffle of my plans - finger trouble at my side in that case.

[LuLuJo]

The bad news is my problem seems to have returned but the good is I think I can see light at the end of the tunnel. I think I have been pushing the boundaries of what the system can do.

First some background: This is a largish layout 24 feet by 13. It has a loop connecting Hincaster junction (a scenic fiddle yard and loco shed) with Carnforth (Loco shed and yards) There is a double track branch to a scale model of Windermere station. There are almost 100 solenoid operated points plus solenoid and servo signals and 80 locos. It has separate buses for track and accessories each with an eLink with 4 amp supplies. The track has a separate power district for Carnforth with a Hornby power booster. The track part of the system has worked fine.

On switching on yesterday I was back to square one with both eLinks showing as functioning properly. The track being fine but not controller 2 on the accessory bus not working to move any points or signals. So I disconnected the accessory bus and restarted the system. Checked that there was voltage on the eLink output. Connected a small section of the bus to the eLink and restarted the system . The points now worked on the section connected. Added the rest of the bus which operated for a minute or so then nothing. Disconnected the Windermere section of the bus and restarted. It worked and continued to work. Today with Windermere still disconnected it loaded up ok.

My conclusion is that on starting up the load on the eLink is too great and it shuts down in self preservation. The accessory decoders for the points all have capacitor discharge to some degree. On starting all have to be charged. My extension to the layout has probably pushed the system over the edge. Windermere is about 30 solenoids and servos and also has some ADSX8 decoders which are known to have a high initial load as there are two capacitors per output. They certainly work very well.

So moving forward I plan:

1. To put in switches around the bus to separate it to allow for a phased switch on and to help find any faults. I have done the same thing on the track bus as removing all the locos to find the one with a defective decoder was a pain
2. I have a spare power booster from when Hattons were selling them cheap and plan to make the Carnforth part of the layout which has around 40 points a separate power district. Hopefully that will solve the problem in the longer term
3. See if the replaced eLink will now work again.

One query if anyone can help? I recall reading that having 2 eLink it can be a problem with reliable start up if the last 4 lines or the ini file are all set to 1 and it was more reliable line if one was set to 0. But can't now find it. Anyone know if thats right.

[Chrissaf]

Your start up charging load deduction is spot on. This issue has been discussed before on the forum and particularly affects the DCC Concepts ADS range CDU fitted decoders. The recommended solution has always been to either add switches to phase the start up charging, or use "Boosters" to create multiple power domains for an all autonomous start-up solution.

As you have two eLinks then the "Alternative coms & Check controller lines" are best set to 1. A 0 [zero] value whould usually only be used on an Elite.

With RM version 1.74.x the key eLink .INI line setting is to have "Reset eLink on start=0". I'm wondering if this is what your memory is thinking of with regard a zero value.

[LuLuJo]

Thanks for confirming that I am on the right track with the issue of initial load on start up. There are four DCC concepts later type decoders in the system and one more to add.

I have read that the eLink and select both have protection built in against an excessive load.

I am convinced I saw a post which said that there can be an issue with two eLinks were the final 4 lines are all set to =1. It said that it was more reliable if one line was =0 but there was no fixed best choice of which line was best set to =0 and to try trial and error. I was surfing the net at night last week whilst looking for solutions but cant remember where I saw it or find it again. It was not the reset elink to =0 issue.

Once I have sorted things out I might try some combinations to see if it seems to make any difference. I did find an old backup version from 2017 which had this solution in the ini file

Mike

[Jim-R855]

I'm not fully up to speed with all aspects of possible way to operate DCC layouts, and what I'm thinking is probably a massive change to the architecture of your layout(s), but...

Do the CDUs need to be powered from your 2nd e-link? Obviously each decoder has to be on the e-link circuit for data and it's own power, but doesn't it simply facilitate switching of the CDU? Could you have another power bus, or even individual power supplies connected to the power terminals of the CDUs, with the decoders connected to the switch terminals? i.e. the decoder effectively becomes a smart relay switching CDUs that are powered completely independently? That way charging the CDUs doesn't draw anything from the e-link. Maybe if you have CDUs with decoders built in this isn't possible?

I'm guessing a bit here, I don't have solenoid point motors, only Cobalt Digital IP ones which for now (small testing layout) I am running from the track bus, but I have looked at a lot of different wiring suggestions before deciding that I could keep it simple for now ,which is where I got the idea about powering such devices seaprately from the DCC control. of course I may be getting mixed up with hybrid systems where the points are all controlled by push buttons or panels and not by DCC at all...

[Chrissaf]

@Jim-R855

In the products the OP is using, the CDUs are fully integrated into the Accessory Decoders and there are no facilities for powering the CDUs independently of the decoder itself. If there was, the OP, as an experienced DCC user, would have used it.

Your Digital IP motors don't use CDUs and work on a completely different principle.

There are solenoid point motor Accessory Decoders on the market that have a 'DCC in' for the control commands and a separate 'power in' to operate the solenoid motors, but these Accessory Decoders ONLY have the separate power facility because they do not have any internal integrated CDUs.

To do what you are suggesting, would need an Accessory Decoder whos output is a relay contact which could then be used to work with an external CDU. I am not aware of ANY commercial manufacturer making such a product designed specifically for solenoid points, as the relay contacts need to be pulsed and not latching. You need an Accessory Decoder with metallic relay contact outputs as all commercial CDUs are designed to function with a metallic contact trigger path [whether that be relay contacts or passing contact manual switches]. Normal solenoid Accessory Decoders output a semi-conductor driven voltage / current pulse which cannot be connected directly to an external CDU.

[Jim-R855]

OK, that all makes sense, as I say I've only worked with servo point motors which have very different requirements - should have realised that you only need a pulse for a CDU. And obviously if all LuLuJo's decoders are integrated it was always a non-starter of an idea.

[Fishmanoz]

Which returns us to Chris having suggested more domains to limit inrush current, with Mike having confirmed, and Chris also having specified the correct INI setting for a 2 x eLink setup, Mike having acknowledged but also saying he might try alternatives to confirm.

Sounds to me like this one is solved?

[LuLuJo]

Final update.

I have tried putting back the initial elink as controller 2 and it now also works fine again. I have split the bus into three sections with switches to allow for section one and either or both of two and three to be on or off. Having any two sections on at start up seems to load up okay. All at once with all three is iffy. But take it in any sequence and there is no problem.

I have also invested £12 in a couple of ac digital voltmeters which I plan to permanently connect on the outputs of both elinks. I am printing a housing for them at the moment. Quick tests shows both outputting 13.5 volts.

So I am pretty confident that this issue is put to bed and it looks like if you go above 60 to 70 cdu operated points problems may occur on start up.

Thanks for all the comments and suggestions

[96RAF]

Be aware that the AC reading you will see on your meters is not a true DCC voltage value as the waveform the meter measures against is not the same as a DCC waveform. Those panel meters are indicative only.

To measure DCC accurately you need an RRAMP meter which as seen shows 15.4v.

[LuLuJo]

I appreciate the readings may not be absolutely accurate but it does let me see in an instant what is happening from anywhere in the room. The first picture shows the position on start up with Controller A to the right and B to the left and just one section of layout for controller B switched in. The second with a second part of the layout added. It then stays at 12.8 with all three sections added. I am pleased with the meters and the 3D printed housing has worked well as it was very quickly designed and printed. I am thinking of adding a pair of ammeters as well.