Why can't I get locos to run smoothly? Is it the interface, the decoders, cables or locos?

[RFS]

1 hour ago, Allegheny1600 said:

Hi there,

I have heard a lot of bad reports about using these (& a few good ones too!) - I think the problem is they will always make a good connection with the bus due to it's larger size BUT may not make a good connection with the dropper due to the dropper being a much smaller cross section.

I solved this by ensuring that each dropper end was folded upon itself sufficient to make it approximately the same physical size as the bus wire. The clamping pin then hits each wire equally and current will flow.

The shed I was building this in is 32 x 12 and I had no problems.

 

An earlier build saw me getting a similar fault to yours despite my droppers being soldered connections. An electrical expert friend advised me not to blow on the just soldered connection to speed up the cooling, apparently that caused the solder to 'freeze' - creating a crystalline structure that blocked the signals from getting through.

 

Good luck,

John.

 

I used these too but not for individual droppers. Instead I use much thicker wires and connect these instead to Wago connectors, mostly 5-way ones. In this way multiple droppers in a busy location can be connected just to the Wagos, with each port able to accept 2-3 wires of the same gauge. So only a few Scotchloks needed and no soldering under the baseboard.

[WIMorrison]

13 hours ago, RFS said:

 

It certainly can cause issues with trains not stopping in the correct location. A diesel loco, for example, with pickups on both bogies will not be detected entering a monitored section from an unmonitored one (most usually turnouts) until all of its pickup wheels have entered the monitored section. Dread to think how my 4TC+33/1 combo would behave given that the 4TC has pickups on every single coach axle to feed into the decoder for lighting etc. 

 

I'm using Traincontroller which uses "dead reckoning" to calculate where a train is in a block based on its speed profile. 

doesn’t cause an issue in iTrain as anything that draws power shows instantly (or within the sensor switch on delay) as occupying the Feedback irrespective of whether it is coming from an unmonitored or monitored bit of track. Number and length of pickup points on the train also have no impact upon recognition of a train entering a feedback - it is as simple as something drawing current causes the sensor to report occupancy and that occupancy may cover one or more feedback sensors.

[RFS]

1 hour ago, WIMorrison said:

doesn’t cause an issue in iTrain as anything that draws power shows instantly (or within the sensor switch on delay) as occupying the Feedback irrespective of whether it is coming from an unmonitored or monitored bit of track. Number and length of pickup points on the train also have no impact upon recognition of a train entering a feedback - it is as simple as something drawing current causes the sensor to report occupancy and that occupancy may cover one or more feedback sensors.

 

This problem is one of electronics and not software. If your occupancy detectors are using voltage drop diodes, and a loco straddles a monitored and unmonitored section where the latter is wired directly to the bus, then the sensor will not see the loco until all of its pickup wheels are in the monitored section. No way that either  iTrain or TC can see the loco unless the sensor has reacted.

 

What occupancy detectors are you using?

[Whetstone Bridge]

1 hour ago, RFS said:

 

This problem is one of electronics and not software. If your occupancy detectors are using voltage drop diodes, and a loco straddles a monitored and unmonitored section where the latter is wired directly to the bus, then the sensor will not see the loco until all of its pickup wheels are in the monitored section. No way that either  iTrain or TC can see the loco unless the sensor has reacted.

 

What occupancy detectors are you using?

 

I'm using the DR4088CS.  Would you recommend connecting turnouts to detectors?  I was considering doing this for critical turnouts close to stopping points on iTrain train routes?

[RFS]

45 minutes ago, Whetstone Bridge said:

 

I'm using the DR4088CS.  Would you recommend connecting turnouts to detectors?  I was considering doing this for critical turnouts close to stopping points on iTrain train routes?

 

Assuming that the DR4088CS is using diodes for sensing, then for accurate stopping you need to feed unmonitored sections such as turnouts either via a spare port on the DR4088CS or via a set of diodes. On my layout, which uses LDT RS8s,  I have set up connections via 4 diodes to Wago connectors at various places, and all turnouts in the area take feeds via that and not direct from the bus.

 

You could ask Digikeijs about the issue, and they should tell you which diodes to use. 

 

This is the advice from LDT for RS8s: for DR4088CS it will be the same but possibly with different diodes.

 

page_497.pdf

 

[Whetstone Bridge]

7 minutes ago, RFS said:

for accurate stopping you need to feed unmonitored sections such as turnouts either via a spare port on the DR4088CS or via a set of diodes

Thanks for the advice.  I've probably installed too many feedbacks in short sections (as recommended in the Rocrail guidance), so I'll simplify those blocks and reallocate the ports to turnouts.  I'm with iTrain now and it doesn't seem to need so many feedbacks.

[WIMorrison]

6 hours ago, Whetstone Bridge said:

 

I'm using the DR4088CS.  Would you recommend connecting turnouts to detectors?  I was considering doing this for critical turnouts close to stopping points on iTrain train routes?

 

You don't have to put detectors onto the turnouts as iTrain will perform perfectly without feedbacks on the turnouts. It also doesn't need diodes or any other device between monitored and unmonitored bits of track - I don't know anyone who does this and not one of them has issues, certainly not with iTrain but I can't answer for TC because that is a very different beast.

 

I also know that a loco can traverse 2 detectors and it will show occupation on both detectors, it will also show occupancy on a detector when one half of it is in a detected section of track and the back end is on an undetected section of track e.g. a turnout.

 

That said if you want to place detectors on turnouts that is good practice - I do that on all the turnouts that I cannot see, however turnouts that I can see I don't bother having feedbacks on them as I can see if anything is fouling the turnouts. As for place diodes or bridges to reduce track voltage in unmonitored track, it probably wont do any harm but experience shows that they aren't needed either.

 

 

 

[Crosland]

22 hours ago, Allegheny1600 said:

I have heard a lot of bad reports about using these (& a few good ones too!) - I think the problem is they will always make a good connection with the bus due to it's larger size BUT may not make a good connection with the dropper due to the dropper being a much smaller cross section.

 

They've been discussed plenty of times on RMWeb. The bad reports are due to improper use. It's imperative that the correct connector is used (colour coded) for the wire being used. When joining droppers you need a splice connector, designed for exactly this sort of situation, that takes two different wire sizes. Still only works if the dropper is the appropriate size.

 

22 hours ago, Allegheny1600 said:

I solved this by ensuring that each dropper end was folded upon itself sufficient to make it approximately the same physical size as the bus wire. The clamping pin then hits each wire equally and current will flow.

 

That's really not the way to use scotchloks

[RFS]

1 hour ago, WIMorrison said:

 

Itrain also doesn't need diodes or any other device between monitored and unmonitored bits of track - I don't know anyone who does this and not one of them has issues, certainly not with iTrain but I can't answer for TC because that is a very different beast.

 

 

Sorry Iain, but that is just not true. If you still think this is a TC rather than iTrain problem then you are not understanding it. If your occupancy detectors are using diodes to induce a voltage drop for detection, then having a loco straddling a monitored section (that has a lowered voltage) and an unmonitored section (that has full track voltage) will cause problems because the voltage drop is "shorted out" until the whole loco is in the monitored section. Only when the sensor detects the loco will it then upload status to the software which will think the loco has just entered the section. 

 

I uploaded this diagram from LDT for their RS8 detectors only today on another thread, but the same principle applies to any other detectors using diodes. What software you are using is not relevant.

 

page_497.pdf

 

 

[KingEdwardII]

Regarding "Tap Splice Connectors" (as RS Components dub them), otherwise called Scotchloks after one of the leading brands, I have very good experience with them. There are different sizes of connector - and the right size must be used for the wires concerned.

 

My DCC power bus uses 2.5mm2 flex cores (== 14AWG) and I use the Blue RS Components connectors to tap into the power bus with 0.75mm2 (== 18AWG) wires. These tap wires I typically then connect to a pair of terminal connectors which then have two or more pairs of dropper wires attached - the dropper wires are smaller still, to make the connections on the track unobtrusive. The terminal connectors allow me to detach the droppers if required.

 

Meanwhile the DCC accessory bus uses 1.0mm2 flex cores and I use the Red RS Components connectors to tap into this, linking to 0.75mm2 or 0.5mm2 wires to whatever accessory device is involved, such as the Digikeijs DR4018.

 

Using the tap splice connectors is quick - stick the relevant wires into the connector and a quick squeeze with a pair of pliers and it's done. I've never had one fail on me. The bright colours used are also helpful in finding the connections when messing around in the dim light under the board.

 

The tap splice connectors are really cheap - a recent 50 pack of the red connectors was £1.65.

 

Yours,  Mike.

[WIMorrison]

@RFS what I know is that all the layouts I have helped automate haven't bothered with any of these diodes that you say are absolutely necessary and yet every single layout works perfectly. I also know that my personal layout works perfectly without using any of these diodes.

 

I have only come across one other person who said these are necessary and he admitted that he hadn't every tried to operate without them  - when he did he was surprised to find that they are not required.

 

It may be that the DR4088xx and DR5088RC are different from the LDT units and that LDT needs them,  but what I do know that the DR4088xx works perfectly without them on iTrain - but whether TC can work without it I cannot say.

 

 

[Michael Hodgson]

Surely the reason for the advice to fit these diodes is to do with the design of the particular detector?  I infer from the LDT paper that their product incurs a small voltage drop as some current sensing circuits certainly do.

 

You want the same voltage to reach all of your track regardless of whether there is a detector in the circuit.  So if you have 2 diodes worth of voltage loss internally to each detector you ought ideally to introduce a similar voltage loss to unmonitored sections, so that the whole layout is fed with the same voltage.  Yes, of course it should still work without these extra diodes, but you would have a slightly higher voltage on your unmonitored tracks than on those that are detected.  

 

If you were using detectors which do not incur such a voltage loss (eg those which use a current transformer like the MERG DTC-02), such diodes would be inappropriate.

[Michael Hodgson]

I should add the BY251 diodes they recommend are rated as 3A, which I assume is the same as the booster.

 

I would translate Spannungsgleichheit as Voltage equivalence rather than Current equivalence.

[RobinofLoxley]

Despite the fact that in my working career I have seen many examples of really simple things overlooked in the design of industrial machinery and equipment, could it just be that the designers of some equipment items realised that detected blocks would be different to undetected ones ( a sort of DCC version of the uncertainty principle) and either utilise that or work around it?? The solution is never to say here's my kit to do X, but to make it work properly you also need to do Y. 

[BoD]

5 hours ago, RobinofLoxley said:

 The solution is never to say here's my kit to do X, but to make it work properly you also need to do Y. 

 

Some do.  LDT for one.  Whilst this may not be the problem here I have attached their diagram (freely available) in case it is of use.

 

equalizing occupation detection sections.pdf

[RobinofLoxley]

2 hours ago, BoD said:

 

Some do.  LDT for one.  Whilst this may not be the problem here I have attached their diagram (freely available) in case it is of use.

 

equalizing occupation detection sections.pdf 101.65 kB · 3 downloads

I have seen this item fairly recently, if not in the same thread. I also note the document is from 2008, I wonder if its still valid, that is still required on 2020 versions

[BoD]

1 hour ago, RobinofLoxley said:

I have seen this item fairly recently, if not in the same thread. I also note the document is from 2008, I wonder if its still valid, that is still required on 2020 versions

 

Yes, it is posted above.  Apologies for not checking recent posts.

I think that the voltage 'dropper' is fairly generic.

[RFS]

1 hour ago, RobinofLoxley said:

I have seen this item fairly recently, if not in the same thread. I also note the document is from 2008, I wonder if its still valid, that is still required on 2020 versions

 

Yes it's still very much valid because it applies to the LDT RS8s and I have 8 of these in use. 

 

There's a long thread from 2012 on the Traincontroller forum discussing the same issue but with Digitrax BDL168s. According to one post there are different types of detector - 

 

1. One diode voltage drops (0.7 V). Lenz, BDL168 etc
2. Two diode voltage drops (1.4 V). Littfinsky RS8s etc
3. Current transformer types (almost no voltage drop, but can still be shorted out by a loco passing to and from an undetected section).

 

This article predates Dijikeijs equipment so I can't say what type it is and the level of voltage drop if it's using diodes. Plus there can be differences depending on sensitivity, and with smaller locos such as N gauge.

 

However I must emphasize that the issue is hardware-related and what software you are using is not relevant: if the loco is not detected crossing between monitored and unmonitored sections then the status update to the software will be delayed. What effect that has depends on how dependent the software is on exact notification of trains entering blocks. 

[Michael Hodgson]

Yes, it's hardware dependent.  You should consider your controller voltage as well as the choice of detectors.

 

Or to put it another way, if you want to put x volts to your track, and you are using RS-8s your controller/booster would needs to output x+1.4 volts to get x at the track, and you should add the two diodes to any undetected sections. 

 

If you use a mixture of detectors some incurring a one diode drop and others losing two diodes worth, you ought still run your booster x x+1.2 volts, and add single diodes to sections monitored by the detectors that only lose one voltage drop.

 

I don't know what principle the 4088 uses,

I haven't seen the case opened on one, and their site doesn't say anything about its voltage drops if any.

[Whetstone Bridge]

Thanks for all your comments.  Very helpful.  I'm wondering if the loco decoders have a part to play in this, so I'll start a fresh topic.

 

WB

[WIMorrison]

Absolutely certain that they have an effect - some decoders are very good, others not so.

 

Personally I would only use Zimo or Lenz with Zimo as the preference  

[RFS]

On 24/10/2020 at 21:35, Michael Hodgson said:

Yes, it's hardware dependent.  You should consider your controller voltage as well as the choice of detectors.

 

I don't know what principle the 4088 uses,

I haven't seen the case opened on one, and their site doesn't say anything about its voltage drops if any.

 

I thought it would be useful to get this clarified so asked the question of Digikeijs and got this response almost straightaway:

 

"There is always a voltage drop of around 0.7 V at diodes. It doesn't matter which diodes you use for this, as long as they can withstand enough power."
 

As we have strayed somewhat from the original topic I will start a new thread in due course to summarize this particular issue.

[Sir TophamHatt]

Have you tried locos on a rolling road?

While the general consensus does point to the wiring, might be worth confirming it.