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

[Whetstone Bridge]

After many months of research and preparation of loft space, I've finally got a fair amount of track laid and connected up to an interface.  But why can't I get the locos to run smoothly?  They all seem to speed & slow very slightly as they run.  When it's bad, you can see wagons/coaches bunch up then pull apart.  This may be a common/simple issue to resolve but I haven't found a solution anywhere online.

 

I'm using:

• Digikeijs DR5000 (USB connection) & DR4088CS (LocoNet T) connection).  Digikeijs DR4018 (with separate 12v supply to power accessories).  2 feedbacks per block.
• iTrain 5 Pro (trial download) and DR5000 software.  Tried Rocrail but got the same problem.
• 8 pin decoders: Lenz, Hornby, Bachmann.  21 pin decoders: Lenz, Dapol, Bachmann
• Worst decoders: Bachmann 8 pin & 21 pin.  Best: cheap Hornby 8 pin
• Worst loco: nearly new Bachmann Warship.  Best: ancient Bachmann Class 25.

 

Also:

• Running in: I've bought second hand locos that have seen a few miles but all have the same issue
• Track & wheels: spotless
• Happens in all blocks/feedbacks
• Tried a variety of CV adjustments in the DR5000 software and iTrain.  I can mask the issue with changes to speed/acceleration/braking but it's still there. 
• Changes to the BEMF settings make no difference
• Less noticeable on the 1.5% inclines but this is probably due to the extra load on motors
• Code 75 track/electrofrog turnouts/mtb point motors

 

Makes me wonder if it's an issue with the DR5000/DR4088 or the wiring.  These are the only consistent things.

 

Any ideas?

 

WB

 

 

Edited October 20, 2020 by Whetstone Bridge

[DaveArkley]

Is this voltage drop?

Your loft layout is potentially quite large. I didn't see any boosters mentioned so assume the whole layout is powered from the DR5000.

 

It would be worthwhile taking resistance mesurements artrounfd the layout - testing from the DR5000 outouts to each rail. If the resistance varies at differest points around the layout then more/bigger droppers or buses may be required, or even a booster.

 

Cheers

Dave 

 

 

[Whetstone Bridge]

Thanks for the suggestion.  It's not overly large (at the moment) - 6m x 2.5m.  2 loops.  Splits into 6 through a 3.5m fiddle yard.  I've got 26 feedbacks and 15 turnouts connected to the bus.  The number of droppers from the rail via the feedbacks to the DR5000 is greater than the number from the rail direct to the bus.  Perhaps 2 to 1 ratio.  Could this imbalance cause a problem?

[Whetstone Bridge]

Another thought.  I've got 3no. DR4088 in a chain.  That's only 3 connections to the bus.  So 26 sections of track going through 3 connections/droppers.  Perhaps use both connections to the bus from each DR4088 and increase the diameter of the cables?

[DaveArkley]

1 hour ago, Whetstone Bridge said:

Another thought.  I've got 3no. DR4088 in a chain.  That's only 3 connections to the bus.  So 26 sections of track going through 3 connections/droppers.  Perhaps use both connections to the bus from each DR4088 and increase the diameter of the cables?

Yep I'd suggest that for sure.

 

What cable are you using for your bud, and then to each 4088?

 

Cheers

Dave

[RobinofLoxley]

There are public domain recommendations about wire gauge. There are some on here, if I knew where to find them I would post a link.

[Whetstone Bridge]

Assuming there's a problem with resistance and an impact on voltage/current,  how does that cause the problem with smooth running?  Does the decoder just not have enough power (or a strong enough command signal) to turn into a consistent DC flow to the motor?  Or is it constant fluctuation in the supply via the bus that causes a knock on?

 

I'm not an electrician or a DCC expert so I've only got a very simple understanding of how this all works together.  Solving frustrating problems is a good way to get a better understanding though...

[Whetstone Bridge]

26 minutes ago, RobinofLoxley said:

There are public domain recommendations about wire gauge. There are some on here, if I knew where to find them I would post a link.

Oh dear.  I followed some guidance a few months back.  Forgotten where I found it.  Hope I don't need to re-wire the whole lot!

[WIMorrison]

10 minutes ago, Whetstone Bridge said:

...

 

I'm not an electrician or a DCC expert so I've only got a very simple understanding of how this all works together.  Solving frustrating problems is a good way to get a better understanding though...

 

somewhat ironic choice of username then  

[Crosland]

6 minutes ago, WIMorrison said:

 

somewhat ironic choice of username then  

 

Why? Are you perhaps thinking of Wheastone Bridge?

 

There is a Whetstone Viaduct http://www.forgottenrelics.co.uk/bridges/gallery/whetstone.html

 

[Whetstone Bridge]

3 minutes ago, WIMorrison said:

 

somewhat ironic choice of username then  

 

Made me laugh.  Every time I search for info about the viaducts/bridges/embankments in our area, google tries to correct my spelling and tell me all about measuring unknown electrical resistance!

[Whetstone Bridge]

Just now, Crosland said:

 

Why? Are you perhaps thinking of Wheastone Bridge?

 

There is a Whetstone Viaduct http://www.forgottenrelics.co.uk/bridges/gallery/whetstone.html

 

That's the one.  Plus all the other GCR structures around and about.  Basically one big viaduct ploughing across the Leicestershire countryside.

[cravensdmufan]

I had a similar problem when first using DCC.

 

My layout was continuous run "roundy roundy", not end to end.  I installed a continuous bus wire all the way round to which I connected the tracks at numerous places.  Trains would speed up and slow down intermittently, and decoders (all different makes) would sometimes behave strangely.

 

Kevin from Coastal DCC helped me solve it.  He advised me not to have the bus wires continuous, but to split it.  Which I have now conceniently done by means of my lift out section (I used to feed it from both ends, now just from one).  i.e. the bus wires are now in a U shape rather than a complete circle if that makes sense!  It completely cured the problem.

 

I certainly don't know the ins and outs of DCC but possibly the decoders were previously getting digital signals from both directions and were getting confused?!  Others may be able to explain it better.

 

Anyway, delighted with my DCC layout ever since - all decoders work perfectly on it. 

 

I

[Crosland]

2 hours ago, Whetstone Bridge said:

 

Made me laugh.  Every time I search for info about the viaducts/bridges/embankments in our area, google tries to correct my spelling and tell me all about measuring unknown electrical resistance!

 

You can use +Whetstone to force google to include that word. The problem then is you get all the pages where some idiot has misspelled wheatstone

[Crosland]

1 hour ago, cravensdmufan said:

Kevin from Coastal DCC helped me solve it.  He advised me not to have the bus wires continuous, but to split it.  Which I have now conceniently done by means of my lift out section (I used to feed it from both ends, now just from one).  i.e. the bus wires are now in a U shape rather than a complete circle if that makes sense!  It completely cured the problem.

Every time a loco bridges the gap where you removed a feed, you are back where you started!

 

1 hour ago, cravensdmufan said:

possibly the decoders were previously getting digital signals from both directions and were getting confused?! 

 

Doesn't happen. You are up against the speed of light and would need a VERY large layout.

 

It's almost certainly noise causing packets to be corrupted, usually due to poor wiring.

 

To repeat an earlier question, what gauge wires are you using for the bus and droppers? How long are the droppers?

[brossard]

My branch line 0 gauge layout is wired with 22 gauge droppers (from EVERY piece of track) and 18 gauge buss (speaker wire).  Connections are choc blocks (I know there are higher tech connections but these are recovered from a previous layout and they do permit troubleshooting).

 

An example of what I did:

 

 

This is the busiest board, 4 altogether with a sector plate module (21' total).  Tortoise point motors and a pair of Wabbit stationary decoders.

 

I do have the NCE Pro 5A system.

 

Yes, the layout works great.

 

John

 

Edited October 21, 2020 by brossard

[Whetstone Bridge]

16 minutes ago, Crosland said:

To repeat an earlier question, what gauge wires are you using for the bus and droppers? How long are the droppers?

 

Was this a newbie error....

 

Bus:  24/0.2

Droppers:  16/0.2

 

I found this advice on a forum.

 

Droppers generally no more than 300mm.  There are some between 1m and 2.5m where there's some distance back to the feedback monitor that I couldn't avoid.

 

What's the ideal max length before you get problems?

 

[Whetstone Bridge]

Connections to bus with crimp connectors.

[markw]

5 hours ago, Whetstone Bridge said:

Thanks for the suggestion.  It's not overly large (at the moment) - 6m x 2.5m.  2 loops.  Splits into 6 through a 3.5m fiddle yard.  I've got 26 feedbacks and 15 turnouts connected to the bus.  The number of droppers from the rail via the feedbacks to the DR5000 is greater than the number from the rail direct to the bus.  Perhaps 2 to 1 ratio.  Could this imbalance cause a problem?

Does that mean you have some sections with feedback and some without?

I think  the DR4088 is a diode based current detector, this means you will have a lower voltage on detected sections of track.

You need to have feedbacks on all track sections or use some diodes to drop the undetected track voltage to match the detected sections.

[Whetstone Bridge]

All sections (apart from turnouts) have feedbacks.

 

Between turnouts, one rail is continuous (multiple lengths connected with metal joiners) and has 1 dropper.  The other rail is split into 2, for the 2 feedbacks, and connected with insulated joiners.  Therefore one side, there's 1 dropper and other side, 2 droppers.

 

Longest length with a single dropper is 2.5m

[WIMorrison]

This is normal wiring with feedbacks and while some people advocate a need to use a diode or bridge to drop the voltage on unmonitored track sections most people don’t bother and don’t have any issues,

 

you may want to check the track voltage is sufficient to compensate for the dr4088xx drop- 16v works well for me, 17v even better 

[Whetstone Bridge]

1 hour ago, WIMorrison said:

you may want to check the track voltage is sufficient to compensate for the dr4088xx drop- 16v works well for me, 17v even better 

 

Just checked.  Measures 17.8v unoccupied.  Drops to 16.5v when occupied.  Pretty consistent around all areas of layout.

[RFS]

2 hours ago, WIMorrison said:

This is normal wiring with feedbacks and while some people advocate a need to use a diode or bridge to drop the voltage on unmonitored track sections most people don’t bother and don’t have any issues.

 

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. 

Edited October 21, 2020 by RFS

[RFS]

20 minutes ago, Whetstone Bridge said:

 

Just checked.  Measures 17.8v unoccupied.  Drops to 16.5v when occupied.  Pretty consistent around all areas of layout.

 

Have you tried testing your wiring using the "coin test" method? Go round the layout placing a coin across the tracks (or use a piece of wire) and verify that, wherever you do this, your command station trips immediately? If it doesn't, then wiring at that point would need to be beefed up. 

[Allegheny1600]

4 hours ago, Whetstone Bridge said:

Connections to bus with crimp connectors.

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.