Junctionmad

It’s simple 

for single leds the forward voltage drop is typically 2.5V for coloured and 3v ( sometimes more ) for white 

current varies depending on type but these days 5 mA is s good starting point 

hence using ohms law 

(supply voltage - led forward voltage ) / current in amps 

For 12v 

(12-2.5) /.005 = 2500 ohms or nearest. Adjust the resistor value to get difference brightness. 

19 minutes ago, Andymsa said:

it’s covered in the manual.

I read the manual. 
 

the system uses magnetic field strengths to record a pattern of magnetic flux for each given location 

ok it doesn’t then compute position at all. 
 

in reality this is just another “ spot detector “ system. It’s the same as RFID except the decoder is mobile and the tags are fixed. 
 

disappointing really , nothing new to see here and a whole  heap of installation issues 

I can easily see why he has issues with booster track sections. 

Just now, Andymsa said:

according to the railmagic manual the magnets only have a range of 10cms so does this rule out triangulation. Else that’s a lot of magnets  

Ok so the magnet has a simple coding structure to let the loco know it’s at that point. But how does it deal with multiple parallel tracks. 

1 minute ago, Andymsa said:

 

 

this triangulation method has not been mentioned anywhere as far as I can see, if this was the case why are two magnets used for locomotive calibration.

From discussions I’ve seen elsewhere is is not that you position a series of reference magnets and the system uses  a magnetic field sensor to triangulate  

5 hours ago, RFS said:

 

Miniatur Wunderland in Hamburg uses occupancy detectors and IRJs just like we do in Traincontroller and iTrain, so if it works for them it must be accurate enough which is certainly my experience.

 

It's not just about knowing where every train is, it's using that information however obtained to schedule the trains that's important. It's the software that does all the work really, and something Railmagic do not understand in the way they dismiss existing software products. And no, I don't wish to run my layout from a smartphone.....

Not supporting railmagic but a workable system that can accurately determine geographic position has significant advantages over a known fixed point followed by dead reckoning  

1 minute ago, Andymsa said:

I have not seen any such claims about an accurate position once per second. The geographical positioning is exactly the same as current systems that a fixed point is passed and the systems works out how far to travel. 

That doesn’t seem to make sense. It uses a series of reference magnetic fields to triangulate a position it doesn’t need to know “ how far to travel “ as it can Re compute position 

30 minutes ago, NIK said:

Hi,

 

However RailMagic seems to promise an accurate position at least once per second and with no need for position sensors on the layout.

Geographic positioning (X and Y) is not needed if the tracks are fixed, just the distance along the track. 

 

Regards

 

Nick

That means you need a computer tracking progress across the route. That means you need a full iTrain style system. 
 

no reason if railmagic can resolve in X that it can easily resolve in Y. All you then have to do is run a train over every track route and record a series of positions.  

4 hours ago, Andymsa said:

Something I forgot to add, there is also the problem of shoe horning in a further decoder with railmagic, its bad enough in OO/HO but N gauge forget it.

Oh I fully agree. Which is why railmagic now has to produce a full DCc decoder and of sound with no sound either. 

5 hours ago, Andymsa said:

 

Although technically you are correct that geographic positioning is better there are no systems that employ this method, both rail magic and existing systems are Identical in estimating from a fixed point so no advantage to be gained. Additionally railmagic requires an isolated section between power districts. I would say that even thought existing software is estimating the position of the train, it does a pretty good job of it.

I understand railmagic triangulated from a series of magnetic sensors. Hence it should be capable of calculating its current position at snub time and does not need a single reference point. 

I also suspect the DCC is being modified to feed back position information rather like Zimo do with their HULU speed protocol , the issues with coupled track boosters would suggest that the issue as the signal is corrupted when fed through two boosters 

I should add that precise geographical positioning solves many problems that conventional computer automation can’t easily solve 

the first problem is that automation software doesn’t know what’s actually happening , it’s “ modelling “ what’s happening , precise position feedback is actually relating what is really  happening 

precise position reporting is superior to point detection mechanisms and modelling profiles .  The issue is how the system can generate those precise positions and that the proposed solution works and can be handled by ordinary model railway users and compatible with existing systems. 

As I understand it Railmagic works by triangulating positions from known magnets placed in a pattern around the layout. The detector board on the loco has a magnetic field sensor. 
 

what  not clear is how you are expected  to translate that physical position into say a block occupancy signal , surely you would have to map all “ useful “ positions on a layout and record them then access that info and relate it to a track diagram. 
 

secondky if it’s used precision magnetic sensing , surely it will be disrupted by motor fields , uncoupling magnets and so forth. 

On 26/03/2021 at 14:55, AndyID said:

The direction switch is attached to the tiebar. Basically it's a divide by two circuit. Every time you push the button it changes state.

Oh that wasn’t clear from any discussion to date 

how can the circuit illustrated actually follow the tie bar, as I understand it when you switch the " direction" switch the LEDs change immediately , even though the button to fire the point motors hasn't actually been pressed ?

Interestingly, the Roco booster manual is entirely quiet on the issue of grounding **  seperate boosters , even though it’s exactly what  the issue is. 

** by grounding , I mean establishing a common 0V reference on the power supplies to the boosters. 

All the power supplies to any device feeding the rails needs to have a common 0V to prevent return currents finding incorrect return paths. 

My only comment would be that the direction leds will be wrong if the operator ever forgets  to push the “ activate “ button , the usual way these circuits are arranged is to use a latching  relay driven off the coil pulses , hence the led follows the point motor , not the switches 

you could also replace the led relay with a simple two transistor set/reset toggle  from the pm drive which would do the same thing 

We use lots of tam valley juicers on our O gauge club layout. You should see the same voltage measurement as on plain track on the frog ( after a loco has run through the point ) 

I still have my duettte , I still love the big chunky knob. I actually 3D printed some replica ones for other DC controllers. 
 

however the fact is the duette controller itself is very basic , and is easily bested by a simple electronic design from jelly bean components. 
 

good for nostalgia value , that’s about it. 

What do you mean by ” rattling across “. What’s actually happening 

What’s the purpose of the spring on the gauges 

Just in relation to ABC. You can get issues where metal wheels in the stock following the loco bridge the ABC track section and non ABC sections , causing errant behaviour and failure to stop. 
 

I’ve had to place very short insulated sections in between the two sections to prevent such issues , but it’s not a complete success , double heading ,carriage light pickups can all cause confusion 

Just by way of illustration. My friend has a large OO gauge fixed layout , 18by 12 , multiple levels all code 100 insulfrog 

it was DC with conventional isolating sections etc. He connected the dcc feed to the track and closed all the section /isolating switches 

dcc working fine now ( nce ) for several years 

is it best practice dcc wiring , god no. Does it work , absolutely 

dcc does not require u to do anything special if U don’t want to. 

3 hours ago, spikey said:

Gosh, this is all good stuff

 

Ref that points controller circuit, thank you but as I understand it I'd need one per switch, which with 27 switches is hardly feasible.

 

ETA - It's just occurred to me that some of my solenoids are connected to Gaugemaster GM500 relays for frog switching and signal activation.  These are fitted adjacent to the solenoid rather than the switch.  Could this have any bearing on switch failure?

No you just need the transistor circuit, per point , the capacitors can drive as many points as you need , the leds can be removed if not required 

hence you have a few jelly bean components per point 

to derive the advantage of low currents through the switches you do need a 2x transistor for each point 

1 hour ago, spikey said:

 

And where, pray, might a chap find such circuit designs?

 

 

No space on the panels, which as I stated have to be A4, and yes.  It was a waste of time.

All you ever need to know about CDUs. Near the end  there’s  a nice circuit that uses transistors to switch the CDU rather then the switches directly ( choice number 9) 

http://www.talkingelectronics.com/projects/75 Model Railway Projects/75 Model Railway Projects.html#Points PartB

quite a nice circuit as it uses standard toggles and provides mimic panel leds as well 

Shameless self promotion 

you could build my dcc dropper boards, Optional dcc indicator led supported 

£3.50 each plus P&P ( not commercial , just covering costs ) simple to build connectors included etc.