RAF96

Hornby Power bank has 3 x 1F super caps in series, a transistor, a diode, three resistors in parallel, a zener diode and another resistor on its pcb. Charging is also controlled by further components on the decoder and is triggered from the decoder MCU after a time lag.

If you chop off the Ae the decoder will not be able to talk to the app, hence no updates of firmware or sound profiles. If you want to bin the OTA bit then you may as well go with a different make sound decoder loaded with your preferred profile.

Previous talk about excessive run-on into the 'bush' after a derail, reminded me that my layout has a removable section for access, sensibly fitted with a power interrupt isolating the track sections either side of the gap when open, which of course my Hornby power bank (PB) equipped locos with their long non-adjustable run-on totally ignore. The same thing happens if a derail of another loco trips track power, these PB locos carry on regardless until the on-board charge is depleted.

To put it in Noddy terms - the controller shouts out number 3 move it. All number 3s will move. The controller shouts out 667 put your lights on and only 667 reacts. Unless you are running say a power car and dummy there is no need to have decoders with the same address, nor is there any need to have a decoder on address 3 after you have tested it is working out of the box. As your controller is capable of writing any address short or long without drama I suggest you give each loco a suitable associated number

The subject of a separate Ae has previously been discussed in various forum, but the design of the next generation is already set with an improved Ae sizing. Jenny Kirk has done comparative installations and found that she could get these decoders into almost all her locos and she noted that the 21 pin variant is smaller than some big name decoders. Ref her channel for those details. Ref running on DC, this is well covered in the decoder manual and recommended for moderate steady state running-in purposes only, for the reasons stated. General stop/start running is not recommended due to need for the decoder to boot/reboot at each throttle on/off/fwd/rev selection making for poor control, the same forvany decoder equipped loco on DC. Added to this is the chance of some DC controller waveforms spiking a decoder at high throttle setting.

Be aware that the need for a booster depends upon traffic density on a layout. If you split the layout into power districts without taking account of this you can end up no better off - e.g say if all your traffic is still in a large depot yard.

Confused - you will be... If you never alter the default short address by only writing a long address then you will readback value 3 from CV1. Similarly if you never write a long address then a conversion of the values found in CVs 17/18 (values 192/100) will give you the default long address which oddly is 100.

Whilst neat wiring is pretty it can lead to problems with cross talk from one bundle of disparate use wires to another leading to unwanted signals being induced into adjacent looms. Twisted pairs helps but it is wise to keep track bus and points bus separated, and especially from feedback or sensor wiring. If you must run looms adjacent look at routing them so they cross at right angles to each other then follow their own route home.

Generally left and right rail (in the direction of travel) refers to that rail which has the most positive bias, i.e. the rail opposite to the one that has the break that the module is installed to. The diode pack dropping the apparent voltage of this side more than the other ungapped side. Simplest way is to set the decoder to react to both rails for those trains not required to stop one way only.

A similar disparity has been noted with the Hornby HM7000 series decoders twixt Bluetooth running and DCC direct running.

Neat can be a problem; when routing wires try to keep feedback looms away from track bus and point motor bus runs to avoid cross talk. Twisted pairs will help and if close routing of looms cannot be avoided cross each bundle over the other at right angles.

The ESU site has (used to have when I had an interest) a whole series of idiots guides and links to sample files to download and use as a learning aid. In my opinion relying on ‘tubers for info can be the blind leading the blind, especially when most manufacturers sites either have specific product info or a dedicated forum populated by folk who actually know what they are talking about.

CV127 is stop distance default 25 so value 0 would be a dead stop.

Using RM and your eLink in DCC.

Try a decoder reset - write value 8 to CV8 then try to load the Class 08 which is the smallest file and most likely to load thus proving the decoder and loco, etc. If good go back and reload the 56. Hopefully all will be well.

If you cannot delete a decoder try 3 times and you get a force delete option, that gets its attention. It could simply be a duff decoder.

Vin and Vout with four big capacitors suggest a CDU.

You have to delete the decoder from the old controller, which also means unassigning its loco. Then go link it to the new device. A reset will normalise CVs and leave it linked to the old device, but not over-write a sound profile. Are you using the latest sound profile as updates have just been released.

2MT sound profile is on the live server and associated sound function sheets will be loaded onto the support site area for all active profiles. Meanwhile here is the latest status list. Note the version state explanations as amended.

Due in country mid/end Jan, hopefully testing will be complete by then to coincide.

I am sure I have explained this before. The app uses your internet connection to fetch the sound profile from the Hornby server using your login as authority. Your handheld then loads the profile onto the decoder, hence reliability and timing of the former depends upon your broadband speed and the latter upon your handheld device spec. 2MT profile is up for final checks and a future update will show an ETA date for future profiles, which will replace the ‘paper’ list. The app will also make provision for listing international brand profiles alongside the Hornby 00 and TT120 profiles.

Although the DCC address still shows in loco settings after a profile reload it has to be over-typed to write it back to the decoder. If it is a long address the glitch is translating the wrong values into CVs 17/18. You can read these values, back convert them in a long address calculator and use that wrong address to control the loco. As stated it is being fixed.

Any idea where Hornby actually make these components. Maybe they are bought in like everyone else’s are.

You have to be content that the point has changed before setting the signal, as opposed to having been selected to change. The only way to be sure is to control your signal from a switch on the point motor.

Mine is a four channel model and the DCC serial decoder works on that. I wrote a follow on article for Peco-tech using the Elite. https://www.picotech.com/library/application-note/dcc-demonstrating-the-Hornby-elite-using-picoscope The critical part of the analysis is writing a link file which translates the binary signal to plain text.