If you are modelling British H0 it is almost certain that your models will come with a variety of motor technologies, and hardly any will be DCC-ready let alone have DCC factory-fitted. It is also likely your model railway layout will be a pretty modest affair, maybe 10 or 20 feet long, and you will be operating one or two engines at a time. You will expect to shoehorn decoders into your engines and to tweak these decoders to control motors designed before DCC was invented. It would be good if you could limit the range of DCC hardware you buy so you don't have to learn the peculiarities of too many manufacturers or indeed sell on too much stuff "bought by mistake".
This post is my attempt to summarise and rationalise my choices of decoders and controllers for British H0. I have only done about fifteen conversions (though some two or three times over!) and there may well be better solutions.
I converted my engines to DCC in a concentrated effort during August and September of 2018, and posted the details in three sections:
I do not claim to have done the work the best way, but I am trying to keep the installations neat and tidy. The models run, with limitations noted. Some run better than others - my initial impression is a cheap decoder in a good model with a heavy flywheel can work well, and similarly a better quality decoder can bring out the best of a cheap motor ... but the combination of cheap decoder with cheap motor is best avoided. The single most useful configuration option for me has been to disable support for analogue operation (CV #29 bit 2), this improves slow running for everything.
I would like to standardise on one make and model of decoder if I possibly can, simply so I have only one instruction book to look at. The Lenz Standard+ V2 is good candidate for this - it has built-in support for five different types of motor, a built-in slow running mode, and comes with an 8-pin plug on a harness. I have used it in DCC-ready models, hard-wired, and plugged into 8-pin sockets I have added to models,
All of the electronic components of this decoder are mounted on one side of its printed circuit board, so mechanical fixing into the model is easy with a piece of double-sided adhesive foam or tape. The decoder is too long to fit across the width or vertically into the height of a British outline (H0) model, so it has to go in lengthwise, flat or on edge.
I have also had success with examples of the 'Zen' range of decoders from DCC Concepts, details are with my notes for on track plant.
Where I have failed with decoders is with 8-pin to 21-pin MTC adaptors. I've bought a couple of these adaptors and achieved nothing beyond blowing up a decoder. I gave up, the decoders are too expensive to do more experiments. I have two engines with 21-pin MTC connectors, and for these I settled on the ESU LokPilot V4.0 DCC. The LokPilot seems to control motors very well with its default settings.
I encountered only one factory-fitted 6-pin connector, and this received a Lenz 'Silver' decoder. I don't need the extra features of the 'Silver' over the 'Standard' but it is good to have a similar instruction book and the same motor options and slow running feature.
Finally, there are the PluX connectors. These seem to have become a standard fitment for current Roco and REE models. For my three PluX-fitted models I bought decoders by Roco, Zimo and Uhlenbrock. All of these work well enough. The Uhlenbrock one is a PluX12 and I could have saved myself a few pounds by buying three of these instead.
Along the way I have bought decoders with other brandings and enjoyed mixed results. Some decoders will not work properly with some kinds of motors, and some decoders come without instructions. For future conversions its seems best to buy decoders with the widest possible application. For a balance of motor control and value for money (and where, for me, the number of function outputs is rarely of any consequence) I will look first at decoders from Lenz and Uhlenbrock for future conversions.
To my mind, there are three broad grades of DCC controller - the basic and stripped down entry-level equipment like the Hornby Select; mid-range equipment like the NCE 'PowerCab' and the Roco 'z21 start', and high-end equipment designed for large layouts.
Typically, the mid-range equipment can be expanded for a larger system, but the way this expansion happens makes some systems more suitable than others for my postulated British H0 application. If we accept that any DCC controller will provide enough electrical power, then the choice of controller is a balance of its feature set, ergonomics and routes for expansion.
I started with an NCE PowerCab. This is a hand-held controller with all of the electronics built into the handset, a plug-top style power unit and an escutcheon to connect these together and to the layout. The handset is pretty chunky and looks like a piece of test equipment, with no fewer than 34 buttons, a thumbwheel and a two-line LCD display. I decided to buy the PowerCab after seeing four of the things in use at a provincial show where there were four DCC layouts and all of them had a PowerCab.
The PowerCab is functional and it lets you set up all of the CVs in the decoders. There are no obvious limits on how you can use it with a small layout, but the ergonomics are geared towards the internal workings of decoders rather than how you might want to run a railway. As an illustration, there are five buttons and the thumbwheel to control the speed and direction of a train, and the display shows you the numerical value of the chosen speed setting; but the thumbwheel has no end stops and there is no facility to identify models by anything except their decoder addresses. The only real concession to ease of operation is a 'recall' button which cycles through the most recent addresses. The PowerCab has some handy features like a display of track current, but you get to this by pressing a specific button six times in succession. The user interface is there, but it it struggles.
Just before I finished converting my models, I bought a Roco z21 start. This was prompted by seeing a Roco wireless handset in a shop, and realising how good the ergonomics of a DCC controller can be. Unfortunately Roco have taken a rather limited approach to marketing the z21 start: they only sell it in sets with a train and an oval of track; and if you model British H0 there is nothing of interest in the sets except the controller. The best solution seems to be to buy the controller from a dealer who will break sets, and I bought mine in this way through a dealer in Germany. The postage was only ten Euros, and the deal seemed better than buying a complete set in the UK and trying to sell on the unwanted trains. Of course, if you want the engine in the set then the complete package looks a lot more attractive.
The z21 start comes as a box containing the control circuits, a plug-top style power supply, and a wired handset called a Multimaus. The names don't mean much in English, and so again the Roco marketing is geared very much for the home market. The Multimaus has half the number of buttons as the PowerCab, a control wheel with a central 'zero' notch and mechanical end stops, and a more detailed LCD display. The system holds a database of engines; you have to create an entry in the database for each engine and having done this you call up engines by name not decoder address.
I find the Multimaus easier to use 'blind' and I suppose the bottom line from my experience so far is, the PowerCab controls decoders while the Multimaus controls a model railway. If you want a wireless handset (I do), the upgrade path for the Roco system means adding a wireless router. Conversely, the PowerCab is already a wireless handset but you need to buy a whole new DCC controller to use it in this way.
Neither controller has a dedicated output for a programming track and this became a serious shortcoming as soon as I had more than two engines on the layout. To begin with I relied on marshalling my engines onto one baseboard and then unplugging the baseboard feed to isolate them from unwanted programming. After a while I bought an add-on board by NCE to give me switching for a programming track. The board detects when the controller is trying to use a programming track and operates a relay to suit. My programming track is a short length of the main line where I can send an engine from any part of the layout. Electrically, the programming track is permanently connected to the controller while the relay cuts the track power to everywhere else. The NCE board works with the z21 start just as well as with the PowerCab.
After I blew up a decoder I measured the track voltage from the z21 start and saw around 24.5V AC on my multimeter. This was in a no load condition; the equivalent voltage from the PowerCab was around 19.5V. One of my decoders (the six-pin one) is specified for N gauge and to be on the safe side I swapped the power unit from the PowerCab to use this with the z21 start. The no-load voltage from the z21 start dropped to around 14 V, and the trains still ran perfectly. Later I bought a new switch mode power supply for the z21 start so I could sell on the PowerCab as a complete system.
I do find the 'advanced consisting' feature of DCC decidedly not advanced at all. I guess it is really for semi-permanent arrangements of multiple engines as used in North America. I want to use consists when I attach a banker to get a train up the gradient on Shelf Island's main line, but the design of DCC has me jumping through hoops. An operation which would need one isolated section and a toggle switch for analogue control (to hold the train engine) becomes a succession of CV adjustments.
But if consisting is a niggle, I am so glad I made the switch to DCC. I am running trains most evenings instead of most weeks, and much appreciating the ability to control the trains directly rather than through controlling the track. With the decoders I have described above, motor control is as good as with my Kentrol Feedback (which I always rather liked), and often better. This was a good move.