Hornby B17/6 Gilwell Park


Another loco aimed mainly at our extended "Little Histon" layout, to enable post war period running through the station.

Hattons photo of 61646

Likely configuration:

If the motion is smooth enough, I would like to use a Zimo decoder..... in the absence of a dedicated B17 sound project, I'm considering use of the Paul Chetter A3 protodrive software as previously fitted to my Scotsman. The B17 also has 3 cylinders, with similar valve gear to the A3 and I expect, uses a similar whistle etc, so except for the impact of the single chimney, the sounds should also be fairly similar.

A 28mm speaker will be fitted in the tender, probably with a Zimo super cap stay-alive device. I'm hoping to add a "stopping train" front lamp (top centre) plus red & orange LEDs in the firebox. 

I'll also add a footplate crew.


Another Hattons  shot


Issues to sort out?

There seems to be a common issue concerning slipping driving wheels on Gilwell Park. (An enthusiast friend sent back 3 units to Hattons, before finally receiving one that did not display this problem.)

Surprise, surprise!  My unit suffers from the issue too!   However, the problem was simply due to restricted vertical movement of the front bogie, resulting in the front and middle driving wheels lifting.  Removing the bogie and bracket assembly, revealed a bolt passing through the slot in the bracket, surrounded by a helical spring between the bracket and a set of plastic washers, which rested on top of the bogie frame, while the bolt was secured in a captive nut embedded in the centre underside of the bogie frame.  On flat track, the spring was in full compression, with the bogie unable to retract upwards further due to the thickness of the washers. There were two plastic washers, so I removed one of them...... and that seems to have resolved the problem.

The tender is very light despite the usual Hornby ballast weight, so I can anticipate rather unreliable electrical pick-up performance from the tender wheels. I will endeavour to add weight when I fit the speaker and decoder inside.


Tender Modifications:

Its not easy to obtain good isolation between the front and back of the speaker if using the Hornby speaker mounting arrangements. So instead, a 28mm speaker is mounted, downward facing in a new sealed enclosure, built into the upper rear section of the tender. The speaker wires pass through sealed holes into the coal hopper.

The original 8 pin socket is discarded. The motor and track connection wires are connected instead to an ESU 21 pin adapter PCB, mounted in the coal hopper. Motor and track wires are routed through a hole cut into the bottom of the hopper. The speaker wires are also soldered to the 21 pin adapter board.

Additional ballast weight is added using "liquid gravity" steel balls.


Showing the original motor and track wires plus the new speaker cables, soldered to the 21 pin adapter PCB.


Later, a sub-miniature 4 way connector socket will be glued to the underside of the tender to provide lighting connections to the locomotive. The socket will be connected to Common positive, headlight and FO1 pads on the 21 pin adapter board (via the hole in the bottom of the coal hopper).

But first, with the 21 pin interface in place, I can plug my Zimo Scotsman decoder in and find out if the B17 is compatible with its motor control system....... If I find jerky slow speed behaviour, I may need to consider using a LokSound V4 solution instead!

Scotsman Zimo MX644D plugged in.


Well, that went pretty well!  With the controller set to 128 speed steps, just select speed step 40, and with a fairly high setting of CV3, the loco starts and smoothly accelerates, with just an initial hesitation if its just changed direction, while the motion slack is taken up. There is an occasional random glitch, which I think may be due to the lack of stay alive capacitance (no room in the small tender) or maybe just indicates a need for running in?  I'll look for space inside the loco that might take a stay alive capacitor, when I've given the loco a run around the loft layout.

Very successful loft layout run!  With 4 Hornby Mk1 coaches, the loco runs very well!  Good smooth behaviour, almost as good as the Scotsman...... so that's now 2 out of 4 Hornby steam locos running well with a Zimo decoder!  (and to be fair, my D16 is only marginally better using its lokSound V4.)

So the next step is to source a second Scotsman Zimo decoder for the B17. While I wait for that to arrive, I'll take a closer look at the loco, to see how I can add the lighting and if its possible to accommodate some stay-alive capacitance within the boiler, as there doesn't appear to be room in the tender......

...........No luck there!  A ballast weight is firmly glued into the rear of the front boiler barrel. I can't shift it to gain access inside the smoke box, so there's no room for a cap in the loco either.

Hmmm....... maybe there is just room to slip a Zimo 6800uF super cap in beside the decoder. I'll have to cut a small hole where the corner of the cap needs to penetrate the speaker enclosure, then seal the gap between the enclosure wall and the capacitor. Had to hack the coal load, but I think it will JUST fit!

Capacitor glued in place.


B17/6 Circuit Diagram.

A zener diode protection circuit is used to prevent excessive charging voltage being applied to the capacitor.

(A pair of 6V8 zener diodes, in series with a normal silicon diode restrict the charge volts to circa 14.5V.)


New connector, protection circuit and decoder now on board.


Showing new connector socket, glued to tender underside.


Modified coal load fitted. (Just the Hornby brake linkage moulding to add now.)




Lighting Arrangements:

The boiler barrel weight stops short of the funnel, so there is room to run LED wires from a centre top oil lamp, through the smoke box and down through holes drilled through the smoke box base to the loco underside.

There is also room behind the firebox door, to mount orange and red LEDs, plus all the LED series resistors. The tender micro-connector wires can be routed to the rear under the footplate floor.

Oil lamp glued to top bracket.

A DCC Concepts LNER style oil lamp was used. However straight out of the pack, light leaks through the sides of the white plastic moulding. So before use, the lamp casing was painted matt black, then over painted with matt white, to provide a light proof barrier. Finally, a drop of yellow water colour paint is applied through the lamp lens to give more of an oil lamp colour to the otherwise pure white LED light.  The red wires visible below the lamp, will be touched in with matt black.

Showing the lamp wire routing.

The chassis is a flush fit below the smoke box, so the wires are run through shallow grooves, cut into the plastic of the upper body shell, until they emerge below the running board, where they are glued in place and routed to the rear of the cab underside.

The red & amber LEDs are mounted in a small hole cut into the firebox doors.

The red LED provides a continuous red glow to the cab, while the brighter amber LED, is synchronized with the coal shovelling sound, and flickers every time this activates. 


Showing the other side of the cab front wall, with the loco inverted.

I have used the small space between the rear of the motor and the cab front wall, to mount the three LED series resistors and to interface with the 3 way cable that goes across to the micro-connector on the base of the tender.


The new tender connector plug (with one spare way) before matt black paint is applied to visible part of the cables.


Decoder CV Adjustments:

I've sourced Paul's active drive A3 sound project on an MX644D from Digitrains for the Class B17.  A few CV changes are required to bring the motor control into line with my original Flying Scotsman decoder that I used to check the B17's behaviour.  Also I needed to change the lighting CVs to enable the headlight output to stay on in reverse. Finally, I discovered that contrary to the info sheet, the Safety Valve lifting sound was non-operational via button 10 (although it was included in the random sounds).... a few more CV adjustments resolved this issue.

Not listed below, I also had to add 1 to CV29 to reset the direction in line with the controller.

Drive adjustments:

CV3 increased to 90

CV9 set to 87

CV57 set to 120

Oil Lamp on in both directions:

CV34 set to 1

CV126 set to 0

Assign Safety valve release to key 10:

Set CV540 to 143

Set CV541 to 45


Set CV395 to 90 (max volume)

Set CV1 to 47  (DCC Address)

Reduce CV267 to 54 to achieve 6 chuffs per wheel revolution (Gresley 3 cylinder beat).


Adding the bagged items and re-testing:

Having added the front steps, wheel guard struts, cylinder drain pipe assemblies and brake hoses, I resumed testing.


Motor control:

The loco behaves very well in the forward direction, but exhibits regular "stuttering" when moving slowly in reverse. The stuttering seems to be related to the driving wheel rotational position. I think this points to cyclic resistance from the rods and other aspects of the motion. I've carefully lubricated all the motion linkages and I can't detect any "pinching" within the motion as the train moves slowly in reverse. An initial period of running in has not made any significant improvement. So the actual reason for the brief instances of hesitation remains a mystery. Particularly as the problem is absent when the loco is driven forward at similar speeds. (The stay alive capacitor system is working well, so the cause is not wheel contact related.)

The loco is travelling at what I would regard as full speed, with the throttle only set to around half way, so I will experiment with CVs 57, 3 and 4 to limit the top speed of the loco while maintaining similar rates of acceleration and deceleration.


The front oil lamp and firebox LEDs all behave as planned.

Negotiating the curves:

Although the loco can successfully negotiate the curves on the loft layout outer circuits, without fouling of the bogie wheels on the cylinder drain pipes, manually sliding the loco through the tightest curve on the layout (no power here yet) resulted in a bogie wheel derailment. Plan A is to improve the geometry of the problem corner, to increase the worst case curve radius. Plan B is to truncate the pipe assembly just behind the forward steps, to allow the bogie wheels to move the flexible drain pipes outwards when necessary.

Postscript: Plan A done (tick) The worst case curve radius is now much larger. But.... plan B also done (tick) to minimise danger of similar issues on our village layout.


You've read the book, now watch the video!..........

(Click the line above to see the You Tube video.)



Supplier website links:
Hattons    B17 Sourced from Hattons
Digitrains    Source for Zimo stay alive super cap and DCC LNER headlights


Click to move back to Model Rail Index Page