Hornby LNER D16

 

The D16 was initially sourced for possible use as an Analogue DC controlled loco on our newly acquired village layout (Little Histon).  However, it will have limited interest in its original condition, so I have decided to fit DCC sound and lights after all.

 

Likely configuration:

DC control of the locomotive is quite good. Very smooth in the forward direction but a shade lumpy when in reverse. I think its good enough for the motor control on board a Zimo DCC sound decoder to transform into near perfect behaviour.

Digitrains offer Paul Chetter's Active drive D16 Zimo sound project.  There is provision inside the tender to mount a 28mm diameter speaker, so I think this will be the approach. 

The application will be stopping passenger services, which have a single lamp mounted centrally near the top of the boiler front. I will fit a modified DCCC LNER lamp in this location. (Modifications consist of a coat of black paint to stop light leakage, followed by a coat of white to restore the correct external colour. Also a drop of yellow water colour paint over the lamp lens to give the correct oil lamp tint.)

Fixed Red & flickering amber LEDs (synchronised with coal shovelling sounds) will be fitted behind holes drilled in the fire box door.

 

Hornby D16/3 as received.

 

A few issues to sort out:

First a minor cosmetic issue: Paint chip on the left hand cab side, immediately under the cab roof, over the doorway (visible in the photo above). To be fixed with careful application of red & black paint.

The Westinghouse brake cylinder on the right hand side of the boiler just forward of the cab is tilted to the side..... successfully bent back into a vertical position.

The centre wheels of the tender do not rotate when the train is in operation. This one proved a bit more tricky to resolve. The wheel set relies purely on gravity to drop down within its mounting slot. Even with minimal side pressure, the electrical contact fingers provide sufficient friction to stop the wheel set dropping reliably on to the track. I concluded that because of this, there was no chance of the contacts actually drawing current from the track, so bent them away from the wheels. Without the contact fingers, the wheel set now drops down to the railhead, makes contact with the track and hence all the tender wheels now rotate (at least some of the time) when the loco moves.

A test run with a DC power supply revealed smooth forward motion starting with circa 2 volts applied but rather less smooth reverse motion starting at around 2.7 volts with a little hesitation. However continued reverse movements seemed to indicate a small improvement, suggesting that a period of running in may sort this issue out.

I think the behaviour is within the ability of the Zimo BEMF motor control to make good, even without further improvement.

The detail parts were then glued in place:

 

The D16 with detail parts glued into position.

 

 D16 on DC load pull test. (3 Mk1s on the flat but only 2 Mk 1s up the slope without wheelslip. Including a brake coach with extra resistance due to wheel contacts on both bogies.)

May investigate options for adding weight inside the loco body moulding, when I fit the front oil lamp and firebox LEDs!

Yes, you are quite right! Mk1s weren't around before the war, but at present I have no appropriate LNER teak stock.

     

Next the sound and lighting modifications.....

 

Function Map:

Key Number Function
0 Front oil lamp & firebox red glow
1 Sound on
2 Brake
3 Short whistle
4 long whistle
5 Light engine
6 Coal shovelling + amber firebox flicker
7 Steam injector
8 Blower
9 Flange scrape
10 Safety valve
11 Handbrake
12 Tender water filling
13 Coupling
14 Buffering
15 Cylinder drain cocks open
16 Air Pump
17 Whistle
18 Guards whistle
19 Fader
20 shunt mode
   
27 Volume down
28 Volume up

 

Some CV adjustments :

Issue CV Number Value
To keep oil lamp & firebox glow on in reverse direction 34 1
Reduce volume of long whistle 523 100
Reduce volume of air pump 559 15
Reduce volume of  whistle 562 100
Improve very slow running (BEMF sampling) 9 95
Improve very slow running (start volts) 2 10
Improve very slow running (40kHz) 112 32
Improve very slow running (BEMF params) 56 32
Improve very slow running (extended sampling) 147 0
Improve very slow running (experimental) 148 30
Reduce acc delay to camouflage jitter CV3 40
Reduce dec delay to camouflage jitter CV4 40
Make speed characteristic linear 5 250
Make speed characteristic linear 6 125
Reduce max speed 57 80
To reduce vol of random air pump: read 558 for sound # =82    
check:744,747,750,753.756 etc for sound # the +1 for vol 744 15

The sounds all work very well and once the speed is up to 8/128, the loco runs smoothly.

But: There is pronounced jitter at extremely low speeds (<8/128). Although I'm unlikely to want to operate the loco that slowly, the jitter is noticeable when the loco accelerates away from a standing start or returns slowly to a stand still.  I can camouflage the effect to a degree, but I haven't found a way to prevent it. This is bad news as its a similar result to my experience with my re-motored Class P2 steam engine. Based on a sample of two, I'm now having my doubts about the Zimo decoder's ability to control Hornby steam loco drive mechanisms at very low speed.

In the end, I had to fit an ESU LokSound V4 to the P2 to cure this problem, and at present I think its likely that I will have to take similar action to get my D16 to run smoothly over the entire speed range. Which is a great shame as the sound projects in both Zimo P2 and Zimo D16 are very good!

I shall continue the D16 story when I've listened to the only V4 D16 sound project I could find. Depending on the result, I will either change decoders or continue to work on camouflaging the Zimo low speed vibration.

 

Very simple lighting circuit.

 

Front Oil lamp

 

Reconstructed Smokebox door handles

Firebox red glow

 

Additional flickering amber during coaling

 

 

Trying a LokSound V4 decoder in an attempt to reduce the start/stop judder:

I've now fitted an Olivia's Trains D16 sound project equipped ESU LokSound V4. The slow speed start characteristics are far from perfect, but I think they are a sufficient improvement over the Zimo results, to mean that a change to LokSound is justified.

The LokSound V4 has two CVs dedicated to minimising jitter at extremely slow speeds. (CV51 & CV52). This is a feature that Zimo do not offer.

Having optimised the motor control CVs, the minimum LokSound speed step is significantly faster than the minimum offered by the Zimo solution..... and perhaps because of this, there is much less jitter as the loco slowly accelerates from a standing start, or when it comes to a halt.  However, some of the problem is probably due to the Hornby mechanics, as the performance is not entirely consistent !

Another peculiarity of the LokSound V4 is interaction of the steam sound chuff rate and the motor control settings. In particular, if CV2 or CV53 are set too low, the chuff rate can not be brought down to the correct frequency.

Olivia's Function Map:

Key Number Function
0 Front oil lamp & firebox red glow
1 Sound on
2 Playable Whistle
3 Fixed duration Whistle
4 -
5 Coal shovelling * + amber firebox flicker
6 -
7 Guards whistle
8 Whistle toot
9 Buffering *
10 Coupling *
11 Flange scrape
12 Tender water filling
13 Rail clack
14 Bird Song
15 Reduce volume
16 Safety valve
17 Steam injector
18 Westinghouse pump *
F5, F16 & F17 random Volume via 2/451 (0-128)
Auto Brake squeal Volume via 2/459 (0-128)

* = volume wound down from original settings

 

Motor control & related CV settings :

Issue CV Number Value
Start voltage 2 3
Acc delay 3 150
Dec delay 4 90
Speed linearity 5 255
Speed linearity 6 128
motor control frequency (40kHz) 49 19
BEMF control low speed 51 3
BEMF control low speed 52 20
Overall speed range 53 50
BEMF control low speed 54 20
BEMF control low speed 54 18
BEMF active speed range 55 128
Variable motor frequency (enabled) 124 4

 

Performance in the reverse direction is now pretty good with little sign of vibration..... However, behaviour in the forward direction is not quite as good with a little judder apparent at around the 6/128 or 2/28 speed step. However if this speed is avoided as a steady setting, all appears well. The different behaviour depending on direction, suggests a mechanical issue occurring at the critical speed described above, in the forward direction only. I'm continuing to exercise the loco near the problem speed area, in the hope that behaviour may improve as a result of further running in.

After running for a while under PC control, forward starting and stopping is now reasonably smooth in a straight line, but is still a little jittery on a curve.

The values applied to the motor control optimisation CVs are not very critical. CV51 has the biggest effect, dramatically reducing step 1/128 and 2/128 vibration.

 

Adding an electronic flywheel: (AKA Stay-alive capacitor)

Not enough room in the tender for conventional electrolytic capacitors of the kind of size needed......but room for a tiny Zimo SC68 super-cap (6800uF but only 15 Volts maximum).

Because of the low maximum voltage, I've added a simple protection circuit, to ensure that if the D16 is powered from a higher voltage DCC controller than mine (up to the NMRA limit of 22 Volts) the capacitor will never be charged at a voltage as high as 15 Volts.

If I had a 14 Volt zener diode, the horizontal 1N4007 (cathode to Cap +) would not be needed.

Unfortunately I had to select the lowest value 15 Volt zener left from a pack of 50 x 5%, 1.3W types and it was still a shade too high.

The extra diode drop keeps the capacitor voltage within specification, even if a DCC controller with a higher output voltage than my own Lenz unit is used.

 

Glued into the rear of the tender.

 

Click here for a brief You Tube video demonstrating the D16 on the loft layout.

 

January 2019 Postscript:       A complete cure for the motor jitter problems.....

A second Hornby D16/3, this time in early crest BR livery, was obtained for use on our village layout. It exhibited similar jitter characteristics at very slow speed, which was quite evident during acceleration or deceleration near zero speed. A Google search revealed an RMweb forum entry describing how another modeller had substituted a drop-in Hornby K1 motor into his D16 and completely cured a similar jitter problem.

Sure enough, having sourced two X6970 replacement motors from Peter's Spares, I can confirm that at least on my two D16 locos, the jitter effect is cured by fitting the new motors. Unlike the original motor, the X6970 has a worm gear that incorporates a brass flywheel. The D16 motor cradle includes provision for the flywheel, so it really is a drop-in replacement! The new re-motored D16 is using a Zimo MX645 decoder, equipped with a Digitrains Paul Chetter sound project, which is now just as smooth as the now re-motored original LokSound equipped D16. The photos below show the original and replacement motors in place on the chassis:

 

Original Motor

Replacement K1 motor in place

   

 

 

Supplier website links:
Rails of Sheffield    Sourced from Rails' Hornby sale
Digitrains    Source for Zimo sound decoder, speaker and fire box LEDs
Olivias    Source for the LokSound V4 D16 sound decoder

 

Click to move back to Model Rail Index Page