And Now For
Something Completely Different!......
Couldn't resist obtaining a model of Gresley's impressive P2 locomotive. However, to fit in with my modern image layout, it is about to undergo a minor transformation into new build P2 number 2007 "Prince of Wales", currently under construction at Darlington, by the P2 Steam Locomotive Company.
I came across a picture of P2 "Wolf of Badenoch" in an old railway book as a youngster. This massive 2-8-2 variation on an A4 body shape, left a lasting impression. I've never understood why these powerful locomotives were not moved South to the East Coast mainline during the war, where they could have pulled the massive war time train loads with comparative ease! Then, instead of fixing the remaining technical issues (e.g. the front pony truck etc) Mr Thompson decided to create a much less successful and rather ungainly pacific from these thoroughbreds, well before I would get the chance to see them!
However, thanks to the P2 Steam Locomotive Company, an improved Gresley P-2 will hopefully be back where it belongs on the East Coast main Line, within the next few years.
This gives me the opportunity to get slightly ahead of reality and introduce my own LNER Number 2007, " Prince of Wales" alongside my otherwise contemporary locomotive fleet. This will also form alternative motive power for my Northern Belle Pullmans. (Indeed, somewhat more up-to date than the now defunct Class 47s to be released from Rail Exclusive!)
Adding lighting to the P-2:
I have ordered some DCCC LNER style front oil lamps from Digitrains and I plan to fit red & amber chip LEDs in the firebox door to produce periodic glowing, synchronised with coal shovel sounds when the loco firebox is being fed (stationary or on the move).
The Sound System:
I'll run initially with the onboard Hornby decoder and speaker arrangements. First look at the manual suggests a lot of button pressing and not much automated content, but I'll get the controlling PC to deal with those issues. (The three available function outputs should enable appropriate lighting control, although I'll need to diode feed the front lamps from both white and yellow function wires to keep them on regardless of loco direction setting and.... I'll need to come up with a suitable flicker circuit for the firebox lighting effect.)
If the sound and/or running characteristics leave something to be desired, the current favourite for a Plan B upgrade is Paul Chetter's Zimo 645 based sound project, but I'll run with the Hornby system to begin with.
Oh Dear !!!!! Not good! Not good at all !!!!!
Using the default motor control algorithm number 1, Lots of jerky behaviour at low speed! Checked the adjustment CVs and they were no where near the defaults quoted in the decoder leaflet...... reset them to the defaults and..... even worse! I subsequently found via RM Web that Hornby have changed the recommended settings to those that were programmed as received ...... but frankly its a disaster with completely unacceptable low speed characteristics.
Tried using motor control algorithm number 2 and this behaves rather better, with jerky behaviour predominantly at the lowest speed steps. However, this is still unacceptable!
Do I have a sticky motor? Its quiet and seems fairly smooth on the rolling road, so I doubt it. Only one way to be sure is to fit a DC operation 8 pin plug in place of the TTS decoder in the tender, and try it on a DC PSU........
........Well, its not perfect, but its not half bad. The loco rolls at around 2.5 volts without any major drama. I think its good enough and will probably be better still when properly run in. So what's the problem???? Its looking rather like a decoder motor control issue. Not the first I've come across with a Hornby decoder!
Although the TTS motor control seems to be rubbish, the sound is more promising! I particularly like the chime whistle options and the only own goal seems to be the distorted guards whistle, which I'll wind down in volume and recheck.
So where next? I have a spare used ESU LokPilot V4, which I'll try, to see if it solves the motor control smoothness problem...... if its successful, I can set a 2.5 second start delay on the LokPilot, to mirror that on the TTS decoder, then I can run them in parallel on the wheel contact wires, with only the LokPilot driving the motor.
Right, after a bit of a struggle, I've got the P2 behaving pretty well under the control of an early ESU LokPilot V4. I've plugged into the tender 8 pin socket, but I've had to drop the motor control frequency to a fixed 20kHz to get the desired result. I'm guessing there may be TV suppression caps within the main locomotive body. If so, I'll take them off and try 40kHz again. I've got a 2.5 second start delay programmed in whenever button 1 is on (i.e. when sound from the TTS decoder is active) to synchronise with the TTS start delay. I've also programmed slow shunt mode when button 20 is pressed for really slow motion. I'll wire up a pair of wheel contact twin connectors inside the tender, so that each decoder can be accessed independently for programming.... On second thoughts, if the LokPilot will fit within the locomotive boiler, that puts it in the right place for the lighting connections and I can simply separate loco and tender for independent programming!
The suppression came in the form of a single 10N cap and a twin ferrite bead. (Now removed) I've fitted the LokPilot on the loco and tinkered with the CV settings again and its running O.K. with the settings tabulated below: The operating frequency is now 40kHz and the motor is extremely quiet.
LokPilot V4 hard wired to the motor and track contact wires. I'll feed the function wires up to the upper body shell via a 4 way connector behind the front screw boss.
The original motor wires from the tender are left in place, with the loco end insulated and tucked under the flywheel, embedded in Blue Tac in case of a future use.
Motor control CV settings for the LokPilot Decoder.
Base of the tender, showing the TTS decoder & speaker locations.
Oops! Not quite that easy! No chuff noises from the TTS decoder now its motor leads are open circuit! I'll stick a resistor in place of the motor and see what happens......
No dice! must be using the motor to determine the chuff sounds......so I'll try the TTS again, with a low value resistor across the motor, otherwise, its back to plan B and a new sound decoder! A lowish value resistor across the motor doesn't improve motor control, its still rubbish!
Back to plan B! At least I've confirmed that a LokSound solution would provide satisfactory motor control, so I'm reasonably confident that a Zimo decoder will behave just as well! I've got a spare Zimo 644D, which I'll fit via a spare ESU 21 pin interface board in the tender. I'll have to run 3 extra wires between loco and tender to supply the lights, but I'll run the new wires directly into the loco upper body shell to avoid an extra internal connector. I'll test the motor drive first and if its OK, I'll send the Zimo off to Digitrains for a Paul Chetter P2 Project reblow on Monday.
Yes, Zimo motor control is pretty good, just a hint of hesitation on the minimum speed step which is a really slow crawl! Things get better when the loco is under load and I think it will probably improve further as it runs in. Its not quite perfect but a huge improvement in motor control compared to the Hornby TTS performance!
On both the Zimo and the LokPilot, I was predominantly fighting jerky behaviour, running light engine at absolute minimum speed (using the motor control optimisation CVs). The motor is turning very slowly and the BEMF systems are struggling to maintain control. I'm using previous diesel experience as a gold standard and maybe I'm being unrealistic to expect the P2 motor to run that slowly without a bit of occasional jitter. I suspect the smaller diesel wheels make the corresponding minimum diesel motor speed somewhat faster, perhaps making life easier for the BEMF systems to maintain smooth running at the minimum speed step? Also, on the P2, 8 drivers joined by coupling rods located asymmetrically on each side, probably results in some cyclic variation in the forces transmitted back to the motor as the wheels rotate, unlike an all wheel drive diesel, where the equivalent forces are transmitted via a common set of meshed gear wheels, which I think would result in a steadier load on the motor.
Zimo 644D motor control settings
21 Pin adapter ready for the reblown Zimo 644D .
Zimo Function List & Possible
Paul Chetter's listing from the Digitrains website is as follows:
I'll try the motor CVs in their default form first and switch to the values quoted above if low speed jitter is an issue....... Motor CVs now updated in line with my earlier table!
I should be able to keep the front oil lamps on during reverse, by using the headlight FO and changing CV34 from 2 to 1....... Done!
To establish which FO Paul has used for the firebox flicker, I need to take a look at the value programmed into CV40 .... its 4 which indicates FO1
Simplified lighting hardware & Firebox
I came up with a couple of possible hardware solutions based on astables with ramping capacitors, to add firebox flicker to the TTS output, but now I've had to change the decoder to obtain smooth motor control, I can also use the built-in Zimo flicker effect to drive the firebox lighting directly and I can avoid having to use a couple of diodes to keep the front oil lamps on when travelling forwards or backwards, by making the F0 function output from the Zimo non-directional.
Nothing complicated this time!
Fitting a sub-miniature 4 way
connector for the loco lights, to the tender:
I used a sub-miniature TCS 4 way connector from Digitrains to provide the extra interconnect for the loco lighting. The socket is glued to the inside of the tender front step and the wires are fed through a slot cut beside the adjacent wheel well, & routed to emerge below the NMRA 8 way socket. The wires were trimmed and soldered to the appropriate pads of the 21 pin adapter PCB.
New wiring and decoder in place, prior to taping the assembly down to the speaker sides
Showing the location of the new socket.
Adding the lighting:
1) The front Oil Lamps:
Oh Dear! Another nearly there, but not quite! solution, this time from DCC Concepts who usually do better!
I bought some LNER style oil lamps from DCC Concepts (via Digitrains) but although they are good reproductions of the lamps, the LEDs are pure white, rather that yellowish (not a problem as I can colour them with water colour paint) but more seriously, the casing is translucent! So unless the LED light level is absolutely miniscule, white LED light escapes through the body of the lamp, looking rather spooky and definitely not prototypical! (See photo below which has the two lamps in series connected to a DC PSU via 80K of series resistance.)
This is why I paint my own home made rear lamps black overall, then overcoat them with matt white!
So.....out with the enamel paints to add a light proof coating!.......................
That's better! The matt black undercoat has made the case light proof and the yellow water colour paint has produced a more realistic yellow tint to the light output. The water colour tint has cut down the light intensity, which has been restored by reducing the common series resistor to 22k.
2) The firebox glow:
I've drilled and filed out the firebox door, then fitted a red LED and an amber LED to an angled plasticard back plate, directly behind the door opening. The idea is to run the red LED continuously to give a background glow within the cab, while the amber LED will be made to come on whenever the coal shovel noises are produced, with a flicker drive waveform. If this does not work well, I may rethink and simply run the red LED from the same drive voltage as the amber.
With it all wired up, the front oil lights only come on in a forward direction (looking OK). The firebox red glow works well (but only in the forward direction) and the amber LED comes on when the coal shovelling is selected, but there is no flicker from the amber LED. So its all got potential, but needs some programming adjustments!
Ah Ha! Found the problems.... I've reset CV125 & CV126 to zero making button zero non-directional. That's got the oil lamps & red firebox glow to stay on in reverse.
I've reset CV127 from zero to 8 which has introduced random flicker to the amber LED.
Quite pleased with the results!
Just need to adjust the cable length between tender and loco and bond down the wires behind the firebox wall....
Lighting mods complete and
|The Finished Loco:
Essential operational mods to the cylinder drain cocks:
The cylinder drain cock detail parts run inside the buffer beam step detail parts, severely limiting sideways movement by the pony truck wheels. In this condition, the loco could not reliably negotiate the second radius curves associated with the Peco medium points on my test track. The problem was successfully dealt with (at least for slow speed movements) by shortening the drain cocks so that they terminate just behind the buffer beam steps. This enables the pony truck wheels to slightly deflect the flexible drain cock plastic if contact is made. It may be necessary to permanently bend the pipes outwards if problems persist at higher speeds or on any tighter turns in the loft layout.
Showing the truncated pipes just making contact with the pony truck wheel on a medium radius Peco point.
I notice that the pipes are shown running outside of the buffer beam steps on the Hornby box photo, which would achieve the same result. (Although Hornby's detail parts assembly instructions show the pipes on the inside of the steps, which is how the real locos were constructed.)
|More on the real P2:
The December 2014 edition of Railway Magazine includes a revealing article concerning the testing carried out on the P2 prototype in the then new Vitry rolling road test facility in France. It seems that in order to maintain the overall oil film between axle and white metal wheel bearing, a degree of track induced vibration is essential. Unfortunately, this was not realised when Cock O' The North became one of the first large engines to use the new Vitry facility, which incorporated a precision ultra smooth rolling road. Testing was therefore repeatedly interrupted due to wheel bearings degrading due to lubrication failures. When evaluated on real French track, the results were impressive with no hint of bearing problems, although the French engine crew struggled to maintain steam pressure, commenting that their own coal shovels had twice the capacity of the LNER ones and the firebox door was far too small!
Never occurred to me to check the temperature of the Hornby wheel bearings when running my own P2 on the DCC Concepts rolling road :-)
|P2 Protodrive and software
October 2015 update by Paul:
Revised Function List:
Note from Paul on RMweb forum: set CV281 to 100 to reduce the wheel slip sounds.
|P2 Slow speed judder:
Its almost acceptable....... I keep trying to ignore it....... but it wont quite go away!
After seeing a Youtube "Mikes Movies" video, where Mike replaced the original motor with a 5 pole skew wound type. (Hornby X9108) I thought I would try the same approach. The motor (sourced remarkably quickly from Peter's Spares) looked promising, with far less hysteresis between start voltage and stop voltage than Hornby's original motor, when not loaded. However, performance on the loco was somewhat less impressive. It's better than the Hornby original, despite the lack of a flywheel, but with the Zimo decoder optimised, there is still judder when at absolute crawl speed. It maybe more of an academic issue as normally when the loco accelerates or brakes, the speed passes through the problem zone in the wink of an eye and even if you are looking critically for signs of judder, it is rare to notice anything untoward....... However, my recent Colas 37 has a similar Zimo decoder (and a 3 pole motor) but has no hint of judder even when barely moving. Now the P2 running gear is very different to that of an all wheel drive diesel. Perhaps this presents a variable load just when the motor and BEMF system can least deal with it, or..... have I swopped one bad motor for one that is only slightly better, despite the extra poles?
So...... I could live with it........ but lets see if its possible to get a further improvement.
Trying some alternative decoder BEMF systems:
First I tried an ESU LokPilot (which did give quite reasonable results with the original motor........ Curious! No better than the Zimo and the settings that minimised low speed judder, had the undesirable side effect of flattening the initial speed curve.
Next I tried an ESU LokSound V4..... Now this includes additional low speed optimisation via CV51 which is not available in the LokPilot.
With a small value in CV51, there is an immediate improvement!
After some experimenting with CV51 and CV52, there is an incredibly slow first speed step (1/128), followed by a smooth and linear increase in speed as the controller speed is increased by 1/128 at a time, onward throughout the speed range....... This looks very promising!
The Control CV settings are as follows:
So, one solution might be to switch over to an ESU LokSound V4 based sound system.......But, I can't find a P2 sound project available from any of the usual sources on that decoder type. So the only way forward in that direction, might be to use a double chimney Gresley A4 sound project for LokSound V4. I expect this should be very similar to the P2. (Classic Gresley 3 cylinder 6 beats per revolution and similar chime whistle.)
If I decide to try this route, it would enable me to give South West Digital's recently introduced "Udrive" offering a try........ However, there really isn't a lot wrong with the Zimo approach, where the ultra low speed jerky behaviour is rarely apparent in normal operation. I'll ponder, and maybe pursue in the new year when SWD re-open after their extended Christmas/New Year break.
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 plan to add a simple protection circuit, to ensure that if the P2 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 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.
|Trying the South
West Digital LNER double chimney A4 LokSound V4:
The decoder arrived on schedule with a useful set of notes describing operation of the sounds and outline advice about setting up the essential CVs.
First I entered the motor control CV values that worked well on my experimental installation with the spare diesel LokSound:
With the sound-off, motor performance was very acceptable with smooth take off from an extremely low speed start point.
Static sounds were also good, with the A4 chime whistle sounding just as I expect the P2 would sound, and a good array of other supporting peripheral sounds...... I had to add reverse operation to the front oil lamps but the aux 1 flicker was already applied to the firebox, so this synchronised with coal shovelling and firebox door activation without any additional work from me.
However! With sound on and the locomotive moving........ the chuff rate was dramatically too fast and CV57 / CV58 adjustments failed to tame it.
I made small changes to a number of the relevant motor control CVs, looking for possible settings that might be provoking the chuff issue. Eventually I discovered that CV2 was the main issue, with some less significant additional interaction from CV53.
With CV2 set to 1, it was impossible to reduce the chuff rate to the required 6 per driving wheel revolution via CVs 57 & 58. Unfortunately when CV2 was changed to 2, although I could get the required chuff rate, the minimum speed step (1/128) was considerably faster and the lowest speed steps became non-linear, with a plateau between 1/128 and circa 5/128, where an increase in speed step did not increase locomotive speed. The good news was that all speed steps still exhibited smooth operation without judder.
The best compromise I could achieve was with CV2 changed to 2 and CV53 reduced to 100, to slightly reduce the minimum speed step. This still had a short plateau but at least produced smooth movement without judder. CV 57 had to be set to 148 to produce 6 chuffs per revolution and CV58 went to 45 to maintain this rate across higher speeds.
CV2=2, CV57=130, CV58=40
Table showing "as received" CV values on the left, interim values shown in brackets and final values in red on the right :
Original CV settings: (Interim settings) final settings
Comparing with Zimo..... The low 128 speed steps are smaller and slower on the Zimo decoder than on the ESU LokSound! However the Zimo is not smooth at minimum speeds, while the ESU V4 is jerk free.
CV56 set to 128 to disable bemf above half speed and CV249 set from default 30 to 60 to improve sound of fast chuffing
Looks about the best we can do!
Final settings shown above in red.
Performance assessment next day : Very smooth! this will do nicely! Although the early Zimo speed steps are slower than on the ESU V4, they include judder, which is absent from the somewhat faster minimum speed steps on the ESU V4. The end result is a clear win for the ESU LokSound V4. I'll miss the Zimo brake key with high CV4 and the station sound effects, but smoother movement due to the superior low speed motor control optimisation on the SWD ESU LokSound V4 carries the day!
ESU V4 SWD Sound Project individual sound slot volume optimisation:
Activating random sounds: CV61 to 40 and CV62 to 100.
I'm not able alter the movement linked changes to relevant sounds such as the chime whistle, via the function mapping CVs, so I guess these must be handled within the sound project.
Trying the random sounds was disappointing: The random sounds are not altered by the individual sound slot volume adjustments. Instead, they are are all controlled by CV451, and the relative level of the random sounds is not user adjustable, so e.g. the random injector is too loud while the random safety valve is too quiet......... So I've ended up disabling the random sounds! (By resetting CVs 61 & 62 to zero).
So in summary, the SWD U-Drive A4 sound project provides a LokSound V4 solution for the P2, but it has its limitations: The odd interaction between the CV2 setting and the minimum chuff rate means that an ideal motor control CV solution is not in fact achievable. However an acceptable compromise was accomplished. The lack of any user CV control of the relative volume levels of the random sounds led to a decision to switch these off , which is a shame as their effect could add positively to the user experience. Otherwise the sound system worked well.
And that is finally the end of the P2 enhancement saga.........(I sincerely hope!)
In May 2018 I added a very impressive LokSound V4 based sound system to a Hornby Class A4. This was my first experience of a "Locoman" sound project and it not only sounds excellent, the driving experience is equally good, with stunning low speed motor control.
Taking a closer look at Locoman's motor control arrangements, two aspects looked worth trying on the P2 LokSound.
1) Locoman managed to get 6 chuffs per revolution with CV2 set to 1 and CV53 set to 110, using an extremely high setting for CV57 and a correspondingly higher than normal value for CV58. This resulted in an extremely slow initial speed step.
2) Locoman used the 28 point DCC speed characteristic option in place of the simpler 3 point option that I normally use. Locoman's speed characteristic involves an initial very gradual increase in the speed difference between adjacent points until at around a quarter speed, the difference becomes constant, resulting in a linear characteristic at medium and higher speeds.
Using both of the above techniques resulted in an equally impressive driving experience on the P2, albeit with not such outstanding sound quality from the SWD decoder software, as heard on Locoman's A4. This has now been remedied by means of a new Locoman A4 reblow of the the P2 LokSound V4 decoder. So it has all ended well, with a superb sound and driving capability now fitted to Hornby's P2.....
|Supplier website links:
(More to be added as the sources become apparent)