class 455 and camshaft controler

[nschichan]

Hi,

I have recently been trying to script a realist camshaft controller for the Class 455.

I have mainly been using as reference the Current versus Speed graph that is available in the EE507 traction motor thread.

Right now I implemented it as follow:
- The TractiveEffortVsSpeed dscv allows for a constant 70 kN from 0 to 90 Mph, this is to simplify scripting.
- I tweaked MaxForce in the Simulation bin so that the Ammeter shows the same currents as in the EE507 traction motor thead.
- I tweaked the vehicle drag coeficient so that WEAK FIELD balancing speed is at about 95 Mph (there were some testimonies on this forum that someone did bring a class 455 unit to 90 Mph and that it was still accelerating).
- PARALLEL balancing speed is at about 72 Mph.
- SERIES balancing speed is at about 48 Mph.

Right now the acceleration seems a bit weak when in SERIES notches (especially on an uphill gradient), but once you start notching into the PARALLEL notches and then WEAKFIELDS, you end up doing 75 Mph in no time.

Here are some graphs:
- http://columbia.libpthread.so/rw/cl455- ... VsTime.png speed is in km/h and time in seconds
- http://columbia.libpthread.so/rw/cl455- ... VsTime.png current is in Amperes and time is in seconds
- http://columbia.libpthread.so/rw/cl455- ... VsTime.png tractive effor is in kN and time is in seconds
- http://columbia.libpthread.so/rw/cl455- ... VsTime.png acceleration is in m/s^2 and time is in seconds

There seem to be disagrement whether you can notch down without first moving the power handle to OFF, I made however the choice to replicate this behavior.

I have however a question regarding the case when you put the power handle back to SERIES while notching into the PARALLEL resistor notches:
- should the resistors continue to be shorted as speed is rising until the last resistor for the PARALLEL mode ?
or
- should the controler stay in its current state and not continue to short resistance regardless of the speed build up ?

I'd like to get also your input on how this graphs looks with respect to the real thing.

I'd like to release the modifications, but this seems to be a great period of uncertainty regarding the publication of modified DLC files, and to be honnest, my modification may break the AI trains in scenarios, and I have other modifications (modified AWS test and DSD) that are not yet ready for prime time.

Regards,

[Kariban]

The one in DaveB's 455 did a reasonable job of simulating it's operation ( although iirc has far too many notches ). As far as I remember if you're in parallel and you go back to series, provided you've not run out of series resistor notches then it will stop where it is - you can use this to notch up one parallel notch at a time. That obviously means you can't run down the notches either, just go back to the beginning.

Glad you got round to doing this, I kept meaning to but just never found the time for it.

I also vaguely remember a 455 will get to 40 very fast, and then acceleration drops off a lot. There's at least one 455 driver on the board though so it'd be better if he can chime in.

[nschichan]

The one in DaveB's 455 did a reasonable job of simulating it's operation ( although iirc has far too many notches ).

I'll have a look to the script of DaveB's 455 then

As far as I remember if you're in parallel and you go back to series, provided you've not run out of series resistor notches then it will stop where it is - you can use this to notch up one parallel notch at a time. That obviously means you can't run down the notches either, just go back to the beginning.

Ok, I'll look into implementing that.

I also vaguely remember a 455 will get to 40 very fast, and then acceleration drops off a lot. There's at least one 455 driver on the board though so it'd be better if he can chime in.

Well in my model, WEAK FIELDS are taken at around 44 Mph, so at this time the accelleration is at its peak. To get from 0 Mph to 40 Mph, it takes 73 seconds, I feel it is a bit slow, especially compared to the original physics of the class 455 DLC.

Here is the acceleration vs speed graph: http://columbia.libpthread.so/rw/cl455- ... sSpeed.png

Unfortunately I do not live un england so I probably will never have the occasion to ever see a real class 455. Where I live, we tend to bury our 3rd rail lines below the ground

Most of the tweaking of my script is also (mostly uneducated) guess work, the relationship between speed and regulator position (or current for that matter) is proportionnal to 1 / (x^2), with some tweaks to have the regulator at fixed positions at the beginning and end of a notch.

[Kariban]

It looks a reasonable approximation of a series wound motor, that's probably as complicated as I would have got anyway dug through my PMs, 0-40 is "40-45s" apparently, not easy to measure anything faster because of so many stops. They are suburban trains so the focus would be on acceleration rather than top speed.

Camshaft operation:

Shunt ( all resistors held in circuit ), 3 steps in series, 3 in parallel, then weakfield; each notch will let the camshaft run up to it's end and then hold there, so if you go straight to weakfield from starting it'll just run all the way; if you hold it in series it'll notch up to the end of series and stick, then you can move to parallel until it notches up once or twice & go back to series again and it will stick in the first notch in parallel. Or you can hold it in parallel and it'll run up until the notch before weakfield. You're not meant to hold in shunt very long or it will overheat and possibly catch fire, which would be a fun but silly thing to implement

[linuxbot]

Hi,

I have recently been trying to script a realist camshaft controller for the Class 455.

I have mainly been using as reference the Current versus Speed graph that is available in the EE507 traction motor thread.

Right now I implemented it as follow:
- The TractiveEffortVsSpeed dscv allows for a constant 70 kN from 0 to 90 Mph, this is to simplify scripting.
- I tweaked MaxForce in the Simulation bin so that the Ammeter shows the same currents as in the EE507 traction motor thead.
- I tweaked the vehicle drag coeficient so that WEAK FIELD balancing speed is at about 95 Mph (there were some testimonies on this forum that someone did bring a class 455 unit to 90 Mph and that it was still accelerating).
- PARALLEL balancing speed is at about 72 Mph.
- SERIES balancing speed is at about 48 Mph.

Right now the acceleration seems a bit weak when in SERIES notches (especially on an uphill gradient), but once you start notching into the PARALLEL notches and then WEAKFIELDS, you end up doing 75 Mph in no time.

Here are some graphs:
- http://columbia.libpthread.so/rw/cl455- ... VsTime.png speed is in km/h and time in seconds
- http://columbia.libpthread.so/rw/cl455- ... VsTime.png current is in Amperes and time is in seconds
- http://columbia.libpthread.so/rw/cl455- ... VsTime.png tractive effor is in kN and time is in seconds
- http://columbia.libpthread.so/rw/cl455- ... VsTime.png acceleration is in m/s^2 and time is in seconds

There seem to be disagrement whether you can notch down without first moving the power handle to OFF, I made however the choice to replicate this behavior.

I have however a question regarding the case when you put the power handle back to SERIES while notching into the PARALLEL resistor notches:
- should the resistors continue to be shorted as speed is rising until the last resistor for the PARALLEL mode ?
or
- should the controler stay in its current state and not continue to short resistance regardless of the speed build up ?

I'd like to get also your input on how this graphs looks with respect to the real thing.

I'd like to release the modifications, but this seems to be a great period of uncertainty regarding the publication of modified DLC files, and to be honnest, my modification may break the AI trains in scenarios, and I have other modifications (modified AWS test and DSD) that are not yet ready for prime time.

Regards,

Coincidentally I have been tweaking the 455 over the past couple of days as well! I've been basing my script on Daveb's code, but trying to eliminate the kludges used for balancing speeds and tweaking it to fit the graph.
In your case, I believe your problem is that the motor current from the graph in the previous thread was total current drawn from the third rail, which needs to be halved for the parallel and weak field notches to get the motor current.

[coolhand101]

The camshaft class 455 have the same motors and equipment as the EPBs. There are six resistances in series and six in parallel and one in weakfield.
To notch up in series, power controller moved to 2(series) then back to 1(shunt) six times. To notch up in parallel, power controller first move to 3(parallel) then back to 1(shunt), then notching up using 2 to 1, again six times.
Transition from series to parallel = 15 mph. Transition from parallel to weakfield 30 mph. All this happens between 28 to 30 seconds if the power controller is in notch 4.
The 455 should get to 40 mph in 40/43 seconds on the level, with an initial acceleration of around 1.25mphps in notch 4(weakfield). 1mphps in other notches.( the shunt notch initial acceleration drops sharply after a couple of seconds).
Quoted balacing speed at 675 volts is 63 mph, however this is more likely to be at or slightly above 70mph at 750v. Doing 90 mph in a class 455 is possible, but only on a long downhill stretch with at least 1 in 100/80. Starting on an uphill grade, the CLR should notch up quicker due to the amp notch being reached earlier because of the quick fall in voltage. ie weakfield transition called be as low as 25mph on a heavy grade. I've seen a 4VEP transition from parallel to weakfield at 35 mph on a 1in75/80 grade. Level is 42mph. All this is from memory, so there might be a few errors/omissions in my info.

Thanks

[nukem]

The camshaft class 455 have the same motors and equipment as the EPBs. There are six resistances in series and six in parallel and one in weakfield.
To notch up in series, power controller moved to 2(series) then back to 1(shunt) six times. To notch up in parallel, power controller first move to 3(parallel) then back to 1(shunt), then notching up using 2 to 1, again six times.
Transition from series to parallel = 15 mph. Transition from parallel to weakfield 30 mph. All this happens between 28 to 30 seconds if the power controller is in notch 4.
The 455 should get to 40 mph in 40/43 seconds on the level, with an initial acceleration of around 1.25mphps in notch 4(weakfield). 1mphps in other notches.( the shunt notch initial acceleration drops sharply after a couple of seconds).
Quoted balacing speed at 675 volts is 63 mph, however this is more likely to be at or slightly above 70mph at 750v. Doing 90 mph in a class 455 is possible, but only on a long downhill stretch with at least 1 in 100/80. Starting on an uphill grade, the CLR should notch up quicker due to the amp notch being reached earlier because of the quick fall in voltage. ie weakfield transition called be as low as 25mph on a heavy grade. I've seen a 4VEP transition from parallel to weakfield at 35 mph on a 1in75/80 grade. Level is 42mph. All this is from memory, so there might be a few errors/omissions in my info.

Thanks

You are correct the traction motors, camshaft control system and even the compressors was from the 1936 Sub stock and yes there are 6 resistors in the Series and Parallel notches, however the Series setting will get the units up around 20mph depending on the state of the motors (I've had one that would hold 20-23mph another that would struggle to get above 15mph so there are variables), Weakfielding happens at around 38mph in a 455 but again this can vary. The hand notching proceedure is as you have described above I have made a couple of quick diagrams which explain in basic the camshaft control and how it works and gives an insight into handnotching.

Shunt & Series
http://www.flickr.com/photos/74309819@N ... hotostream
Parallel & Weakfield
http://www.flickr.com/photos/74309819@N ... otostream/

Not brilliant quality but hope this helps explain things a little,

Alan

[coolhand101]

Weakfielding happens at around 38mph in a 455 but again this can vary.

That does seem quite high for a 455, even downhill. Not sure about the few "chopper" 455s.
On the level, SR EPB was 26-27 mph, BR EPB was 28-29mph. The only variations of the WF transition that i experienced was up and down grades.

Thanks

[nukem]

Weakfielding happens at around 38mph in a 455 but again this can vary.

That does seem quite high for a 455, even downhill. Not sure about the few "chopper" 455s.
On the level, SR EPB was 26-27 mph, BR EPB was 28-29mph. The only variations of the WF transition that i experienced was up and down grades.

Thanks

That' what I was told when I did my traction on 455s

Alan

[nschichan]

Hi,

Sorry for the delay in my reply

In your case, I believe your problem is that the motor current from the graph in the previous thread was total current drawn from the third rail, which needs to be halved for the parallel and weak field notches to get the motor current.

You are completely right, sorry I missed that, I corrected my model to have more or less the same throttlesetting/current/tractiveeffort regardless of whether the motor is in series or parallel mode.

I also tweaked the model so that starting accelerarion is about 1.25mphps (0.55 m.s^-2). Keeping the same start/end current ratios for a given step the same as the one in the graph, I manage to get to 40 Mph in 43 seconds, which would be not far from the reported performance of the real thing. I also tweaked the weakfield step so that the balancing speed is about 75Mph on level, but maybe that is a bit overpowered (but not as much as my previous settings).

Concerning the hand notching capabilities, this is something I did not knew

I'll try to implement it soon.

Just to be sure I understood the behaviour of this here is what I intend to implement:

- when moving the power handle from the 0(off) position to the 1(shunt) position, and keeping it there, no resistor is shorted and the speed balances at about 4/5 Mph. Fire risk may occur if that position is held for too long (well I could play with the pantograph control and cut power in that case), but that's outside the scope of the simulation. moving the power handle back to 0(off) position cuts power and resets the whole thing.

- when moving the power handle from the 0(off) to the 2(series) position and keeping it there, SERIES manual notching is enabled, then moving back and forth from 1(shunt) to 2(series) shorts a resistor from the grid, thus raising the voltage on the motor. in that mode, moving the power handle to 3(parallel) and 4(weakfield) does nothing, moving the power handle back to 0(off) cuts power and resets the whole thing.

- when moving the power handle from the 0(off) to the 3(parallel) position and keeping it there, PARALLEL manual notching is enabled, then moving back and forth from 1(shunt) to 2(series) shorts a resistor from the grid, thus raising the voltage on the motor. in that mode, moving the power handle to 4(weakfield) does nothing, moving the powerhandle back to 0(off) cuts power and resets the whole thing.

- when moving the power handle from (0)off to 4(weakfield), automatic notching is enabled, and the controller walks every resistor steps in SERIES, then every resistor steps in PARALLEL and finaly switches to weak field. moving the power handle to 1(shunt), 2(series), 3(parallel) pauses the notching up until either the power handle is moved back to 4(weakfield) in which case the automatic notching resumes, or the power is moved back to 0(off) in which case the power is cut and the whole controler is reset.

Concerning the number of resistors, are there any resistors that are taken out immediatelly when pulling away in 4(weakfield) ? The current graph would seem to indicate that there are 3 resistor steps in SERIES and 4 resistor steps in PARALLEL. That would also explain the initial acceleration difference:

- 1.25Mphps initial accelearation in 4(weakfield) with for instance 3 resistor out.
- 1.00 Mphps initial acceleration in any other notch with 0 to 2 resistors out.

Regards,

[nukem]

Just to be sure I understood the behaviour of this here is what I intend to implement:

- when moving the power handle from the 0(off) position to the 1(shunt) position, and keeping it there, no resistor is shorted and the speed balances at about 4/5 Mph. Fire risk may occur if that position is held for too long (well I could play with the pantograph control and cut power in that case), but that's outside the scope of the simulation. moving the power handle back to 0(off) position cuts power and resets the whole thing.

Yep that is correct for the Shunt position (1)

- when moving the power handle from the 0(off) to the 2(series) position and keeping it there, SERIES manual notching is enabled, then moving back and forth from 1(shunt) to 2(series) shorts a resistor from the grid, thus raising the voltage on the motor. in that mode, moving the power handle to 3(parallel) and 4(weakfield) does nothing, moving the power handle back to 0(off) cuts power and resets the whole thing.

Not quite no, if you move the power controller to from off(0) to Series (2) and leave it there the camshaft will automatically start to knock out the resistors. To hand notch you move the power controller from off(0) to Series(2) and then back to Shunt(1) doing this will knock out one resistor and this can be done upto 5 more times to achieve full series. If you moved the power controller to Parallel the unit will configure to the parallel settings etc. With the camshft control system you can select a higher power position but to select a lower setting you have to shut off power then re select the lower setting eg if you were in parallel and wanted to select series you have to first shut off the power controller and then select the series position.

- when moving the power handle from the 0(off) to the 3(parallel) position and keeping it there, PARALLEL manual notching is enabled, then moving back and forth from 1(shunt) to 2(series) shorts a resistor from the grid, thus raising the voltage on the motor. in that mode, moving the power handle to 4(weakfield) does nothing, moving the powerhandle back to 0(off) cuts power and resets the whole thing.

Again same as above if you move the power controller from off(0) to straight to parallel the camshaft will automatically knock out the resistors until it has done all 6 and full parallel has been achieved. To hand notch you will need to first move the power controller to gain full series so from off(0) to series(2) and hold it for a short while for the camshaft to knock all resistors out, then move the power controller into parallel (3) and straight back to shunt, doing this has reconfigured the motors and knocked out one set of resistors and again like above this can be done 5 more times to achieve full parallel.

- when moving the power handle from (0)off to 4(weakfield), automatic notching is enabled, and the controller walks every resistor steps in SERIES, then every resistor steps in PARALLEL and finaly switches to weak field. moving the power handle to 1(shunt), 2(series), 3(parallel) pauses the notching up until either the power handle is moved back to 4(weakfield) in which case the automatic notching resumes, or the power is moved back to 0(off) in which case the power is cut and the whole controler is reset.

yes this is correct the weakfield position gove quicker acceleration as the camshaft will knock all resistors out as quickly as it can, which differs from if you moved the controller in to each notch separately.

Concerning the number of resistors, are there any resistors that are taken out immediatelly when pulling away in 4(weakfield) ? The current graph would seem to indicate that there are 3 resistor steps in SERIES and 4 resistor steps in PARALLEL. That would also explain the initial acceleration difference:

- 1.25Mphps initial accelearation in 4(weakfield) with for instance 3 resistor out.
- 1.00 Mphps initial acceleration in any other notch with 0 to 2 resistors out.

This I can't say, as a Driver I have been told that there are 6 resistors and the camshaft knocks them all out as quickly as possible in the weakfield position, so that's all I know I'm afraid.

I have a camera built into a set of glasses so next time I'm driving a 455 I'll see if I can create a quick video (providing it's safe to do so) showing what as a driver I do to control the power of the train and show first hand how to hand notch a 455.

Hope this clears things up a bit,

Alan

[coolhand101]

On the EPBs, the motor coaches had an amp meter. Even in notch 4(weakfield) there was always 6 resistances in series and 6 in parallel. They where cut out very quickly in weakfield. The CLR(current limit relay) had two amp settings, can't remember the actual numbers but it was around 25% higher than the shunt , series and parallel setting.
The weakfield resistance could be caught with hand notching in camshaft BR stock but was too quick in SR contactor stock.

Thanks

[dp123]

- when moving the power handle from (0)off to 4(weakfield), automatic notching is enabled, and the controller walks every resistor steps in SERIES, then every resistor steps in PARALLEL and finaly switches to weak field. moving the power handle to 1(shunt), 2(series), 3(parallel) pauses the notching up until either the power handle is moved back to 4(weakfield) in which case the automatic notching resumes, or the power is moved back to 0(off) in which case the power is cut and the whole controler is reset.

This is quite spectacular in a Class 73 (especially with two in multiple). You can feel when the loco transitions from Series to Parallel to Weak Field (certainly when it WF's, the ammeter flies right up and it accelerates like a boss). The default control for a 73 is hand notching, rather like an AC electric with a tap changer, but with three Run Up notches, Run Up (Series), Run Up (Parallel) and Run Up (Weak Field). This gives compatibility with a 4 position EMU controller. You can hand notch from a standing start and go up one notch at a time, which knocks out the resistances individually, or run up in "blocks" of resistances (i.e. Series, Parallel and Weak Field). There's also a Run Back position (akin to Run Down on a tap changer) that means you don't have to shut off power and reselect a lower position. It allows for very fine control of power application. Because the loco is short, you have to run back the amps at gaps to prevent snatching of the train when the power drops off, like an AC would at a Neutral Section. You don't need to shut off, but anywhere 500 and below is OK I gather. At complex junctions, the ammeter practically bounces off the bottom stop as it power is cut off at gaps.

[nschichan]

Hello,

- when moving the power handle from the 0(off) to the 2(series) position and keeping it there, SERIES manual notching is enabled, then moving back and forth from 1(shunt) to 2(series) shorts a resistor from the grid, thus raising the voltage on the motor. in that mode, moving the power handle to 3(parallel) and 4(weakfield) does nothing, moving the power handle back to 0(off) cuts power and resets the whole thing.

Not quite no, if you move the power controller to from off(0) to Series (2) and leave it there the camshaft will automatically start to knock out the resistors. To hand notch you move the power controller from off(0) to Series(2) and then back to Shunt(1) doing this will knock out one resistor and this can be done upto 5 more times to achieve full series. If you moved the power controller to Parallel the unit will configure to the parallel settings etc. With the camshft control system you can select a higher power position but to select a lower setting you have to shut off power then re select the lower setting eg if you were in parallel and wanted to select series you have to first shut off the power controller and then select the series position.

Thanks for the explanations, so this means that if I'm coasting at a speed sufficient for the CLR to knock all the resistors in parallel, if I put the power handle to 3(parallel), the camshaft will iterate on all the steps, from the start. this means that the camshaft will take some time to walk all the steps and the driver can stop the camshaft by selecting 1(shunt) when the camshaft has selected the desired step.

I took the liberty to enforce an interval of 0.75 second between any notch up event in my code, I don't know how prototypical that is but that means that when coasting at 50 Mph, it will take about 9 seconds for the motor to deliver full power. this gives a confortable margin during manual notch up.

I also wanted to know what happens if the driver notches manualy too fast ? is there a risk of tripping the power supply due to the too high amp rise ? I am currently setting the PantographControl to 0 in that case. If this happens the driver should put the power handle back to 0(off) and try to regain line volts using 'P'.

Anyway, here is the code (beta):

http://columbia.libpthread.so/rw/pub/class455-0.9.zip and
http://columbia.libpthread.so/rw/pub/ns ... ua-0.9.zip

Please note that right now the only tested input method is via the keyboard, I did not try to control it with the mouse (and Raildrivers are probably broken as well).

Requiremements:
- Thomson's Class 455 pack
- JustTrain's Class 153 pack if you intend to use the DSD.

Installation procedure:

- backup the Assets/Thomson/Class455Pack01 directory.
- unzip both zip archives in a temporary directory.
- backup the Assets/Thomson/Class455Pack01 directory.
- Copy the Assets directory of both zip archives in your railworks directory, and select overwrite in the window prompt that will appear.
- Clear the blueprint cache.

If you have the JustTrain's Class153 and want to use the DSD, then:
- Copy the file "Assets\JustTrains\Class153\Audio\RailVehicles\Diesel\Class153\Cab\jt_153_dsd.dav" to "Assets\Thomson\Class455Pack01\Audio\Electric\Class455\DTOS\Cab\jt_153_dsd.dav"

Otherwise, in "Assets\Thomson\Class455Pack01\RailVehicles\Class455_9\Default\Class455_DTSO_EngineScript.lua":

Remove at line 6:

Code: Select all

	dsd = new_dsd({
			      aws_reset = "AWSReset",
			      dsd_alarm = "DSDAlarm",
			      emergency_brake = "EmergencyBrake",
		      },
		      {
			      AWSReset = 0,
			      VirtualRegulator = 0,
			      VirtualTrainBrakeControl = 0,
		      })

And remove at line 399:

Code: Select all

		dsd:handle(time)

Aside of the camshaft controler, I added a 4-step reverser (with the subtelty that on the driver desk, the reverser handle position does not know the 'off' position so visually it is in 'neutral' when it should be in 'off'). AWS Self test will sound the first time the reverser is put in neutral after being set to off. the DSD Self test will sound the first time the reverser is put in forward after being set to off. DSD can be canceled by pressing 'Q'.

For the curious, most of the interesting lua parts are in "Assets/nschichan/lua", the modifications to the original EngineScript are minimal and only consist of the object creation/configuration in Initialise() and the call of the objets :handle() callbacks in Update()

Well that's it and I hope I did not forget anything.

Should you experience a problem wit those modifications, do not hesitate to communicate me the output of the "Script Manager" tab of LogMate along with a description of the problem.

Regards,

[Kariban]

I don't think it's really manual notching is it? it's more stopping the controller from running up, ie if you go back & forth from shunt to series & back really quickly it won't grab notches until there's enough voltage drop no matter how much waggling you do; but if you pop it back in shunt & the voltage drops below where it'd have notched up normally, then it will just stay where it is instead.

BTW rename "VirtualTrainBrakeControl" to "VirtualBrake" everywhere, and it'll pick up the F4 control.

Hurray for someone doing OO lua I have been terribly bad at getting round to moving in that direction.