Experiments in Throttle Position Speed Matching...

[learmoia]

I haven't posted any of my Youtube stuff lately, but recently I have been experimenting with speed matching where I match the locomotive speed to the throttle percentage..

I came up with a system using the following components:
DCC with a computer interface.
JMRI of course
Any throttle that shows Throttle Position as a number (usually a percentage of 0-100)
any speedometer product.. I chose a Train Speed 4 from TSC but really anything will work..

In the video below.. I speed matched the following locomotives to work in unison..
Atlas Fast Motor GP40
Atlas Slow Motor GP38s (2)
Kato SD40-2

At the end of the video they all run in consist in unison.. and I know the speed by looking at the throttle.

I started working on a 5th loco but the top speed was only 32.. so that one needs some work..

Once I iron out the process I'll do a How To Video to clearly show what I'm doing..

But if you want to watch the raw footage of messing with it..
https://www.youtube.com/live/rGtvcaQZ70g


~Ian

[OrangeAnon]

I do this as well, but on an NCE system (and TCS wifi throttle away from home) so I can get into the 120s if appropriate. I also limit locos to their prototype capabilities; if I MU an FEF-3 and an ES44 for an excursion run, I better not go above 70 or some wheels are gonna slip! But to your point, by doing this I can MU anything in my roster, and I know exactly how fast I'm running, it's great.

[learmoia]

I do this as well, but on an NCE system (and TCS wifi throttle away from home) so I can get into the 120s if appropriate. I also limit locos to their prototype capabilities; if I MU an FEF-3 and an ES44 for an excursion run, I better not go above 70 or some wheels are gonna slip! But to your point, by doing this I can MU anything in my roster, and I know exactly how fast I'm running, it's great.

Out of curiosity are you using the the 3 step basic table or the 28 step table?..

When I started I was trying to match every CV Step to a throttle position and speed... but it was so granular I wasn't getting repeatable results.
Then I tried 3 step, but with all the Atlas Slow speed motors I can only use half of the speed table.. (Start and Mid)..

[OrangeAnon]

I use the 28 step table. Most of the time it ends up pretty linear and I only need to make slight adjustments, but I've come across a motor or two that definitely needed fine control over the lower end to make it speed match right.

[NtheBasement]

I have in the past matched locos with speed steps but I've found that they drift; after a dozen hours of running they don't match anymore.  I've since reverted to high-mid-low.  Still drift but less effort to correct.

[Sumner]

    I've started to speed match my locos to the 128 speed steps on the throttles I use. Speed step 10 = 10 mph, 45 = 45 mph and so on up to speed step 65 which is 65 mph. For most all of my locos that will be the top speed. For some locos used for passenger service I might up that to speed step 90 which would be 90 mph. Even with the passenger locos having a higher top speed 65 would still be 65 mph.

    I feel there are a couple advantages to speed matching to the speed steps vs. to another loco which is a common practice. If all the locos are speed match to the DCC speed steps they will be closely matched to each other. If you consist 3 locos and run the throttle to speed step 50 they will all be going 50 mph. Also I like knowing that if I see 22 on my throttle screen I know the loco is going a scale 22 mph.

    I haven't so far found it very hard to speed match them like above with JMRI's DecoderPro. My ESU decoders have easy settings to use for the max, middle and low end of the speed range for a loco. Setting those in a few minutes by setting the throttle on 65 and then messing with the max number so that is a scale 65 mph and then doing that for the mid and low end usually ends up with the loco being very close to scale mph from 0-65 mph (speed step 0 to 65). There are other settings in DecoderPro to fine tune it but so far doing the above has proved to be just fine. The Digitrax decoders I have done this with so far have only 2 settings but still they seem to match up pretty close after a few minutes.


I use one or the other of the speedometers shown above.  You can build one for under $50 and all the build info is on my site ( HERE )

Sumner

[Dupesy]

I'm doing basically the same; throttle position to scale speed.  I settled on 100% throttle being 40 SMPH, 50% throttle 20 SMPH etc, and every engine is programmed that way.  Most difficult was a couple of older Kato engines that just love to run fast.  The setting on the highest speed step is very very low.

[Maletrain]

Some of thoughts on this:

1.  Setting speeds in scale miles per hour to match the % of full throttle tends to produce larger jumps in speed between steps - which can make matching less precise.  And, if you go from Digitrax to NCE with a set of locos, you will then need to deal with the throttles showing actual speed step numbers, not some percentage of their maximum (14, 28 or 128).  My preference is to make the top speed occur in speed step 28 of the 28 step table.

2,  I have found that I need to run locos for abut 15  minutes before their speeds settle down to consistent times for a lap around my test loop.  Difference from cold to hot can be substantial.

2.  I have found that some locos run at substantially different speeds in forward and reverse, at the same throttle setting.  So, especially for diesels that may run in reverse in consists, I try to use the forward and reverse trim CVs to get them running at the same speed in both directions before I start speed matching.  (To reverse then, I physically turn the locos around so as to run them in the same direction on the test loop, in case there are directional issues with the loop itself.)

3.  Older decoders seem to have lot of glitches in their speed step algorithms.  Some have very course jumps in top speed with CV5 settings, forcing me to use the 28 step tables.  Some don't seem to actually do what their manuals claim they do, such as really changing BEMF at some speed value despite its control CV being set to to the value that should make BEMF be "always on". (It still cuts off at a speed step CV value of 127 and above).  Auto changing the BEMF setting introduces a substantial shift in both speed and rate of speed change with CV value where the BEMF change occurs.)

4.  Speeds in 128 step mode sometimes do not track as expected when set in the 28 step table, especially at low speeds.  In one case I found, the loco actually stops at speed step 4 in 128 step mode, but crawls well at the same speed in 128 mode steps 1,2,3 and 5, before going to higher speeds.

These glitches were experienced with old model decoders in my club's N scale loco roster, and may not be an issue for more modern decoders.  (But they are driving me nuts with a bunch of older, mostly TCS decoders.)

[learmoia]

My notes based on several of the comments (and basically a summary of the process in video)
3 Step vs 28 Step...
3 Step definitely gives you a close enough result, and is easy to manage..
BUT.. when your dealing with Kato locos and Atlas Scale Speed motors, either you limit everything to the Scale Speed motor's capability (everything is capped at 50-60smph) 
OR slow locos use half the table, while fast locos use the full table.   

I opted to use specific steps on the 28 step table in a hybrid solution.
Using the 28 Step Table I chose the following steps:
Step  2 = 3mph/Start movement.
Step  9 = 25mph
Step 17 = 50mph
Step 25 = 75mph
Step 28 = 100MPH
.. This roughly correlates to the Throttle% matching Speed on a Digitrax Throttle using Digitrax Decoders..
I've got some TCS decoders that I'll recheck the throttle response related to CVs on a Digitrax System..

I don't have an NCE command station to check there..
 

The steps I'm following are as follows... (Assumes you have JMRI and a Speedometer at the ready..)
Plus..
-- I made a spreadsheet to calculate the CV values for all throttle positions across a default table in step 2.
Email me at inmrlive@gmail.com if you want a copy.
--- If you don't have the spreadsheet, you only need to do Step 1, 3, 8, then skip to 11, but it will take longer to find the correct CV value for the speed in Step 11.

1: Make sure the Locomotive is set to 28/128 step throttle response.
2: Set the Locomotive to run on the 28 step table AND set the 28 step table to a linear curve (Step 1 = 0, Step 28 = 255)
3: Run the loco at 80% for approx. 15 minutes to warm it up.. (or until the speed stabilizes)..

You'll need the spreadsheet for Steps 4-7
4: Slow the loco to 50MPH and note the throttle %.
5: Slow the loco to 25MPH and note the throttle %
6: Slow the loco to where it barely moves and note the throttle %
-- I made a spreadsheet to calculate the CV values for all throttle positions across a default linear table.
Email me at inmrlive@gmail.com if you want a copy.
7: Enter the throttle positions in Step 4, 5, 6 into the table - this will calculate the new (approximate) CV Values to match speed to throttle position.
8: Lock Speed Steps 2, 9, 17, 25, and 28 on the table. (This allows JMRI to calculate the CV Values for the steps between your target points.).
- The alternative you have to type all of the 'New CV' values into the table by hand.
9: Enter the New CV values into JMRI on the corresponding locked steps.
10: Write the changes to the sheet.
11: Retest the throttle positions and adjust the CVs to fine tune the speeds.
Set throttle to 50 and tweak Speed Step 17 until your running at 50mph
Set throttle to 25 and tweak Speed Step  9 until your running at 25mph
Set throttle to 3 and make sure locomotive is still crawling.. (Adjust Speed Step 2 as needed)
 
For Fast running/Passenger locos you would add Speed Step 25 with a target of 75mph and Speed Step 28 with desired max speed.
for Slow locos, I set Speed Step 25-28 at 10 above the value of Speed Step 17

I haven't worried about reverse trim yet but it's on my radar.. so..
-- I haven't done this next step yet.. but
12: Set loco to reverse and repeat step 11 using the Reverse Trim CV.

Now that I have several under my belt, I can get a loco tuned in about 5 minutes after the brake in period.

Definitely not an exact science, but its nice to have a Kato and Atlas Loco run together vs a 50mph speed difference..

~Ian