Speed Matching Projects

At last nights operating session I managed to pick up a couple of speed matching projects from the host Mike. First there is a Proto 2000 SD60, this is the same as my Oakway (just different paint and decoder).

Since my SD60 is by far my smoothest runner it really didn’t take long to get this one matched, I ended up matching it to my SD60 (per owners request) in about an hour which I thought was a pretty good time. In the same period of time I also got some work on my module’s fitter rails done.

The other project is a GE C44-9W, I’m not sure who the manufacturer on this one is however.

This locomotive has been a bit more of a cantankerous beast in regards to speed matching. I’m fairly confident it can be done but this one will take a bit longer.

More Speed Tables

In my previous attempts at speed matching I had set them up as CV19 consists for matching since Digitrax decoders change BEMF settings in consist mode. With BEMF off I noticed some of my locomotives were not very smooth at lower operating speeds so I decided to take a stab at leaving BEMF on and consisting using the functionality built into the Digitrax command station (I believe it’s called Universal consisting). Using this approach lets me have BEMF and makes it easier to build up and break down consists while operating.

This meant that I had to reconfigure the speed tables with BEMF on. As I’ve mentioned earlier my method for building speed tables is to take assign a target scale speed for each of the 28 speed steps. Then I use my Bachrus Speedometer to find a value for each of the 28 speed step that is as close to the target speed as possible.
(Example table)

Speed Step	Scale Speed	Loco 1	Loco 2
1	2	15	18
2	4	20	23
3	6	25	28
…	…	…	…
28	56	100	105

The Bachrus speedometer is intended to be used with their roller sets, however due to their higher cost I only have the speedometer so I prop the locomotive up like this (with alligator clips connecting it to track power).

After I’ve got it all setup and rolling on the speed0 properly I keep up the locomotive on and set the speed to 100%. I find this is easier than trying to make sure it’s in a specific step. I fire up JMRI and make the speed table a flat line at my starting speed and as I get the value for a step I move to the next one till I’ve built a full curve.

In this screen shot (speed table programmer in Ops mode on left, speedo output on right) I’m working on Step 11 which I have defined as 12 scale mph, based on the Speedo reading I’m as close as I’m going to get. Generally I get as close to the MPH as I can (I don’t try to match increments smaller than 1 mph). This does leave a bit of push in pull between the locomotives but it would be possible to get that precise and that’s what happens on the prototype anyway. I run all metal Kadee couplers on my locomotives so they seem to stand up to the push and pull of operation (at least so far).

Speed Matching Tips and Tricks Part 2

In my last post I wrote about some observations I’ve made in my experiences using the Bachrus Speedometer to do speed matching. This time I’m going to describe the steps I go through for my locomotives.

I start by setting up a spreadsheet to keep track of the speed table in scale miles per hour that will be used to map the CV values to actual speeds. It would look something like this (completely made up numbers):

Speed Step	Scale Speed	Loco 1	Loco 2
1	2	15	18
2	4	20	23
3	6	25	28
…	…	…	…
28	56	100	105

The first column is the speed step as it maps to the actual CVs, the second column is the target scale speed. Then there is a column for each locomotive that is going to use this table. The values correspond to the CV values that result in the desired scale speed.

To get these values I start with a locomotive and put it on the speedometer, before I start figuring out CV values I run it for a while on the rollers to make sure the motor and drive train are warmed up. Once I’m ready to start I use JMRI and OPS mode programming to set all of the speed steps to the same value which should be relatively close to the target speed (largely a guess) and then I take a throttle and set it to the maximum speed and adjust the steps until it matches the target speed. At this point the speed table is flat.

Next I pick the next speed and move adjust every step except the first one until I find the second step. I basically repeat this for each step, leaving each find value in as I go so by the end you’ve got your full table. I also copy this to information to my spreadsheet as I go for safe keeping.

Once you have more than one item set up you can start testing them together. You’ll want to pay attention to the acceleration and deceleration programmed into the decoder, if you find a locomotive behaves differently going in reverse you can tweak the forward and reverse trim and if a locomotive just needs a little help getting started you can program in some kick-start.

Speed Matching Tips and Tricks

I’ve been watching the statistics on what’s driving traffic to my blog and noticed a lot of it is searches for Bachrus and/or JMRI so I thought I’d post an update.

There are three components I use in speed matching, the Bachrus Speedometer, JMRI, and a Digitrax PR3. As I have now successfully matched three very different locomotives from three different manufactures with three different decoders I have some tips and tricks.

1. A good drive train can make all the difference. My Kato SD38-2 and Proto 2000 SD60 were very easy to speed match as they run well at low speeds and across the range are smooth runners. My Athearn SD45T-2 however has been a bit tougher, it varies quite a bit even at constant voltage so it can be frustrating maintaining constant speeds to set up your speed table.The takeaway here is that the easy of matching can vary quite a bit between locomotives so don’t give up if things aren’t going smoothly.
2. If you are using a Digitrax PR3 and JMRI keep the PR3 plugged in to the USB port even when not in use. This solves two problems I’ve run into, first it helps ensure that the correct driver is always loaded and second it ensures the COM port assigned is the same.At least on Windows 7 I’ve observed that in some cases plugging in the PR3 will cause Windows to load a generic USB to Serial driver which will not work.If Windows assigns a different COM port to the PR3 or Bachrus Speedometer JMRI won’t function until you open it, fix the COM port assignment and restart the app. While this is simple it’s tedious.Once I switched to a dedicated desktop rather than my laptop these problems have disappeared.
3. Use the latest version of JMRI and keep up to date, since I started I’ve gone through a couple of versions of JMRI and it has gotten better as versions progress.
4. If you are going to use advanced consisting (CV19, 20, and 22) it’s probably a good idea to do your speed matching while your locomotive configured for an advanced consist. Digitrax, and possibly other manufactures, alter the BEMF behaviour in consist mode so a table that may work individually may not work the same in consist mode where it really matters.
5. You probably aren’t going to get a perfect match, even with identical locomotives with identical decoders and identical usage so expect a bit of push and pull between the locomotives. This happens with the prototype and so the point is to get close but you are just going to drive yourself mad if you try to get it perfect.

Speed Matching

One of my ongoing projects has been speed matching all of my locomotives, so far I haven’t been too successful. It appears the crux of my problem was BEMF, it appears Digitrax decoders turn it off when you use the advanced consisting (CV19). To go around this I just set up the locomotives in a consist and set up my speedtable that way. The lower speeds are much harder to set up without BEMF on (an amazing difference). I shot a short video (only 46″ of track to test it on).

The first locomotive (DM&IR 210) is a Kato SD38-2 with a Digitrax SDH164K1B (sound decoder) and the second (EMDX 9003) is a LifeLike Proto 2000 SD60 with a Digitrax DH165.

Bachrus Speedometer & JMRI tip

As I mentioned in a previous post I had picked up a Digitrax PR3 and Bachrus Speedometer (http://bachrus.com/speedometer.php) for some speedmatching and I got the version thats hooks up to a PC and DecoderPro. I downloaded the latest version of JMRI, or so I thought, and found it odd that it didn’t match what the JMRI notes for it look like (http://jmri.org/help/en/html/hardware/bachrus/index.shtml). It turns out that I had downloaded the latest stable version (2.10) but the latest development version (2.11.4) has numerous improvements specifically for the Speedometer.

Mean while I got speed tables built for my DM&IR 210 and 413 but somehow they are way off in terms of scale. By that I mean if I modify the forward and reverse trims on each I can get them lined up so I’ll have to figure out what happened there.

Speed Matching Attempt 1

I recently attempted speed matching a couple of my locomotives with a Digitrax PR3 and Bachrus Speedometer (Got a good deal and great service from TrainTekLLC.com) with the intention of speedmatching since I don’t have a large enough space to run them in parallel. I’ve got two of the fairly close, but more on that later.

If you find yourself purchasing a Digitrax PR3 and don’t already have a Loconet cable you should probably remember to add one to your order. The RJ12 connector isn’t something you will find at Radio Shack. If you do find yourself without a cable check the O-scale section of your local hobby shop as Atlas makes a cable that works just fine.

On to speed matching… I got two of my locomotives pretty close but did all of my work by taking the shell off, putting it on the rollers and using alligator clips to hookup directly to the decoder which worked well but I discovered afterwards that one of the locomotive’s wheelsets is either having pickup problems or is enough out of gauge to not have good electrical pickup.

I also installed a rotary beacon on this troublesome locomotive (I didn’t realize it was troublesome when I started) and while it works great I’m having trouble remapping the functions on it. I installed it as F1but ran into problems when consisting it with my sound locomotive. When the beacon is on the sound locomotive’s bell is on too. I tried using JMRI but didn’t get it working. I guess I’ll have to hit the drawing board on that one.