JMRI Speedometer

[John]

I have started to fool around with speed matching my locomotives. Most 1:1 loco's have a top speed of around 70-75 mph.

Previously, I used 4 foot modules and a stop watch. This process involved running a loco back and forth over the closed course, trying to time the point it entered the course, and the time it left the course. Using 4, 8, 12 foot courses, you could come to a reasonable estimate of the loco speed.   Never leave a simple process alone, unless you can make it better through complicated technology.  ;D

JMRI includes an application, that allows you to use your existing block detectors, a ruler, and an computer interface to your layout to do the same thing. This app takes advantage of the fact that (in loconet at least), when a block detector goes active, a message and a time stamp is sent over the network. The JMRI application requires you to use 3 blocks, which can be of  variable lenght, to perform automated timing operations. In my case, I used blocks of 4600 and 3500 scale feet. As the loco enters the first block, the timer starts. When the loco hits the second block, the first timer stops, and a speed calculation is made. The second timer is now started. As soon as the loco enters the third block, the timer once again stops, and another calculation is made. This is a great way to determine the current top speed of a loco, then adjust it down from warp drive to a more reasonable value



More information is here
http://jmri.sourceforge.net/help/en/html/tools/speedometer/NCE-Speedometer.shtml

[ednadolski]

I've found that speed calibrations tend to drift over time.  It's no fun to have to periodically re-calibrate a fleet of locos.  Anyone else see that issue?

[davefoxx]

I've found that speed calibrations tend to drift over time.  It's no fun to have to periodically re-calibrate a fleet of locos.  Anyone else see that issue?

Perhaps that has to do with a locomotive breaking in or, over time, becoming more in need of lubrication and/or cleaning.

DFF

[John]

I've found that speed calibrations tend to drift over time.  It's no fun to have to periodically re-calibrate a fleet of locos.  Anyone else see that issue?

Perhaps that has to do with a locomotive breaking in or, over time, becoming more in need of lubrication and/or cleaning.

DFF

I think thats a pretty good thesis ...

[Ed Kapuscinski]

Ken McCory had a speed trap on his layout.

I like the idea.

[Day One]

you can easily take this a step further.
In the Yahoo JMRI Users Group there is a script posted for automatically speed matching loco's.
http://groups.yahoo.com/group/jmriusers/files/Script%20examples/ and look for the Speed Table script and panel.
Using a loop of track (kato unitrack with insulating joiners works great) divided into 4 blocks and connected to a BDL168 and interfaced to JMRI, you can load the script and it does the speed matching automatically. It takes roughly 15 min to do it's magic but it's all hands off. Put the loco on the track and start the script. It will warm the loco up for a few min in both directions, then it runs it at different speeds while figuring out it's speeds and writes the appropriate CV's to the speed table. When you are done, you can then readback the CV's into a decoder roster file to save it if you like.

I've seen it in action from the person who designed and wrote the script and am currently building my own as well.

[Day One]

I talked to Phil who wrote this script tonight and the more, small blocks you use the better. Hence the BDL168. Using 8 blocks on a small loop of track it takes roughly 15 min to match a loco. Using 4 blocks on the same loop of track will take much longer. Possibly up to an hour.

[Ian MacMillan]

How does one get Phils script work? I tried to get it running last night but kept getting an error and the read me is mostly just changelog info.

[bsoplinger]

#   Hardware tested with this script
#   Command Station - Digitrax DCS100, Digitrax Zypher
#   Train Detection - BDL168 (Board Address 1),Team Digital SIC24 with DBD22's
#   Computer Interface - MS100, Locobuffer II, PR3

Seems rather expensive here unfortunately.
DCS 100: $250
Zephyr: $170
BDL168: $150
MS100: $40
Locobuffer II: ? (Locobuffer USB: $70)
PR3: $65

Now I assume I'd only need one of each category, so that'd be:
Zephyr: $170
BDL168: $150
PR3: $65

That's the cheapest in each category and uses a USB computer interface. That's still almost $400 for a system just to do automated speed programming. Yikes!?!?!?

[Ian MacMillan]

#   Hardware tested with this script
#   Command Station - Digitrax DCS100, Digitrax Zypher
#   Train Detection - BDL168 (Board Address 1),Team Digital SIC24 with DBD22's
#   Computer Interface - MS100, Locobuffer II, PR3

Seems rather expensive here unfortunately.
DCS 100: $250
Zephyr: $170
BDL168: $150
MS100: $40
Locobuffer II: ? (Locobuffer USB: $70)
PR3: $65

Now I assume I'd only need one of each category, so that'd be:
Zephyr: $170
BDL168: $150
PR3: $65

That's the cheapest in each category and uses a USB computer interface. That's still almost $400 for a system just to do automated speed programming. Yikes!?!?!?

Well it would also run your main system, you just switch it over with a DPDT....so your actual cost for just automated speed programming is an extra BDL168, which you can get online for $119.

[Day One]

Well, it took almost 1 complete year but I finally have my programming track built and operational.

I built it using the following items:
*SPROG II - used for DCC command station
*Locobuffer-USB - used to monitor BDL168/Loconet
*BDL168 - used for 8 sections occupancy detection
*Kato Unitrack - 11 3/8 radius -8 segments per circle

I wanted this to fold in the middle for storage and transport and I used small radius track to keep this setup as small as possible while still allowing it to run 6 axle diesels and big steam engines. So far I've ran an Athearn SD-45 and a LL 2-8-4 Berkshire with no issues.

It took a fair bit of tweaking for me to get it running! I spent a lot of time pouring over the manuals of all 3 devices to get each configured properly to work together.  I'm not sure if all the functions I want to be able to use with this programmer are working right now either. JMRI seems to be having a few hiccups while running 2 USB devices. I'll be spending more time testing things out with this setup. If I'm missing some of the things I want to be able to do thru JMRI, I may ditch the SPROG and LocoBuffer setup and pull the Digitrax Chief system off my layout. The layout's dismantled right now anyways, so it's not going to be running trains anytime soon.

These pics are a day old and as such, do not show the finished wiring nor the complete wiring setup.
I'll be finalizing the wiring layout after I know it works and once I figure out how to add in my Lok Programmer.

[TiVoPrince]

Seems rather expensive here unfortunately.
DCS 100: $250
Zephyr: $170
BDL168: $150
MS100: $40
Locobuffer II: ? (Locobuffer USB: $70)
PR3: $65

Now I assume I'd only need one of each category, so that'd be:
Zephyr: $170
BDL168: $150
PR3: $65

That's the cheapest in each category and uses a USB computer interface. That's still almost $400 for a system just to do automated speed programming. Yikes!?!?!?

Plan
for me is simply to reuse/repurpose what I own.  Using DCS100, BDL168 and MS100 (PR3 on order because next gen JMRI needs it) that I already have. 

My bigger expense will be the loops of specific HO and N track for the project...

[Ian MacMillan]

Forget the PR3, get the Locobuffer USB. I've heard better reviews with JMRI with the LB'r.

I love how Digitrax boasts that the PR3 is a USB2 device like its a selling point...its Loconet for crying out loud, you are not transfering 1GB video files at high speed.

[Day One]

OK, I never thought this thing would take so much effort to get sorted out. After spending a few days trying to get the SPROG and LocoBuffer setup to work, I decided I'd wasted enough time on it. I couldn't get the setup to write anything to the loco on the track and couldn't get any decoder programmers to function properly. I could drive the loco, read and program the BDL and even run the speed match script fine. But I just couldn't verify that the values were written to the loco.  So I pulled the DCS100 from the layout and put my chief system on the programming track. Well, this has turned out to be a PITA as well. I don't know what's going on anymore LOL. The Chief system will only half way turn on. All indicator lights on the booster are on but the DT400 only 1/2 way works. I can select the different modes but turning the nobs and pushing the number buttons does nothing.  I think I'm going to let this project cool off for a while before I re-visit it.

[TiVoPrince]

Forget the PR3, get the Locobuffer USB. I've heard better reviews with JMRI with the LB'r.

I love how Digitrax boasts that the PR3 is a USB2 device like its a selling point...its Loconet for crying out loud, you are not transfering 1GB video files at high speed.

Sage
advice.  But the PR3 is already on order.  The MS100 was a good ride but timing errors, dropped or unreadable packets mean replacement.  Excellent for programming but I was concerned that it could become a problem when automating.  I'm looking to automate processes, not annoy myself further...