Trains.com

Subscriber & Member Login

Login, or register today to interact in our online community, comment on articles, receive our newsletter, manage your account online and more!

Getting slow speed on DCC

17254 views
44 replies
1 rating 2 rating 3 rating 4 rating 5 rating
  • Member since
    April 2020
  • 1 posts
Posted by Negosav Stosic on Saturday, April 18, 2020 2:05 PM

I'm less than a year into the hobby so still learning a lot.  My locomotives are mostly LokSound 5 and a few with LokSound Select (no Tsunami anymore).  My Walthers Proto and Atlas Gold units run perfectly right out of the box (only change was momentums).  However I have a couple of Athearn Genesis locomotives even after extensive break-in that were giving me problems especially as they came to a stop.  The would slowly come to a stop and then jerk forward slightly a couple of times until they finally did stop.  I put CV54 (BEMF effect) from 50 to 255 and CV55 (inertia) from 100 to 10 and it solved the issue.  It runs great now!  I will continue to test and see if I can back this 255 number down or increase the 10 value (not sure if I need to although these numbers are far from the settings in the Walthers and Atlas units).  CV51 and CV52 are left at the original settings of 0.  Additionally I changed CV5 (top speed) to 100 and CV6 (midspeed) to 30 to enhance low speed control performance.  CV 2 was at 1 from the factory and left that way.  I also changed the momentums CV3 and CV4 which were both at 20 and now set to 80.  The range of stock values for CV54 for my locomotives range from 24 to 53.  CV55 stock settings range from 27 to 100.  These wide ranges tell me there are many combinations that can achieve prototypical movements.

  • Member since
    February 2002
  • From: Reading, PA
  • 30,002 posts
Posted by rrinker on Saturday, September 2, 2017 11:01 AM

 In a video where Matt Hermann demonstrates the CV54 auto configuration, he makes several mentions of how you need to be careful that the track is clear, and that some locos will go further than others while configuring - the one he demonstrates with goes only about 1/4 of the way around a simple loop of one of the plastic based sectional tracks. 

 One would PROBABLY expect a 24 volt motor to take off at half the speed of a 12V motor. But the total distance traveled could possibly be a lot longer since part of the BEMF adjustment requires the decoder to collect a sufficient number of samples.

                            --Randy

 


Modeling the Reading Railroad in the 1950's

 

Visit my web site at www.readingeastpenn.com for construction updates, DCC Info, and more.

  • Member since
    May 2004
  • 7,500 posts
Posted by 7j43k on Friday, September 1, 2017 11:26 PM

Useless?  Well, maybe.

I've got some motors tucked away that are 24V.  As opposed to 6V.  They're the famous, or not, 1" cube motors.  Perhaps Dave's experience might be of use on the great day when/if I "make them go".

 

Ed

  • Member since
    July 2006
  • From: Bradford, Ontario
  • 15,797 posts
Posted by hon30critter on Friday, September 1, 2017 10:51 PM

You can likely file this under "Totally Useless Information", but there was one incident where I was glad I had a few feet of track set up when I did the CV54 thing. Like others have mentioned, the procedure usually takes only about 12" of track, but on one occassion it was all I could do to catch my scratchbuilt railtruck before it launched itself off of the end of the three feet of track at something approaching the sound barrier. I wondered what the heck was wrong, but I couldn't figure it out. When I posted about the incident someone suggested that I probably had a 6V motor in the thing. That made sense. The motor came with a very primitive and apparently ancient kit which I used as fodder for building. I had assumed that it was 12V because it came in an HO scale kit. There was no name on the package.

Changing the motor would have required that the truck body be partially destroyed, so I decided to play with CVs 2, 5, and 6. I was pleasantly suprised that I was able to get the railtruck to run very nicely by setting the three CV values very low. I can't remember the exact values but I think they are all in the single digits.

Since I doubt that many of you are planning on using 6V motors, the above information is, as I said, quite useless. Funny, but useless.LaughLaugh

Dave

I'm just a dude with a bad back having a lot of fun with model trains, and finally building a layout!

  • Member since
    May 2004
  • 7,500 posts
Posted by 7j43k on Tuesday, August 29, 2017 7:50 PM

Randy,

Thanks for the thought.  For now, I think I'll let this go.  It IS pretty darn nice the way it is, right now.  I've got a Free-mo event to get ready for.  I'll bring it with me, and see what "the folks" say.  Also, since it's an informal event, I thought I'd bring my programming stuff and maybe play with it when I ain't running trains.  Gotta let the other guys play too, ya know.

 

Ed

  • Member since
    February 2002
  • From: Reading, PA
  • 30,002 posts
Posted by rrinker on Tuesday, August 29, 2017 5:16 PM

 By "working too hard" do you mean it sounds like it's revving up the prime mover way too much for alight engine move? Try reducing CV3 AND also turning up the throttle slower. Or just the latter. If you crank it open and rely on the momentum to accelerate slowly, the prime mover will run wide open like it's startting a heavy train. If you don't want it to go right into a high notch, you need to bump the throottle up slowly, like you are trying to accelerate slowly without the help of momentum. Part of what triggers the higher notch on the prime mover is the difference between commanded speed and actual speed. Dunno if that is still the case with the Full Throttle feature, I have no place to run anything and I haven't loaded the updated sounds into any of my locos.

                                    --Randy

 


Modeling the Reading Railroad in the 1950's

 

Visit my web site at www.readingeastpenn.com for construction updates, DCC Info, and more.

  • Member since
    May 2004
  • 7,500 posts
Posted by 7j43k on Tuesday, August 29, 2017 3:30 PM

I will say that I am not at all unhappy with how this little guy runs.  Now.  And I appreciate both the advices and the learning opportunities provided by all.

What is really quite neat, as I run between the computer and the layout, is that the CV's can be changed almost in an instant.  Over and over and over.  That is NOT true for a locomotive paint job.  Or, for that matter, a DCC/sound install.

 

Ed

  • Member since
    April 2004
  • From: Ontario Canada
  • 3,574 posts
Posted by Mark R. on Tuesday, August 29, 2017 3:22 PM

I personally don't like using the Auto Tune feature (unless I HAVE to, to get an engine running right) for just the reasons you are experiencing. All my engines run just fine without having to resort to that.

I'm sure there are ways to tweak it manually with the four BEMF CVs, but I don't understand them fully enough to actually KNOW what I'm adjusting.

There are no CVs for adjusting the sound transition points on the Select decoders. In the V4.0 decoders, you can go into the sound schedule and adjust the points.

Mark. 

¡ uʍop ǝpısdn sı ǝɹnʇɐuƃıs ʎɯ 'dlǝɥ

  • Member since
    May 2004
  • 7,500 posts
Posted by 7j43k on Tuesday, August 29, 2017 2:57 PM

Strictly speaking, I didn't say "reset".  I said "rearranges the output".  The "output" being what gets sent to the motor.

So, I just wandered into the train room and checked:  CV5 is currently at 255.  So I don't see CV5 being increased.

Actually, I cranker er up to full throttle, and the top speed looks just fine-ish.

I guess what I'm seeing is that a light engine is sounding like it's working way too hard for the speed and acceleration I'm seeing.

Are there CV's for that?????

I did, by the way, change CV3 and 4 from 80 to 40, to see what would happen.  Not much change, to me.

 

 

I am impressed with ESU decoders.  Not so much some of their instructions.  Which is an incredibly common problem with electronics.  Instructions should be designed for people who DON'T get how something works just by touching it.

 

Ed

  • Member since
    April 2004
  • From: Ontario Canada
  • 3,574 posts
Posted by Mark R. on Tuesday, August 29, 2017 12:54 PM

7j43k

In addition, doing the CV54 automatic thingy kind of rearranges the output of those numbers.  Too.  Swell.

 

 

 

 

No, the Auto Tune does not reset ANY of the user-defined speed table settings. It adjusts the four CVs used in setting the BEMF performance. BEMF is most beneficial in the lower range speeds, so it is adjusting the parameters so the engine will perform optimally at lower speeds - hence you noted how much slower the engine would crawl. Since these parameters affect the entire speed range, it only makes sense that the upper levels would be affected as well. Again, that can easily be compensated for by increasing CV5 without affecting the speed curve you already created.

Starting to see the beauty of how Loksound does it yet ? Smile, Wink & Grin

Mark.

¡ uʍop ǝpısdn sı ǝɹnʇɐuƃıs ʎɯ 'dlǝɥ

  • Member since
    April 2004
  • From: Ontario Canada
  • 3,574 posts
Posted by Mark R. on Tuesday, August 29, 2017 12:47 PM

Loksound has five pre-set speed curves as well .... Exponential1 / Exponential2 / Linear / Logarithmic1 / Logarithmic2 in addition to just creating your own.

Mark.

¡ uʍop ǝpısdn sı ǝɹnʇɐuƃıs ʎɯ 'dlǝɥ

  • Member since
    May 2004
  • 7,500 posts
Posted by 7j43k on Tuesday, August 29, 2017 12:33 PM

maxman

 

Whoa, big guy, no insult was intended. 

 

OK, then.Smile

 

As has been hinted at by myself (and observed by others) I'm a beginner.  The only other decoder I've messed with "extensively" has been a Soundtraxx.  First, keep in mind that I had a "keep alive" installed, so I couldn't read the CV's (learned THAT one pretty quickly) (and, yes, I realized I could have temporarily disconnected the "keep alive"--with no small amount of work).  Anyway, it was then obvious that I had to actually do some planning and record keeping.

As delivered, the Soundtraxx installation had far from ideal running traits (remember: gas electric).  I had heard of the speed curve adjustment.  And I figured (perhaps wrongly) that the pre-installed speed curve was dead linear.  So I came up with numbers based on that and what I wanted, and input new CV's for 67-94.  And 29, etc.  Worked BEAUTIFULLY right outa the box.  For 67-94, I made a number column as if there was a linear array of increases (since that appeared to match how it was running). Then I made a number column that reflected how I thought the speed curve should be.  And input my new numbers.

Hey, I'm a programming wizard!!

So my next one was an ESU equipped ATSF GP9.  I approached the programming exactly the same way as the gas electric.  Assuming, incorrectly, that DCC decoders had a standardized program.  Seemed reasonable at the time.

And here I am.  But enough about me.....

 

Your comments have been and are helpful and useful.  But the way ESU set up the various speed "choices" is not straightforward.  To me.  And I certainly don't think they did a great job of explaining things in their official manual.

What it appears is so is that IF you wish to utilize a custom speed curve (your own or pre-installed), that speed is ALSO affected by CV's 2 and 5.  Hence, one uses CV2, 5, and 68 (!) through 93 (!) with 67 always automatically fixed at 1 and 94 always automatically fixed at 255.  Thus, although CV 94 stays constant, the top speed of the model may vary, caused by changes in 5.  And intermittent 

In addition, doing the CV54 automatic thingy kind of rearranges the output of those numbers.  Too.  Swell.

 

Please be reminded that when I programmed the Soundtraxx decoder and changed CV 94 to 156, it just did it.  And the model matched that nicely.  Without an opinion.  

 

Whatever ESU did, I'm not saying it's a bad idea.  I'm saying they are not terribly clear on how to use what they did.  Which is a problem.

 

 

Ed

 

 

  • Member since
    February 2008
  • 8,877 posts
Posted by maxman on Tuesday, August 29, 2017 11:41 AM

7j43k
I have to ask, why would you ever think to ask if someone would pick speed curve values at random? Do you really think anyone would do that? Do you really think I might? Do you see that a person might see that as insulting? Did you mean it that way?

Whoa, big guy, no insult was intended.  The reason I asked the question was because you were speaking about a speed curve and changing speed curve step values but it wasn't clear if you were changing only the CV's for steps 1 and 28 or others also.  It also seemed that you were not aware of the fact that the CV2 and CV5 values defined the end points of the "curve", and that CV6 had no effect on the curve.

In addition, I have a strong suspicion that the "curve" installed by Litchfield was really nothing other than the curve that was the LokSound decoder default, unless you specifically asked them to do something other than just download the decoder files.  And I also believe, although I haven't looked, that the default "curve" is a straight line defined by whatever values were programmed into CVs 2 and 5.

When you make the changes to CV29 (if that is the one) that allows you to use speed curves, that usually means "user defined speed curves".  So if the only defined speed curve is a straight line, that's what you're going to get. Now in the case of a Soundtraxx decoder (yes I know that's not what you have) there are a group of speed curves that you can pick from, including one that is a straight line.  There is also a curve that can be totally owner generated.  Thus you can pick any of the pre-loaded curves or the one that is owner generated by changing another CV value.

So anyway, that was why I asked my question.  Again, no insult was intended.  However, if you really do feel insulted, let me know and I'll add you to my list of those to whom I should not offer any assistance.

Have a nice day!

Dick (oh, and that happens to be my name, not an insult)Smile

  • Member since
    May 2004
  • 7,500 posts
Posted by 7j43k on Monday, August 28, 2017 8:17 PM

Thank you, Mark.

 

I'll go in and muck around a bit more, later.

 

 

Ed

  • Member since
    April 2004
  • From: Ontario Canada
  • 3,574 posts
Posted by Mark R. on Monday, August 28, 2017 8:02 PM

I've found the CV54 Auto Tune always resulted in my engines running overall more slowly .... no idea why, it's just the way the decoder interpolates the BEFM from the motor and optimizes the settings. DO NOT ask me HOW it does that ! Smile, Wink & Grin

As you discovered, having set your speed table to where you liked it, the Auto Tune now makes it run too slow. Going back in to CV5, you can pick up the pace again. If you have the speed table activated, CV6 has no bearing on it - just CV2 and CV5. If you have the speed table off, then you use the three point curve utilizing CV2, CV5 and CV6.

Mark.

¡ uʍop ǝpısdn sı ǝɹnʇɐuƃıs ʎɯ 'dlǝɥ

  • Member since
    May 2004
  • 7,500 posts
Posted by 7j43k on Monday, August 28, 2017 7:18 PM

maxman

 

 
7j43k
You may also define your own speed curve: simply enter the desired values in CV's 67 to 94. The decoder will superimpose these 28 values onto the real speed steps." To me, that sounds like what I'm hearing, here. The "real speed steps" would seem to be CV 2, 5, and 6.

 

You still didn't answer whether or not you are using JMRI to do the speed step CV changes.  Again, I think that the only two speed step settings that carry over are CV2 and CV5.  I've not heard/read anyplace the the mid-speed CV (6?) carrys over.  You said you are using the complex speed curve, but exactly what other CVs are you changing?  Are you just picking the other speed curve values at random?

 

No, I'm not using JMRI to do the speed step CV changes.  The CV's 67 to 94 are whatever Litchfield Station installed.  I haven't changed them.  I DID change CV2, 5, 6.  And I have since changed them back, also to what was supplied by Litchfield.  So, no, I'm not picking other speed curve values at random.  I'm not picking them at all.  The Litchfield folks picked them.

As far as I know, the only CV that is different from what they shipped is CV54 (which was shipped as 50 but self-adjusted to 96, and 29, which was 14 (??) and I changed it to 50.

 

Ed

 

I have to ask, why would you ever think to ask if someone would pick speed curve values at random?  Do you really think anyone would do that?  Do you really think I might?  Do you see that a person might see that as insulting?  Did you mean it that way?

  • Member since
    February 2008
  • 8,877 posts
Posted by maxman on Monday, August 28, 2017 6:58 PM

7j43k
You may also define your own speed curve: simply enter the desired values in CV's 67 to 94. The decoder will superimpose these 28 values onto the real speed steps." To me, that sounds like what I'm hearing, here. The "real speed steps" would seem to be CV 2, 5, and 6.

You still didn't answer whether or not you are using JMRI to do the speed step CV changes.  Again, I think that the only two speed step settings that carry over are CV2 and CV5.  I've not heard/read anyplace the the mid-speed CV (6?) carrys over.  You said you are using the complex speed curve, but exactly what other CVs are you changing?  Are you just picking the other speed curve values at random?

  • Member since
    May 2004
  • 7,500 posts
Posted by 7j43k on Monday, August 28, 2017 6:42 PM

Paul,

I hadn't tried the CV54 thing yet.  So I did.  After first recording the original number.  Minimum speed dropped to .1MPH.  And it was pretty smooth.  Bad news was that the fella goes REAL SLOW now, at least according to the sound.

When I first put the loco on the track a couple of days ago, with sound, it felt like the loco was going way faster than the engine sound.  Then when I changed the low number CV's an hour ago or so, it felt like the engine sound matched pretty nicely the loco speed.  Now it sounds like the sound is way up there, compared with the speed.

So the CV54 adjust was, all in all, anti-rewarding.

So it still needs work.  I do wonder what would happen if I restored the low number CV's and then did a CV54.

OK, I did restore the low number CV's to "as delivered".  It's "less negative".  Meaning it still seems like the loco is working way too hard (soundwise).  But less than just awhile ago.

So, I'm pretty happy with things.  Certainly enough to leave it for now.

DCC is SO neat.  And SO much better than the olden days.  Though a little intimidating.

 

 

Ed

  • Member since
    January 2010
  • 2,616 posts
Posted by peahrens on Monday, August 28, 2017 5:59 PM

I would try the CV54 procedure again.  See where it runs slow better (not necessarily slowest), at CV2 set at 1 vs 2, with the Cv54 procedure done for each.  You might have a choice to make; i.e., slower at CV2 equals 1 vs. smoother at CV2 = 2.  Of course if it runs equally smooth at 1 or 2 value the choice is easy.

Would enjoy the answer if you don't mind experimenting more.  All this is informative as there's a learning curve.

BTW, if you have not tried the Loksound Yahoo User Group it is worthwhile monitoring.

Paul

Modeling HO with a transition era UP bent

  • Member since
    May 2004
  • 7,500 posts
Posted by 7j43k on Monday, August 28, 2017 5:38 PM

More experiment:

I checked some more CV's (keep in mind that this decoder was reprogrammed by Litchfield Station from a 73800 to a 73836):

CV2 was 3, I changed it to 1

CV5 was 255, I changed it to 130

CV6 was 88, I changed it to 50

 

Keep in mind that I've (supposedly) got CV29 set for the complex speed curve.

 

Results:  Speed step 1 went from 1.3 mph down to .2 mph.  Uh, WOW!

I'll note that you can see a bit of jerkiness from the motor cogging.  And I suspect that's fixable a bit, too.  Not for me now, but I got a feelin'........

 

I also noted that, when I tried changing CV 25 from 0 to 16 (which I heard somewhere I should do), the decoder wouldn't accept it, and stayed at zero.

 

But, I'll tell ya, I am indeed happy at the improvement.

 

Thanks, all for helping out.  And a couple of extra thanks for Mark.

 

 

Ed

 

PS:  This little loco is a zebra stripe Santa Fe GP9 (P2K), destined for road switcher service.  So she'll run solo, mostly.  Besides the decoder, there's a speaker and "keep-alive", all by Loksound.  And as much lead as I could conveniently fit.  A real sweety.  And I ain't even a Santa Fe guy.  Well, a little, I guess.

 

  • Member since
    February 2002
  • From: Reading, PA
  • 30,002 posts
Posted by rrinker on Monday, August 28, 2017 4:39 PM

 The "real speed steps" are the actual steps issued by the command station, not the steps from CV 2, 6, and 5.

                          --Randy


Modeling the Reading Railroad in the 1950's

 

Visit my web site at www.readingeastpenn.com for construction updates, DCC Info, and more.

  • Member since
    May 2004
  • 7,500 posts
Posted by 7j43k on Monday, August 28, 2017 3:58 PM

More big fun (and confusion):

From page 21 of Loksound Users Manual:

"5.2.2.3. Speed curve

You may also define your own speed curve: simply enter the desired values in CV's 67 to 94.  The decoder will superimpose these 28 values onto the real speed steps."

To me, that sounds like what I'm hearing, here.  The "real speed steps" would seem to be CV 2, 5, and 6.

But:

"When this mode is active, the settings in CV 2, CV 5, and CV 6 have no influence."

 

THAT seems to be saying what I thought in the beginning:  Simple and complex speed curve settings are unrelated.

 

The Loksound Manual is surely being unhelpful.  As, indeed, many manuals are.

 

 

Ed

Moderator
  • Member since
    June 2003
  • From: Northeast OH
  • 17,249 posts
Posted by tstage on Monday, August 28, 2017 3:27 PM

Ed,

I mentioned 3' in my previous post.  However, when I actually ran the automatic motor tuning feature with my H20-44 it didn't even travel 12" before it came to an abrupt stop.  I think 3' should be plenty...unless you plan to run the feature using a 4-8-8-4 Big Boy.

To answer your other question: I think a decoder reset is what you'd need to perform an "undo" of the motor tuning feature.

Tom

https://tstage9.wixsite.com/nyc-modeling

Time...It marches on...without ever turning around to see if anyone is even keeping in step.

  • Member since
    May 2004
  • 7,500 posts
Posted by 7j43k on Monday, August 28, 2017 3:26 PM

peahrens

 

I'd say 3-5 ft.

 

 

Thanks Paul.

I've got much more than that.

 

 

Ed

  • Member since
    May 2004
  • 7,500 posts
Posted by 7j43k on Monday, August 28, 2017 3:25 PM

I thought I'd try a test:

(remember, this is a Loksound decoder)

CV 67 was set at 1, CV 68 at 5.  I changed both of them to 40, to see if there's be a sudden bout of quickness on start.  No noticeable change.

I DID see that CV67 did not accept the new value--it stayed at 1.

That would appear to endorse what Mark has said.

Well, there you are.  Or, more correctly, here I am.  I guess.

 

I do kinda get the point.  Sorta.  In particular, at least for Loksound, there isn't straightforward choice between simple and complex speed curves.  As Mark says, some of the data from the simple speed curve is incorporated into the output of the complex speed curve.

Thanks, Mark.  I'll have to think on this some more.

 

Ed

  • Member since
    January 2010
  • 2,616 posts
Posted by peahrens on Monday, August 28, 2017 2:54 PM

7j43k
They say be sure to have plenty of track available for the auto-tune.  Since I don't have a full loop available, just how much track are we talking about?

I'd say 3-5 ft.

Paul

Modeling HO with a transition era UP bent

  • Member since
    February 2008
  • 8,877 posts
Posted by maxman on Monday, August 28, 2017 2:40 PM

7j43k
But. It appears you can only do one or the other. If you are using CV2, for example, you cannot be using the speed tables. And if you are using the speed tables, you cannot be using CV2.

No, that is not what they are saying.  What they are saying is that if you set CV2 (min) and CV5 (max), the values you end up with there are carried over to the speed tables for speed step 1 and speed step 28.  So you can set steps 2 through 27 but you can't set, by using the tables steps 1 and 28.

I haven't actually looked recently, but from other forums I think you will find that steps 1 and 28 are greyed out if you are looking at the tables in JMRI.  And if you are not using JMRI to adjust the speed table values, you are going way out of your way to get the job done.

  • Member since
    April 2004
  • From: Ontario Canada
  • 3,574 posts
Posted by Mark R. on Monday, August 28, 2017 11:10 AM

See the full speed curve section here (1/3 down) ....

http://www.intermountain-railway.com/customerservice/dccwebpage/ESU-FAQ-Page.html

Bit 4 is the value 16 you see mentioned.

Regarding the Auto Tune - If you are not happy with the results, just do a factory reset (CV8 = 8 > power cycle). If you don't want to do a full reset, note the values in CVs 52, 53, 54 and 55 before you try it. These are the four values affected during the Auto Tune.

Mark.

¡ uʍop ǝpısdn sı ǝɹnʇɐuƃıs ʎɯ 'dlǝɥ

  • Member since
    May 2004
  • 7,500 posts
Posted by 7j43k on Monday, August 28, 2017 11:05 AM

Paul and Gary,

Yup, the "automatic tuneup" sounds interesting.

Two questions do come up:

 

If I'm not happy with whatever the "thing" does, how do I undo it?

They say be sure to have plenty of track available for the auto-tune.  Since I don't have a full loop available, just how much track are we talking about?  My available trackage has a finite length.

 

Ed

Subscriber & Member Login

Login, or register today to interact in our online community, comment on articles, receive our newsletter, manage your account online and more!

Users Online

There are no community member online

Search the Community

ADVERTISEMENT
ADVERTISEMENT
ADVERTISEMENT
Model Railroader Newsletter See all
Sign up for our FREE e-newsletter and get model railroad news in your inbox!