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!

CV changing value BY ITSELF?

4874 views
33 replies
1 rating 2 rating 3 rating 4 rating 5 rating
  • Member since
    January 2014
  • From: Chicago area
  • 335 posts
CV changing value BY ITSELF?
Posted by Arto on Sunday, February 4, 2024 6:43 PM

This is sort of related to my QSI post.

I moved on to speed match three BLI E8 Pragon2 locos.

I've only been changing CV5 values. On several occasions now, somehow, CV1 (loco address) seems to be changing to some random value and therefore changing the locomotive's address.

SO, I'm wondering if the wire gauge I'm using to go from The DigiTrax PR3 or LokProgrammer the Programming Track is too small. For sake of convenience I'm temporarily using a couple of sections of Bachmann E-Z Track with one of those "grade crossing/connector" track sections. The wire is Bachmann E-Z Track, about 10' long. Seems pretty thin. Using a wire gauge I'd say it's 26awg.

I think it's kind of weird that CV1 has changed "by itself" while changing only CV5, on two different locos, same brand, same model.

  • Member since
    March 2011
  • From: Westford MA
  • 542 posts
Posted by Tophias on Sunday, February 4, 2024 7:19 PM

Arto, I don't think the wire conditions you mention have anything to do with the problem.  Hate to bring a negative comment, but those old QSI decoders are junk (ask me how I know). I would bite the bullet and just replace it with a current manufactuers' decoder (Tsunami, LokSound, TCS)

Regards, Chris

  • Member since
    September 2004
  • From: Dearborn Station
  • 24,281 posts
Posted by richhotrain on Sunday, February 4, 2024 8:36 PM

Tophias

Arto, I don't think the wire conditions you mention have anything to do with the problem.  Hate to bring a negative comment, but those old QSI decoders are junk (ask me how I know). I would bite the bullet and just replace it with a current manufactuers' decoder (Tsunami, LokSound, TCS)

Regards, Chris

Another QSI hater.
 
It wounds my heart with a monotonous languor.  Smile, Wink & Grin
 
I love QSI decoders in spite of their quirkiness. 

Rich

Alton Junction

  • Member since
    March 2011
  • From: Westford MA
  • 542 posts
Posted by Tophias on Sunday, February 4, 2024 10:07 PM

Sorry Rich, one of the very few times we'll have to agree to disagree. It's not that I'm a QSI hater, but they have always had that wonky "go to sleep" issue and they often need to be reset. Plus, their sound file for diesel just doesn't sound anything like a EMD nor an Alco. That said, they originated many years ago in the infancy of DCC and were life changing. It's just that nowadays they're a dinosaur. Just my $.02

Regards, Chris

  • Member since
    September 2004
  • From: Dearborn Station
  • 24,281 posts
Posted by richhotrain on Sunday, February 4, 2024 10:10 PM

Ahh, I was just joshing, Chris.

Truth be told, I am stuck with a whole bunch of QSI-equipped locos.

Rich

Alton Junction

  • Member since
    March 2011
  • From: Westford MA
  • 542 posts
Posted by Tophias on Monday, February 5, 2024 8:06 AM

Yeah, I realize that Rich. I just wanted to express why I was suggesting to Arto to consider replacements. In fact, my biggest issue with them is the sound file. One of my very first threads I started here 11 years or so ago was that the sound file, at least to my ears, sounds nothing like a diesel locomotive. IIRC, I think I mentioned it sounds more like a marine vessel engine. But at least certainly not an EMD nor an Alco. Motor control seems fine, now that I know the three CV sequence to reset it I can live with that, but the sound file . . . . Sad

  • Member since
    December 2008
  • From: Heart of Georgia
  • 5,406 posts
Posted by Doughless on Monday, February 5, 2024 8:36 AM

I beleive that some of the CVs won't hold their changes and revert back to default setting after either a period of running or after power down and power up cycles.  Same thing happened to me with older Soundtraxx Tsunami's, which are every bit of "junk" as the QSI.  Just a function of the technology of the era.

There is no doubt that the new Loksound V5 and Tsunami2 are better than their predecessors, but dealing with old technology can be rewarding too if you know the limitations.

I just keep the individual turbocharger and prime mover sounds turned down some on the QSIs so that the hisses, spurts, flange squeal, bells, and whistles/horns over power the PM.  Which is probably more prototypical anyway.

- Douglas

  • Member since
    January 2014
  • From: Chicago area
  • 335 posts
Posted by Arto on Monday, February 5, 2024 8:54 AM

Arto

This is sort of related to my QSI post.

I moved on to speed match three BLI E8 Pragon2 locos.

 

Sorry about the confusion. To be clear.......this thread I started is about three BLI PRAGON2 locos, not the BLI QSI equipped loco.

There are up to four locos I use in this consist. One BLI w/QSI, plus three BLI Paragon2. Obviously the QSI unit has way different performance characteristics than Paragon2. The QSi unit's top speed is MUCH faster than the Paragon2 units, so I tried to to reduce it's max speed first.

I decided to put the QSI loco aside for now, and focus on the three Paragon2 units.

Sometimes, when I change CV5 (using JMRI) on any of three BLI Paragon2 locos, the loco's address changes. I suspect that's what happened with the QSI loco described in my other thread/post.

When I put the loco back on the mainline with the changed CV5 value, it doesn't respond. Eventually I discovered the loco wasn't responding because CV1 had somehow also changed. The CV1 change does not show up on JMRI - UNTIL I do a "Read All Sheets". Then I can see what number the address has changed to. And it's always a three digit number. All my locos use the short two digit address. For all four of these locos (1 QSI + 3 Paragon2) the short address I use is 99 (all four units road number start with 99).

Hope this explains it better.

(PS: Chris, I agree with you. I avoid QSI equipped locos at all cost nowadays, and replacing the decoder is probably the best option)

EDIT: I don't know if this makes any difference, but.....FWIW

I have both a DigiTrax PR3 AND a LokProgrammer connected to the PC at the same time, PR3 on USB Com3, LokProgrammer on USB Com4. I manually change the wire connection from PR3 & Lok at the Programming Track. The PC is a brand new HP Windows 11 machine.

  • Member since
    March 2011
  • From: Westford MA
  • 542 posts
Posted by Tophias on Monday, February 5, 2024 10:43 AM

Arto, my bad. It clearly states in your second sentence that they are Paragon equipped. I just didn't originally read it carefully enough. I'll do better! Big Smile

  • Member since
    May 2020
  • 1,057 posts
Posted by wrench567 on Monday, February 5, 2024 10:52 AM

  Arto.

  Make sure you have DC disabled in CV29. 

 As for the QSI decoder. Some had a magnetic read switch that was used for resetting. Sometimes they get accidentally reset during a derailment or going too fast into a sharp curve and I accidentally reset one with a magnetic uncoupling tool.

   Pete.

  • Member since
    January 2014
  • From: Chicago area
  • 335 posts
Posted by Arto on Monday, February 5, 2024 12:09 PM

wrench567

  Arto.

  Make sure you have DC disabled in CV29. 

 As for the QSI decoder. Some had a magnetic read switch that was used for resetting. Sometimes they get accidentally reset during a derailment or going too fast into a sharp curve and I accidentally reset one with a magnetic uncoupling tool.

   Pete.

 

Pete, what value should I use for CV29 to disable DC? It looks like CV29 is related to a lot of other functions.

AND, now guess what (again)?

I did a "Read All Sheets" with JMRI one of the locos that I'm trying to slow down using CV5 (CV5=190). It showed CV2=2 (should be CV2=1) and the loco address (CV1) showed 99. Changed CV2 back to CV2=1. I put the loco back on the mainline - no response. Put it back on the programming track, "Read All Sheets" again, and CV1=127 - address changed again!

EDIT: Yep. Confirmed! CV1 changed back to 127 AGAIN. A repeat of what I posted above

 

  • Member since
    December 2015
  • From: Shenandoah Valley
  • 9,094 posts
Posted by BigDaddy on Monday, February 5, 2024 12:36 PM

Arto
It looks like CV29 is related to a lot of other functions.?

https://www.digitrax.com/support/cv/calculators/

Use the CV29 calculator.  I have no idea why your CV's change themselves

 

Henry

COB Potomac & Northern

Shenandoah Valley

  • Member since
    January 2014
  • From: Chicago area
  • 335 posts
Posted by Arto on Monday, February 5, 2024 12:46 PM

OK, thanks Henry. CV29=2 (which is supposed to disable DC mode operation)

  • Member since
    May 2020
  • 1,057 posts
Posted by wrench567 on Monday, February 5, 2024 1:53 PM

  Arto.

  CV1 should equal 3 if you're using a long address over 0127. Anything under that is considered a short address. 

  CV29 configures a few things. DC disable, Direction, long address enable, and speed step 14/28-128. Use a CV29 calculator to easily get the proper setting. CV29 calculators can be found online if you Google. Once you enable long address then the decoder will ignore CV1 settings. If the decoder is not responding. Look at CV19. It may have a consist address present. Sometimes for some reason will not clear CV19 when breaking up a consist. Some of my Paragon and an NCE will keep CV19 for some reason.

   Pete.

    

  • Member since
    January 2014
  • From: Chicago area
  • 335 posts
Posted by Arto on Monday, February 5, 2024 2:37 PM

Well, I'm NOT using a long address.

And why should CV1=3, for any address?

Using JMRI, on the CV list shown under the CV Tab, CV1=99, the address I use for that loco. On the JMRI "Basic" tab, "Short Address" is selected, Active Address=99, Primary Address=99. Extended Address=128 for some reason. And "Address Format"=Short (one byte) address.

What am I not understanding?

  • Member since
    January 2014
  • From: Chicago area
  • 335 posts
Posted by Arto on Monday, February 5, 2024 3:11 PM

Well, now the other two Paragon2 are not responding on the mainline.

I'm at a complete loss as to how and why this is happening. 

I guess that's probably why I stopped bothering to change CV except for the address years ago. I thought I'd give this a try again but this is insane. All the documentation is inconsistent and in many cases outright wrong or conflicting.

I have support ticket into BLI. But I suspect this is going to be an ongoing/reccuring issue. Right now I feel like getting rid of all this stuff for all the time I've wasted. 

  • Member since
    September 2004
  • From: Dearborn Station
  • 24,281 posts
Posted by richhotrain on Monday, February 5, 2024 3:27 PM

Arto

OK, thanks Henry. CV29=2 (which is supposed to disable DC mode operation) 

CV29=2 translates to 28/128 speed steps, DC operation Off.

CV29=34 should be used as the value to accept long addresses.

Rich

Alton Junction

  • Member since
    September 2004
  • From: Dearborn Station
  • 24,281 posts
Posted by richhotrain on Monday, February 5, 2024 3:30 PM

Arto

AND, now guess what (again)?

I did a "Read All Sheets" with JMRI one of the locos that I'm trying to slow down using CV5 (CV5=190). It showed CV2=2 (should be CV2=1) and the loco address (CV1) showed 99. Changed CV2 back to CV2=1. I put the loco back on the mainline - no response. Put it back on the programming track, "Read All Sheets" again, and CV1=127 - address changed again!

EDIT: Yep. Confirmed! CV1 changed back to 127 AGAIN. A repeat of what I posted above 

Try programming the QSI decoder on the Programming Track and without JMRI.

Rich

Alton Junction

  • Member since
    February 2002
  • From: Mpls/St.Paul
  • 13,892 posts
Posted by wjstix on Monday, February 5, 2024 4:55 PM

I'll second what Rich said. I've had several decoders (not QSI)  where it seems like Decoder Pro changed the ID when I was adjusting other CVs. I have a separate programming track hooked up to my DCC system, and using that to change the ID back would work.

Stix
  • Member since
    January 2014
  • From: Chicago area
  • 335 posts
Posted by Arto on Monday, February 5, 2024 5:34 PM

Guys....forget the QSI decoder for now. I've put that loco aside.

The problems I'm having now are regarding the Broadway Limited E8's, all Paragon2.

Something keeps changing CV values that I have not changed! I've verified this repeatedly. And there really isn't any consistency to what I'm seeing, except that the CV values keep changing without me doing anything. Oh, excuse me. The only thing consistent is the loco Address keeps changing to 127.

Like now for instance. I put loco #9967 on the Programing Track. Read the settings from the JMRI Basic tab. It says Primary and Active Address is 99. Save and Close the JMRI window. Change the PR3 Mode back to Interface. Put loco on mainline. Power On. Select loco 99 on the throttle. Loco powers up, sounds come on. But no throttle response. Prime mover sound is on but other sounds (bell, horn) do not respond. I put the loco back on the Programing Track, JMRI CV tab, Read Full Sheet. Sure enough. Loco address changed to 127. HOWEVER....when I had the loco on the mainline and it wouldn't respond to address 99, I tried address 127, and got no response either.

All of this brings up the question again, of whether or not the 10' long 26awg wire I'm using connecting the PR3 to the Programming Track is adequate

If I have to make ANOTHER (3rd) connection to the Programing Track, directly from the Command Station, bypassing the PR3 and LokProgrammer, this just adds yet another layer of complexity, works for some locos, not for others. I'm really getting tired of this. If things don't work, thats fine. But what is really bugging me is there is absolutely no consistency to what's going on.

  • Member since
    September 2004
  • From: Dearborn Station
  • 24,281 posts
Posted by richhotrain on Monday, February 5, 2024 7:10 PM

Arto
.

Like now for instance. I put loco #9967 on the Programing Track. Read the settings from the JMRI Basic tab. It says Primary and Active Address is 99. Save and Close the JMRI window. Change the PR3 Mode back to Interface. Put loco on mainline. Power On. Select loco 99 on the throttle. Loco powers up, sounds come on. But no throttle response. Prime mover sound is on but other sounds (bell, horn) do not respond. I put the loco back on the Programing Track, JMRI CV tab, Read Full Sheet. Sure enough. Loco address changed to 127. HOWEVER....when I had the loco on the mainline and it wouldn't respond to address 99, I tried address 127, and got no response either.

In an earlier post today, you said that you are not using long addresses, but 9967 is a long address.

What value is in CV29?

CV29=2 turns off long addressing. If you are, in fact, using long addresses, then CV29=34.

Rich

Edit Note:  Are you using 2-digit addresses?  99, not 9967?

Alton Junction

  • Member since
    May 2020
  • 1,057 posts
Posted by wrench567 on Monday, February 5, 2024 7:16 PM

Arto

Well, I'm NOT using a long address.

And why should CV1=3, for any address?

Using JMRI, on the CV list shown under the CV Tab, CV1=99, the address I use for that loco. On the JMRI "Basic" tab, "Short Address" is selected, Active Address=99, Primary Address=99. Extended Address=128 for some reason. And "Address Format"=Short (one byte) address.

What am I not understanding?

 

   Okay. So all your locomotives have addresses less than 127?  Most people run a long address. My bad. Disregard the changing of CV29 to long address.

  Still look at CV19 when you find the locomotives don't respond. And sometimes CV1 will change when you advance consist. Especially if it is the lead or rear unit. 

     Pete.

  • Member since
    February 2007
  • From: Christiana, TN
  • 2,134 posts
Posted by CSX Robert on Monday, February 5, 2024 9:19 PM

Arto
Select loco 99 on the throttle. Loco powers up, sounds come on. But no throttle response. Prime mover sound is on but other sounds (bell, horn) do not respond.

That sounds like it's in a consist, make sure CV19=0 and that they aren't in a command station consist.

Arto
I put the loco back on the Programing Track, JMRI CV tab, Read Full Sheet. Sure enough. Loco address changed to 127.

Is is always 127 that it changes to? I suspect that it's not actually changing, but that the command station is having trouble reading the CVs (not uncommon with sound decoders, especially older designs). The first thing I would do is replace the programming track wire, if the command station is having trouble reading the decoder, the long thin wire is not going to help.

Here's what I would do:

  1. Replace the programming track wire.
  2. Reset the command station (so you know they are not in a command station consist).
  3. Reset the decoders.
  4. Attempt to program and read the address only using the command station (not JMRI)
  • Member since
    February 2002
  • From: Mpls/St.Paul
  • 13,892 posts
Posted by wjstix on Wednesday, February 7, 2024 9:18 AM

Did you buy these Paragon 2 engines new, or used? I'm just wondering if a previous owner might have one set up in a consist. Paragon 2 uses CV19 and CV230 together to set up consists.

Maybe check that CV15 and CV16 are both zero. If they aren't, it could be the ID number is "locked".

Just to kinda "try everything", maybe set CV1 to something other than 99 (like 67) and see if it saves it and responds. Maybe try setting a long address and see if that works. See if it responds to 9967; maybe even try 0099 as a long address. 

On some decoders, you have to change the CVs you want and then power down and then power back up to get the changes to take effect.

I would mention again trying to change CVs using a programming track connected just to your DCC system, rather than JMRI. I've had situations where DecoderPro would allow me to read and change all the settings OK, but then would screw up the ID. I would have to put the engine on a programming track connected to a Digitrax Zephyr to get the address corrected.

Stix
  • Member since
    November 2013
  • 2,774 posts
Posted by snjroy on Wednesday, February 7, 2024 2:44 PM

Yup, I also resort to trial-and-error in these situations. In at least one case, I thought JMRI changed the CV, to find out later that it hadn't. I think it turned out to be a CV29 issue. 

JMRI is supposed to do that math automatically, but in some situations, it just would not work on some of my locos.

As a user of Digitrax decoders (among others), I also use the SoundLoader program. There is a tab to change CVs, which works on many decoders, not just Digitrax. I kinda remember a situation where that program worked, and JMRI did not.  And like others have mentioned, I resorted to my DCC system to reprogram a CV, successfully,

Simon

 

  • Member since
    September 2004
  • From: Dearborn Station
  • 24,281 posts
Posted by richhotrain on Wednesday, February 7, 2024 4:07 PM

Haven't heard from Arto since Monday.  Anxious to hear what's going on.

Rich

Alton Junction

  • Member since
    December 2004
  • From: Pa.
  • 3,361 posts
Posted by DigitalGriffin on Wednesday, February 7, 2024 4:14 PM

All NAND flash memory devices can lose their charge in as short as 10 years.  That means they start forgetting what settings they had in the memory.

Modern day SSD/NVME storage controllers intentionally "top off the charge" of old cells to make sure the contents are not lost.  (When powered on obviously)

Don - Specializing in layout DC->DCC conversions

Modeling C&O transition era and steel industries There's Nothing Like Big Steam!

  • Member since
    March 2010
  • From: Sherwood Park, Alberta, Canada
  • 252 posts
Posted by CNR378 on Tuesday, February 13, 2024 8:32 AM

My experience with Paragon 2 decoders in SW7s isn't great. They are worse than any QSI decoders I've programmed. These SW7s had a max speed of 25 smph and ran well together so other than setting the address I've left every other CV alone. Playing with other CVs always caused other issues (I don't recall what they actually were) requiring a full decoder reset.

I would suggest you try and speed match using the 127 address. When that is done change the addresses back to 99 to see if they maintanined the speed match.

Upgrading your wire from 26 agw to say 18 or 16 would be a good idea but I doubt it causing this issue.

Peter

  • Member since
    January 2014
  • From: Chicago area
  • 335 posts
Posted by Arto on Wednesday, February 21, 2024 12:48 PM

TEST

  • Member since
    January 2014
  • From: Chicago area
  • 335 posts
Posted by Arto on Wednesday, February 21, 2024 12:51 PM

Why can't I copy and paste to reply to my own post? When I do, I get a "Forbidden 403" error

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!