Mike, that was a potentially critical step to swap the decoders. The one important thing that it revealed is that both decoders are apparently OK since the 540 decoder worked in the 541 loco. The fact that the 541 decoder acted up in the 540 loco clearly points to the 540 loco itself as the source of the problem.
There is one more test that you ought to perform. Take both locos off the main layout and reprogram each on the programming track. Leave the decoders in place so that 540 has 541's decoder and 541 has 540's decoder.
Start out with each loco on the programming track by resetting its decoder to factory defaults. Then, reprogram each loco and add back the long address on each loco. Do not program CV2, CV5, or CV6. Check to be certain that the value of CV19 is zero. Check and record the value of CV29. Then return both locos to the main layout.
At this point, neither loco, 540 and 541, is part of a consist. Now run the two locos around the layout, one behind the other, and let us know what happens. If either loco malfuctions, let us know before returning the locos to the programming track.
I just want to be sure that consisting is not part of the problem.
Rich
Alton Junction
OK, decided to eliminate one variable, so swapped decoders between 540 and 541. The result? 540's decoder is fine in 541's chassis. 541's decoder in 540 indicates it's likely not the decoder, as it immediately started acting up. In this case, I tried to use POM to adjust speeds between the two locos running freely apart by adjusting the VMAX, CV5 on 541 as it ran to get them closer together. Then I stopped the consist -- although it wasn't coupled and tried to reverse the direction. 541 (with 540 decoder) ran fine. 540 (with 541 decoder) stopped, but did not start in the opposite direction.
A check of CV19 in 540's chassis showed it had gone from 248 to 255 and CV29 went from 34 to 0. This accounts for the failure to restart.
I usually only use POM for sound adjustments, so not sure if that has any significance to the change in CVs. In previous changing motor/control CVs, I typically haven't used POM, so doubt that is the cause here.
So, tried it again and after stop and attempted reversal, 540 (with 541) decoder showed CV19=9 and CV29=0.
So reset those values again, to 248 and 34 respectively. Then sent one chasing the other again. This time I simply stopped, then started in the same direction. Sometimes 540 chassis would follow, other times it would hesitate, reverse briefly then slow as I lowered the throttle, then it again started moving in the correct direction. I could actually get it going back and forth just with twirling the throttle setting -- completely without using the F/R keys. It seemed to keep the CVs, or at least I could still mostly control its direction, untill I finally tried reversing the consist. Then 540 got really stupid, so I pulled it onto the programming track. It showed this as the final settings.
CV19=223
CV29= 14
I am now considering the factory lightboard the decoder plugs into as the potential source of the problem. I don't see how the motor could be the problem, but enlighten me if you think so. The lightboard has various diodes/circuits that could be a source. I'm not wedded to the idea, just that the decoder board swap did not solve 540's problems, just complicated them.
At some point in this low drama, I did pull the drive wheels and check them with the VOM for continuity and everything was good there, so not the wheelsets or their insulation
Anyway, time to give it a rest for now. What does the DCC brain trust think about this eliminating the decoder itself as the problem?
Mike Lehman
Urbana, IL
CR 2032 installed successfully. I tend to read the instructions and definitely did not want to reenter a bunch of macros, etc, etc in addition to consists. At least I got that right. I took the VOM and measured the old battery at 3.21 volts, which I think means it's likely still good and thus may not have been an issue.
maxmanThis sounds like the motor is slowing down as it warms up, leading me to jump to a possibly erroneous conclusion that the motor may be bad.
Yeah, I've been thinking about the motor myself. I did check it and it's a Buhler.
maxmanThere is something going on with the consisting that I don't understand. Or else you are reporting the data incorrectly. Or else the value being read back by the PowerCab is incorrect. And I'm wondering exactly how many consists you really have stored in the main command station. You originally said that the consist number you read from the 540 unit decoder was 61. Then after playing with it some more you said that the consist number read back as 50.
Those consist numbers were changed by whatever mysterious force is at work -- demons, I think, although I am a skeptic about spiritual visitations -- as I never have had that many consists in the system. Typically, the lowest assigned is around 110. Now, the PowerCab may be reading them back wrong or something. But I'm only reporting what I'm seeing.
maxmanI would also like to point out if it hasn't been mentioned already elsewhere that at one point you reported that the consist address you read on the 541 unit, which I believe is the unit that works correctly, was CV 19 = 248. Well, this can't be since the consist addresses can only go up to 127.
Yeah, that is interesting . Hadn't thought about that, but it's what it read out on the PowerCab. I would think that it and the PowerPro would speak the same language so that there's no possibility that could happen, but all I know is that's what it said. I'm certain I didn't manually invoke it, because I just let system decide the consist address for me.
Just to check it, I polled it again and sure enough it's still 248, despite me making and breaking that consist with 540 several times. I manually entered CV19=120 to match what is in 540 and found it wanted to run together, but in opposite directions? So not sure what that means, other than if I then re-enter CV19=248, they run together OK (except for 540's issues, which continue.)
I checked all consists in the system and neither 540 or 541 is entered in any other consists.
mlehmanWill catch up with things a little later in more detail. Going to install the new CR2032 right now.
I hope that you remember to do this with the power on. If you don't, you will lose all your existing consists. Unless the battery is already bad in which case you will have already lost all your consists.
mlehmanBy 541 catching up with 540, what happens is 540 gradually slows as it runs. They are speed-matched very well before that happens, but now after that first lap when running apart, 540 is already slowing enough to make the match irrelevant.
This sounds like the motor is slowing down as it warms up, leading me to jump to a possibly erroneous conclusion that the motor may be bad.
mlehmanI will check the consists (yes, I mean using the Browse Consists function for the consist stack) on the layout to be certain that 540 isn't in some consist it's not supposed to be in, but very unlikely as I never run the FTs except with each other.
There is something going on with the consisting that I don't understand. Or else you are reporting the data incorrectly. Or else the value being read back by the PowerCab is incorrect. And I'm wondering exactly how many consists you really have stored in the main command station. You originally said that the consist number you read from the 540 unit decoder was 61. Then after playing with it some more you said that the consist number read back as 50. The consist number is assigned by the main command station which starts looking at the available consists starting at 127 and picks the first available next lower address. If you were reconsisting the units after every attempt to do a decoder reset, then the command station thought that 61 was the first available opening and then the next was 50. Assuming that you have been diligent in deleting all the unused consists as you said and don't actually have that many consists (that you are aware of), the command station should have been assigning consist addresses up in the high numbers. So I have to ask...when you are consisting the locos you should notice that the main command station assigns a consist address and asks you if you want to use it. That gives you the oppertunity to use something different if you want. Are you looking to see what consist address the command station is assigning before you start entering the first and last loco information, or do you just blow past this point without looking? In other words, did you happen to see if the command station assigned consist addresses of 61 and 50, or did you just not look at that and then place the locos on the PowerCab programming track and then attempt to read what value was in CV 19?
I would also like to point out if it hasn't been mentioned already elsewhere that at one point you reported that the consist address you read on the 541 unit, which I believe is the unit that works correctly, was CV 19 = 248. Well, this can't be since the consist addresses can only go up to 127. It may be possible to manually enter a consist address higher than 127, but I don't know if the command station will accept that value and then just choose to ignore it. In any case, you either read the number incorrectly or the PowerCab reported it incorrectly.
richhotrainWhen Mike said that his programming track was "separate" from his main layout, I assumed that the programming track was wired back to the main layout. It is beginning to sound like the programming track is just not wired at all into the PH-Pro booster, so that the Power Cab is used to power the programming track.
Rich,
That's it exactly. No tie-in between the main layout powered by the PowerPro and the programming track powered by the PowerCab.
To power the programming track from the PowerPro would require the leads to be around 40' long, kinda impractical.
I will check the consists (yes, I mean using the Browse Consists function for the consist stack) on the layout to be certain that 540 isn't in some consist it's not supposed to be in, but very unlikely as I never run the FTs except with each other.
By 541 catching up with 540, what happens is 540 gradually slows as it runs. They are speed-matched very well before that happens, but now after that first lap when running apart, 540 is already slowing enough to make the match irrelevant.
Will catch up with things a little later in more detail. Going to install the new CR2032 right now.
ah, maxman, I think I see what you are driving at.
When Mike said that his programming track was "separate" from his main layout, I assumed that the programming track was wired back to the main layout. It is beginning to sound like the programming track is just not wired at all into the PH-Pro booster, so that the Power Cab is used to power the programming track.
richhotrainWith all of the issues presenting themselves on Mike's layout, my suggestion would be to use a Pro Cab on the programming track even if it is separate from the main layout. For all we know, the Power Cab may be the culprit.
If he is using the PowerCab and its associated transformer for his programming track, then he cannot use the ProCab. That's why I'm trying to understand how he has the programming track set up.
I agree that the consisting should be done on the main. However, the PowerCab can be used to operate consisted units on an isolated piece of track so long as he understands that he will have to use the alias address to make that happen. Just because he calls the PowerCab setup the "programming" track doesn't mean that that track can only be used for programming.
But again, I'm still trying to understand how his setup really is.
The proper way to set up a consist is to do so on the main layout, not on the programming track. All locos to be consisted should be on the layout and on the same section of the layout,not on a non-powered section of the layout. This is necessary to receive the consisting commands.
With all of the issues presenting themselves on Mike's layout, my suggestion would be to use a Pro Cab on the programming track even if it is separate from the main layout. For all we know, the Power Cab may be the culprit.
While it is certainly possible to get hung up on CV19, it surely seems to be associated with the overall problem, as it controls consisting. Which raises an interesting question, and one that I cannot recall if it was covered previously. What if the consist were killed, and CV19 set to zero on each of the two locomotives. Would 540 run correctly in that case?
P.S. Mike, have you checked Browse Consists to see if 540 is included in another consist?
Oh, one other thought. You said that when you start the two units the 541 catches up with the 540 and starts pushing. Is this just a speed matching issue, or does the 540 visibly start to slow down in a drastic manner?
The reason I ask this is because I did a lot of speed matching for a friend and he had a couple units where the motor speed seemed to change on its own. Eventually this was traced back to a motor problem. The motors in question where Kato. Did not some of the Stewart units come with Kato motors? Just another thing to check.
richhotrainThe Power Cab and the Pro Cab are not the same, but a Power Cab can function as a Pro Cab on a PH-Pro layout. I just raised the issue because I wondered why you just don't use your Pro Cab on the Programming Track.
I can't be completely certain, but from what I'm reading I think he has a separate programming track on which he uses the PowerCab, including the PowerCab power supply.
This business of the CV 19 value changing is confusing. The NCE 5 amp command station assigns the consist number unless the operator manually selects some other number. So if the command station assigns a set of units to consist 100, for example, I don't see how that might change to 50 by itself.
Now if a set of units has been consisted on the main and then taken over to a separate PowerCab programming track, I'm not certain what the PowerCab will read as the CV 19 value. I would like to think that it would read the same value on CV 19 as was entered by the main command station. But if the PowerCab somehow thinks that it should be a different number because the units were consisted one time on the programming track, maybe it will read out differently.
To the OP, the Hammerheads will not "show" 127 consists, unless you are using the browse consists function. Maybe that's what you meant.
There is a way that you can run the consisted units on the programming track after consisting them on the main. Consist them on the main and manually give them a consist address that you can remember and that you know is available, say 99. This will be the alias address that gets entered into CV 19. Take the units to the programming track, set the PowerCab to not be in programming mode, select loco 99, and see if the two units run together. Note that they will not run using the cab numbers because they will want to run on the consist address, and the PowerCab will not be able to find that address because you have said that you have cleared out all consists from the PowerCab memory. But if you just select loco 99, the PowerCab will find that.
I guess what I'm trying to see is if we're getting hung up on this CV 19 changing thing when what there really is is a conflict between the main command station and the PowerCab.
The other CVs changing might be another story.
Yeah, I'm thinking that the battery will solve the QSI issue, maybe. I happen to have one on hand, so will replace later today (much later most likely).
cacole,
Will check the axles. All seem to be in gauge, as original, and no issue there I can think of that baling wire won't fix Will put the meter to things to ensure that's so.
Have tried CV30=2 multiple times with nothing but brief improvement. After reset and reconsisting, a couple of laps of the main, then the loco slows and 541 catches up and starts pushing 540.
FTA to FTB connection is a plastic drawbar. Design of plastic coupler boxes and mounto to metal frame seems to preclude an issue with metal KDs. And the issue happens when not coupled to 541, so doubt that's it.
A 8-pin plug is attached via solder points, plus additional soldered leads to output 3 to drive the Mars light, so not a plug and play swap or I would've tradding decoders between 540 and 541. May still do that, as we're exhausting the other possibilities.
The mystery continues, but will check for poetntial wheel continuity later.
Remove each axle and double check your wheel gauge on the problem locomotive, and check it with a VOM to insure that the axles aren't touching within the axle gear.
If you've changed any of the wheels, it's possible that the replacement axle length is enough to let them touch inside the gear if the Stewart trucks are clones of Athearn's.
Even if you've never changed the wheels, it's possible that there's a manufacturing defect in one axle allowing an intermittent short.
Try a factory reset of the problem decoder (CV30=2 on older ones).
Another, but bizarre possibility is, do all the locomotives in the consist have Kadee metal couplers that are touching the frames through mounting screws? This could be causing intermittent shorts as the coupler faces all touch at the same time.
If the problem decoder is installed with a JST harness, change to a different decoder and see what happens.
The battery in the command station is always a possibility. So, I suppose that it wouldn't hurt to replace it. It is a coin-shaped battery, CR2032.
But, if the battery were the problem, I would think that other decoders would be adversely affected as well. The other thing to consider is that the battery permits the system to remember data when power is off. But, loco 540 seems to be scrambling data while the system remains powered on.
Mike, you mentioned that some QSI sound decoders lose their long addresses. I, too, have had problems with QSI sound decoders on my PH-Pro system, but those problems are mostly restricted to consisting where I have to manually add and deleted the consist value in CV19. I blame that on the QSI decoders, and I have discussed that issue with Larry at NCE.
Maybe others will chime in today with some more thoughts on your 540 problem.
OK, had to do a test run. Same results. Even with baling wire FTB off the tracks, the same issue arises. The 540 FTA starts slowing and you find that CV19 has changed and is no longer the same as 541. CV29 also .changed from 34 to 32 on 540. So it remains a genuine mystery.
For maxman,
I don't do any consisting on the programming track with the PowerCab. I checked what was in it's consist stack and there was only only I must;ve inadvertently put in by mistake. It wasn't 540/541 though. I deleted it.
Have a great Christmas!
Guys,
It is Christmas and I have to get up in the morning to attend to things, probably offline a lot of the day although will check first thing in the morning. I suspect maxman may be onto something with the command station battery, although I'd think it would affect all consisting, not a select few. And why is 541 oblivious to all this drama? very strange, but it's the wonderful world of DCC.
I really do appreciate the help. I think we'll get it fixed eventually. Maybe a good nite's sleep and some holiday cheer will bring something else to mind. I'll check in once the fstivities have subsided and see If I can add anything.
Seriously, I really am not making this up...
Those CVs are really changing on their own. I rechecked the decoder and it's safely isolated with no stray strands anywhere I can see...But I can't see whhy 541 is mOK and 540 is not, until got to thinking I'd better double-check something...
It's the back truck from the unpowered B unit, so not sure if this is it or not. I suspect it's sparking very, very little, but enough that maybe somehow the decoder in the A unit is picking it up?
Yep, baling wire. I think this may just be it, now that I've had a relaxing shower and a peek at this pic I remembered. I thought I'd ensured the two sides remained isolated, but I suspect there is just enough intermittent contact to scramble CVs, but no other indications of shorting. I'll see about this and advise ASAP, but likely will be some hours.
Thanks very much for all the help on Christmas Eve, Rich and maxman. I suspect that command station battery is at the root of the issues with my QSI sound units losing their long addresess, so this has really not be wasted effort. Happy holidays and will update when I know more about the FT.
EDIT: Couldn't help my curiousity, had to look at that truck. I'm leaning against it being the issue. The baling wire doesn't contact metal on either side of the truck. So there is no direct short, which would explain there being no obvious indication from fault protection. However, the wire does form two loops in close proximity to the rails, so maybe some inductive effects could be present?
The Power Cab and the Pro Cab are not the same, but a Power Cab can function as a Pro Cab on a PH-Pro layout. I just raised the issue because I wondered why you just don't use your Pro Cab on the Programming Track.
I suppose that it just could be a D13SR decoder gone haywire, but my sense is that something is amiss elsewhere.
richhotrainI have the PH-Pro wireless system, but not a Power Pro. I just wonder if there are storage and/or memory conflicts when using both systems together like you are doing. Do you Browse Consists on both the Pro Cab and Power Cab?
I think PH-Pro and PowerPro are the same? maybe not.
The PowerrCab and the main layout consists (whatever it is) are completely separate. The PowerCab is standalone and I haven't graduated to JMRI yet. So I can't run consists off the layout on the programming track unless I zero it out of the unist involved first. That helps me keep thinsg cleaned up...I thought.
maxmanDo you have a separate programming track with the Powercab?
yes.
maxmanAre you doing your consisting with the ProCab on the main? Or are you doing your consisting with the PowerCab and then moving the locos to the layout and expecting them to run consisted there?
All consisting is on the main. I know when units are in cosnist they won't run on the progTrack/PowerrCab, but with 540, everything is scrambled and the consists has to be zeroed out to make it run on the PowerCab, which happens anyway under the circumstances.
maxmanYou mention that you were clearing out the consists on the hammerhead. All this does is clear out the controller loco recall slots. I think there are only six slots. It does nothing to clear out multiple consists from the command station, where you can have up to 127.
No, six slots is the recall stack on NCE. The hammerheads will show up to 128 consists. I usually only have about two dozen entered though. It starts assigning at 128 and works it's way down. I never have consists lower than 100 because I am conscientious about deleting unused/superseded ones. That why I asked about them potentially still being in the command station.
For Rich, no consist 50, that's just where it decided to go on its own...
I have the PH-Pro wireless system, but not a Power Pro.
I just wonder if there are storage and/or memory conflicts when using both systems together like you are doing.
Do you Browse Consists on both the Pro Cab and Power Cab?
Do you get identical readings for consists on both cabs?
maxmanSo if the unit is in a consist it will want to run on the consist address, contained in CV 19. If the consist addresses are lost because of a dead command station battery you can enter the cab number all day long and the unit won't run because the command station looks for that number amongst the consisted units which no longer exist.
It's been awhile since I changed the caommand station battery, but I have changed it once sicne purchasing the syustem 5 years ago.
The QSI units started having problems about a year ago. The FTs recently. I have other consist with neither in them and they're all OK, no problems.
richhotrainWhen you use the Power Cab on the Programming Track, do you then run your locos on the main using the Pro Cab or the Power Cab?
Power for the main bus is in the corner of the layout room on opposite side from the wall that separates staging from the main layout. Programming track is with staging. There's just no room for a programming track on the layout and would crowd the aisle to use it even if there was.
I use the PowerCab just like any other controller on the main part of the layout, other than the other throttles are all radio cabs and the PowerCab isn't.
maxman And since the posts are now starting to get crossed, please go back and read my post above concerning the command station battery.
And since the posts are now starting to get crossed, please go back and read my post above concerning the command station battery.
Mike, is that command station battery fresh?
What is your latest EPROM update?
You should use your Pro Cab on the Programmng Track.
Browse Consists to see if loco 540 is being used in another consist.
mlehman Zeroede out CVs 19, confirmed CV29 was 34, deleted old consist from stack and confirmed no others present. Put both back on the track. 541 is fine as frog's hair. After I found it wouldn't respond on the main, back to the programming track and found 540 reverted to CV29=1 and CV19=50...
Zeroede out CVs 19, confirmed CV29 was 34, deleted old consist from stack and confirmed no others present. Put both back on the track. 541 is fine as frog's hair. After I found it wouldn't respond on the main, back to the programming track and found 540 reverted to CV29=1 and CV19=50...
Do you have a consist #50 elsewhere on the layout?
When you Browse Consists, is loco 540 included in another consist?
mlehmanI have a PowerCab on the proggramming track. The layout is a PowerPro 5 amp with a 5 amp booster.
Okay, now I'm more confused. Do you have a separate programming track with the Powercab?
Are you doing your consisting with the ProCab on the main? Or are you doing your consisting with the PowerCab and then moving the locos to the layout and expecting them to run consisted there?
You mention that you were clearing out the consists on the hammerhead. All this does is clear out the controller loco recall slots. I think there are only six slots. It does nothing to clear out multiple consists from the command station, where you can have up to 127.
mlehman richhotrain Edit Note: Picking up on maxman's comments, while 540 is on the programming track, set the value of CV19 to zero to clear any prior consist. Do the same for loco 541. After the last fail, checking CV19 shows it = 61 on 540 and = 248 on 541. That probably explains why they weren't playing well together at that point. I suspect whatever is changing CV29 is also affecting CV19?
richhotrain Edit Note: Picking up on maxman's comments, while 540 is on the programming track, set the value of CV19 to zero to clear any prior consist. Do the same for loco 541.
After the last fail, checking CV19 shows it = 61 on 540 and = 248 on 541. That probably explains why they weren't playing well together at that point. I suspect whatever is changing CV29 is also affecting CV19?
mlehman I have a PowerCab on the proggramming track. The layout is a PowerPro 5 amp with a 5 amp booster.
I have a PowerCab on the proggramming track. The layout is a PowerPro 5 amp with a 5 amp booster.
When you use the Power Cab on the Programming Track, do you then run your locos on the main using the Pro Cab or the Power Cab?
richhotrainEdit Note: Picking up on maxman's comments, while 540 is on the programming track, set the value of CV19 to zero to clear any prior consist. Do the same for loco 541.
mlehman I have a PowerCab on the proggramming track. The layout is a PowerPro 5 amp with a 5 amp booster. I move things back to the pogramming track when 540 starts stumbling, bceause I can't read back using Programming on the Main. I also am more comfortable doing the resets there, as I can check my work. Otherwise I'd never have known about the weird values showing up in various CVs.
I move things back to the pogramming track when 540 starts stumbling, bceause I can't read back using Programming on the Main. I also am more comfortable doing the resets there, as I can check my work. Otherwise I'd never have known about the weird values showing up in various CVs.
Going back to Mike's initial post on this thread, it does seem clear that the CV values are changing without user reprogramming. Very weird.
As far as creating and killing consists goes, I have had recent problems in this regard and I started a thread about it a few days ago.
If it were me, I would check the values in CV19 (consist) for both locos to see what is stored in each decoder even if a consist does not show up on Browse Consists.