I've been doing a long-term DCC conversion for a lifetime collection of DC locos. As a result, I like to take advantage of the bulk packages from a well-known DCC vendor. However, over time, I'm beginning to notice as much as a 25% defect rate: I install a decoder, and maybe one time in 4, when I apply first-time power, I can't turn the headlights off (a typical problem). Resetting the decoder to default doesn't help.
Is this typical? Frankly, I hate to go to the time and effort of a hardwire install only to have to toss another decoder into a drawer until I get fed up and send a package back.
What are other people finding for a defect rate?
My average is quite lower, John
I just took inventory of decoders that I installed and the total defects is 1%. One out of a hundred. And that was an old Digitrax DH121 that a friend of mine had used and removed from one of his engines so I don't know what may have happened to it before I got it.
My manufacturer spread is about 60% Digitrax, 15% TCS, 10% Soundtraxx, 10% ESU and 5% NCE.
I would take a hard look at your installation methods OR at least test your installs on a programming track (a Sprogg is handy here) before applying track power.
Out of sixty factory installed decoders, most sound, I don't recall any defective ones there, either. Maybe I've just been lucky?
Ed
I've only installed about thirty decoders but have found no defects. I recommend you use the decoder tester sold by Ulrich. Hook up your decoder to the tester before installation. If it works on the tester but not after your install...(need I say more?).
I don't have problems with Digitrax. However, the one I use mostly is the 5% you give.
Let me ask this: I'm not sure how you can go wrong on, say, a DA-SR drop in on an Atlas Classic loco. Remove Atlas light board. Disconect truck power leads from board. Solder same to tabs on DA-SR board. Disconnect motor wires from Atlas board. Solder same to motor tabs on NCE board. (When needed, cut resistor loops on DA-SR for Atlas LEDs). Disconnect LED or bulb leads from Atlas board. Solder to bulb lead tabs on DA-SR.
One in four typically has a problem. Is there something obvious I'm not doing right here?
It is possible the manufacturer had a bad batch. Have most of your defective decoders been of the same type and brand? If so, did you purchase them all at the same time? If so I would contact the manufacturer to see if other customers have had the same problem.
With regard to the DA-SR are you sure you cut the correct traces? When you solder wires to the board what size tip are you using. Anything other than a pencil point or narrow chisel tip may overheat the contacts. Ditto if you are using a high wattage iron. With tinned wire, clean surfaces and a small dollop of rosin flux you shouldn't need more than a second or so of contact with the tip.
JOHN BRUCE III Let me ask this: I'm not sure how you can go wrong on, say, a DA-SR drop in on an Atlas Classic loco. Remove Atlas light board. Disconect truck power leads from board. Solder same to tabs on DA-SR board. Disconnect motor wires from Atlas board. Solder same to motor tabs on NCE board. (When needed, cut resistor loops on DA-SR for Atlas LEDs). Disconnect LED or bulb leads from Atlas board. Solder to bulb lead tabs on DA-SR. One in four typically has a problem. Is there something obvious I'm not doing right here?
Rich
Alton Junction
Your failure rate is way, way too high. You're obviously doing something wrong, but without knowing exactly how you're installing the decoders its hard to say what.
I've installed close to 100 decoders, both sound and non-sound, and have had only one failure, an MRC decoder (4 out of 5 were bad right out of the package, so I'll never install another one no matter what.)
My choices are NCE D13SRJ, TCS T1, TCS MC2, and LokSound Select if at all possible.
The DA-SR point from Rich may go some way to answering the question just previous.
I've used 10 or so DA-SR and have no failures. Your soldering iron could be the source of a lot of failures, especially if it's one of the really cheap Radio Shack pencil irons -- sometimes they are not properly grounded and introduce voltage into the decoder that can destroy a microchip. I use a Xytec controlled-temperature soldering staiton with a very fine tip.
Ditto
JOHN BRUCE III I don't have problems with Digitrax. However, the one I use mostly is the 5% you give.
My NCE breakdown (no pun intended) is 6 DA-SR; 2 N14IP; 2 SW9SR and 1 D13SRJ1... None have given me any troubles.
My DA-SRs are all installed in Athearn Genesis F-3s and F-7s but that shouldn't matter.
Good Luck, Ed
One of the problems with the DA-SR is that the function output tabs are too close together. So, one small stray piece of stranded wire from an adjacent tab can short the function output or the motor tab or the power pick up tab.
One time, I complained about DA-SR on the NCE-DCC forum and got shouted down, LOL, but later it was conceded that NCE had a bad production batch. There were some unexplained failures, especially earlier on, with that decoder.
Something I have found to be a huge contributor to decoder failure is excessively high track voltages. DCC track voltage should be around 14.5 volts. I've seen some as high as 18, and even approaching 20 volts ! While the decoders ARE designed to handle these higher voltages, it's no different than 12 volt bulbs .... a 12 volt bulb will last a lot longer if you don't run it at or near its maximum rated voltage. Same goes for any electronics.
There are always voltage spikes present in a DCC buss. The higher the main voltage is, the higher those spikes are going to be, and that is what wears on and can ultimately damage a decoder.
It's surprising how many people have no idea what their track voltages are.
Mark.
¡ uʍop ǝpısdn sı ǝɹnʇɐuƃıs ʎɯ 'dlǝɥ
I've installed several dozen decoders in my locos and others. I have yet to have a decoder that was bad out of the box, and none of the ones I've installed has ever failed. Among these were a 10 pack of NCE D13SJRs, a couple of early TCS T1 and A4's, a whole bunch of newer TCS T1s (the old ones, pre BEMF, the new ones, with the auto adjust BEMF), a couple of Digitrax board replacement decoders, an aftermarket QSI Iinstalled for someone, a Loksound Select, an old Soundtraxx DSD, and a couple of others
One thing that is useful to have - a system with a low current dedicated program track (generally, any of those that offer readback on the program track). The current here is too low to fry the decoder. So you test on the program track, and if you can read and write the decoder, you can be reasonably sure that at least the track to decoder and decoder to motor wiring is correct and there are no shorts. Only then should you try the loco on the full power main track.
--Randy
Modeling the Reading Railroad in the 1950's
Visit my web site at www.readingeastpenn.com for construction updates, DCC Info, and more.
My guess is that the leading cause of decoder failures is the user's fault, lack of sufficient caution when installing, inexperience, etc. In the case of my DA-SR decoders, that was the likely problem on some of them. Soldering iron tips that were too big, too hot, stray wires, lack of an Optivisor to spot problems.
The next leading cause, IMHO, is the decoder design. I blame some of the DA-SR decoder faults on tabs being too close together, no room for error.
Another cause is sensitivity to voltage. It doesn't seem to take much to fry a function output, usually the lighting.
I don;t know about it being a design flaw - the TCS A-series looks to have exactly the same spacing. Since these are board replacement decoders made to fit a specific form factor,t he maximum width of the entire board is contrained by this, and the spacing of the tabs is such that they are wide enough to re-use those plastic caps for a true plug and play installation (but we all know those plastic caps are hopeless, and soldering it in is MUCH better). Actually, I think it's easier to mess up using the caps, as a stray wire strand can easily be pushed into contact with a neighboring tab when the cap is pushed into place.
richhotrain One time, I complained about DA-SR on the NCE-DCC forum and got shouted down, LOL, but later it was conceded that NCE had a bad production batch. There were some unexplained failures, especially earlier on, with that decoder. Rich
Seems like just another day on the World Wide Web.
I'm using an NCE Power Cab, which doesn't use a separate dedicated program track. The options for "use program track" on prog/esc are 1 - std 2 - cv 3- reg, which I now discover means "register". The problem is that any use of options 1 and 2 puts you back onto a regular dcc controlled track at full voltage when you're done. I don't know enough about "register" to know if this will let you set up a program track with reduced voltage for testing.
Also, thinking this over, one problem (if we're looking at the bad batch possibility, since almost all my problems have been with DA-SRs, although a bad Athearn light/decoder board in an RTR loco has also led to some disgruntlement) could be that a decoder will in fact work briefly when tested either with a tester or a decoder track, but if the board is part of a bad batch, any application of normal voltage will cause problems.
I appreciate the advice. I've returned a bunch of DA-SRs to NCE and will see if they have any input -- but I haven't had the same issues with Digitrax, and I like their design better.
Part of the learning curve of DCC, I suppose.
The program track power on the PowerCab should be limited, doesn;t matter which mode (except programming on main - that's write-only programming with full track power active). A very useful accessory for PowerCab users (and those with the Digitrax Empire Builder/DB150) is the auto-switch that NCE sells. Basically, the track connections from the powerCab panel connect to the input of this device, and it has 2 sets of outputs. One goes to your normal rest of the layout, the other goes to an isolated section of track that becomes your program track. Whenever you go into one of the program mode, it turns off power to the main track and activates the program track - so you don't accidently program all locos sitting on the main.
The issue of the lights being on and not being able to turn them off is a pretty sure sign of a blown function output.
Try the TCS A-series - they have BEMF like the Digitrax ones but it's a better BEMF, works smoother and no fooling around with CVs to configure it. Also have a goof-proof warranty so even if you make a mistake wiring it, you're covered for a replacement.
Then here's a related question with the PowerCab programming track. Normally I test a loco on DC before I do a decoder install, just to be sure the basic parts are working. Then I do the decoder install. The only thing a PowerCab programming track will do is program all the decoder features or go straight to a CV. There's no way to put it on the track at reduced voltage to see if the decoder will work on address 3. (Or is there?)
So, unless there's a feature I don't understand, the only way I can check to see if the loco will move and the lights will work under DCC is to put it on the track on full DCC power, with a PowerCab. Basically the Auto Switch simply protects you from mistakenly selecting use program track on your Power Cab without disabling the other blocks on your layout that may have locos on them -- as far as I can see, it does not in itself provide a way to test a new install at reduced voltage.
Do more conventional program tracks allow you to operate a loco on DCC at reduced voltage? And even then, what's the advantage? My new decoder works fine at reduced voltage, but when I put it on the main, it blows its function output anyhow!
Or is there a way around this?
JOHN BRUCE IIISo, unless there's a feature I don't understand, the only way I can check to see if the loco will move and the lights will work under DCC is to put it on the track on full DCC power, with a PowerCab.
No, that's just wrong. The PowerCab will definitely test at reduced power and stay there UNTIL you step out of the Program Track mode. Yes, if there was a short detected, you do need to remove the loco/decoder first before stepping back to full power. The menu is a little funny like that. I'd prefer that it did not go back to full power directly after short detection, but it does give you the message that it's going to do so. Before you go to that next screen, take the loco off the track if you get the short detected warning. Otherwise, yeah, a short onboard will then fry. You wouldn't take a loco that had a message like that on the programming track and take it over to full power on the layout, expecting the short to magiclly disappear along the way. Well, that's essentially what you're doing when you step out of Programming Mode after you get the short detected message.
Instead, remove the loco from the programming track, while stepping the PowerCab back to full power, then call up Progrmaming Track mode again. Take the loco to the workbench and tend to the short in order to correct it. Then when you go back to the Programming Track, it's already in low power mode and ready to go back to porgramming in case you forget. When you do installs in brass, you do this a lot until things get right, but you must do it or suffer the consequences.
As for decoder experience with bad outta the package, I've had none in something like 75 installs. Most are NCE, with a few Tsunamis. Lots of DA-SR. I'll agree you need to check your work carefully with these, they aren't the only ones with this close spacing. It's tight in there, folks, and it really is necessary in certain installs.
Where I've run into problems with a couple of DA-SR is in Genesis F units. The side terminals sit very close to the chassis casting/weight, so are subject to shorting until insulated and/or prevented from touching the metal that's close by. I will say that every decoder that has turned out to have an issue I've been able to trace back to my own error, one way or the other.
That's not to say you couldn't have gotten a bad batch. That happens with any manufactured product. NCE is good about followup, so call them and sort things out. It amy even be the case here that it may be partly you, partly the decoders. However, I will say that a failure rate that high is very unusual and you need to get to the bottom of the problem.
Mike Lehman
Urbana, IL
Hi,
I've installed about 20 decoders, and have another 20 or so that were factory installed. Never has one failed. All the ones I installed were NCE, but I have no reason to say they are better than others.
The thing is, these are delicate, unforgiving electronic components. A crossed wire or inadvertant connection and "poof", it is gone. So if you have a high incidence of failure, I would review the installation process first.
ENJOY !
Mobilman44
Living in southeast Texas, formerly modeling the "postwar" Santa Fe and Illinois Central
I am a big fan of NCE. My layout is powered by the 5 amp PH-Pro wireless system, two NCE boosters, a Mini Panel, a Switch-It and a Switch-8.
Most of my non-sound decoders are NCE. Out of the box, I cannot recall a single failure, at least not initially. But, a few did experience subsequent failure.
Most of the NCE decoders, including the D13, D14, and D15 series, and the P2K-SR, are 8-pin and 9-pin decoders. As long as the board is isolated from the metal frame, success is virtually assured. And, those four decoder series make up the bulk of the NCE HO scale decoders.
That leaves the DA-SR as one of the few, if not the only, NCE decoder that needs to be hard wired, requiring that the wires need to be soldered to the tabs, excpet for those foolish enough to secure the wires with the plastic holder clips. In the absence of extreme care, the DA-SR can easily be shorted out. So, I have fried a few, early on, due to my inexperience and unintended carelessness.
But, I have had a couple of DA-SR decoders fail once in service on the layout. I blame those failures on poor design.
I've only had one failure out of the box. It was a Soundtraxx. I carefully documented the failure, contacted them and they replaced it. They wanted it back, though, and they tested it to see what had gone wrong.
I, too, heartily recommend testing your installation on the program track before letting it roll on the main.
It takes an iron man to play with a toy iron horse.
Yeah, a Tsunami was my only decoder I thought failed from the box. I sent it back. It was my very first brass install and went fine up to the end, where it suddenly quit because I'd just redone so much along the way to get that Short Detected message to disappear. Thought I'd let the smoke out in a moment of carelessness. They found that I had pulled a wire lose with all my tugging around (it was hidden under the shrink), the decoder itself was fine. They resoldered and sent it back N/C.
Funny experience: I just got a new Digitrax DH126D that I ordered with a new loco from a well-known on line supplier. I installed it -- not much way to screw it up, the 9-pin plug can only go in one way, the socket is a factory item. On the first test run, the headlight blinked on and off intermittently, couldn't turn it on full on F0, couldn't turn it off full. I tried to program it, it was a presumably new decoder, but it already had a long address programmed.
I happened to have another DH126 on hand, swapped this one out, and the loco worked as specified, programmed with no long address in the default. No soldering for me to screw up.
I doubt if the well known retailer played any games with it -- looks like this somehow came bad from Digitrax for whatever reason.
Given some of the explanations here, this sounds like it has a bad output function, and I've e-mailed Digitrax. I am wondering if an already-returned bad decoder was somehow mistakenly recycled there, since it already had a long address.
However, in the context of this thread, I am having a hard time thinking that the less than 1% defect rate some guys have mentioned here is usual.
Setting CV08 to a value of 08 will reset the decoder to factory defaults.
I assume that you have tried that.
A new decoder out of the box should not have a long address pre-set.
Rich,
Sounds like a specific CV for the reset you quoted. Others vary, so depending on what's going in, checking the docs are always a good thing. I think NCE is CV 30 to 2, for instance.
I agree about the decoder not having a preset long address. The only one in there is usually the short address 3. I think that could well be a bad one that got recycled to a customer somehow from the sounds of it.
CV 8 = 8 is the universal reset to factory default on all Digitrax mobile decoders.
I do not own or op NCE powercab but I do know full programing mode can be used.
There is, of course, no separate programing track as such but I know a separate programing track can be isolated from main layout/track.
DPDT centre of switch is installed between system and layout.
Two wires from system go to the 2 centre terminals on DPDT switch.
Two wires go from 2 terminals at one end of switch to main layout.
Two wires from 2 other end terminals on switch go to separate programing track.
Switch one way puts power to main layout and program track is isolated, switch other way connects system to program track and main layout is isolated.
To test, program, or read CVs put switch so only program track is connected to system and main layout/track is isolated IE no power to it.
Before putting loco on program track put system into program mode which will make sure no power is on P/track.
Now put loco on P/track and read address and or some CVs.
You can put switch to centre off so no power is going to either main track or program track then put loco on P/track.
Put system in program mode then switch to program position.
I emphasise that system must be put in program mode before connecting program track to system
If short or problem is shown do not give decoder full power as disaster will most likely happen.
I know a few NCE powercab user have their system set up this way and works very well for them.
Some even have the programing track incorporated in their layout as a dead end siding or spur line so thay can drive the loco on to main track when all is OK and programed the way they want.
The above does take a little thought before doing any programing etc.
Cheers
Ian