I picked up an Atlas U30B Silver HO loco at a local train show (new, unopened box) recently and had time to install an appropriate DH126P last weekend. Removed the shell, popped out the jumper plug, plugged in the decoder, put it on the test track and surprise, surprise, surprise. The immediate symptoms were delays in the loco responding to throttle changes. Running at a moderate forward speed keying reverse on the DT400 the loco may or may not change direction - then 3 to 4 seconds later it would change direction but sometimes at a different speed and then 3 - 4 seconds later it would speed back up, or down, to the set throttle speed. The bottom line is the loco runs erratically all over the place, forward and backward. Lights will always change immediately in response to change in direction but the motor will not. Plugged in a different decoder (DH166P) - same result. Any ideas as to what is going on? Any suggestions for a fix? Return to Atlas? If its something in the lighting board I can't find any spares poking around in the usual places on-line. Thanks.
Did you test the loco on DC before you added the decoder? If not, I would remove the decoder, replace the plug and test with DC power. If it responds correctly on DC then it's not something mechanical in the loco. I would then re-install the decoder and do a reset. If it runs OK after a reset you're home free. If not it could be something in the light board or the decoder.
Joe
WHat other equipment do you have? This sounds like it could eb the result of having the same address on another throttle, specifically a potentiometer type such as the Zephyr console or a UT4.
Reversing while still moving is just as bad in DCC as it is in DC. It's hard on the decoder and the loco - if the motor is turning and you reverse direction, this will cause the motor to draw maximum current as it tries to stop mootion and restart turnign the opposite way. This could be nearing or exceeding the decoder's maximum capacity even though under normal conditions with eve a long train coupled on it would not draw too much currnt. You can get away with it in DCC by setting some acceleration and decelleration momentum (cv3 and 4) which menas when you flip the direction while the loco is moving, it will first slow down and then speed back up in the opposite direction rather than try to jump from one direction to the other.
--Randy
Modeling the Reading Railroad in the 1950's
Visit my web site at www.readingeastpenn.com for construction updates, DCC Info, and more.
This problem sounds VERY similar to one I read about yesterday.
That person's solution (yup, SOLUTION) was based on "Packet time-out value". I suggest you do a search on the subject and see what you can find out.
In his case, he reset CV11 from 1 to 0.
It solved his problem.
If yer not in the mood for research, just do the above change. If it works, you win. If it doesn't, you can change the number back TO THE ONE YOU WROTE DOWN BEFORE YOU CHANGED TO 0. And you'll be no worse.
I recommend doing the search, first. Just for fun.
Good luck and please report back,
Ed
Or a full reset, CV8=8 for Digitrax. The default values for things like CV11 in Digitrax decoders definitely work with Digitrax command stations. If it works fine oin address 3 but then fails when the address is changed back, it's probably a duplictae throttle or corrupted slot, in which case a command station reset will clear it. If it still acts up on address 3, then it may be a bad decoder or an issue with the factory board, in which case the initial advice to swap the decoder back to the dummy plug and testing on plain DC will isolate.
Joe, Randy, Ed - thanks guys. I'll try to cover everyone's replies. Indeed, the loco ran / runs like a clock with its original analog light board. As mentioned, I tried a couple of different decoders with the same result at address 03 and the assigned address of 846. I'm prety much automated with Traincontroller software and seldom have any throttles at all plugged in except 1 for this purpose of implementing a new loco / decoder. Otherwise, all command station slots remain empty and locos are dispatched when I unplug my DT400.
Now, here's what I did, given my impatience with the matter and lack of anything else to do: suspecting something in the lighting board affecting only the digital signal (i.e. noise, waveform clipping, distortion, etc.) I disconnected the power leads coming from the trucks and disconnected the motor leads from the lighting board and wired them directly to the corresponding leads to the decoder which I had unsoldered from the 8 pin plug. Then, because the lights worked fine during all of this, I plugged the 8 pin plug which now was only supplying the lighing common lead and the fwd & rev light wires back into the lighting board socket. I then fired everything back up on the test track and BINGO! Worked like a miracle. Smooth as silk. I'm very excited! Put the shell back on, set the cv's like I want 'em, put it back on the layout and let 'er rip. After a couple of hours of break in could be one of the smoother and quieter locos in my roster. Hope this feedback helpful to others.
-geoff
7j43k Good luck and please report back,
And so you did.
Thanks. So, "bad board". I'll keep that in my brain bin.
Nice bit of detective work. Glad that you are up and running. It would be interesting to trace what the actual fault is in the board.
There's only a handful of the DC locomotives that I have converted that still have the original light board in them. The Atlas AEM7 comes to mind since they thoughtfully designed it so the plug-in decoder fits nicely under a roof hatch. No removing couplers and prying shells off!
Randy has been preaching—and I fully agree—that completely eliminating the original light board and hard-wiring the decoder is nearly always the best choice. Also, many of those boards rely on the poorly designed plastic wire clips to keep the wire in place... no good.
I have a shoebox "bluebox" full of old diode light boards!
In some cases like Bachmann and my Con-Cor MP54s the motor RF suppressor is still in the circuit and that will mess up any back EMF or other motor control the decoder may offer.
Regards, Ed
This is why I usually hard wire most of my decoders in a similar fashion - also I get just the basic decoders with leads, as they tend to be cheaper. Just hook the track pickups and motor wires right to the appropriate decoder wires, and wire in lights (I always use LEDs). That way I always know the electrical path - wheels to decoder input, decoder output to motor, with no 'extra' stuff in between potentially.
This would not be the first loco that has issues witht he factory board, nor will it be the last. Yes, the hard wiring is a bit more work than just plugging in an 8 pin plug, but it also results in fewer odd frustrating problems.
Ed, Randy, I hear you guys! Believe it or not I have at least a dozen older (very) Athearn (mainly), Life Like and Bachman locos that I have always practiced what you are preaching. Indeed, most often it was mechanically and electrically expedient to direct wire my decoders which is why I keep a few common 'P' and 'PS' types in the parts bin. But NO! Every now and then I like to tinker and since this board lent itself to keeping the lighting circuitry because of the resistor arrangement and the SM LED used for the nose headlight which had its own resistor, I made the decision and paid the small but interesting price. I'm pretty confident there is some stray capacitance on the board that was warping the DCC waveform which was difficult for the decoder to process. If I think it through that theory supports the symptoms 100%. The DC voltages I checked were always right on so it wasn't there. Probably need a digital signal guru to corroborate but until then I'm sticking to my story. Cheers.
PS - as a bonus, the Atlas U30B Silver loco does really run extremely well and looks good to boot.