Thanks to those who responded with possible solutions to my question about the slow speed start on the Walters H10-44. Nothing suggested made much of a difference so after quite awhile on the QSI programmer, I am assuming it is just the minimun speed of the motor/gear combination. My layout is point-to-point so I plan to put it on a roller stand to see if additional break-in could help. My only other QSI decoder at present is in a BLI SW-7 with a newer decoder (with BEMF) and it has better slow speed initial movement. Thanks again!
John C.
John.
Have you tried disabling the regulated throttle control? If RTC is enabled the decoder seems to disregard whatever is programed in CV2. My BLI I1s seems to run better on standard throttle than RTC. Also adding a bit of momentum in CV3 seems to keep the loco from jack rabbit starting. Down loading the latest manual off the QSI web site will explain the RTC and BEMF.
Pete
I pray every day I break even, Cause I can really use the money!
I started with nothing and still have most of it left!
After reading the original thread I decided to check in JMRI.
To change from CV2 and 5 to use the speed tables you need to change CV29.
CV25=3 is fast start for the speed tables, I suggest CV25=1 which is a user defined table. Then change CV67-92 to get your profile.
CV2 and 5 are disabled when using speed tables and CV3 is for acceleration.
Have a look at JMRI and even if you don't use it directly to program decoders you can use it to see what values to use for your liking.
Also 2 smph isn't that bad and will probably improve as the locomotive breaks in.
Peter
Thanks to you who responded with suggestions about the slow speed start-up issue with my Walthers H10-44. I have tried all the suggestions with no success. I could get low-speed start-up speed to be worse but never better. Once moving, the engine can creep forever at 2 smph but it is the little "mini-lurch" when it starts that is the issue and I have concluded that it is the nature of the beast. As mentioned before, I plan to run it on a roller-stand when mine arrives. I expect this will help. My other QSI equipped loco has similar tendencies as the H10-44 but to a much lesser extent. The H10-44 is excellent in every other way. The lurch is just an annoyance and is magnified when compared to my four Lok-Sound equipped locos, which begin to move almost imperceptibly and run smoothly throughout their range. As my wife tells me, "get over it!" I guess I will!
David,
The H10-44 runs excellently otherwise and the QSI sounds are good. The only issue is the "mini-lunge" at start-up, which is not a major issue, just annoying. I run the engine quite reqularly and its qualities far outweigh this one issue. I just hoped perhaps I could "tweak" the settings a little to eliminate this. As they say nothing ventured, nothing gained.
Mine runs fine, but I have other QSI equipped locos tha exhibit this issue.
I haven't tried, but it seems there ought to be an adjustment for "kick" frequency, or something.
Maybe a good ole reset is in store for the decoder. I have a unit that decided to go weired one day. It would lunge and stop etc until the throttle was half way, like an old DC loco with dirty track and wheels. A reset of the decoder fixed the problem. QSI requires 3 cv's to be changed for a reset.
Springfield PA
Yep, I did a re-set with no effect (except the labor involved in restoring appropriate CV's to my preferences). I have played with CV 2 and the speed tables and now have the minimum speed down to about 1 smph, making the starts a little better. Thanks for the suggestion, though--a good thought.
If it's running consistently at 1smph right from the start, you aren't likely to do any better - that's about as far from 'leaping into action' as you can get. Maybe you're being a bit too picky? If it runs that well you should be able to couple cars without jostling them, and you should get a nice even takeup of slack when starting out with a train behind you.
--Randy
Modeling the Reading Railroad in the 1950's
Visit my web site at www.readingeastpenn.com for construction updates, DCC Info, and more.
Yeah, Randy, I am a little picky about how my engines start as my layout is small (8'x6'x8) and point to point (U-shaped) and my preferred operational mode is switching, "Lance Mindheim style." As noted earlier, the ESU (LokSound) decoders have excellent low-speed control, despite their idiosyncrasies and expense, so it is an "apples and oranges" thing. I am satisfied with the speed response now and have stopped obsessing! I have put my QSI Programmer away! After running an ESU-equipped engine, the QSI engine seemed "a little jumpy" I guess! I'll drink a decaf coffee and chill!
So say it jumps when it starts to move right. After that jump, does it run the way you like? I know no where as much as the other people that have gave you ideas, but to me it sound's like there is to much lash in the worm gears.
With the shell off, turn a flywheel and see how longs it takes for the gear to turn. I have never owned a Atlas, but I am guessing the gears are set up the same as Athearn and Proto. Athearn fix is to install Kadee Washers in front or rear of the worm gear. It would be the same with Proto's, but I have never had the problem with one.
Cuda Ken
I hate Rust
John -- don't give up on your principles!
I have the same issues with two, nearly identical Atlas Dash8-40Bs. Both were upgraded. One has an ESU LokPilot Basic decoder and is an absolute delight to operate. Starting speed is 0.25 SMPH at speed step 1. The other has an upgrade to the QSI q1a chip. Slowest starting speed for it is 1.5 SMPH and if you were to compare the two, you would say it "jumps" to a start. Worse yet, there is no perceptible speed difference between speed steps 1 and 10 per the instructions that accompanied it! Granted, there is one difference: the QSI unit has sound. I have no idea what difference that makes, if any. I had both decoders installed just to get the benefits of BEMF.
I do not understand the P.I.D. info in the QSI manual, nor am I interested in spending hours trying to input various settings without the presence of a specialized controller or JMRI interface. I'm about ready to bite the bullet and trash the QSI in favor of another ($100+) ESU (Loksound).
The PID stuff for BEMF is actually simialr to what Loksound has. The P in QSI is basically what Loksound refers to as K in the manual, the I is the same. Loksound doesn't have the D available for the user to change. There's a good explanation of what PID is at Tony's: http://www.tonystrains.com/technews/introto_bemf_pid.htm
Seems QSI uses 7 sets of PID values depending on the speed step - this is so you can have BEMF drop off above certain speeds so it doesn't act as cruise control. This is also a good reason to use DecoderPro to program stuff like this. WHat you have is a somewhat overreactive BEMF setting which is why it jumps more noticeably. You need to turn down the resposne at low speed, but not totally disable it or it probably won't continue runnign at such a slow speed. There's no magic here, it's a trial and error process. Euro decoders generalyl have superior BEMF, which I've more than once heard is because they tend to do a lot more with automation so it is VERY critical that the loco moves when it is expected to and goes at an expected speed. There really is no comparison when it comes to motor control between the top Euro designs like Kuehn and CTE and typical US decoders like Digitrax and TCS (NCE doesn't have BEMF in theirs). QSI's is pretty much up there - you just have to tune it properly for the loco and motor it's connected to.
Glad you got it more to your liking.
IMO, the H10-44 is the best QSI implementation yet. Better than any other I have.
...but now we see the real advantage of ESU starting to show it's self, ther's really no comparison in motor control.
I also LOVE the sound ramping up BEFORE the motor, just like real life!
I don't get people who claim they don't like it.
If only the didn't just increase to TRIPPLE the cost of Digitrax sound decoders!