Kato N scale FEF TCS DCC problem

[bobdobbs]

I am having issues with slow speed control on DCC.  with Factory defaults, speed step 1 is way too fast.  Disabling BEMF (Set CV61 to 0 ) makes it run even faster.  Setting CV66 (motor trim forward)  to 1 actually produces the slowest results (with BEMF CV61 set to 1), but nowhere near a creep like I can get from any of my other DCC locomotives regardless of the type of decoder or company of manufacture.  This is my first locomotive on the layout with a coreless motor, and I am not sure if this has something to do with it.  Beyond that, the locomotive surges randomly when BEMF is engaged, and if I turn BEMF off, it does not surge, but once again slow speed suffers.  This is a beautiful model and does run really well, I just can not get slow creeping starts like I have seen on DCC.  I am using a MRC Prodigy advance wireless system, and I have never had a problem with slow speed operation before, even my Bachmann S4 runs incredibly well at very slow speeds.  I have contacted TCS and they suggested turning BEMF off, and that does not help.  Turning to 128 speed steps does not produce slower starts or speeds, nor tweaking CV 2 5 or 6.

[jdcolombo]

I am having issues with slow speed control on DCC.  with Factory defaults, speed step 1 is way too fast.  Disabling BEMF (Set CV61 to 0 ) makes it run even faster.  Setting CV66 (motor trim forward)  to 1 actually produces the slowest results (with BEMF CV61 set to 1), but nowhere near a creep like I can get from any of my other DCC locomotives regardless of the type of decoder or company of manufacture.  This is my first locomotive on the layout with a coreless motor, and I am not sure if this has something to do with it.  Beyond that, the locomotive surges randomly when BEMF is engaged, and if I turn BEMF off, it does not surge, but once again slow speed suffers.  This is a beautiful model and does run really well, I just can not get slow creeping starts like I have seen on DCC.  I am using a MRC Prodigy advance wireless system, and I have never had a problem with slow speed operation before, even my Bachmann S4 runs incredibly well at very slow speeds.  I have contacted TCS and they suggested turning BEMF off, and that does not help.  Turning to 128 speed steps does not produce slower starts or speeds, nor tweaking CV 2 5 or 6.

Well, I don't think I'm going to solve your problem directly, but perhaps indirectly.  This will be a long story.

I got tired of waiting for my TCS "drop in" decoder, so I decided to hard-wire an ESU LokPilot (no sound).  The LokPilot is small enough that it fit under the rear of the stock Kato board, and I used the traces on that board as solder points for the power pickup, motor leads and front LED.  So far so good.

But when I started running the loco, I had the same problems you did.  Speed step 1 wasn't slow enough for me and the engine ran very "jerkily" (is that a word?) until I got up to about 5 on my Digitrax DT400 throttle - then it smoothed out OK.  Thinking that maybe it just needed a little break-in time, I let it run around the layout for about an hour, then coupled it up to a 30-car freight.  Not only did I have the same problems at speed step 1, but with the load of the freight cars, the engine now "surged" at low-medium speeds and upward.  I was not happy.

Then I remembered Victor's thread and the apparently new coreless motor, and began to wonder if I needed to adjust the BEMF parameters in the LokPilot for this motor.  The ESU has 4 such parameters, which I won't go into here, but it turns out that three of the four parameters needed some major manual adjustment to get smooth "creep" at speed step 1, and to avoid the surging.   After about 30 minutes of trial-and-error programming of these four parameters, the engine now runs like a dream - creeping at step 1, smooth as butter in the rest of the throttle range.

Now here's the problem - unlike ESU's LokPilot (or Lenz's Gold/Silver series), it doesn't appear that TCS has user-adjustable BEMF parameters.  I just downloaded their general programming manual, and I don't see any CV's for BEMF parameter adjustment.

So here's my conclusion.

1.  Yes, your problems are related to the BEMF implementation.
2.  If the individual BEMF parameters are adjustable, like on my ESU LokPilot, you can smooth everything out and get tremendous performance from this engine.
3.  The TCS decoders do not appear to have individually-adjustable BEMF parameters.
4.  If I'm right about 3, you're basically screwed, unless you are willing to do what I did, which is buy an ESU LokPilot decoder and hard-wire it into the standard circuit board.  If you want to do this, I can lead you through the steps (which require cutting some traces on the board, scraping some insulation here and there for solder points, and so forth).  It's a couple of hours of work and needs some modest soldering skills, but this will solve your problem.

Note, finally, that I'm not an expert on TCS decoders, and it may be that I'm missing how to adjust individual BEMF parameters for these decoders.  Perhaps someone more familiar with TCS's BEMF implementation can chime in.

John C.

[Bendtracker1]

Bob,
If you set CV61 to "0" on a TCS decoder, you are actually shutting BEMF and are toggling over to Dither Control.  As JD said, there is no adjustment for BEMF on a TCS decoder, it is automatic.

To help slow things down you might want to try adjusting CV56 & CV57.  This is explained in this link  It's from TCS's page and explains how Dither works.

http://www.tcsdcc.com/Customer_Content/Technical_Info/Tech_Info/FAQPage.php?q=21

This is also from TCS's site and and shows a little more how to adjust the Dither Control

http://www.tcsdcc.com/BEMF.pdf

I had a similar problem with a TCS chip I installed an Atlas VO-1000.  The closest thing I could figure out was that the chip had both the BEMF and the Dither on at the same time, basically the chip was scrambled and i was told by a friend to do a factory reset and then I adjusted the dither.  The unit runs like glass at low speed now.

It takes a little playing with until you get use to how the values work and how they react.   Not sure if this will help, but it 's always good to know how to set the CV's

[bobdobbs]

Thanks for all the outstanding info.  The situation sucks for a few reasons.  The Factory Installed DCC FEF does not include the DC board, which I feel you pay for when you buy the engine so cant hardwire anything to it, and TCS did not design firmware or circuit design or both to take advantage of this incredible Loco, I am a SP modeler, but this engine is awesome.  What is sad is my 2 Intermountain cab forwards creep at speed step 1  With digitrax 8 pin decoders and so do my Key Mt4 and Mt5's with hard wired "run of the mill" DZ 123's. 

The best solution I have is CV61 to 3 (allows BEMF cut out with fn 6), CV66 and 95 both values of 1 (forward and reverse motor trim)  Dither adjustments didn't seem to do anything dramatic.

Thanks for the analysis, and I will post any updates of my tinkering.

[jdcolombo]

Thanks for all the outstanding info.  The situation sucks for a few reasons.  The Factory Installed DCC FEF does not include the DC board, which I feel you pay for when you buy the engine so cant hardwire anything to it, and TCS did not design firmware or circuit design or both to take advantage of this incredible Loco, I am a SP modeler, but this engine is awesome.  What is sad is my 2 Intermountain cab forwards creep at speed step 1  With digitrax 8 pin decoders and so do my Key Mt4 and Mt5's with hard wired "run of the mill" DZ 123's. 

The best solution I have is CV61 to 3 (allows BEMF cut out with fn 6), CV66 and 95 both values of 1 (forward and reverse motor trim)  Dither adjustments didn't seem to do anything dramatic.

Thanks for the analysis, and I will post any updates of my tinkering.

The stock circuit board is available as a part from Kato for $5.  Might be worth it.  Then get a Lenz, Zimo or ESU decoder to hardwire.

John C.

[Mark W]

This is a bummer to hear.  I also have the TCS board and am having the exact troubles Bob has. 

[C855B]

This is a bummer to hear.  I also have the TCS board and am having the exact troubles Bob has.

I'm going to hold off on installing mine, then. This sounds like the firmware needs to be revised to accommodate the coreless motor, or, at least, TCS reviewing the problem in-house for more comprehensive instructions on adjusting CVs, eventually incorporating the revisions into default values.

That the board is application-specific and not some manner of adapted generic design covering multiple models leads me to edu-conjecture that a firmware change will be the response.

[RAILCAT]

Wonder why Kato did not follow Fox Valley and fit a 6 pin DCC socket.

[peteski]

Wonder why Kato did not follow Fox Valley and fit a 6 pin DCC socket.

Good question - maybe they felt that 2 functions (only forward and reverse headlight) would be limiting. Maybe it was a space issue (plug-n-play board replacement decoder is the most compact option available). 6-pin plug and add-on dummy board or decoder takes up more room).

I was wondering about the ability of the decoders to properly drive the coreless motor. Looks like they aren't doing too well. There are some decoders out there (I don't exactly recall - maybe Zimo) that are designed specifically to handle coreless motors.

[nstars]

Although most people don't know it, I think this is a very clear example that not all motor - decoder combinations will result in a good performance. This is the reason why we have several brands of decoders on our workbench. In general european decoders tend to work a little bit better with coreless motors as coreless motors are more common in Europe. Besides de Lokpilot we have had very good results with either the CT Elektronik or the Zimo decoders. What a coreless motor requires is a very smooth DC current and these decoders can provide it. With the CT Elektronik decoder we normally don't need an adjustment of the decoder for perfect performance.

Marc

[jdcolombo]

Marc,

Thanks for the info.  I ordered a CT Electronik DCX33D and a Zimo MX621.  The CT Electronik decoders are harder to find in the US than the Zimo or ESU's, but Streamlined Backshop had one (they are now out of stock; I think I got the last one).  Going to try each of them in the FEF and see how they compare to my tweaked ESU LokPilot.  Will report on my findings (probably next week).  It may be that we're just going to have to be creative to get the best performance out of this new offering from Kato, but since I expect this motor to find its way into other Kato offerings in the future, it seems worthwhile to figure this out.

John C.

[peteski]

Marc,
I doubt that any DCC decoder provides smooth DC to the motor.  That would mean that the motor driver was an analog circuit. Those circuits waste all the unused power as heat.  N scale decoders cannot afford big  heat sinks.

I suspect that the coreless motor friendly decoders still use PWM (and switching transistors) for driving the motor outputs. But the decoder firmware is specifically written to deal with the specific power requirement of those coreless motors. I don't know what specifically: Maybe lower or higher pulse frequency and maybe shorter duration pulses).  The BEMF feedback loop routine is most likely also written to deal with the specific BEMF output of a coreless motor.

[bobdobbs]

Thanks for all the great insight from you guys!  I am glad this is not a problem with my unit exclusively.

[Mark W]

Is anyone here using the Digitrax board?  Any similar troubles, or did they build it properly?

[carlso]

I have a question / comment, and I am not attempting to start a war, but where was KATO or TCS QC in this deal. Did either one of them test the installs before releasing them?  Did either one of them suspect a problem with the coreless motor? Did someone think the model community would accept and shake off the situation?
Did the employees, who might have tested them, know what they surely saw, like the problems you guys are seeing?

Just curious, inquiring mind, you know. After such a beautiful release, and the loco is fantastic, this is a real bummer.

Carl