Speed profile

[WIMorrison]

I have just installed a MX617f into a loco and I can’t for life of me get a decent speed curve :(

 

from 0 to around 35 it rises steeply and then from 35 to top speed it is pretty well constant.

 

i created a customer speed curve dropping very quickly and the a very gradual slope but introduced a lot of spikes and variations in the curve.

 

any suggestions on how to creat a straight line or a decent S curve?

[WIMorrison]

This is the best I can get using this speed table

 

 

which ends up with a speed profile looking like this

 

Ignoring the slight difference between forward and reverse what suggestions are there to tweak CVs to make this better ?

 

I am sure I am missing something simple, but I just cant see what it is

 

Thanks

 

 

[Dungrange]

I'm assuming that you are looking for the red and green lines to be straighter.  If that is the case then I think you're looking for lower values for the first 40% of the CV table before gradually returning to your higher end values.  I note that these currently rise in steps of three, so I'm going to suggest you go for steps of two up to the 40% mark and then jump up in threes and then fours.

 

Current value   New Value

10                       10

12                       12

15                       14

18                       16

21                       18

24                       20

27                       22

29                       24

32                       26

35                       28

38                       30

41                       33

44                       36

47                       39

49                       43

52                       47

55                       51

58                       55

61                       59

64                       63

67                       67

75                       75

80                       80

85                       85

Thereafter just leave your values unchanged.

 

[34theletterbetweenB&D]

Did you first try adjusting CV6 to see if the preset speed curve could be made satisfactory?

 

What's the loco  - or motor/gearbox combination if not RTR?

[WIMorrison]

Yes, but custom curve is always better.

 

It is a Bemo V51

[Izzy]

Do I take it that the speed profile has been generated by iTrain and the ideal for automated running is to have as linear speed line as possible? How do you normally generate custom curves, via decoder pro? I would have thought adjusting each individual step relative to the profile (fairly easy with DP) and then re-doing it to see the results is possibly the only way of doing it. 
 

The current profile does seem to replicate what might be expected from the average motor, where max torque arrives at about 50/60% of rpm and then flattens off or falls a bit. Depends a lot on gear ratio and loading of course. All ROT ( rule of thumb).

 

Izzy

[Crosland]

It looks like the output from JMRI and a Bachrus speed interface to me.

 

Problems like this are why we never tried to "close the loop" and have the software auto tune the decoder to match a desired profile

[RAF96]

When using auto-profile (decoder capable dependant) does the algorithm take account of any load on the motor. E.g the usual method (after setting various CVs) runs the motor along a test track at max chat for a couple of seconds, then stops it. This is obviously at no real load, but what if you tipped the test track into an incline would the algorithm associate this as a loaded train or would it just screw up the basic setting.

I don’t have any auto-cal decoders so cannot test the theory.

[Greg Elmassian]

Not real helpful, but I have found quite a few decoders where the custom speed table produces very strange results, where setting a value "early" in the curve creates differences way "later".. over here in the USA, the TCS decoders are (in)famous for this, it's poor math in the firmware, and apparently some other glitches.

 

By all rights, a smoothly increasing speed curve should result in smoothly increasing motor rpm. All you can do is keep tweaking.

 

Greg

[WIMorrison]

1 hour ago, Crosland said:

It looks like the output from JMRI and a Bachrus speed interface to me.

 

Problems like this are why we never tried to "close the loop" and have the software auto tune the decoder to match a desired profile

It is DecoderPro and iTrain 

 

what software are you talking about that Autotunes a decoder? I have never heard of this being mentioned before on here and it sounds interesting 

 

 

[Andymsa]

2 hours ago, WIMorrison said:

It is DecoderPro and iTrain 

 

what software are you talking about that Autotunes a decoder? I have never heard of this being mentioned before on here and it sounds interesting 

 

 

Hi Iain, 

 

I could be wrong but I think the auto cal that is being referred to is done by the decoder, but sets up the motor parameters not the speed curve itself. This function is available on the ESU v4 decoders and expect also the v5 ones, I not looked at other decoders for this setting.

[34theletterbetweenB&D]

20 hours ago, WIMorrison said:

...It is a Bemo V51

 

Little Mabuchi type motor, same type as in Hornby's ex-Dapol tank locos. I tried 'bemo v51 mechanism' in google and found a few German language threads relating to remotoring with a selection of alternatives to improve performance.

 

20 hours ago, WIMorrison said:

... custom curve is always better.

Yet it's the custom curve result that you don't like. The option of trialling the pre-set curve with CV6 values of around 30 might be worth exploring, if you haven't already thoroughly worked through this? It's quick and easy and doesn't prevent going back to a custom curve if it really isn't a goer.

[WIMorrison]

I have 5 of these chassis and this is the only one that doesn't perform, plus the conversion to a coreless motor isn't that simple and if 4 others work with lovely speed curves why should this one not work?

 

Strangely enough I had tried it with the standard settings - as I do will all locos -  before I went to a custom curve.

[34theletterbetweenB&D]

3 minutes ago, WIMorrison said:

I have 5 of these chassis and this is the only one that doesn't perform, plus the conversion to a coreless motor isn't that simple and if 4 others work with lovely speed curves why should this one not work?...

That's information that would have been helpful up front. Suggests either a motor or a mechanism that's well off the norm for the type. I would have the decoders out of this loco and the best of the rest, and see how they compare on 'vanilla DC'.  If the performance is clearly deviant on DC and you are so inclined, you could swap the motor around with the proven good one, and see if the poor performance travels with the motor or stays with the mechanism.

 

 

[Crosland]

21 hours ago, WIMorrison said:

what software are you talking about that Autotunes a decoder? I have never heard of this being mentioned before on here and it sounds interesting 

 

The software we didn't write! "we never tried to "close the loop" and have the software auto tune the decoder".

[WIMorrison]

Ah, I interpreted what you said as that you didn’t write it but that someone else had.

 

I can now read it the way you intended with the clarification

[WIMorrison]

 

Same loco, same decoder but now with a decent speed curve - and this the default curve from the MX617f 

 

 

And how was this acheived I hear someone ask?  I changed the motor for a coreless motor supplied by TramFabriek (took 10 mins to change), reset the decoder and adjusted for a coreless motor.

 

Happy bunny