How do you set up Zimo decoders?

[melmerby]

Hi all

I like Zimo decoders for their value and smooth slow running.

However I have a hell of a job setting up a straight speed profile for Traincontroller with the 3 point settings.

With a Lenz decoder if you set the start, midpoint and max to 1 - 127 - 254 the midpoint speed will be half the top speed and zero to max will be linear.

With a Zimo, it's anything but and even forward and reverse can be different.

The latest attempt with a MX618N18 for an Accurascale Manor ended up so far out I cancelled it.

 

The forward speed reached max at half throttle and in reverse it hardly moved until about a third of the way and then went up steeply☹️

 

 

 

[jpendle]

Hi,

 

All my Zimo decoders behave the same as yours in that they tend to reach max speed at around speed step 120.

All mine are diesel/electric so no difference running forwards or backwards.

I'm using iTrain and when I measure a loco's speed I tend to see an S shaped speed profile with CV's 2,5,&6 left at their defaults values.

 

I usually adjust either CV5 or CV57 to ensure that speed step 127 matches the top speed of the loco.

I don't bother adjusting the rest of the CV's as itrain can handle the bottom half of the speed curve.

 

Regards,

 

John P

[WIMorrison]

Keith,

 

I also recognise this challenge, and I find it even worse with the MN series, but I have been able to resolve it for my locos. 

 

When I have a flat line at the max speed I start by reducing CV57, usually down to around 100-120, which is nominally 10-12v to the motor. Then I will play with CV5 for top end and CV2 for bottom end which I like to be between 2 and 4kph for Speed Step 1. my next adjustment is CV 6 which I start as the value that is ⅔ the difference between CV2 and CV5. This is usually a pretty good start point, which I *may* tweak, but rarely do.

 

With the MN decoders I am finding that tweaking is a lot more involved, perhaps that is simply because I was so used to the MX series which I could get setup in 10mins, the MN are taking me 60mins or more and annoyingly CV2 is around 50 or 60 🤒, according to Zimo the speed algorithm has been changed from linear to logarithmic which seems to give a very concave curve out of the box.  I have also found that I have been playing around with the EMF CVs on the MN series to give me a similar performance to the MX, whereas I never touched them before.

[34theletterbetweenB&D]

12 hours ago, melmerby said:

I like Zimo decoders for their value and smooth slow running.

So have I.

 

5 hours ago, WIMorrison said:

With the MN decoders I am finding that tweaking is a lot more involved, perhaps that is simply because I was so used to the MX series which I could get setup in 10mins, the MN are taking me 60mins or more and annoyingly CV2 is around 50 or 60 🤒, according to Zimo the speed algorithm has been changed from linear to logarithmic...

We must charge Zimo with a serious infringement of KISS.

 

But then matters likely are different in the Germanosphere, where at least one in ten comprehend graphs, but in the Anglosphere I am pretty sure we are down to 2%  of the population, and that's for linear in both axes only...

[Pete the Elaner]

30 minutes ago, 34theletterbetweenB&D said:

 

 

But then matters likely are different in the Germanosphere, where at least one in ten comprehend graphs, but in the Anglosphere I am pretty sure we are down to 2%  of the population, and that's for linear in both axes only...

 

Also in the Austriaosphere? 😁

[melmerby]

Since I posted originally I've had another go.

I did a full reset

I decided to try the speed table instead. I set the first step to 1 and last step to max with a straight line inbetween

I reduced the voltage using CV57 to get the top speed I wanted and did a full speed profile which took ages (I just left it and did something else in the mean time)

It now has a gentle S shape, similar for both forward and reverse. It is now acceptable but not straight by any means.

 

I wonder why Zimo have gone this way when Lenz (and others) decoders have a perfect or near perfect linear response?

 

I've got a couple of MNs on the way, it'll be interesting to see how they shape up.

 

[melmerby]

37 minutes ago, 34theletterbetweenB&D said:

at least one in ten comprehend graphs

I had enough of graphs at work, what with standard deviations and calibration error plots, I never thought I'd encounter them in model railways.☹️

 

Believe it or not I have a pad of metric graph paper just here! (Actually it's X/Y plotter paper but it serves the same purpose!)

Edited August 11, 2023 by melmerby

[Nigelcliffe]

Has anyone who needs/desires a linear response raised a question with Zimo's tech support ?   An update might be a CV to switch on "straight line" (for automation)  rather than "S-curve" (nice for manual driving).  

[RFS]

I've adopted a simple approach with my Zimo decoders. I adjust CV 57 downwards until I get a top speed that I'm happy with. I adjust CV 2 until the loco just starts to move, then set CVs 5 and 6 to 1 and do the TC speed profile.

 

Here's an example of a Hornby N15 with MX616R decoder. It was profiled with 128 speed steps, but TC only subsequently lists it as 28 so the line is not completely straight. But the performance is perfect. I have 14 Zimo decoders (MX600R, MX616R, MX623R, MX634 and MX638) and get similar results with all of them. 

 

 

 

Edited August 11, 2023 by RFS

[34theletterbetweenB&D]

2 hours ago, Pete the Elaner said:

 

Also in the Austriaosphere? 😁

It's included within the accepted usage of 'Germanosphere', covers all countries where German is commonly used.

2 hours ago, melmerby said:

I had enough of graphs at work, what with standard deviations and calibration error plots, I never thought I'd encounter them in model railways.

So it wasn't you on one of the early and now long gone MR websites who suggested that my proposal for collecting data to obtain a weight on driven wheels to traction correlation should be confined to that place where the sun proverbially shineth not...

[melmerby]

16 minutes ago, 34theletterbetweenB&D said:

 

So it wasn't you on one of the early and now long gone MR websites who suggested that my proposal for collecting data to obtain a weight on driven wheels to traction correlation should be confined to that place where the sun proverbially shineth not...

Eh?

I've only ever been on RMWeb.

[WIMorrison]

@Nigelcliffe it was discussion with Zimo about the MN problems I was having because CV12 was incorrect that led me to discover the change from linear progression to logarithmic.

 

I have to say that now I have ‘mastered’ the MN I am now getting towards, actually very close to, the standard CV set that I will use as a base for new or reset locomotives. I have this for the MX series and I simply ‘pour in’ the standard set of values and this reduces the setup time to 5mins, followed by 20 mins speed measurement, giving 30 mins to a new loco running on the layout fully measured and calibrated to stop within 5mm of desired stopping position 😃

[34theletterbetweenB&D]

4 hours ago, melmerby said:

Eh? I've only ever been on RMWeb.

Sorry, I forgot to put a winky icon after my question...

[5BarVT]

15 hours ago, WIMorrison said:

@Nigelcliffe it was discussion with Zimo about the MN problems I was having because CV12 was incorrect that led me to discover the change from linear progression to logarithmic.

 

I have to say that now I have ‘mastered’ the MN I am now getting towards, actually very close to, the standard CV set that I will use as a base for new or reset locomotives. I have this for the MX series and I simply ‘pour in’ the standard set of values and this reduces the setup time to 5mins, followed by 20 mins speed measurement, giving 30 mins to a new loco running on the layout fully measured and calibrated to stop within 5mm of desired stopping position 😃

That’s impressive for stopping accuracy.  It’s clear I have a long way to go once I get my layout operational again.

For better or worse I standardised on DCC Concepts so will have to learn how to manage them..

Paul.

[Torper]

Do you have to set them up?  I've usually adjusted CV29 to the simplest setting possible (I think it would be 2 if it wasn't for the long loco numbers that push it up a bit) and my locos seem to run fine being driven manually.  I suspect that one would probably have to do rather more with sound decoders.

[melmerby]

5 hours ago, 5BarVT said:

stop within 5mm of desired stopping position

 

5 hours ago, 5BarVT said:

That’s impressive for stopping accuracy

I get similar using TC and mainly Lenz decoders, although the Manor is now close to that using the Zimo decoder.

[melmerby]

On 11/08/2023 at 17:36, WIMorrison said:

@Nigelcliffe it was discussion with Zimo about the MN problems I was having because CV12 was incorrect that led me to discover the change from linear progression to logarithmic.

 

I have to say that now I have ‘mastered’ the MN I am now getting towards, actually very close to, the standard CV set that I will use as a base for new or reset locomotives. I have this for the MX series and I simply ‘pour in’ the standard set of values and this reduces the setup time to 5mins, followed by 20 mins speed measurement, giving 30 mins to a new loco running on the layout fully measured and calibrated to stop within 5mm of desired stopping position 😃

Hi Iain

What sort of CV motor settings (non sound) have you ended up with?

 

I've just put one of the new MNs in my Dapol Mogul

Straight out of the pack with factory settings (about 14.4v DCC) top speed is 133mph (due to Dapols rubbish gear ratio).

It will go at any speed from zero to 133 depending on the throttle setting

 

My usual trick is to then reduce the MAX voltage from the decoder with CV57.

Unfortunately this doesn't seem to work as before, as with a slight reduction to say 12v the loco still goes like a bat out of hell for a short distance and then stops dead.

Reducing the throttle setting a little it will then start again.

I got CV 57 down to 7v and has cut the top speed noticably but it still stops dead after a short distance.

I've tried it with the speed table and the 3 point setting and it is the same.

With less than max throttle it will run continuously, anything over about 3/4 throttle and it stops after a short distance.

 

What's going on?

[Kaput]

The manual for the MS/MN series suggests using CV5 as the top speed control again instead of CV57.

 

Quote from said manual:

 

Quote

CVs #5, #57 (top speed, Reg’ref.): CV #5 gets the original configuration to reduce
the top speed back, which was in CV #57 in MX decoders. CV #57, therefore,
is used to select, whether the speed depends on running voltage or if voltage
variations can be eliminated by references to a lower value (e.g. 14 V). NOT
in SW. 4.50! Chapter 3.6

 

Although I don't entirely understand the part about "Not in SW 4.50" since the MS/MN software has only reached version 4.229 or so.

[Nigelcliffe]

13 minutes ago, Kaput said:

 

Although I don't entirely understand the part about "Not in SW 4.50" since the MS/MN software has only reached version 4.229 or so.

 

Version 4.fifty   and version 4.two hundred and twenty nine    

 

( The JMRI team have the same problem with some people misunderstanding things written this way ).  

 

That said, the Zimo manual is far from clear.  Whilst the introduction does say use CV5 rather than CV57,  the chapter 3.6 wording indicating that CV57 does control the motor speed much as it used to in MX-decoders.   Double-checking what the German language version says might be useful (not done that yet, but in the past it has untangled things which made no sense in the English version).  Then ask Zimo tech support to clarify what appears to be a contradiction in the manual.

 

- Nigel

 

Edited August 17, 2023 by Nigelcliffe

[WIMorrison]

 

Keith,

 

This is for a Zimo MN170 in a Tillig 2291 which is a small loco, but reassuringly expensive for the small size 😊 You can see the speed curve that I managed to get in the lower image.

 

My biggest problem is at the lowest speed because I set the locos to move at 2-4kp at speed-step 1 and this needs a CV1=100!, but it is a very nice smooth start and run at this setting. making CV1 lower will cause the loco to move but an imperceivably slow pace which is not what I want because it is too slow to accelerate for exhibition use and these settings are still realistic for the public, but giving much better reliability for starting and stopping.

 

Hope that they help.

 

Iain

 

 

 

 

[Vecchio]

Speaking about speed curves and Zimo - How do you all do your speed measurements? 

What I do is using a test circle with a rather large curve radius and some straight set track and time the speed per regulation step manually (stopwatch) over 2 meters. I am pretty sure the curves have an influence on the speed. But I do not own a garage or shed which would allow using only straight track for the measurement (5.5m).

 

The procedure described above is not very precise. Has anybody a better method? I thought of triggering a stop watch either by an optical/electrical input or just by an electrical (track) sensor. 

I am not using automatic breaking in front of signals at the moment, but I am using an ABC shuttle.

 

The main reason for me to get speed curve input are double headers (to sync the speed between different locos). 

I am using Z21 which has the table regulation step / time to sync locos. I am using 8 values in the table (Step 16,32,48,64 and so on)

 

If this has been already discussed somewhere in the forum, I am happy for a hint where to find it. 

 

By the way - I read the MX manual in German. CV5 is described as maximum speed of the 3 point speed curve if CV29 bit4 =0 

CV57 is described as an alternative to CV5, especially for cases where the track voltage is not stabilised.  

 

[melmerby]

15 minutes ago, Vecchio said:

Speaking about speed curves and Zimo - How do you all do your speed measurements? 

In the program.

I'm using TrainController and a fixed length straight section (in my case 72") to measure a speed, using block detection.

Iain is using iTrain and it works in a similar way.

Edited August 17, 2023 by melmerby

[melmerby]

6 minutes ago, Vecchio said:

I read the MX manual in German. CV5 is described as maximum speed of the 3 point speed curve if CV29 bit4 =0 

CV57 is described as an alternative to CV5, especially for cases where the track voltage is not stabilised.  

 

CV 57 sets the absolute output voltage which if used instead of Vmax will allow the full range of speed steps to be used

I have used it on MX618N18s and it works fine, which allows a max speed setting to be 255 whilst only reaching the speed you require.

The trouble with not using CV57 with something like the Dapol Manor, which is ridiculously over geared, will mean setting V max to around 100, wasting more than half the decoder's range.

[Jim Easterbrook]

1 hour ago, Vecchio said:

Speaking about speed curves and Zimo - How do you all do your speed measurements? 

What I do is using a test circle with a rather large curve radius and some straight set track and time the speed per regulation step manually (stopwatch) over 2 meters.

I also use a test oval but measure circuit timings with an Arduino and an optical proximity detector. A simple program gives me the scale speed for each lap. (The lowest speeds take 20 minutes or so per lap, so I can go and do something else for a while. Then come back and find the loco stopped on a bit of dirt.)

Edited August 17, 2023 by Jim Easterbrook

[melmerby]

There's always Rudy Boer's projects:

Speed measurement using an Arduino

https://rudysmodelrailway.wordpress.com/category/arduino/page/2/