Enough PC-power??

[Per Olsen]

Hello again.

I am beginning to think I might have too little computer-power for my layout...?  Look at the screenshot, is this good enough for WinDigiPet 2012.1 ?  How many trains/feedbacks/turnouts etc can this PC manage safely?

Best regards,
Per.

[Stefan Lersch]

Hi Per,

this is round about what I had until some months ago. I only changed the PC because it got slower and slower - you know, the old Windows problem. I also use this PC to arrange my movies and there I need more power. For the model train this is sufficient. I had up to 6 trains running at the same time and round about 40 trains on my layout. For WDP switches the turnouts only when nessessary this is not much. I don't know wich cental staton you have. I have a Tams MC wich is very fast and I never had any problems.

[Per Olsen]

Hello !  Thanks for answer.

Yes, I did think I had enough power, but I am not sure..  On my old, much smaller, layout, I had the "problem" of the trains not stopping exactly at the same point every time (on each stop-point).  I wanted more precision, so I could be sure exactly where the front of the loco would stop EVERY time. 

So, to eliminate as many problems at possible, when we now had to make a new, and bigger, layout, I did all I could to minimize this problem.

I bought a Tams controller, for speed.
I bought 2 Littfinski USB Hi-Speed interface for the S-88 bus.
I installed 2 big transformers each 240 VA, feeding 6 seperate boosters.
I have a 3 big shadow-stations, divided in 3 booster-sections.  I have room there for about 50+ trains.
All S-88 wiring is seperate from all other wiring.  So no RFI and no false occupy-messages.
All feedback decoders are brand new (40 together, most of them in the shadow-stations).
All turnout-decoders are brand new (48 together, most of them in the shadow-stations).

Now, when testing, I have max 5-6 trains running at the same time (but the result is the same with only 2-3 running at the same time).  No other trains on the layout.

I have started to test out the Intelligente Zugnummerfeld.  The precision seems better.  I have so far only testet with "stop at signal".  But no 2 trains stop at the exact point.  Even if they have all been speed-calibrated first.  I think they should all stop at the exact same spot?

Any help is greatly appreciated.

Per,
Bergen,
Norway.

[Stefan Lersch]

Hello Per,

are they calibrated with 15 point? Have you already used the Bremskorrektur? After I calibrated a loco I immediately test (always at the same iZNF) where they stop. Then I set the Bremskorrektur. All my trains stop (there) within a 3mm range. They also stop at the other ZNFs and iZNFs more precisly, maybe not within a 3mm range but a bit more.

[Per Olsen]

Hello again !

Thanks a lot for answer !!

Yes, I use 15 points, always.

What do you mean by "Bremskorrektur" ?  I searched the PDF-files (Handbuch and Update) for that word, but cannot find it.  If you mean the "Abbremsen" in "Fahreigenschaften", they are always on full (18).  Should I correct this, to make all locos stop at the same spot?  With IZNF I thought it shouls always be set to max?  I am new to IZNF, so I may have misunderstood something 

I use the German version of the software, but my German is not so very good, as you may have understood.  My english is much better.... I still prefer the German versions, as the english versions always come later, and the english docs are not always correct.  Anyways...

Best regards, and thanks a lot, so far !!!

Per..

[Stefan Lersch]

Hello Per,

attached you find a picture of the Fahrzeug-Datebank. In the corner to the bottom, right you find the Bremskorrektur. What does it do? If you have a loco that doesn't stop at the right place but drives too far (at an iZNF) then use the Bremskorrektur. The higher you set the value the earlier the train stops.

[Per Olsen]

Hello again.

Last post today, my bed is calling (early to work tomorrow..

Thank you, I have used it now.  I did need to correkt the braking.  I have a straight distance of 3m that I use for the calibrating.  In fact, the same distance I use for calibrating the speed of the locos.  The measured distance is 311cm.  All straight.  Also the line before and after, is traight.

I adjusted one of my best locos.  It now stops exactly at the right spot (signal) on the test-track.

After that, I tried it on another place of my layout.  It is in a station.  It enters the station at 50km/h, and then IZNF's down to stop.  But here it misses with around 10cm.  Each time.  Stops 10cm too short every time.  I cannot see any mistake I have made here...

Have to check more.  Don't know what this problem is...

Best regards,
Per.

[Stefan Lersch]

Hello Per,

if your loco once stops too early and some place other too late then please let me ask you if both ZNFs are really iZNFs. This only works with iZNFs. Additionally this works only when you don't use "Stop am Kontakt" but the other options. Also the length of the iZNF has to be long enough that the train can stop.

[Per Olsen]

Hello.

Quick reply before I run to work:

Yes, both are real IZNF.
I use "Stop at signal" in both places.
The distance is around 3m (don't remember, have to check).

Must run..

Per.

[Stefan Lersch]

Helle Per,

is your decoder old? Do you have the same experience with other locos, too? Have you entered the length of your loco in the database and also the distance between buffer and the first feedback axle? If you have Marklin it also could be distance to the slider.

[Per Olsen]

Hello again.

I took some time to do some more careful testing.
I have one big station with 4 identical tracks with each a length of 316,8cm from start of first RMK to end of the last RMK.  This is identical for the 4 tracks.

I set up a new calibration-section, using the same station-area.  I calibrated again one norwegian EL17 loco and one Marklin Thalys and one TGV.  The EL17 uses LokPilot 3 decoder, the TGV and Thalys uses MFX.

I used the same station-area to program the braking of each loco.  I used IZNF where I put in the exact lenght of the track, and put in 49cm from the signal to the end of the track.  I created a route for this, and put in there to "Stop by signal".  I measured exactly 49cm from the end of the track, to find the point where the trains should stop.  In the loco database, I put in the lenght of each train (which in this case should not be critical) and also the exact lenght from the front of each loco to the first wheel with feedback.  Next I calibrated each locos braking till they stopped at exactly 49cm.

The EL17 seems to stop at the exact same point every time (as many as I have had time to test so far..).
TGV and Thalys sometimes stop up to 7cm later.  Don't know why.  Maybe because of MFX?

But next I changed the 49cm to 29cm (meaning the trains should run exactly 20cm longer down the track before it stops.
Result:  None of them travel 20cm further ! 
EL17 runs 15cm further.
The others run 10-17cm further.

Why do not at least 1 of them, stop exactly 20cm later?

What could I be doing wrong..?

You ask for the age of the decoders.  They are all brand new.  All feedback-decoders are type Bolls.
I use Marklin C-tracks, all are brand new for this layout. All are very well cleaned, so all locos run very well.

Bottom line:  I need to be able to program WDP to stop all trains exactly where I want them to stop.  This is why we made this huge effort to create as "perfect" layout as possible this time.  And why I think IZNF is perfect for me.  But I must be sure the trains really DO stop where I want to, every time. 

I do wonder if I need a faster PC, but as you said earlier, this one SHOULD be OK.

Best regards,
Per.

[Stefan Lersch]

Hello Per,

unfortunately I have no experience with mfx as I run DC. How many different feedback contacts do you have in your 316 cm long track? E. g. I have in my 94 cm long track (N gauge) 5 feedback contacts. The more you have the preciser the locos will stop. WDP calculates at the beginning of the first contact the way it has to slow down the locos. With contact of each new contact within the iZNF WDP recalculates.

Is there a way you can - for a moment - encrease the number of contacts within the iZNF. I mean by deviding  your now existing contacts.

Second idea: do you have the possibility to borrow a much faster PC for a while and test it?

3rd idea: I've seen that you use an HSI. Please try to connects the s88's directly to the Tams MC. Does that help? If yes you might have a problem with your HSI driver.

I will ask Markus if he's got more ideas for  he knows the Marklin decoders.

[Per Olsen]

Hello again.

I really do appreciate your help !!!

Within the 316 cm, I have a total of 6 RMKs.  Should be plenty....  It was planned for non-IZNF usage

No, I do not want to perform the big job of installing extra RMKs, sorry..

Yes, I will borrow my sons PC this weekend, this was already planned, to see how that goes.

It is a very big job to re-program all the feedbacks, if I take away the HSIs.  I will not do this, not for the moment, maybe if I get REALLY desperate..

Thanks again !

Best regards,
Per.

[Stefan Lersch]

Hello Per,

Markus shares my opinion. Additionally he says that a mfx decoder should not be the problem. However Markus's got an idea: have you reduced the deceleration rate in the decoders? You should set them to 0 or 1 (some decoders understand 0 as "default" and not as "nothing").

[Per Olsen]

Hello again.

After many hours testing today and yesterday:

I do have more variable results with trains with MFX decoders, than with LokPilot and old Marklin decoders.

So, to the results of the testings..:

First with my own computer, I calibrated all locos to stop 20cm before the end of the track, adjusted with the Bremsekorrektur, and with the deoders own braking set to 0 (not default).
I then changed the value from 20 to 5cm.  2 locos stopped always at 3-4 cm.  2 stopped at 3-5cm.
Last I changed to 60cm.  One stopped at 58, one at 57, one between 56-59 and one between 56-59cm.

But i do not like much the running of the locos with all decoder-braking OFF, so I put back the values I normally use for braking.  (Re-calibrated, of course).  It did not seem to make a very big difference.
At 5cm the same 2 stopped at 2,5-4cm.  The other 2 stopped at 3-7cm.

Next, I tested with my son's PC, it is 3GHz and 4Gb memory.  I copied everything exactly to this computer, so all should be the same.  On both PCs, nothing else is running, not even a firewall or virus-scan.  Nothing.
The results were exactly the same (as well as I could measure, within the hours I spent).

So I will give my son his computer back..

As you may see, at 5cm, none of them stopped there
And at 60cm, none stopped there either.
Close, but not exactly there.
I guess I just must live with this, and allow for more variations then I had expected?  This is really only a problem for the longest trains, that need to use the entire available tracks at some of the stations.  Smaller trains are no problem.  I look forward to program the smaller trains to stop at the platforms, by the way..  But one thing at the time

This software is really GREAT.  I have used WDP for many years now, and like it better and better.  The long braking with IZNF is fantastic.

Best regards,
Per.