[Matej Pacha]

You can observe the impact of the adhesion model by disabling/enabling the advanced adhesion model. The main point is that the force is not computed from the power but it is computed from the wheelslip. Thus, if the wheelslip is oscilating, the tractive force is oscilating too. The jitter is here because of some coupler slack. Without the slack the mass of the rest of the train would "filter" this jitter. Anyway, just try to run MS FlightSim X on low powered PC and you'll get into the oscilations too...

[roeter]

 Matej Pacha, on 01 February 2013 - 07:38 AM, said:

You can observe the impact of the adhesion model by disabling/enabling the advanced adhesion model. The main point is that the force is not computed from the power but it is computed from the wheelslip. Thus, if the wheelslip is oscilating, the tractive force is oscilating too. The jitter is here because of some coupler slack. Without the slack the mass of the rest of the train would "filter" this jitter. Anyway, just try to run MS FlightSim X on low powered PC and you'll get into the oscilations too...

With all respect, I haven't got a low powered PC and I think an FPS rate of around 100 is NOT a low value. Yet I still get these jitters. So, it has NOTHING to do with the impact of low FPS. It has a lot more to do with multiple power in a single consist.

Regards,

Rob Roeterdink

[Genma Saotome]

For my part, thank you Matej for your answers, they certainly help me understand what I observed. Comparing the two runs by advanced slip vs. basic slip it was clear that everything I saw that was "odd" occurred under the advanced slip algorithm, operating at ~ 30fps. I will look for a situation where I can run that consist someplace else w/ higher fps and will report my findings afterwards.

It sounds to me like Rob is observing something different -- perhaps related, perhaps not and so I think the correct question back to Rob is to ask: have you run the same activity using the basic wheel slip and seen the same thing?

[Matej Pacha]

 roeter, on 01 February 2013 - 07:51 AM, said:

With all respect, I haven't got a low powered PC and I think an FPS rate of around 100 is NOT a low value. Yet I still get these jitters. So, it has NOTHING to do with the impact of low FPS. It has a lot more to do with multiple power in a single consist.

Do you observe this with the advanced adhesion model disabled? As I wrote - two different systems - low and high FPS - and two different experiences. MU consist is the case but I have no such experience with tripple DASH9...

Best regards,

Matej

[roeter]

 Matej Pacha, on 01 February 2013 - 09:31 AM, said:

Do you observe this with the advanced adhesion model disabled? As I wrote - two different systems - low and high FPS - and two different experiences. MU consist is the case but I have no such experience with tripple DASH9...

Best regards,

Matej

The fact that it does or does not occur with advance adhesion on or off does not mean it is caused by FPS.
Yes - I fully agree that low FPS causes problems with advance adhesion, and that will cause jitter. But that statement can not be reversed. It is not only low FPS which can cause jitter - nor is all jitter caused by low FPS.

This is the scenario :
run a single DMU (with one power unit) - no jitter.
run the same route, same activity, same actions with two DMU coupled (two power units) - jitter.
Now tell me the difference is FPS???
What's more : the jitter occurs with power off, brakes full on. The jitter disappears when the brakes are released. I know wheelslip can occur on heavy breaking. But I don't believe that such wheelslip can actually move the train backward and forward with the brakes fully applied for several minutes. That sounds very much like the tale wagging the dog.
The problem is worse when the power units are at each end of the train. Best to try is a top-tailed train, with engines either end, and about half-a-dozen passenger cars in between. Perform a brisk brake application with this consist and see what happens.

Regards,

Rob Roeterdink

[Matej Pacha]

OK, I really need to investigate this. I'm not sure but the advanced adhesion model is released maybe a year but there is no such a report but the low FPS machines - until now. It will need to record time steps and the variables conected with the force. Maybe there is a single long time step following by a set of normal short steps causing this. But I'm afraid that I won't be able to dive into this - I'm a bit loaded at work lately.

Regards,

Matej

[roeter]

 Matej Pacha, on 01 February 2013 - 12:53 PM, said:

OK, I really need to investigate this. I'm not sure but the advanced adhesion model is released maybe a year but there is no such a report but the low FPS machines - until now. It will need to record time steps and the variables conected with the force. Maybe there is a single long time step following by a set of normal short steps causing this.

It's not new, even though it may not have been reported earlier. Anyway, I could not have reported on it before september last year as it was then I got my new computer and finally got some decent FPS.

 Matej Pacha, on 01 February 2013 - 12:53 PM, said:

But I'm afraid that I won't be able to dive into this - I'm a bit loaded at work lately.

Regards,

Matej

OK - appreciate that. Just keep it in mind.

Thanks,

Rob Roeterdink

[Lindsayts]

If this has got anything todo with this thread I do not know, but it sounds very simliar.

One of the locos I have, A comercial offering the Victroian Railways H220, consists of three separate items the locomotive, the crew and the tender. Now the crew is setup as a another loco withs its own eng file with the TE and power set very low. Quite some time back, this on release of a certain version of OR, stopped working properly. It would not accelarate over around 4 mph and the motive force varied wildly over a large range. Recently on spec I got rid of the crew and of course the second eng file, after that the machine performed perfectly. I will give the original a go with the advanced adhesion turned off a see what happens.

Lindsay

[Lindsayts]

It was the Advanced adhesion model causing the problem, turned off worked OK.

[Genma Saotome]

Matej, I ran the locomotives using advanced wheel slip in circumstances where I could get 100fps. I did not observe any of the strange jumpy data that I reported earlier, so It appears that difference is entirely the consequence of high vs. low frame rates -- just as you said.

I would like to note something else I saw in today's test: while the data changes were all smooth the values seen in the Motive column were very often not within a narrow range of values for all three units. The middle unit, again was often an anomaly... I'd say most, but not all of the time with values + or - 15-20% different from the lead and trailing locomotives (motive values these two were always very close). None of the motive values changed very much until there was a grade change and then they'd settle down again. But as I said values for the middle locomotive were, for most of the time, well off the other two... sometimes higher, usually lower.

If there is another test you'd like me to run to monitor and report other data, let me know.