Speed matching - how close is good enough?

[davetheroad]

OK, before making a consist you have to speed match the locos but just how close do those speeds have to match?

Reading everything i can find on the subject does not give any clues to what is 'good enough'. DCC wiki says 10% will be good enough. Are they kidding? that is a 10 foot variation on a 100 foot run!.

 

The idea is to occasionally double head pairs of locos of the same class, 2 x Hornby 2P, 2 x Bachmann 3F, 2 x Bachmann Ivatt 2MT etc.

 

What would be the theoretical accuracy of speed matching, given that the motor response to voltage change was linear?

[Adrian Wintle]

It depends on how you intend to run them. 10% is reasonable if you are going to run a consist at a reasonable speed on a roundy-round (as long as the faster one is the lead loco).

 

Theoretically, you should be able to get in the 2% range, with a lot of caveats. You have 28 speed steps to play with, so getting within half a step would be about 1/50 or 2%. This is assuming that the motor characteristics are repeatable - i.e. both the motor response and the gear drag aren't temperature-sensetive and that both locos respond to the added drag of curves the same way. Note too that you should probably speed-match with BEMF turned off, as the throttle response will be different. As an example, Digitrax decoders disable BEMF by default in a consist, so if you speed-match your locos with BEMF enabled, they might not be well matched when consisted.

 

Adrian

[34theletterbetweenB&D]

Never mind theory, just try it out. I do my speed matching on a scale half mile test circuit and find no difficulty in matching within an inch over that distance. So that's circa quarter percent. As I also use pretty large CV3 and 4 values - rarely less than 50 -  I never bother consisting, just send the inside loco off at speed step whatever, then the lead loco. The inside loco may manage a quarter turn of wheelslip, before the lead loco starts is all.

 

When setting up to match the new loco against the standard, remove couplers or you are forever having to uncouple. What is quite fun is that it makes the effect of curves very visible, the lead loco into the curve slightly caught by the trailing loco until it too enters the curve, then the lead loco pulling ahead slightly as it enters the straight, and so  on.

[HLT 0109]

34C's method certainly saves a lot of time but I am not sure it would work with a banker.  I do agree with him that you should be able to get a lot close than 10%.  Your comment "10 feet ... in 100 ..." seems less dramatic somehow than 1" in every loco length!  I found that it matters little whether your are trying to match models of the same type of loco from the same manufacturer.  If you get a good match and start off with a very slow crawl, things work well - even with two sound locos.

 

Harold.

[Torr Giffard LSWR 1951-71]

.....a sensible differation between speeds of each class of train is, in my view, equally or more important than accurate scale speed as it shows understanding and awareness of the prototype.

 

Dave

[SHMD]

Speed matching isn't all that important!

OK, you wouldn't couple up a shunter to a High Speed Train AND there can be some strange/unwelcome effects when things (the control circuits) started fighting (hunting) each other.


Take, for instance, the following scenarios:

You have two identical locos made from the same batch, and indeed, run pretty much identically Lets call them loco A and Loco B.

Now, run them with a 6” gap between them and, inevitably, they will either drift apart or come together. Does it matter, if after a short time, they come together – couple up – and the continue? Obviously one, the faster one, will be doing more work than the other. Does THIS matter? Not really - BUT it does in our minds!

Now, with the same locos, start off with the same 6” gap BUT this time loco B has 36 loaded hoppers behind it.
Clearly, when run, loco B will VERY quickly fall behind. There will be a much bigger speed differential between the two locos. Even if BEMF is in operation. (Remember, BEMF works on wheel speed and NOT on track speed.)
Now repeat but with the two locos swapped around. Now loco B will run away.
NOW repeat, but this time, couple them all together.
In all three scenarios each loco will behave differently BUT will still be quite 'happy'. After all, this happens ALL the time!

Ah, but what about BEMF?
Yes, this can cause cause problems, but if all works perfectly then the speed difference will result in wheel-slip only. NOW, does this matter?

Well I would say that this is irrelevant at model railway scales!


Get the speeds close enough and, as long as there are no hunting effects, then all will be OK.


%10?
Yeah, %10 +/- 1/2 a yard.
Just try it and see!


Kev.

[Nconsistent]

I am very happy with my NCE powercab, but I am having trouble getting a Bachmann 4CEP and Bachmann MLV to run at similar speeds in a consist. Both have 21-pin chips: 4CEP Bachmann, MLV Hattons. The 4 CEP which I bought second-hand arrived running just as I want it, the MLV bought new has always been slow to move off and then accelerates too jerkily for my tastes. For the same throttle position it is about 25% slower than the 4 CEP

I have read the manual and threads about this subject repeatedly and have modified the CV2, 3, 4,  and 6 of the MLV in an attempt to match the 4-CEP, to no avail. Whatever I do, it doesn't seem to affect its speed very much .

I can run a 'fake' consist with the 4 CEP running at throttle position12 and the MLV at 16 for a realistic cruising speed, but this is far from ideal.

Am I doing something obviously wrong?

[bgman]

I had similar problems "speed dating" once is that the same ?

[johnb]

I'd suggest that you try using the same manufacturer of chip in each unit that you are trying to consist. Yes in theory differing chips should work OK together but ......

 

I rarely consist things, but when I have I have used the same type of chip. I've had no issues with a pair of Bachmann chips in consisted 2EPBs. Haven't tried my pair of MLVs, but both have Bachmann chips and run very nearly at the same speed with the same throttle settings.

 

If you are getting jerky running on the MLV, is it set to a dfferent number of speed steps? 14 steps can be jerky on some motors. If it's a new one have you run it in for 30 mins each way? New motors can be pretty sticky at first. A used 4CEP is likely to be well run in.

 

I've had no problems with either Bachy or Hattons chips, but others have, as recorded elsewhere on here. I do tend to prefer TCS chips though.

 

I disable DC running and if possible BEMF (or at least set it to the minimum possible) on every chip that I install.

I only build small layouts so generally run at little more than shunting/switching speeds.

 

I too am a well satisfied  NCE Powercab user.

 

HTH

 

John

[Ron Ron Ron]

Thank goodness for those new digital control systems that incorporate automatic real-time speed matching. 

[Nconsistent]

I had similar problems "speed dating" once is that the same ?

Speed dating is much, much quicker than what I've been up to.

[Nconsistent]

I'd suggest that you try using the same manufacturer of chip in each unit that you are trying to consist. Yes in theory differing chips should work OK together but ......

 

I rarely consist things, but when I have I have used the same type of chip. I've had no issues with a pair of Bachmann chips in consisted 2EPBs. Haven't tried my pair of MLVs, but both have Bachmann chips and run very nearly at the same speed with the same throttle settings.

 

If you are getting jerky running on the MLV, is it set to a dfferent number of speed steps? 14 steps can be jerky on some motors. If it's a new one have you run it in for 30 mins each way? New motors can be pretty sticky at first. A used 4CEP is likely to be well run in.

 

I've had no problems with either Bachy or Hattons chips, but others have, as recorded elsewhere on here. I do tend to prefer TCS chips though.

 

I disable DC running and if possible BEMF (or at least set it to the minimum possible) on every chip that I install.

I only build small layouts so generally run at little more than shunting/switching speeds.

 

I too am a well satisfied  NCE Powercab user.

 

HTH

 

John

Thanks very much for your thoughts, I have run it in properly, and checked the speed steps, so next step is to swap in a second Bachmann chip. I was surprised at the big difference in performance as I imagined that both units use the same Bachmann power bogie.

[johnb]

Thanks very much for your thoughts, I have run it in properly, and checked the speed steps, so next step is to swap in a second Bachmann chip. I was surprised at the big difference in performance as I imagined that both units use the same Bachmann power bogie.

 

No problem, pleased to offer some thoughts. A further thought though, do both the CEP and MLV run in similiar ways and speed on DC, if they match speeds (well ...  more or less) then I'd suspect the chip in the MLV.

 

The power bogies are probably very similar, but  ..... if they run on widely differing speeds on DC then you will need some work to get them balanced. A SPROG/JMRI setup can present the speed table graphically which would help in balancing things.

 

One other thing, don't necessarily expect ostensibly the same mechanisms to work similiarly, manufacturing tolerancies come into play particularly over time.

 

John

[Andrew F]

If changing the MLV's  CV's 2,3,4and 6 is having no effect then are the CV 29's on both chips set the same? Is advance speed step activated on CV29 on the MLV and not the CEP and are they both using the same speed steps?

[Judge Dread]

Re speed matching 4-CEP's. I adjusted the speed settings on mine and ran two power cars on the layout uncoupled, i.e. separately. I noted if they drifted apart or came together. First trails with a full eight coach train did have a couple of derailments if the power cars where at opposite ends of the formation. However, when they ran in the centre of the formation, the problem was solved. At an exhibition, I had one chip loose it's settings and to keep a service running, I removed that power car. The other one moved it'self and the remaining six coaches without a problem.

I acquired a 2-EPB from Bachmann with a duff motor which I removed along with the gear train. It runs with another 2-EPB as a four car train and this makes me think, you may solve your problem with your MLV by removing it's motor. I run two unpowered MLV 's as trailing loads on my layout, one is a "Southern Pride" example and the other a "Replica Railways". 

[10000]

You can get them pretty close. I did my American H0 by first setting the bottom end of a speed curve to ensure they all moved off on step 1 of 128. Then using a digital stop watch I timed them around a circuit and adjusted the top of the speed curve to set them all to the same top speed, (OK I used Decoderpro to adjust them its easier than doing it by CV)

 

End result was the ability to run a consist of 2 at the front and 1 helper two thirds along with no derailment of freight cars.

 

Be sure to run every loco for a least 20 minutes on its own before adjusting anything because as they warm up they get faster.

[HLT 0109]

In my experience, satisfactory speed matching is only possible using CVs 67-94.  As Nconsistent has discovered, every loco/DMU responds differently to the throttle so you have to do the matching in detail rather than with the "broad brush" approach of CVs 2,5 and 6.  Using CVs 67-94, it matters not at all that the chips used are of the same version (though mine are all Loksound) and, whilst an approximation might be OK for double-heading, greater accuracy is essential for banking - especially if long curves are on the route.  Incidentally, I found that setting CVs 3 & 4 to zero is effective for turning off BEMF rather than setting it to "off" in CV 49 (or is it 29).  When I tried the latter with a Loksound v4, the loco wouldn't run at all.

 

Having said all that, achieving good results using just the NCE Powercab handset is extremely time consuming and frustrating - I have recently found it much easier in conjunction with with JMRI software as the two CV tables can be on screen at the same time (or easily switched between).  Once the 28 speed steps have been matched, it is beneficial to test each of the 128 steps as a small adjustment of some of the 28 might give further improvement.

 

I have for some time consisted a Bachmann 9F (Loksound v3.5) with a Hornby L1 (Loksound v4) but have today speed matched those locos with a Bachmann 2MT tank (split chassis with Loksound v3.5).  Whilst I know that the first two work well together on my long, steep, curving incline (one front one back) I have yet to try the 2MT in the mix in order to haul a longer train but I am not expecting any problems as the speed match is within a couple of inches over a distance of about 30 feet.

 

Harold.

 

Edit to update:  Completely successful, although the 2MT tank seems to be quicker when the locos are cold - so better no to use it as banker.  I was going to put it at the front but found the front Kadee on the 9F had broken, so the 9F was at the front and the L1 was banker.  With sound enabled, the 2MT started to move first, so it was necessary to leave the throttle on speed step 1 until all three were just moving.

[Nconsistent]

If changing the MLV's  CV's 2,3,4and 6 is having no effect then are the CV 29's on both chips set the same? Is advance speed step activated on CV29 on the MLV and not the CEP and are they both using the same speed steps?

i have been dragged away from the trainset for a couple of weeks, but i will check this when i get back in the loft. .thanks to everyone for the the range of useful suggestions. my knowledge is expanding, just need my ability and confidence to follow.

[shibushe]

I was told once this method. Run Loco 1 over a set distance at the speed you are looking for  get Loco 2 to chase it then once they are running together you have your consist.

Not tried it myself, I await comment from out there will it work or not ????

[34theletterbetweenB&D]

If you want to make a consist with both locos running at much the same speed at any selected speed step, it isn't enough to be able to make them run much the same at one speed. One might be on speed step 10, the other on speed step 20. So then  it isn't possible to enter a speed step for the consist which will drive both locos at the same speed.

 

What is required is a close match at all speed steps you expect to be able to operate as a consist. That might mean a match across the whole speed step range, say for a double header, or if the consist is only going to be applied over low speed perhaps for banking or path saving groups of light engines, the first quarter or third of the range.