Best Of New Brass EP-2 assembly & Mod clinic

[u18b]

Given the dual motor situation, would it be better to go with two decoders or just one? I guess with two you'd have a shot at speed-matching the two cabs if it turns out that the motors don't run quite the same.

Cheers,
-Mark

Mark,

I'm not quite at DCC yet, but I have given it a lot of thought and planning.

The EP-2 presents a curious case study for DCC.

In actual reality, we are dealing with two locomotives.  Each Cab is a self-contained locomotive.  The boiler room just drawbars them together.

Sooooooo....... I have two options.

1.  Use one decoder.  The amperage rating is probably sufficient to run two locos off of it.

But, there would need to be SEVEN (well, maybe six) wires that would need to run between the two locos.  Yikes, that's a lot.

But more importantly is the problem of performance.

I have discovered that the absolute critical aspect of running an EP-2 (this new one, or the old 1983 original) is that the two Cabs MUST run identically.  If one runs faster than another, performance degrades.  And that degradation can be so bad, that the full loco can barely move itself.

Now in analog mode, I need to try and get the two Cabs to match as closely as possible.  With one decoder, I would have to do the same.

2.  But if I install TWO decoders, then the game changes a bit.

First, there will be no huge bundle of wires running between Cabs.

Second, I would have the option of controlling each cab separately.
So if I have just done EVERYTHING I know to do and simply cannot get the two cabs to run well as a matched set, then I could program unique speed tables for each Cab and GET them to run the same.

Treating them as totally unique locos, then I would run the "EP-2" as a consist.

The other option is that if they run OK together, then I could just place both Cabs on the programming track and program them as if they were the same loco.
So if the address is 50, for example, the BOTH cabs with be programmed identically with 50.

With two decoders, I obviously don't need all the LED wires.  So the FRONT decoder will only need the blue and white wires for lights, and the rear decoder will only need the blue and yellow.

Hope that helps.

[nkalanaga]

Since you've already done that "junction box", this is probably no help, but I'll suggest it anyway.  Instead of twisting the prongs together, you could take a piece of very fine copper wire, wrap it around both prongs several times, then solder that.  Place it near the connector and you can cut off most of the prongs, making a smaller connector.  The whole thing can be insulated with "liquid insulation", "liquid rubber", or a good thick coat of paint, as there's little danger of the short stub being bent.

[peteski]

The whole thing can be insulated with "liquid insulation", "liquid rubber", or a good thick coat of paint, as there's little danger of the short stub being bent.

I use 5-minute epoxy as insulator/strain-relief on my home-brewed connectors.  It works quite well for both functions.  I apply it as it starts to harden (so it doesn't just drip off the connector).

[robert3985]

Although it's a worthy goal, I will bet that both engines do not run "exactly" the same no matter how much you attempt to make them the same.  When I consist my ABB Kato F's (they're pretty precisely made wouldn't ya say?), no two run exactly the same, and I have to adjust the CV's accordingly.

Two identical decoders is the way to go IMO, but don't give them the same address.  Consist them so you can adjust the CV's to get them to run as exactly the same as possible.

How are you going to be able to tell if they're physically connected to each other?  You disconnect them from each other, run them on parallel tracks, and when you've got them running the same, re-connect them.  That should do the trick.

Oh...remember to write down your CV settings for each engine.

Interesting project and really nice work!

[peteski]

Two identical decoders is the way to go IMO, but don't give them the same address.  Consist them so you can adjust the CV's to get them to run as exactly the same as possible.

Since each decoder holds its own speed tables, I don't see anything wrong with assigning both decoders to the same address (once they are speed-matched).  But while matching the speed, then each need a separate address.

I also agree that no matter how much you try, you'll never be able to speed-match even 2 identical models.  I also don't see anything wrong (operationally) with running both motors from a single decoder.

There is a very simple solution for this problem. I have a Roco double-unit German diesel loco.  Each unit has a motor. Solution: connect both motors with a shaft (with universals of course). That way the motors share the load, and are perfectly speed-matched at all times. 

[u18b]

All good suggestions.

Now, I don't do a whole lot of consisting on my layout and thus have not fiddled with speed tables before.

So Peteski, if I understand you correctly,  you're saying that I should give the two Cabs, say addresses as 5 (say for loco number E5) and 100.  Program individual speed table to get them to run the same.

Then take #100 and (by itself on the programming track) change its address to 5.

Now I have two locos with the address of 5---- but they have different speed tables.

Have I got it?

[VonRyan]

All good suggestions.

Now, I don't do a whole lot of consisting on my layout and thus have not fiddled with speed tables before.

So Peteski, if I understand you correctly,  you're saying that I should give the two Cabs, say addresses as 5 (say for loco number E5) and 100.  Program individual speed table to get them to run the same.

Then take #100 and (by itself on the programming track) change its address to 5.

Now I have two locos with the address of 5---- but they have different speed tables.

Have I got it?

You got it. It's called "simple consisting".
And to make speed-matching even easier, I suggest getting a Sprog and have your programming track be an oval of unitrack. Makes the process a whole lot easier.

[u18b]

Time to get this thing totally running in analog mode.

When I first drew up the idea, I was thinking I'd need to sacrifice a male plug and jumper the red/orange  and black/gray  wires to run in analog mode.

But then I thought-- heck, that is wasting good money.  Why not just jumper the socket itself if all I'm doing is tuning and testing.

So start with a long piece of stripped decoder wire.



Gently twist it.



Put a little flux on it and tin it with solder.


Make a short loop and cut the loop off.


 Bend the curved end up a little so it does not touch the charged heat shield.  And insert into the socket.  Fits fine.



Here, both are inserted.  As you can see, I'm jumpering the black and gray ...... and then the red and orange wires.   You can also see that I have inserted the tip of the red wire from the pilot truck into the junction box.



Ready to tune and test run.

[mmagliaro]

I am having a hard time understanding the reason to consider 2 decoders.
You can speed match all you want, and there will still be points in the curve where those wheel speeds will not
be identical, and will they really be any more "identical" than they already are, given the manufacturing
tolerances of two of the same motor?

In straight DC mode, those two motors are going to run independently.   We are assuming that they
will run close enough in speed that the engine should run well. 

I don't see how running the two motors off a single decoder is any worse or more difficult than running
the two motors on straight DC.

I also don't see how that's any different than
lashing up, say, two independent F unit diesels on a train.  Their geartrains and motors are always assumed
to be similar enough that they run okay together.  People have done that on straight DC for ages and it works.

Why is this different?

[peteski]

Have I got it?

You got it!

However I'm with Max. I don't see a need for 2 decoders in this situation. Both mechanism will ready be pretty much fine tuned in DC model.  I would use the KISS principal (which to me is minimal amount of electronics and tweaking the electronics).

I have been running multiple loco lash-ups for ages (lots of them in in DC days) on N Trak layouts.  They ran fine, even with sightly mismatched speed.  They pulled some long trains without problems.  I thing this quest for speed-matching perfection might be bordering on obsessive, in the end, without much gain.

[robert3985]

You got it!

However I'm with Max. I don't see a need for 2 decoders in this situation. Both mechanism will ready be pretty much fine tuned in DC model.  I would use the KISS principal (which to me is minimal amount of electronics and tweaking the electronics).

I have been running multiple loco lash-ups for ages (lots of them in in DC days) on N Trak layouts.  They ran fine, even with sightly mismatched speed.  They pulled some long trains without problems.  I thing this quest for speed-matching perfection might be bordering on obsessive, in the end, without much gain.

Y'know, Peteski and Max are correct.  I've run multiple units many many times on DC, totally non-speed matched and although one might be grinding away, or another might be being pulled along, they DID work together and pull that long train. 

That said, probably two decoders is overdoing it.  However, if you want each engine to be as perfectly matched as possible without a differential, I'm still for the two decoders...speed matching them first, then consisting them however you choose.  This allows for compensation later on when this locomotive is broken in, and the running characteristics of each engine have changed (which WILL happen). 

I'll be interested in what you decide to do.  Very interesting thread here.

[u18b]

I appreciate all the people following this thread.
Also the compliments, discussion, questions, and suggestions.

I did have an original plan to wire this thing.  (not going in blind).
But that does not mean the plan cannot change.

Here are a couple of my concerns.
1.  I'm planning on cutting a hole in the shell (yikes!) in the vestibule so that I can run jumper wires through the boiler room in the middle.  That will provide ALL axles conductivity to BOTH motors for maximum electrical pickup (without going to the trouble of adding pickups to the insulated wheels).

So with both Cabs tied into each other, the whole loco should be pretty electrically reliable.

2.  In my current plan, I only need two wires for this jumper- red and black.  That is all analog needs.  And if I add two decoders, that's still all I need.

3.  But if I go with one decoder, then I will need to jumper SIX wires.  Red and black as before for best pickup.  Then the orange and gray will need to go from the cab with the decoder to the motor in the other cab.  Also, the blue wire and the yellow wire (assuming the decoder is in the front of the whole EP-2-- so the white wire will be right there).

So that bundle is a lot bigger.

Also, while TCS does make a 6-wire plug and socket in the micro size...
http://tcsdcc.com/public_html/Zen/images/6-Pin%20Colored.jpg

I bought one for experimentation.  And while I like it, plugging in 6 pins is a whole lot tougher than 2.  It will put a lot more wear and tear on the bundle.

4. Also, (we are jumping a little ahead here), but there is not a lot of space in the shell.

With my original setup, only one wire winds its way along the roof of the shell (a red wire).

But with one decoder, I'd have to wind 5 wires (the black is soldered to the shell).  This may not be a problem, but then again, it might.  Remember, this kit has the extra weights that fit in the little cabs and a bit forward-- which is right over the gearbox for maximum traction.

5.  And lastly, I've already barely touched on this.  Some earlier experience would say that:
A.  Getting these to run at pretty close to the same exact speed is crucial for maximum traction and minimum noise.
B.  It is quite a challenge to make that happen.

Remember, it is still a ways off, but I do have some extra motors to experiment with.  But the drive train must be slightly redesigned again to get them to fit.  It will be a while before I get to where I will try it.

So all of this is to explain why I was leaning toward 2 decoders.

But I will definitely explore using one.

[wcfn100]

I can understand going with two decoders to reduce the wiring.  I went with three decoders for my Aerotrain.


Jason

[u18b]

For the sake of visualization, here is some drawings I worked out with my original plan.
This is the analog version.
I still have to do some drawing for the DCC version.
My apologies ahead of time for the size.  If it is a problem, I'll only use a link.

[u18b]

Y'know, Peteski and Max are correct.  I've run multiple units many many times on DC, totally non-speed matched and although one might be grinding away, or another might be being pulled along, they DID work together and pull that long train. 

Bob, your observation here is crucial.

If these two parts of the EP-2 are not balanced as equally as possible, some grinding (or maybe crackling) will occur.

I've been able to get one running B+ quality or maybe A-, but I confess I'm shooting for A+.

Call me foolish, but I am actually aiming for this to be one of the very best running brass locos around.

And as a side note- there is no point in a sound decoder because the EP-2 electrics were actually famous for how QUIET they were.  A sound decoder would ONLY add a horn.

So given that historical backdrop, I really want these to be quiet.