Have you added any earth ground wires or metal tape in the troublesome areas?  See  http://www.trainfacts.com/trainfacts/?p=317 for more information on the earth ground component of the TMCC signal.  Capacitors and more track signal WILL NOT solve your problem!!  It may actually make it worse!

All locos and all antennas are not the same.  Some will be more or less sensitive to track signal. 

Yes Dale, I have earth ground in a number of locations

 

Can you please describe what you have done here?

 

Do you have a volt/ohm meter?  If so, measure the resistance between an outer rail that is fed by the Base (not by a capacitor) and the earth ground.  There should be an open circuit (infinite or very high ohms.)

 

Please confirm that you are running TMCC, and not Legacy.

That makes a lot of sense Marty. I have a few pairs of locos that are the same model and they behave the same way, some good some bad. 

 

Sounds like the choices are:

• "Tune" the track
• "Tune" the locos
• Don't buy any more of the problem models
• Get rid of the problem locos

I guess the universal solution would be to "tune" the track and try to get the bad ones to run right.

Can you please describe what you have done here?

 

In many locations with parallel tracks, I've strung a single bare wire between tracks and tied it to earth ground. In my helix I've used foil and screen tied to earth ground to shield the stacked tracks. However, the section of track where I have been testing these locos has no earth ground application. Looks like I'll have to add some.

 

Do you have a volt/ohm meter?  If so, measure the resistance between an outer rail that is fed by the Base (not by a capacitor) and the earth ground.  There should be an open circuit (infinite or very high ohms.)

 

I just checked that. Bad news, direct connection. I have no idea where or how to break it.

 

Please confirm that you are running TMCC, and not Legacy.

 

Just TMCC right now. I own a Legacy base, but it isn't connected yet. I tried it, but discovered a different problem. My power supplies and the base are on a single power switch and all locos take off at full speed. That suggests that there is no command signal in the combined TMCC and Legacy configuration. I may need to turn the power on to the base before applying power to the rails.

Upon further investigation, it seems that the house neutral and the safety ground are tied together at the main house panel. An earth ground stake and a water pipe bond are also tied to the house neutral.

 

Everything is tied to earth ground!

Let's work on the resistance between the outer rail and the earth ground that shouldn't be there:

1.  Presumably you have a power source connected to the track.  Please describe what you are using.

2.  You mentioned that you have isolated rails.  Do you have anything connected to them for accessories or signaling?  This would be a likely spot for a connection to earth ground.

3.  What value of resistance did you get between the rail common and earth ground?

 

With Gargraves track, you shouldn't have any problems with ties shorting to your foil, screen and ground wires.  What do you use for switch machines, and do you have them wired for anti-derailing?

1.  Presumably you have a power source connected to the track.  Please describe what you are using.

My power supplies are not traditional model railroad transformers. They consist of three 6.3V 20A transformers connected in series. I have three of these units in service, and three more to come online in the future. Each gang of three has a grounded power cord, though the ground is not connected to anything. There are fuses on the hot primary, and the secondary is tied to a terminal strip for distribution. Each block is protected by a 6A polyswitch.

2.  You mentioned that you have isolated rails.  Do you have anything connected to them for accessories or signaling?  This would be a likely spot for a connection to earth ground.

At this time there is nothing connected to the isolated rails. Eventually they will be used for detection, and tied to a computer with a system called CMRI, but that's way down the road. There are no signals or accessories on the layout.

3.  What value of resistance did you get between the rail common and earth ground?

Zero.

With Gargraves track, you shouldn't have any problems with ties shorting to your foil, screen and ground wires.  What do you use for switch machines, and do you have them wired for anti-derailing?

 

My switch machines are custom made screw drives. They have no electrical connection to the track, and no anti-derailing. They are connected to the switch with bell cranks and RC pushrods. None are operational.

 

The goal is to get this debugged before making it more complicated.

That's a good thing, since that's called for by the National Electrical Code!   I've never seen a main panel where the neutral and earth ground weren't common at the panel.

My next step would be to firmly attach the ohmmeter to the outer rail and an earth ground, and then start disconnecting earth ground wires and unplugging power plugs one at a time until you find something that affects the resistance value.

Thanks, I wired the entire house myself, from the meter down to the last outlet. The state inspector made me do it right. I did pull the main panel cover last night, just to refresh my memory.

 

Earth ground is universal, right down to the outside rail.

I unplugged the power supplies and even remembered to unplug the command base. When I said zero before, I was using the 10K scale on my meter. Even after unplugging, I was still getting that reading. When I dropped down to the 10 scale there is some resistance. The needle barely moves. 

 

What value am I looking for? I have a very large "haystack" to search for a needle.

You're saying Earth ground is universal, right down to the outside rail. This implies the earth ground is connected directly to the outside rail of the track. This cannot be right. If it is the TMCC signal is effectively shorted as the receiver picks up the difference in potential between the signal on the outside rail and the earth ground. You have to look for ANY direct connections between the track common and the earth ground and eleiminate them. This will give you an ohmmeter reading of infinity or very high on the scale.

Ron

Alright, I've disconnected the ground plane wires from the conduit, and inspected the entire length of the conduit for any stray common wire contact. No luck. Nothing is plugged in. Now every connection to the house is gone. Am I missing something obvious or looking for the ultra obscure? 

 

I recently had a short in one block. It took a while to figure out it was an intermittent short in an MTH Superliner car.

 

Update: I found something. A serial cable connecting my command base to my computer is a partial source of the grounding. Meter readings went from zero to around 500K. Still not to infinity, but moving in the right direction.

First off, any conduit wiring would seem to be far from this issue.  The issue is on the layout if the earth (entry panel) ground is shorted to track common.  Even if you short the neutral directly to earth ground at your outlet, that should not affect running trains.  Neither earth ground or the 110V neutral should be connected to the track common.

 

Keep looking around the layout, that's where the problem is.

you said "I've disconnected the ground plane wires from the conduit...."

 

you don't have the ground plane connected to the outer rail do you.  the ground plane should only be beside the track and connected to earth ground.

I mentioned the conduit because that is a perfect source of ground and it runs around the layout. Testing found no unintended connections. Disconnecting the existing ground plane wires from it showed no contact there either.

 

I'm going to start reassembling things, testing as I go. Hopefully meter readings will stay at current levels, and I'll keep looking for that last piece.

 

By definition, the primary and secondary of a transformer have no direct connection, so even though the house neutral is connected to earth ground on the primary, it can't pass through to the rail.

No, there was no deliberate connection, and accidental connections were ruled out when they were disconnected from the conduit.

 

Thanks to all who have lent their expertise on this. Still a work in progress.

please post what you find wrong.

A Command Base can probably drive a shunt of 500 ohms between track common and ground without much signal drop, but I would want to know exactly where and how the common and ground are connected together.

Dale, I'm not quite sure what you mean when you say, "A Command Base can probably drive a shunt of 500 ohms between track common and ground without much signal drop."

I did find the bulk of the problem when I disconnected that serial cable. Not sure what I'm going to do when I want that connection back for operation, but for now I can leave it off. I'm going to try and get downstairs tonight to look for that last small contact.

If your ohmmeter reads 500 ohms or more when connected to the outer rail and earth ground, you are probably not losing much of the track signal.

 

What devices are connected at the other end of your serial cable?

Well then maybe I can stop looking, because on the 10K scale it was reading 50. If my math is right, that's half a million ohms. Still not infinity where we'd like it, but if you say it's OK, I'll call off the dogs. 

 

That cable was attached directly to my computer. The end goal is to have the computer control some of the trains, but that's way off in the future.

 

I have to put everything back together. I'll watch that meter at each step of the process.

You can find opto-isolated serial couplers if you end up using the computer interface.

I wasn't aware of such an item. I'll have to look into that. Thanks.

I am having a problem visualizing why a computer would create a bridge between the outer rail and earth ground - unless something else from the layout is also connected to the computer.

 

The ground in the 9-pin serial connector on the Base carries through to the safety ground U pin on the Base's power supply.  That ground would also connect with the ground from the computer coming in on the serial cable from the computer.  If the computer is a desktop (rather than laptop), the computer's case should also be attached to the computer's U pin at the wall, and the serial cable ground should also be connected to the computer's frame.

 

None of this should bother the track signal coming out of the Base's U terminal.

One of a number that are available:  Port-Powered RS-232 Isolator, 9 Pin, 2 channel

Dale, all of those readings were taken with an old analog meter. I just bought a new digital meter. I'm going to retest the layout tonight and get a baseline. It should be well over 500 ohms.

 

I don't know if this makes a difference, but I am using a USB-serial adapter. Would that have anything to do with more ground getting through? Also, that serial cable was attached to the Y cable that came with the legacy base. Doubt that has any effect.

 

I did reconnect everything, except that serial cable. Some of the problem locos are a little better. The one I mentioned at the beginning of this topic, is still not going to make it around the loop. It only wants to "listen" to commands when I have my hand over it.

 

Thanks for the link John.

Have you tried connecting a wire to the ground pin of a three-prong outlet and laying it alongside the track at the problem spots?  It would be interesting to see if this is all you need.

Big Boy, I have seen, in some instances, where folks with ginormous complex layouts have signal reception issues with select diesels. The cause of this issue is the Railsounds Power Board. The locos giving you fits most likely has no grounded shield over the inductor on the power board. When the shield is missing it can create a high frequency signal that interferes with the 455khz command signal. This is a difficult thing to diagnose, because the loco will run flawlessly on 97% of the layouts out there but only act this way on your layout (sorry, your layout is in the 3% of "unique") The solution is to install an RS power board in the loco with a shield. They are in stock under product number 691RSPRA00. Mike Lionel

Mike, is it possible to replace that shield?  I did notice a couple of my RS boards without it, never gave it much thought before.

Thanks Mike, that makes the most sense, and totally explains why the service station says it's just fine, but when I get it home it doesn't work. I think "ginormous complex" is a fair description of my layout, and as construction continues, it will only get worse. No need to apologize for my layout being in that 3%. I like being different. I guess there is a price to be paid for that though.

The locos in question are mainly early 2000's vintage, and there could be a half dozen or more that need this fix. I'm guessing that newer models have the shielding. You wouldn't happen to know the time frame when the change was made? It would help me to avoid buying too many more that need this fix.

WOW!!! 

I ran to the hobby store and saw my old buddy Roundhouse Rick. Told him what "Dr" Mike had prescribed, and it wasn't his fault he couldn't find the problem. The test track just isn't complicated enough.  

A quick trip to the back room, and $40 later, the shielded board was in. Took it home and set in on the rails. What a difference! Like a new engine, runs great! Now I know what to do when this happens.

Thanks to all who have helped with this. Even though the issue ended up not being the track, I have gained a greater understanding of TMCC in the process. Given what I am trying to do, I need all the knowledge I can soak up.

Extra special thanks to Mike, who nailed it.

Well, if Mike couldn't get it, we'd all be in trouble!   It's good information for a lot of us I expect.

When Mike Reagan retires........... I'm out of this hobby.

Mike looks pretty young on the videos, I'm guessing I'll be out of the hobby before he retires.

On this page at the bottom: https://www.lionel.com/custome...;submitButton=Search