I just got some more batteries and fuses for my multimeter. I will double check the voltage all around the rails, and coming out of the booster and psu and check to see if it drops at all.

Just measuring the voltage around the layouts won't tell you anything.
You need to draw current, preferably close to the 10 amps.
No current, no voltage drop - Ohms law, E=IR, if I=0 then E (voltage drop) will be zero no matter how high the resistance is.

What's the model number of the Radio Shack meter?

Knut

 

Greg,
I do understand the wisdom in to using too many amps and I agree with it. I have a tendancy to want to make things simple... ie one 20 amps supply is easier to hook up then 2 10 amp supplies. That said I don't want to melt my trains either.

I think the best option is to invest in a second PSU and a second booster. I considered buying a one bigger 20 amp psu and using two boosters but I have heard there can be issues with that sort of set up.

krs,
I couldn't tell you how a maximum of 18 VAC relate to 32 VDC. That is what the the unit is rated for. That is to say I can use up to an 18V AC psu or up to a 32 V DC psu to power the system. I am not sure what you mean by asking how one relates to the other.

For the voltage all I can say is I have not observed any voltage drops and I have checked. And if the voltage is dropping its not doing so durring operating conditions at or below what the system is rated to.

I only get poor performance when I push the system above and beyond what it is rated for by drawing too much amperage. That is what leads me to believe it is a lack of current, not voltage causing the issue.

I just got some more batteries and fuses for my multimeter. I will double check the voltage all around the rails, and coming out of the booster and psu and check to see if it drops at all.

-Pete

 

Peter -

I'm having a problem with some of the numbers you posted:

The front of my unit says "Max input. 18 V AC/ 32 V DC". I believe my manual says that too.

How does a maximum of 18 VAC relate to 32 VDC?
If the maximum is a pure voltage constraint (not power as well), an 18 VAC maximum should relate to a 25 VDC maximum (the peak of 18 VAC)

I also don't understand why you think this is a current (lack of amps) issue rather than a voltage issue.
If the NCE unit can't deliver 12 amps which is what you are thinking, what do you think happens to the DCC voltage at that point?
You posted that Larry mentioned that the NCE unit does not limit the current when it gets too hot, so it either delivers 12 amps (or whatever is required by the trains) or shuts down and delivers no current - there is no in-between.

 

On the previous boosters, 27 volts was the maximum... I've had a number of friends contact NCE about getting their systems modified to output 24 volts (requiring at least 27 volts in), so maybe they changed the design a bit... will have to call Larry. That 27 volts is specified somewhere in the manual, but it was hard to find as I remember.

Peter, you can get bigger boosters, but when you start putting over 10 amps on the rails (no matter what control system) a short circuit can really melt stuff down. At that point, I think it's wiser (again no matter what control system) to break it down to separate 10 amp power districts.

If you have ever seen what 10 amps can do to a loco that has derailed, you should try 20-25!

At some point it just makes sense to go no higher... Just like your house, mostly 15 and 20 amp circuits, not a 100 amp circuit to run all the rooms... there wisdom there...

Regards, Greg

 

One thing I should clarify is I only get slow downs when I run over 4 locomotives. In this case 6 (I haven't tried 5). I ran run 1, 2 , 3 ,or 4 locos all day long with no issues. That is the other thing that makes me think its a lack of amps more then voltage issue.

Krs,
Yes I did tell Larry that I was using 27VDC 13 amps in. The front of my unit says "Max input. 18 V AC/ 32 V DC". I believe my manual says that too. Larry confirmed the unit can handle it. I know previous versions of the booster did not allow for this from the factory and had to be modified.

As for my multimeter is is a Radio shack brand, I will have to dig out the model number. I purchased it because I used the same one while I worked in a music studio for a short time. It was very accurate and versitile. We had all sorts of voltages and signals and it worked great.

Greg,
I see your point about extra voltage being wasted. I will try turning it down this weekend and see how things work.

If money was no issue, I would invest in the Gwire system tomorrw. Put a big 22 V 20 amp PSU directly to the rails and run the QSI decoders with the reviever upgrade. However I can't afford $250+ per loco to use DCC. I use digitrax decoders because they are decent quality and only about $55 a pop.

 

Thanks for the information Peter, sounds like you are clearing out the issues.

Yes, measuring the voltage inside the booster will give you a good estimate of the DCC voltage to the rails, although this is the first time I have heard that the rails are a volt less. Larry is the guy, though, so what he says is normally 100% right.

I completely concur with your higher voltage, as you probably know from my web site, I had to determine the 3 volt drop from DC input to DCC output, and also how to get my unit modified to go over 20.1 volts output. I can get pretty close to 24 volts to the rails.

But if your system is not modified, you will most likely only get 20 to 21 volts to the rails, and with a regulated power supply, only need 24 volts input. My point is that anything over 24 volts input will be wasted, and converted to heat.

Since you are running the unit close to it's maximum (and I mean it's maximum output to not overheat) then it's not wise to run more voltage. I've considered putting some heat sinks on the back of the case... the only cooling the booster gets is radiation from the case, there are NO heat sinks on the unit...

Yep, you need that RampMeter, and connect it to your booster output, you can see how many amps you are drawing. My locos on heavy grades and heavily loaded can pull 2.7 amps each. That is really close to wheel slip and 50 cars on a 3.4% grade.

The NCE system is (unlike other systems) conservatively rated, and, kept cool, will put out 10 amps all day without complaint, and as you have found, over that...

Regards, Greg

 

Posted By Peter Brayshaw on 01 Jul 2011 11:15 AM

One thing I was mistaken about was the draw of each loco. I always thought they pulled 1-1.5 amps each. I based this off of readings from the amp meter on my elite psu. Last night one thing I learned was that this amp meter is way off. With six locos at full throttle it reads 5-6 amps which is not correct. The locomotives are pulling more like 2 amps each, which is 12 amps which is above what the system is designed for. They pull more then 2 amps when they accelerate. I observed this by watching the multimeter spike as I turned up the throttle, then slowly come back down and level out after a few moments.

-Pete



Pete -

How about the make and model of the meter you are using to measure DCC voltage?
If it's not a "True RMS" meter with a frequency response out to 10KHz or better, it will not read the DCC voltage correctly.

 

Posted By Peter Brayshaw on 01 Jul 2011 11:15 AM
I just spoke to Larry at NCE. They told me that the unit will run at full voltage and full amps and does not limit itself. It can put out 21-22 VDCC at 10 amps continuously. He said it is actually capable of higher amperage (and sometimes voltage) but will shut down once the unit gets too hot.


>>> Pete - Did you tell Larry that you're powering the system with 27 Volts DC?

I have no doubt that the system will provide the full 10 amps at whatever the maximum design DCC voltage is, but the NCE manual states that for the 10 amp system one should use an 18 volt AC transformer and that the absolute maximum input voltage is 22 Volts AC.
From a peak voltage point of view, 22 volts AC will have a higher peak voltage than 27 volts DC, but from a power dissipation point of view it's a different story.
NCE doesn't specify that one can use DC as the input voltage to their Command Station but Greg does, so the CS can handle - question is up to what voltage difference between input and output.> Measuring the voltage with a DC meter on the inside of the NCE unit is fine. I still doubt that you have a meter than can accurately measure the DCC voltage on the track. NCE states in their manual that even a True RMS meter cannot measure the DCC voltage accurately, I was a bit surprised to read that - if the True RMS meter is spec't to the DCC frequency (some are only good to 400 HZ) then it can measure the DCC voltage accurately enough for all practical purposes.

Anyway, if some of your locos slow down when you add more locos on the track, the DCC voltage has to have dropped, no way around that.
Question is why and where - does the output of the NCE CS drop or are you getting large voltage drops in the wiring.> I'm not 100% sure I understand what you mean, but if you mean is running multiple 14 gauge leads from the Central Station to various parts of the layout vs running one 14 gauge lead around the layout with short 14 gauge feeders coming off it, that is definitely not the same at all.


....Boy, I'm having trouble to quote and then comment.


Let me look at the rest separately

 

I just spoke to Larry at NCE. They told me that the unit will run at full voltage and full amps and does not limit itself. It can put out 21-22 VDCC at 10 amps continuously. He said it is actually capable of higher amperage (and sometimes voltage) but will shut down once the unit gets too hot.

He also confirmed that my multimeter is measuring track votage correctly. I set my voltage by two points on the board inside the booster. This is a DC voltage. He said the track voltage will be about 1 volt less on the rails, which is what I measured. Plus the performance of the trains is proportional to the amount of voltage on the rails. That is to say with my old TE system i used to get about 18V to the rails. The trains now run at the same speed, but I have about a 3V loss through the decoder. Greg the reason I have my voltage set high is because at lower voltages the drop throgh the command station and decoder is enough to effect over all performance. Even on rollers on a test track where I know the wiring is in tact.

As far as the wiring goes running multiple leads or one lead with feeders coming off it is essentially the same thing. I confirmed this with NCE, Tonny's and a fellow ECE friend of mine.

The only thing i have not been able to accurately test is to see if there is a voltage drop as a train passes. Or if there is a voltage drop once things are moving. I know under normal conditions there is none. By normal I mean 2-4 locomotives. I have done lots of testing and never seen any significant drop. I should try and measure it while the system is maxed out.

This weekend/ tonight I will put my 6 locos on the rails and check voltage at the tracks as a train passes, at the PSU, and out of the booster.

I don't want anyone to think I am trying to argue. I really do appreciate the information you guys are giving me, and I am exploring your suggestions as much as possible as well as doing my own research behind the scenes.

One thing I was mistaken about was the draw of each loco. I always thought they pulled 1-1.5 amps each. I based this off of readings from the amp meter on my elite psu. Last night one thing I learned was that this amp meter is way off. With six locos at full throttle it reads 5-6 amps which is not correct. The locomotives are pulling more like 2 amps each, which is 12 amps which is above what the system is designed for. They pull more then 2 amps when they accelerate. I observed this by watching the multimeter spike as I turned up the throttle, then slowly come back down and level out after a few moments.

-Pete

 

EXACTLY what Knut said... ditto ditto ditto...

Your wire is not heavy enough, and your topology (wiring scheme) for delivering 10 amps to a train. His point about the voltage drop is very important, those amp ratings you are quoting are for 110v ac supply, not 20v ac.... too much voltage drop.

If it was me, I'd upgrade the wiring. You could also buy a second booster and second power supply and split the layout into 2 power districts, that's actually an inexpensive alternative. I thing the second booster is like $150 and the meanwell can be had for $70-80.

In my case, even with 10 gauge wire right up to the rail clamp, and direct individual runs from each feed point to the booster, I am setting up for THREE boosters, 1 for each main line and 1 for the switchyard.

Greg

 

Posted By Peter Brayshaw on 01 Jul 2011 01:17 AM
Greg,
My power supply is a 27 vdc 13 amp meanwell regulated supply. Do you think just adding a bigger PSU would fix the problem? It seems like I would need another booster as well. Also the wiring going to the tracks should be ok if I am correct. 14 Gauge wire is rated for up to 15 amps. And 12 Guage which is what I use is rated for up to 20 amps. I am not sure if I understand why multiple leads would be better then one lead with multiple connections.

-Pete
I'll throw my thoughts in.

The Meanwell supplies will provide the full 13 amps (actually a bit more) at the voltage you set it ro, so for a 10 amp DCC system which is what I think you have, that supply is fine.

But as Greg said, the voltage you have set on the Meanwell is too high, the NCE system probably limits its current output to prevent it from overheating.
Check the DCC voltage at the output of the NCE system with a very nominal load of an amp or two.
Then set the Meanwell voltage "x" volts higher, whatever NCE recommends, I think Greg said 3 volts higher.
So if your NCE DCC voltage is 22.0 VDCC you want the Mranwell output to be 25 VDC, not higher.
Actually, 22 VDCC sounds pretty high for the NCE system unless you had a special mod done to raise the DCC voltage.

It's pretty important to be able to measure DC and DCC accurately. If you don't get accurate measurements, you are just poking around in the dark wasting your time trying to find the root cause of the problem.
What make and model is your "good" multimeter? Maybe it reads true RMS up to the DCC frequencies, then it would give you pretty good DCC readings; a Rampmeter is a good investment if you have a larger DCC layout, especially one requiring a lot of power.

Anyway - I don't think there is much point continuing until you are sure you get reasonably accurate measurements .

As to the ratings of 14 and 12 gauge wire for 15 and 20 amps respectively.
Those are parameters established for North American residential and commercial wiring using 120 VAC; they don't apply directly in our case.
Simple example - say the length of the cable run gives a 12 volt drop at 15 amps - that's a 10% drop with a 120 volt line - you still get 108 VAC at the load end which is probably adequate to drive whatever is there.
But in our case, with a 22 VDCC source voltage you would only get 10 VDCC at the load end, ie loco decoder, which would be too low.

Now, I assume your runs are not long enough to cause those voltage drops, I'm just illustrating my point.


For our purposes, just look up the resistance of the length of cable you are using, remember there is a leg going out and a return leg, so get the total resistance and then just use Ohm's law to see what voltage drop you should get worst case, say at 10 amps.

But again, before you do that you have to have the means to be able to measure accurately.

Knut

 

Greg,
My power supply is a 27 vdc 13 amp meanwell regulated supply. Do you think just adding a bigger PSU would fix the problem? It seems like I would need another booster as well. Also the wiring going to the tracks should be ok if I am correct. 14 Gauge wire is rated for up to 15 amps. And 12 Guage which is what I use is rated for up to 20 amps. I am not sure if I understand why multiple leads would be better then one lead with multiple connections.

-Pete

 

Peter, you only need 3 volts more input than the output you get... any more differential will be expended in heat, and can eventually limit your output.

For 10 amps, your power delivery system is a bit light, you need independent runs from the booster to the track, not a big loop...

For reference, I run up to 10 amps AND I use 10 gauge wire feeders, each direct from the track to the booster.

By the way, after all the effort please get a RampMeter to measure volts and amps.

Greg

 

Another update. Looks like I figured it out.

I did some more testing tonight. First thing I did was hook up my Aristo elite power supply to the DCC system to see what it would show for amperage under full throttle. The trains ran much much slower with the elite hooked up to the system. Anyhow under full throttle the meter on the elite read about 5-6 amps. Maybe 7 here and there.

Then I wired up a multimeter in line between the DCC system and the elite. Just turning it on it was reading about 1.5 amps in line.

Now I wasn't using my regular multimeter as my good one was out of batteries. The one I had I borrowed from my father and I think it was only about 90% accurate. When I checked the voltage coming out of my elite it read 26 +VDC. The elite only is rated for 22. My other power supply it said was putting out 35-36 VDC or more. This is again above and beyond what the unit is capable of.

That said I am not sure if 1.5 amps with the system just turned on and no trains running was accurate or not. 1.5 seems like a bit much. Anyhow at full throttle under the elite I was getting a reading of about 13 amps.

Back to my other PSU and under full throttle with 6 locos I was getting 10-11 amps. This PSU is rated for 13 amps max.

So it looks like I need more amps if I intend on running multiple lash ups with many engines and long trains.

Anyone have any thoughts on how to remedy this issue? I hear zimo makes a 20 amp booster but I don't really want to change systems at the moment. The other systems that have caught my eye are the G wire system and Aristo's Revo. I like the idea of just hooking a generic PSU directly to the tracks. I also like the idea of having full DCC control on virtually any layout.

-Pete

 

I should add that the NCE manual gives instructions on how to measure/adust the track voltage by taking a reading from two points inside the the booster using a multimeter set to DC. When I measure at these points, and measure the AC voltage coming out of the unit I get almost the same reading, +/- 0.5 volts. My multi meter is digital and self adjusts to the voltage type and range. I am pretty sure its getting the track voltage right. If not its pretty close.

Also incase I did not mention I to have multiple leads supplying power to the rails. I have a continuous line of 12 guage lighting wire runruing under the tracks with feeds coming up every 20-30 feet or so. Under normal opperating conditions my trains run at a constant speed at all points on the layout. I only get slow downs when I run this many large locos.

Ultimately my goal is to be able to run up to 7, maybe 8 of these at once. I like the big 6 axle power and I like mainline railroading. Big consists, big locos, long trains. My friend and I joke around that we collect big locos to help pull all our other big locos around.

 

Update!
So I corrected some wiring issues on the layout itself, and replaced a few failed rail clamps where I noticed they had gone bad. So now the layout seems to be running like clockwork.

Then I decided it was time create a more permanent set up for my control systems.

I have two systems. DCC and the trackside TE. I cannon have both hooked up to the tracks at once as they will short each other out. So I have a 12 guage multi strand wire running all around the layout with leads running to the tracks everu 20 -30 ft or so. The end of this runs into my shed and thus far I have been connecting it to either control system via alagator clamps. Fearing that my shotty wiring might be contributing to the issies I am having I made a point to make a more permanent set up.

Home depot sells standard 120V 15-20 amp heavy duty plugs for extension cords and what not. They pop open with two scres and then they have 3 screw terminals inside for hot, neutral and ground. I purchased 1 male and 2 female plugs. For my track side TE I used a 3 ft strip of 12 guage wire. Attached one end to the screw terminals on the TE and the other I equiped with one of the femail plugs.

An important note is I am using the female plugs on on the the wires coming out of each control system. That way I can't mistakenly plug them into the wall and fry them.

For my DCC system I purchased some multi strand 14 gauge wire as the 12 guage was a bit too large for the screw terminals on the unit. The booster has two positive and two negative terminals for power in and two of each for power to the rails. I took advantage of this and ran two 14 guage leads (4 wires) from the power supply to the boster and two from the booster to the rails. Again the leads going to the rails were equiped with a female 3 prong plug which mates up with the lead going to the tracks.

When I looked at the wires I had been using I was a bit ashamed of myself. I had used 18 guage speaker wire, with a single set of leads for the DCC system, plus a set of alagator clamps to connect to the rails. Not sure what I was thinking.



So after all the wiring was done I did some testing. Same configureation as last time. Two consists totallying 6 disels in all. 4 SD45s and 2 dash 9s. This time I didn't have a train behind either consist however, just the engines. First off things just seem to run a bit better with more robust wiring. Each loco should pull about 1.5 amps. With 6 locos that is 9 amps all together so I should still be undet the limit. But when one consist speeds up the other still slows down. Although now the effect is much much less. Previously both consists at full throttle would max out at abotu 50% of their normal speed. Now at max throttle for both consists they are running at 75-80% of their max speed.

With both running at full throttle I hooked my multimeter to to the output terminals of the PSU and got 30.1-30.3 VDC. At the rails im getting about 21.5-22.0 VDCC. I was not able to measure the amps correctly as when I put the multimeter in line I with the sytem the trains would not move. I am going to have give this another shot durring the week.

One thing in particular that I noticed as I was closing up shopt for the night was both the PSU and the booster were very warm/hot. They were working pretty hard.

-Pete

 

I know you are not arguing... this is interesting stuff to talk about, and it's really probably the only place on the web these discussions take place where many people can read and participate.

Yeah, forgot that many AC scales are WAY high..

Most cheap meters do fine on DCC on the AC scale, in terms of relative measurements.

This is more fun than talking about the weather, and more stimulating!

Greg

 

Posted By Greg Elmassian on 14 Jun 2011 03:19 PM
Even an ordinary AC voltmeter can be used, just for relative measurements
Depends on the specific meter used.

The inexpensive ones only have one or two ranges for AC, I just checked two of the $20.- ones I have kicking around and one has a single 600 VAC range, the other a single 750 VAC range (DC volts, DC amps and Ohms have lots of ranges)

So trying to read a 3 or 4 volt difference on those AC ranges might be a problem.

They are also calibrated to measure a 50/60 Hz sinewave; I have no clue if these meters would even indicate anything useful at the DCC frequency - but I have no experience in that respect.

Those same inexpensive meters have a number of DC scales, all the way down to 200 millivolts, so measuring a few volts difference with the 20 or 50 VDC range would not be a problem and the DCC frequency is no longer an issue.


I'm not arguing with you Greg - my point is really that one has to know and understand what the particular meter one uses is capable of measuring and to use the proper procedure.


Trouble shooting a problem like that via web posts is tricky at best, not having reliable information just makes it so much harder if not impossible.


Knut

 

Even an ordinary AC voltmeter can be used, just for relative measurements, you use whatever reading you get right at the booster, and then see how much it drops (under load) at the "trouble spots"... so if you see 3 or 4 volts drop, even though it's not a DCC rms meter, you know where the problem is.

And the diode bridge makes it exact, since DCC is a square wave.

Sorry Knut, we'll agree to disagree some other time hahahahaha!

Just trying to help poor Pete, he has enough on his mind, so trying not to argue with Dan, but to give Pete the most straightforward, quickest and easiest way to debug the problem. I've seen people go around and around in circles measuring stuff that really does not lead to the root problem, but just to more confusion. There's no reason to connect a completely different DCC system just to measure current, my opinion.

Regards, Greg

 

Posted By Greg Elmassian on 14 Jun 2011 09:08 AM
Let me know if there's a flaw in the logic here......
Sorry Greg, but I have to agree with what you posted.


Everything you said in your last post is 100% correct.
If the train slows down by itself, it's a drop in voltage, not current.

If there is no true-RMS meter available to measure the DCC voltage, one can always use a diode bridge ahead of the meter and use the DC scale of the meter rather than buying a Rampmeter just forthese tests.


Knut

Oh....The reason I'm sorry that I have to agree is because I love to take an opposing position so that I can have an animated discussion - unfortunately not possible in this case.

 

I own the same system Dan. I and many others know how an NCE system works. We know what happens when you pull max current.

He's not pulling 10 amps if there are no defects with the locos. I've run more locos with more load on an NCE system, and I'm sure his grades don't exceed my 60' of 3.5% grade. I know what locos take full slip, believe me.

FYI, the NCE works just like the zimo, and everyone else's, they run to max rated wattage/current and then shut off at overcurrent... there is no tapering off of the boosters if the power supply is a regulated supply and can put out 10 amps minimum. (By the way, if you give it enough current, the NCE will actually supply up to 20 amps for a short time, but it cannot continuously dissipate that much heat)

You do say it must be a power loss, but maybe not in the wiring. I'm confused by what you are saying? Other than the booster, the wiring and the rails, there is nothing else.

I think I've been espousing that you look for power drop at the rails under load, and trace back, and it's either bad rail connectors, or insufficient wire gauge in the feeders. (or both of course)

Anyway, a simple application of a voltmeter to the booster output and then at points further away, under load, will identify the problem very quickly.

Measuring the amps alone really will tell you nothing unless you are actually out of current. I can tell you that is not so (out of current), since even with a short on one consist, there was power to allow the other consist to operate. (the Zimo also cannot put out as much current as the NCE so if you were already out of current with a 10 amp NCE, you would absolutely be overcurrent with an 8 amp Zimo).

Here's the crux, If you were out of current already, then added a short (as indicated) , there would be nothing "left" to run a train on. But he indicated the other train kept running.

That piece of information alone makes it pretty sure it is a power delivery problem between the booster and the loco. The resistance in the system was enough that the short did not take "all" the current in the system. Surely the short circuit drew MORE current than when both consists were moving.

Thus I cannot see how the system is out of current (from the booster), so I cannot see why measuring the current tells you anything really important as compared to the voltage drop, which MUST be occurring.


Let me know if there's a flaw in the logic here, but the pieces of information provided tell you a lot actually.

(This is actually fun, so we can see what was the real problem later... seeing how to fit the data to the "theory") I could be dead wrong....


Regards, Greg

 

How do systems really work? If I draw the max current available from a command station/power supply, what will happen?

Shutdown of one or the other?
Output starts pulsing due to approaching overload? Which unit, power pack or command station?

If current is not near the max then this would not matter, but Peter did let us know he is running 6 diesels with 4 motors each for 24 motors and a caboose with lights. Do the engines all have lights on and what about the smoke units.

Lots of questions, but one thing is certain. When the 4 diesels start up, the 2 diesels slow down. Must be a power loss, but maybe not in the wiring, so every piece must be checked in the power distribution. To assume one thing is bad is not the way to troubleshoot. One must verify each and every part of the power distribution.

 

Sounds like we have a little confussion here. We need to go right to the chase. Get you a ramp meter which will read the volts (DCC AC) and check what is coming out of your system. If the volts read 21 volts at the out put of the booster then you now can start tracing where the power loss from there. Later RJD

 

Dan,
I am about 25-35 minutes from Axel and Joane. I purchased a bunch of track and other supplies from them last summer and this. Actually the new guy there Mike in my buddy from hghschool.

So the Zimo system tells you how many amps are being used?

 

Dan, when I re-read Pete's first post, the first train has a lighted caboose and 2 logos, and indeed, the second train was 4 locos with an unspecified consist.

BUT re-read the part where one train stops and shorts out, and has a burnt smell.

I believe the key here is the other train kept running. I think that's a big clue to not supplying enough current, since it would seem very likely a second train PLUS a short (enough to start burning something) should have tripped a breaker.

Combine that with the other parts of the story of trains slowing strangely and it really sounds like a power delivery problem.

Regards, Greg