Author Topic: A problem that's not a problem?  (Read 6860 times)

Dean

  • Conductor
  • ****
  • Posts: 211
A problem that's not a problem?
« on: January 26, 2018, 08:14:53 PM »
I just finished installing 5 LM-3 decoders in 5 Stewart/Kato F units. I usually run the engines unloaded in both directions for about 20-30 minutes varying the speed with no load. Then I attach a load to the engine ( 1 or 2 hopper cars filled with lead shot ) and run it in both directions for 20-30 minutes and varying the speed. I decided to make a 5 engine consist and do the break-in all at one time. When it came time to add a load, I added 3 hopper cars that were filled with lead shot, a gondola filled with lead, and a flatbed with a 1"x1/2"x3" lead block on it. The 5 engines pulled the cars around the track with no problems, even up the 2% grades.
I was watching the current of the lead engine, on an HC-2, when I noticed a voltage dip as the speed increased. I stopped the consist and the voltage came up to 13.8V, matching the readings I get on a true RMS meter. As I increased the speed the voltage started to drop. At full speed, the voltage was down to 12.8V. Full speed going up a hill the voltage dropped to 12.2V.
The track power comes from a regulated 5 Amp power supply that supplies 26V DC to an NCE booster.The track gets its power through 4 parallel #16 wires, two for plus and two for minus. ( yes, I know it's AC, I am trying to be clear. ) Current draw for the 5 engines is ~2 Amp.
It's not a problem but I don't understand why it's happening. I shouldn't have this voltage drop, right? 
Dean

Alan

  • Conductor
  • ****
  • Posts: 1073
    • LK&O Railroad
Re: A problem that's not a problem?
« Reply #1 on: January 26, 2018, 09:37:07 PM »
My thoughts:

  • You are correct. It is not a problem.
  • RMS meters are designed to measure sine waves. DCC is a square wave of varying frequency and pulse width. Consider the readings approximations.
  • Any electrical circuit with a load on it has voltage drop. That's where the waste heat comes from and why no device/wire/connection is 100% efficient.
  • The voltage regulation of the NCE booster may not be all that tight. It doesn't have to be for correct operation.
  • Depending upon their length and connection integrity, the 16AWG wires and interconnects may be introducing significant resistance. And people poke fun at me for using such heavy buss wire. I say Ha.
  • Lesson learned: On a break-in loop it really doesn't matter. On you permanent layout don't take shortcuts. Wire it correctly.
Alan

LK&O Railroad website

When I was a kid... no wait, I still do that. HO, 28x32, double deck, 1969, RailPro

Dean

  • Conductor
  • ****
  • Posts: 211
Re: A problem that's not a problem?
« Reply #2 on: January 30, 2018, 10:29:08 AM »
I have had my VOM for a couple of years. I went back and checked, it does measure square wave voltages. That's why I bought it.
Two #16 THHN wires in parallel have a capacity of 36 Amps, more than a single #12 THHN. A few years ago I picked up a 500' roll of red and a 500' roll of blue for $20.00 each. For me, it was much easier to run under the layout and to twist the two colors together to try to eliminate crosstalk. I use brass shorting bars for connections.
Dean

Alan

  • Conductor
  • ****
  • Posts: 1073
    • LK&O Railroad
Re: A problem that's not a problem?
« Reply #3 on: January 30, 2018, 09:10:48 PM »
I have had my VOM for a couple of years. I went back and checked, it does measure square wave voltages. That's why I bought it.

I didn't say your meter can't measure square waves. I said it can't measure variable frequency square waves accurately. RMS meters (except ultra expensive heat method meters) measure square waves accurately only when the square wave has a 50% duty cycle which DCC does not. The meter measurement accuracy further degrades as the frequency of the measured wave nears or exceeds the meter's sampling frequency. The frequency of a DCC wave is in constant flux due to the pulse information being conveyed. It can and does change during your meter's sample period. Add to all this, most RMS meters AC couple the signal which causes further inaccuracy. Long story short, a scope is the only truly accurate way to measure power (voltage under load) of a variable frequency square wave.  Here is a video that explains a lot although it stops just shy of getting into variable frequency PWM. https://www.youtube.com/watch?v=ue0wtlrmCJE
Nevertheless, the inaccuracy in your meter is the same during both measurements so it is still a valid indicator there is a voltage drop.

Two #16 THHN wires in parallel have a capacity of 36 Amps, more than a single #12 THHN. A few years ago I picked up a 500' roll of red and a 500' roll of blue for $20.00 each. For me, it was much easier to run under the layout and to twist the two colors together to try to eliminate crosstalk. I use brass shorting bars for connections.

Point of clarification. Ampacity ratings are given based on the ability of the conductor and insulation to manage heat generation. As an example note the difference in ampacity ratings for the same AWG of THWN vs THWN-2. Ampacity ratings themselves do not take into account run length and resulting voltage drop. They are simply maximums for safe operation.

Since you reference 12AWG while in my reply I did not, I am going to assume your read my blog. When I wrote "wire your layout correctly" I did not say use 12AWG. By wire correctly I mean use a wire of sufficient gauge such that it will not exceed your personally set limitation of maximum allowable voltage drop under the maximum load at the furthest point from the power source. That just happened to be 12AWG in my particular layout situation.

If we assume your 16AWG wires are not exceptionally long and we assume your termination methods are of adequately low resistance then the evidence points toward the DCC booster or power supply. It could be sloppy voltage regulation in either component, a result of the DCC signal waveform being broadcast, or a combination of both.

Want predictable voltage levels? Ditch the DCC stuff and use a high quality switch mode power supply preferably one with remote sensing. Or, quit measuring and enjoy running trains!  ;D
Alan

LK&O Railroad website

When I was a kid... no wait, I still do that. HO, 28x32, double deck, 1969, RailPro

Dean

  • Conductor
  • ****
  • Posts: 211
Re: A problem that's not a problem?
« Reply #4 on: February 01, 2018, 07:30:00 AM »
I never read the fine print.   :-[
I would really like to ditch the DCC altogether, but I can't afford that.
Dean