JMRI® connects to...
Digitrax LocoNet®
Supported Hardware
Devices, command stations, networks, and protocols:
Applications
By the community of JMRI.org:
Tools
JMRI tools for working with your layout:
Layout Automation
Use JMRI to automate parts of your layout and operations:

JMRI Help:

Contents Index
Glossary FAQ

Donate to JMRI Donate to JMRI.org

Hardware Support: Measurements on power leads of Digitrax components

I wanted to understand the reasons behind Digitrax's recommendations about separate power sources for PM4 and DS54 boards. This page describes some measurements on my layout, and what I conclude from them.

Conclusions

The DCS100 and PM4 appear to use full-wave rectifier power supplies, while the DS54 uses a half-wave rectifier supply. These two types require different offsets between system ground and the transformer leads, which explains why these three types of components can't share AC auxiliary power supplies.

Specific measurements

In this section, I describe the measurements leading to my conclusions.

The following measurements were made on an operating layout. It is powered by a single DCS100 (Chief) booster, with PM4 units for power subdistricts and DS54 units for various controls. Track power was on, and the DCS100 was in "run" mode, i.e. sending packets on the rails. About 1.2 A was being drawn from the DCS100 booster output. During the measurements, no LocoNetĀ® commands were sent, nor were reversing sections being traversed.

Unfortunately, the camera scope photos wasn't sufficiently sensitive, so they show a lot of blur. In all pictures, the ground reference was centered and the vertical scale was 5V/division. The scope was line triggered. Only a single trace is shown.

Booster measurements

Normally my DCS100 booster's ground connection is connected to the AC ground, hence to earth. I temporarily disconnected this to see what would happen with a floating layout. I found a 60Hz offset voltage of about 9V. A 100K resistor to ground removes this, so the source impedance is very large. The photo was made with a 10MHz bandwidth limit; the visible spikes are about 6V with no bandwidth limit (500MHz scope). The ungrounded layout was probably generating a lot of RF noise...

Booster ground signal without connection to household ground

For the rest of the measurements, I reconnected the DCS100 ground to the power ground. Both noise and offset were then less than 20mV.

With respect to system ground, the Rail A and Rail B lines alternate between 12V and about 0.8V:
RailA signal relative to system gound RailB signal relative to system gound
It appears that the DCS100 internal high-power supply is only positive. To get a better understanding of that, I looked at the two power in lines from the transformer:
power 1 power 2
Note they are both positive at all times, not centered on ground. The lowest voltage is about 1.3V below ground. This is characteristic of a full-wave rectifier supply with one side of the output grounded. The small offset is due to the voltage drop across the low-side diode, while the flat-top on the waveform occurs when the transformer reaches a large enough voltage to charge the filter capacitor.
full wave diagram

PM4 measurements

My PM4 power supply is an isolated transformer. I made no measurements with a DC auxiliary supply. The LocoNet was connected, providing a common ground connection.

The voltages on the PM4 power connections look like:
pm4 power 1 pm4 power 2
Like the booster, this is indicative of a full-wave rectified power supply. The load profile is very different from what's seen on the booster supply. The ugly waveforms may be because my PM4 supply is a cheap little Radio Shack 450mA transformer; it seems to be clearly saturating.

DS54 measurements

My DS54 power supply is an isolated transformer. I made no measurements with a DC auxiliary supply. The LocoNet was connected, providing a common ground connection.

The voltages on the DS54 "black" and "red" power connections look like:
ds54 power black ds54 power red
This is indicative of a half-wave rectified power supply. Note this is different from the PM4 and DCS100.
half wave diagram
Note the ripple in the ground (black) lead. Digitrax recommends that you not connect the DS54 ground lead to system ground; this offset shows why its a good idea to listen to that recommendation. The ground in the cartoon schematic is the DS54 black-wire "ground", but it is not the same as the system (LocoNet) ground.

Don Crano has pointed out that the DS54 is deriving a local ground from the rail connections so that the BD1 block current detector can use the DS54 ground connection. The following photos show that the offset from system ground changes slightly as the pulse-stretching for analog operation moves from one extreme to the other. (Note that these photos are at 1V/division, with system ground in the center.)

connected -99LocoNet connected, analog at -99 connected zeroLocoNet connected, analog at 0 connected +99LocoNet connected, analog at +99
disconn -99LocoNet disconnected, analog at -99 disconn zeroLocoNet disconnected, analog at 0 disconn +99LocoNet disconnected, analog at +99
People on the Digitrax mailing list have recommended that all DS54 units on a layout be wired to the rails and common power supply with the same polarity. (Mine are too.) The offsets visible above show why you need to do this; if two DS54 units were connected oppositely to either the common power or the rails, they would be unable to create a consistent ground reference. Bob Jacobsen

LocoNetĀ® is a registered trademark of Digitrax, Inc.

Back to the Digitrax Hardware main Help Page