DCC booster station output.

T

teewye

Registered
Country flag
Hi,
I am a newbee to this site and have recently returned to railway modelling by building a layout in my backyard.
Having operated on analog dc up until now, I am currently considering moving to DCC control. However a friend of mine also having recently moved to DCC control is experiencing unexplained failures on different decoders fitted on a variety of locos.
The common factor here would seem to be the LGB MTS central station. The LGB transformer output checks OK but the question arises how to check the central station output (i.e the track voltage).
I am an electrical engineer with some experience of data transmission systems as applied to power system monitoring, but the complication here is the data signal is also the loco power. So some questions;
1. How is the loco powered when data is not being transmitted. Nosing around the web I find references to Bi-polar dc which is also referred to as ac.
2. Is this sinusoidal or an alternating square wave.
3. If I connect an oscilloscope to the track what should I expect to see?
4. What is the maximum peak to peak the Massoth decoder expects to see.
5. What is output from the decoder to power the motor. Variable rectified dc or PWM square wave?
 
Welcome o the Forum Tom!

What do you mean by 'unexplained failures'?

Fast, and dirty, you can use a DVM to check there are volts on the track.. Around 19V (allowing for it not being 50Hz AC you are measuring..
I am not 100% sure but would expect you would see 'sine-squared' pulses on an oscilloscope. - Pure square-waves would radiate RF interference from all the harmonics.

Small (short) engines can experience problems with dirty track / bad joints, especially across points.. This can cause the engine to stutter 9if there is enough speed/momentum, or even stop crossing a point.

An older MTS system (and/or) LGB loco's might only respond to 'serial' commands. - Can cause 'strange' results when you first see it.
Speed-step settings, if not matched, can cause lights to 'flash' depending on speed of the loco.

We need a bit more information, really.
 
Hullo and welcome to the forum.

To answer one of your questions, the track power is a modulated hi-frequency signal on a 18 V carrier. I've never 'scoped the voltage of my MTS Central Station, but if I can measure something on my DMM on the AC range near to 18-24 Volts then I'm happy. I have some red LEDs on some of my buffer stops which also indicate to me that the power is on the track, and my colour light signals are also track fed for their power.

I believe the data is constantly transmitted, although I may stand to be corrected. So if the MTS Central Station is commanding a loco's motor to have zero volts, i.e. stationary, it is constantly being told to do so. I also reckon that the decoder output to an LGB is dc, but again I'm not sure on this. I know an LGB loco motor will run on PWM (i.e. an analogue loco with no decoder), but a PWM supplied to the track (by say the Aristocraft Train Engineer set to PWM instead of DC) will fry an LGB/Massoth decoder.

Hope this helps....
 
I've never come across a decoder that outputs variable DC to the motor - they've all been PWM so far. It's probably the easiest way to control the output with programming, gives better low speed control and also allows for back-EMF sensing. Early decoders used quite low PWM frequency with a noticeable growl or whine, whereas modern decoders may go to 32Khz.

IIRC the command station doesn't have to constantly output the same speed instructions to the decoder - that would soon swamp the bandwidth as more decoders are added to the system. However there is data of some kind in the signal all the time. My NCE system has a setting somewhere which alters the regularity at which speed commands are repeated. I notice it with some decoders where sometimes when a loco has a power glitch and the lights go off; they come back on a second or two after the loco gets going again.

Not sure about whether PWM as input to a decoder would actually fry it (why, unless it actually peaks above the max the decoder can handle?) However the pulses are likely to confuse a decoder trying to look for and understand a DCC signal.
 
teewye said:
Hi,
I am a newbee to this site and have recently returned to railway modelling by building a layout in my backyard.
Having operated on analog dc up until now, I am currently considering moving to DCC control. However a friend of mine also having recently moved to DCC control is experiencing unexplained failures on different decoders fitted on a variety of locos.
The common factor here would seem to be the LGB MTS central station. The LGB transformer output checks OK but the question arises how to check the central station output (i.e the track voltage).
I am an electrical engineer with some experience of data transmission systems as applied to power system monitoring, but the complication here is the data signal is also the loco power. So some questions;
1. How is the loco powered when data is not being transmitted. Nosing around the web I find references to Bi-polar dc which is also referred to as ac.
2. Is this sinusoidal or an alternating square wave.
3. If I connect an oscilloscope to the track what should I expect to see?
4. What is the maximum peak to peak the Massoth decoder expects to see.
5. What is output from the decoder to power the motor. Variable rectified dc or PWM square wave?
Click to expand...

Hi there & welcome to the wonderful world of DCC. As you have more than average technical knowledge, you may wish to study the NMRA specifications for DCC data transmission and voltages to satisfy your interest. These are the "rules" on which all contemporary systems, including LGB MTS, are built.
The NMRA standards are here: http://www.nmra.org/standards/DCC/standards_rps/DCCStds.html

However, some simple answers to your questions are:
1) Once a DCC command has allocated a speed step (i.e. a chip output voltage) to a specific chip ID (loco address), the chip continues to take the AC input and convert it to pseudo-DC until it receives another valid command. It "listens" for commands at intervals specified in the recommended practices.
2) Square wave - hence an ordinary voltmeter cannot give an accurate AC reading.
3) A wave form as described in NMRA recommended practice 9.1 and annotated in RP 9.3.1 (how this helps you I'm not sure!)
4) For larger scales, decoders must tolerate peak voltages of 27 volts, though a central station should not put out more than 22 volts. (recommended practice 9.1). Decoders designed for smaller scales are only expected to tolerate 24 volts, but as this is still more than the permitted maximum for a central station in any scale, in theory it should not be an issue for voltage, though amperage demands may mean smaller-scale decoders can't cope.
5) As far as I recall it's all PWM, but check me on that.
 
:) Many thanks for the responses gents.

Just to expand on the history, the decoders were all new, dealer fitted and programmed. They would operate for an hour then fail -burnt out.

Asking around the consensus was that track voltage sgould read between 18 - 24 Volts on a DMM whereas we were reading 31.5 Volts.

Hence the search for data on Central station outputs.

I had checked the NMRA specs prior to posting on the blog and found in Section C the following;

The baseline method for providing the power to operate locomotives and accessories, which shall be
supported by all Digital Command Stations and Digital Decoders, is by full-wave rectification of the
bipolar NMRA digital signal within the Digital Decoder. In order to maintain power to the Digital
65 Decoders, gaps in bit transmission are only allowed at specified times (see S-9.2, Section C). The RMS
value of NMRA digital signal, measured at the track, shall not exceed by more than 2 volts the voltage
specified in standard S9 for the applicable scale. In no case should the peak amplitude of the command
control signal exceed +/- 22 volts. The minimum peak value of the NMRA digital signal needed to provide
power to the decoder shall be +/-7 volts measured at the track.
Reading this it suggests data is transmitted continuously which I find to be surprising.
I had expected as stated above that a carrier wave of some shape would be provided.
The NMRA spec is not specific in terms of how power to the track is provided. This seems to be down to individual manufacturers to implement, the requirement being that they must stay within the limits specified by the NMRA.
Hence the reason for attaching a scope to the track voltage to see if it was nasty spikes or whatever that was caising the failures.

All that said however, the cental station which was purchased as second hand has been returned to an LGB repairer for examination.

Again many thanks
 
The Bi-Polar DC on the track is what provides the power to the loco, accessory etc.

A data stream is encoded within the Bi-Polar DC as specified in standard S 9.1, and thus the data stream is always present.

Packets are placed into the data stream, packets contain messages such as 'loco x go at speed y'. Basic packets are defined in S 9.2. The standard says that packets should be sent as frequently as possible.
 
The voltage you were seeing on the track is/was to high, as you have discovered..
Unfortunate it has destroyed a number of decoders.
 
:)Many thanks for the replies

Am I correct therefore to say that the Bipolar dc shoud not exceed +/-24V to be compliant with the NMRA spec.

Whilst researching on the WEB I found (although I can't get back to it) a reference which said LGB work at a higher voltage. This is what gives them better performance over long layouts and dirty tracks.
I have written to LGB about the 31.5V but have had no response to date.

Enjoyed the videos.

regards
 
The DCC track voltage is 5,000 to 8,500 Hz (depending on how many “0s” and “1s” are in the packets) square wave A.C.

The electronics of a normal multimeter selected to A.C. and assumes that the A. C. voltage being measured is the mains voltage from our  home A.C. voltage supply which is 50 Hz  and the waveform is a sine wave. (DCC is a square wave).

The meter reading is converted with the factors taken into consideration. Reading A.C. voltages of any other frequency and waveform, these “correction factors” will  give an incorrect reading.

Check to see what your digital multimeter shows but don’t use this as an accurate measurement, of say, adjusting your DCC system voltage. Use it as a “guide” only and for comparison measurements.

Using an RMS Voltmeter will give a more accurate reading.
AC is usually measured with qualifiers such as "rms", "average", "peak", and "peak-to-peak."
        DCC is square wave AC.

As to how best measure the voltage, we convert DCC to DC, then the voltage can  be measured by a standard DC voltmeter, and simply referred to  as volts. Simple method as below, on our Club Layout there are 14 of these in place, one to each board and each one has a multimeter attached and sets of flying leads.

As to it`s accuracy, well against a true Rramp meter between 0.25 volts and 0.75 volts, there will always be a difference in the true reading, when you take into account where you actually take the reading from, the track (different metals in the production of the rails) DCC Bus Wire solid core, baseboard interlinks stranded wire, so on and so forth.

Took the view of an average, the lads know when working on the system if the meters show around 10.50 volts to 10.60 volts all is well, they no longer bother to add the " variance, shown in green " to give them a total. The " variance" is the voltage consumed by the circuit.

To verify the accuracy we also  take the Decoder Function Voltage measurement. Measure the function output (blue wire) of the decoder and the white or yellow wire (negative) and select the function on. This DC voltage will be about 0.5 to 1.0 volt less than the DCC track voltage. To date not one single decoder has been fried, or a single LED burnt out.

As to your particular problem, the simple basic tests as outlined above will give you a more accurate indication of what the "true" voltage is. If as you stated it is indeed in excess of 31 volts, remedial action is required, a diode string will enable you to bring the track voltage down to say 22/ 24 volts to within the guidelines.

Average Reading
Z1 1N4001
C1 4.7uF, 50 VDC
R1 10k, 1/4 W
 

Attachments

  • DCC to DC Measurements.gif
    DCC to DC Measurements.gif
    2.5 KB · Views: 45
  • Setup.gif
    Setup.gif
    25.9 KB · Views: 44
  • Diode String.gif
    Diode String.gif
    11.6 KB · Views: 47
If I remember correctly the voltage coming out of a central station is dependant on the AC voltage being put in from the transformer. Has the transformer voltage been checked. Should be about 20 volts.
 
You must log in or register to reply here.
Back
Top Bottom