Taking the pulse(s) of DCC?

[Isambard]

Having read the Digitrax Big Book of DCC and DCC Made Easy by Lionel Strang/Kalmbach, I'm still a bit foggy with respect to the relationship between the digital pulses for command and control, which seem to be well described, and the pulses for loco power, which do not appear to be described.

As I understand it, the loco decoder converts AC pulses into DC to drive the loco motor, while also decoding command and control pulses, in either case at a nominal 15 volts.

Does this mean that there is a stream of AC pulses for power at a certain pulse repetition frequency (PRF) interspersed with bursts of digital command and control pulses at higher PRF's? What does the track voltage look like when no change is being called for i.e. steady state with the loco standing still on a powered track or when running at fixed rate? Does the Command Station still send out bursts of data to that loco to maintain that state? What are typical PRF's in any case?

Is the DC power supplied to the loco motor derived from all pulses received by the decoder, whether for power or for command and control, or otherwise?

I've got my Digitrax system up and running but am interested in getting a better understanding of what's going on with those busy electrons.

[:)]

[rrinker]

The DCC signal is a square wave, not your typical AC sine wave. The width of those square wave pulses is what determines a '1' bit and a '0' bit, 0's being longer and 1's being shorter. The technical details are in the NMRA Standards and RP's which you can download from the NMRA web site.
Older command control systems operated liek you describe, a fixed voltage with a small amplitude pulse of command information superimposed on it. In DCC, the command data itself is also the carried pulse.

From the decoder in, to the loco motor, it is completely different. The motor drive uses a technique called pulse-width modulation - basically, the pusles are alla t the same amplitude, and the longer the pulse duration (wider pulse), the greater the average voltage the motor sees and the faster it spins. The reason for doing this is that switching, say 1.5 amps continuous with 2 amps peak (typical HO decoder specs) using pure DC calls for LARGE transistors to be able to dissipate the resulting heat. Check the size transistor used in the output of a basic transistorized DC power pack - x4 to make the 'reversing switch' to drive a loco motor in forward and reverse. You'd have a decoder that could only fit in a G scale loco but was only capable of driving an HO loco. Using PWM you can use smaller transistors because they are never continuously passing the full motor current because the power is never 'full on'. Early decoders used a relatively low pulse frequency, and this can cause motors to 'buzz' when running even with a decoder. Moder "high frequency" or "silent running" or "supersonic" decoders use a very high frequency (> 22kHz) to elminate this buzzing.


--Randy

[Isambard]

Thanks for the response Randy.

The NMRA site seems to be pretty busy and not accessible whenever I try to access.
Yes, I understand the square wave pulse width modulation approach, and no I wasn't referring to the old command systems where the control signal was superimposed on top of the power pulse. My question in part is whether the control and the power pulses are interspersed, if they are different, i.e. are they part of a single pulse stream and what does the stream look like with and without a specific change being called for.

From what you say, the motor receives square wave pulses from the decoder of varying width, dependent on the speed commanded. Lionel Strang/Kalmbach's explanation is that the decoder converts AC to DC and controls the voltage and polarity that travel through the motor -a bit of a strange description-voltage and polarity travelling, but never mind. Is their explanation a bit of a simplification?

[barrydraper]

The power and control pulses are one and the same. That is the genius of DCC. The command station controls and generates the pulse stream. It just sends continous zeros until you want to change a decoder (or accessory). Then it drops in the approiate ones and zeros for address and data which are read buy the decoders, which only act if the data is addressed to them. The only difference between a one and a zero on the track is the width of the pulse, and both ones and zeros serve to supply power.

[DigitalGriffin]

To add to this...
With the square pulse DCC, one rail is always on. So if one rail is up, another is down. Add them together and you get a constant voltage supply. The DCC decoder can then turn this into a full DC voltage supply for the electrnics section . A seperate DC current is then switched on and off at high speed from full on to full off to the motor to control it's speed.

Quite ingenious solution actually.

[rrinker]

QUOTE: Originally posted by Isambard

Thanks for the response Randy.

The NMRA site seems to be pretty busy and not accessible whenever I try to access.
Yes, I understand the square wave pulse width modulation approach, and no I wasn't referring to the old command systems where the control signal was superimposed on top of the power pulse. My question in part is whether the control and the power pulses are interspersed, if they are different, i.e. are they part of a single pulse stream and what does the stream look like with and without a specific change being called for.

From what you say, the motor receives square wave pulses from the decoder of varying width, dependent on the speed commanded. Lionel Strang/Kalmbach's explanation is that the decoder converts AC to DC and controls the voltage and polarity that travel through the motor -a bit of a strange description-voltage and polarity travelling, but never mind. Is their explanation a bit of a simplification?

Oh, the decoder does indeed convert the square wave DCC signal, which is both the command bitstream AND the propulsion voltage, to DC, but ONLY to power the decoder electronics. The PWM output ont he motor driver controls a 4-transistor array known as an H Bridge, if you look up a schematic of one you'll see why, it resembles a letter H. The H bridge switches tre track power (DCC square wave) to generate the motor pulses. Again, if the decoder recitifed the square wave to pure DC, you's need a 1.5 to 2 amp bridge rectifier, again not a small part. Because of the 4-transistor layout of the H bridge, if you want to move forward, both the positive going portion of the DCC waveform and the negative going part can provide proper polarity to the motor, so it's not like the motor is powered by a half-wave power source.
There a bit of explanation on Rob Paisley's site, at least about H bridges. http://home.cogeco.ca/~rpaisley4/HBridge.html

-Randy

[DigitalGriffin]

Right you are Randy. I went back and corrected my post about the rectifer section.

[Isambard]

Thanks to all for the explanations. Let's see if I've understood correctly:

The command station supplies a continuous stream of 15 V square wave pulses to the track, these pulses being either ones (narrow pulses) or zeros (wider pulses). If no change of locomotive state (or accessory) is called for by the throttle (or other control device), the command station sends a continuous stream of zeros.
If a change of state of a locomotive is called for, the command station sends the appropriate mix of ones and zeroes to that locomotive's address.
Within the locomotive decoder concerned the pulses are rectified to provide DC to operate the decoder itself. The decoder's PWM motor driver circuit controls the H bridge circuit that generates pulses to drive the locomotive motor, the pulse rate determining the motor speed-presumably zero pulses meaning zero speed and speed increasing with increasing pulse rates i.e. the motor is driven at a variable pulse rate (that is separate from the command station-provided pulse rate as seen at the track level). With modern decoders the pulse rate supplied to the motor can be greater than 22kHz

If the above is correct I'm left with two questions:
1) what is the frequency range of PRF's sent out by the command station - presumably the lowest rate is when it's sending a continuous stream of ones and zeros?
2) what is the voltage supplied by the decoder to the motor-presumably fixed?


Thanks for bearing with me folks.

[DigitalGriffin]

QUOTE: Originally posted by Isambard



If the above is correct I'm left with two questions:
1) what is the frequency range of PRF's sent out by the command station - presumably the lowest rate is when it's sending a continuous stream of ones and zeros?
2) what is the voltage supplied by the decoder to the motor-presumably fixed?


Thanks for bearing with me folks.

1) The NMRA DCC effective data rate to the track is approximately 4500 bits per second for 0 bits and 9600 bits/second for 1 bits.

2) Why would you need to restep the voltage? Altering the voltage would result in wasted energy in the form of heat.

Why do I have the feeling you're a big engineering geek like me? [:D]

[Isambard]

n

QUOTE: Originally posted by DigitalGriffin

QUOTE: Originally posted by Isambard



If the above is correct I'm left with two questions:
1) what is the frequency range of PRF's sent out by the command station - presumably the lowest rate is when it's sending a continuous stream of ones and zeros?
2) what is the voltage supplied by the decoder to the motor-presumably fixed?


Thanks for bearing with me folks.

1) The NMRA DCC effective data rate to the track is approximately 4500 bits per second for 0 bits and 9600 bits/second for 1 bits.

2) Why would you need to restep the voltage? Altering the voltage would result in wasted energy in the form of heat.

Why do I have the feeling you're a big engineering geek like me? [:D]

Whatever gave you that idea? See profile.
(B.Eng. '56 -before computers, transistors and almost before electrons)
[:)]

[rrinker]

And before I was born, too [:D]

Here's another tidbit. The NMRA standard allows for the 0 bit to be stretched out well past the minimum length. This is how those DCC systems that support it drive a DC (non-decoder) locomotive. The command station stretches out either the positive going or negative going half of the 0-bit waveform which makes the track 'more' one polarity or the other, which the DC motor will 'see' as a positive or negative DC voltage. See the explanation of "Zero Stretching" in the Standards and RP's. Keep trying - you SHOULD be able to get into the NMRA site and download all the relevant documents in PDF form. And should you REALLY have trouble falling asleep, just print them out and read them [:D]

--Randy

[Isambard]

QUOTE: Originally posted by rrinker

And before I was born, too [:D]

Here's another tidbit. The NMRA standard allows for the 0 bit to be stretched out well past the minimum length. This is how those DCC systems that support it drive a DC (non-decoder) locomotive. The command station stretches out either the positive going or negative going half of the 0-bit waveform which makes the track 'more' one polarity or the other, which the DC motor will 'see' as a positive or negative DC voltage. See the explanation of "Zero Stretching" in the Standards and RP's. Keep trying - you SHOULD be able to get into the NMRA site and download all the relevant documents in PDF form. And should you REALLY have trouble falling asleep, just print them out and read them [:D]

--Randy

Thanks for that tidbit Randy. It answered another question that came to me later.
Is this the same principle used for the motor drive pulses from a decoder in a DCC loco?
[?] [:)]

[rrinker]

Sort of, but not really. When driven fromt he decoder, the motor doesn't get the opposite polarity pulse that causes the buzzing, it only gets pulses in the direction it is supposed to turn. You can't eliminate the negative pulses from the DCC signal while powering DC locos because, well, then there would be no DCC signal.

--Randy

[Isambard]

Thanks again Randy.
I've got enough now to keep my brain buzzing and my locos not buzzing for a while.