Lighting and other non-DCC devices.

[Chris King]

As a newbie to DCC, is the voltage used on this generally AC?
And is it about 18v? Or have I been living on another planet…

Anyway my point or question is, that with a normal train controller
(Select, Elite, GaugeMaster etc), there are certain things that need
still 12v DC supply, signals, maybe some lighting etc, what is the
general way that other model railway users power these auxiliary items,
I know that most controllers have an extra output for AUX or maybe even
12v dc, however if  you are building a fairly big layout or even one
with a lot of 12v DC demands, what is the best way to power these
without putting extra demands on your controller/controllers? Is there
or can you get a dedicated 12v DC controller/power supply over the
counter/bought that is designed or meant to be used as a separate power
unit for these sort of devices and loads? If so what make, what is the
best, and how have you address/resolved this on your layouts? 


--
Chris King
(remove nospam to email direct)

[Rich]

DCC is AC voltage with digital info. Those who do not understand DCC
will say it is not AC voltage.

http://www.loystoys.com/info/how-dcc-works.html

Many are around 12 to 14 VAC. Some higher for large than HO scale
which many run.

You can power many items on a layout if you prefer with DCC. There are
different amperage systems.
Many prefer to use DC to power accessories. Some use AC, some DC.

Running a DC loco on a DCC system. Two or three ddifferent systems can
do this.

http://members.shaw.ca/sask.rail/dcc/DCC-waveforms/DCC_waveforms.html

Lots of DCC info.

http://www.mrdccu.com/

A free eZine with a good set of forums.

http://model-railroad-hobbyist.com/

All kinds of DCC info.

http://www.wiringfordcc.com/wirefordcc_toc.htm

Do a lot of reading. Store the links in Favorites.

R

[Rich]

There is no best, only what you like best.
NCE, Digitrax, MRC, Lenz are a few of the DCC systems. There are about
six or more decoder companies. There are also two or three companies
that have a form of DCC and can run on DC layouts and DCC layouts.
You will have to do a lot of searching for different companies on the
Internet.
The digital control of trains is rapidly evolving.
Many DCC on board, sound and none sound can run on a DC power pack.
Just do not use a DC power pack that has pulse power option.
From you first request, you will have to do a lot of reading to catch
up with the latest in model railroading or you may go down a path you
will not care for.
Stuff in other countries I am not familiar with. Only the USA stuff.

R

[manat...@hotmail.com]

On Jan 29, 12:43 am, Rich <gagnonrc...@netscape.net> wrote:
> DCC is AC voltage with digital info. Those who do not understand DCC
> will say it is not AC voltage.

It can get quite heated at times :-)

> http://www.loystoys.com/info/how-dcc-works.html
>
> Many are around 12 to 14 VAC. Some higher for large than HO scale
> which many run.

Some systems can be adjusted, some you can adjust by using a different
voltage power supply.

If converting locos with filament bulbs designed for 12V it's worth
considering changing to LED as the higher voltage of some DCC systems
will reduce the bulb lifetime.

I would arrange seperate supply(ies) for accessories such as lineside
lighting and point motors. Keep the DCC for loco control.

MBQ

[Jeff]

On 29/01/2012 16:35, manat...@hotmail.com wrote:
> On Jan 29, 12:43 am, Rich<gagnonrc...@netscape.net> wrote:
>> DCC is AC voltage with digital info. Those who do not understand DCC
>> will say it is not AC voltage.
>
> It can get quite heated at times :-)
>

I suppose it depends on what you call an 'AC voltage'. It is not an AC
voltage in the sense of being like the mains or the auxiliary AC output
of a conventional controller ie a sine wave. Also is is not a control
signal superimposed upon another voltage. It is purely square-wave like
digital data sent with a high enough voltage and with sufficient current
capability so as when rectified it will run motors etc.. So it is an AC
signal in as much as it is a pulsed signal with varying mark space
ratio, but it is not a continuous sine wave like the mains.

The contentious bit is calling a pulsed signal AC, when does a pulsed
signal become AC rather than DC that is switched on and off, at what
pulse rate does that change occur. Is turning DC on and off once a
second AC? To some it is, depending on the application and over what
period you are observing the signal!!

Jeff

[manat...@hotmail.com]

On Jan 30, 7:45 am, Jeff <j...@jsystems.com> wrote:

> On 29/01/2012 16:35, manatba...@hotmail.com wrote:
>
> > On Jan 29, 12:43 am, Rich<gagnonrc...@netscape.net>  wrote:
> >> DCC is AC voltage with digital info. Those who do not understand DCC
> >> will say it is not AC voltage.
>
> > It can get quite heated at times :-)
>
> I suppose it depends on what you call an 'AC voltage'. It is not an AC
> voltage in the sense of being like the mains or the auxiliary AC output
> of a conventional controller ie a sine wave. Also is is not a control
> signal superimposed upon another voltage. It is purely square-wave like
> digital data sent with a high enough voltage and with sufficient current
> capability so as when rectified it will run motors etc.. So it is an AC
> signal in as much as it is a pulsed signal with varying mark space
> ratio, but it is not a continuous sine wave like the mains.

The definition of AC has nothing to do with frequency, waveshape or
having control superimposed on it.

> The contentious bit is calling a pulsed signal AC, when does a pulsed
> signal become AC rather than DC that is switched on and off, at what
> pulse rate does that change occur. Is turning DC on and off once a
> second AC? To some it is, depending on the application and over what
> period you are observing the signal!!

DCC is not a "pulsed signal" it is AC.

A decoder sees two connections to the track, lets call them A and B.
During one half of the cycle, A is more positive than B and current
flows from A to B. During the other half of the cycle, B is more
positive than A and current flows from B to A. The current flow
reverses twice every cycle. The current alternates. It is an
alternating current.

DCC is AC. It is not a "pulsed signal", nor is it "bipolar DC" as some
would have it.

Contrast this with a real pulsed DC controller where the pulses simply
turn the track on and off. The current only ever flows in one
direction (unless you flip the reversing switch!). That *is* a pulsed
DC signal and not AC [1].

I think the confusion arises due to people looking at the output of a
command station with reference to the local ground. If you measure the
two outputs A and B independently then they do look like pulsed
outputs. The point is that DCC is defined by the two track signals
alone, there is no ground reference. It is the differential signal
between the rails that is DCC and that is AC.

MBQ

Being really pedantic you could show that it is a combination of an AC
signal and a DC offset but that is taking things to a ridiculous
extreme.

[Jeff]

>
> DCC is not a "pulsed signal" it is AC.
>
> A decoder sees two connections to the track, lets call them A and B.
> During one half of the cycle, A is more positive than B and current
> flows from A to B. During the other half of the cycle, B is more
> positive than A and current flows from B to A. The current flow
> reverses twice every cycle. The current alternates. It is an
> alternating current.
>
> DCC is AC. It is not a "pulsed signal", nor is it "bipolar DC" as some
> would have it.
>
> Contrast this with a real pulsed DC controller where the pulses simply
> turn the track on and off. The current only ever flows in one
> direction (unless you flip the reversing switch!). That *is* a pulsed
> DC signal and not AC [1].
>
> I think the confusion arises due to people looking at the output of a
> command station with reference to the local ground. If you measure the
> two outputs A and B independently then they do look like pulsed
> outputs. The point is that DCC is defined by the two track signals
> alone, there is no ground reference. It is the differential signal
> between the rails that is DCC and that is AC.
>
> MBQ
>
> Being really pedantic you could show that it is a combination of an AC
> signal and a DC offset but that is taking things to a ridiculous
> extreme.

It is most certainly is a pulsed signal, with pulse width modulation
that carries the data. The fact that it is applied to the track in a
differential manner is not really relevant to that fact.

The fact that it is a pulsed signal is what causes the confusion.
Towards the limit if the pulse rate was low enough it would tend to
being called DC (if for example the state changed say only once a
minute), but because the pulse rate is high it would normally be
refereed to as AC.

If you viewed the spectrum of the DCC signal you would observe a classic
pulse envelope, but there again so would you if you observed a DC signal
being reversed every minute, if you viewed it over a long enough sample
time!!

Jeff

[manat...@hotmail.com]

DCC is *frequency* modulated.

> that carries the data. The fact that it is applied to the track in a
> differential manner is not really relevant to that fact.

The fact that it is applied differentially means it is AC as seen by
the decoder.

Are you sure you are not getting confused with the motor drive output
of the decoder, which is PWM and is DC, only reversing when the
direction of the loco changes.

MBQ

[Antti Louko]

On 2012-01-30 14:32, Jeff wrote:

> It is most certainly is a pulsed signal, with pulse width modulation
> that carries the data. The fact that it is applied to the track in a
> differential manner is not really relevant to that fact.

It most certainly is _not_ a pulsed signal.

From Wikipedia, definition of "pulse":
http://en.wikipedia.org/wiki/Pulse_%28signal_processing%29

1) A rapid, transient change in the amplitude of a signal from a
baseline value to a higher or lower value, followed by a rapid return to
the baseline value.
2) A rapid change in some characteristic of a signal, e.g., phase or
frequency, from a baseline value to a higher or lower value, followed by
a rapid return to the baseline value.

The basic feature of pulsed voltage is the existence of the baseline value.

In DCC, there is no baseline value. It has two possible values, +X volts
and -X volts, between of which it alternates. Hey, it is an alternating
voltage!

http://en.wikipedia.org/wiki/Alternating_current

[Jeff]

>>
>>> Being really pedantic you could show that it is a combination of an AC
>>> signal and a DC offset but that is taking things to a ridiculous
>>> extreme.
>>
>> It is most certainly is a pulsed signal, with pulse width modulation
>
> DCC is *frequency* modulated.
>
>> that carries the data. The fact that it is applied to the track in a
>> differential manner is not really relevant to that fact.
>
> The fact that it is applied differentially means it is AC as seen by
> the decoder.
>
> Are you sure you are not getting confused with the motor drive output
> of the decoder, which is PWM and is DC, only reversing when the
> direction of the loco changes.
>
> MBQ
>

I think it is you that is confused with the definitions of AC and DC. A
pulse modulated signal is AC, whether that be pulse width modulated or
otherwise modulated. DC is an unchanging constant voltage. Perhaps Pulse
width modulation is not quite the clearest term for the DCC signal as it
is really a pulse train that caries data, however, the pulses are of
different fixed widths for a one and a zero and so it is the width of
the pulse that encodes the information. There may only be 2 allowable
widths but that does not alter the fact that it is PWM. It is not
frequency modulated as such, although the spectral content for a string
of zeros would be half of that for a series of ones.

Any signal that varies with time is considered to be AC; mathematically
DC is nothing more than a special case where the time component is zero.
Even the PWM motor signal is AC as it is a pulse train and it is not
constant with time.

If you are being really pedantic even the output of a conventional DC
controller is really AC as soon as you make any change to the voltage!!

Jeff

[Jeff]

Of course there is a baseline, a DCC pulse consists of an excursion from
-X to +X and then back to -X, it is merely the length of the excursions
that determine the data. Another way of looking at it is that the source
0 to 5V data has been shifted from 0-5V to -X to +X. Just because you no
longer have 0V as a reference does not mean that there is no reference
at all.

>Hey, it is an alternating
> voltage!

Yes pulse is AC as it has a time varying component.

Jeff

[manat...@hotmail.com]

A decoder has no explicit reference. All it has is the differential
signal on the rails, which is AC. Generally the decoder will full wave
rectify the AC signal to give an on-board DC supply. It is then
creating it's own virtual reference point at the zero-crossing of the
input.

Some specialised DCC circuits half-wave rectify and use one rail as
reference but you need to be very careful if such modules are
connected together.

MBQ

[Jeff]

>
> A decoder has no explicit reference. All it has is the differential
> signal on the rails, which is AC. Generally the decoder will full wave
> rectify the AC signal to give an on-board DC supply. It is then
> creating it's own virtual reference point at the zero-crossing of the
> input.

I am not disputing that, but it does not get away from the fact that
the AC signal is Pulse width modulated, albeit that there are only 2
positions that the data can sit at. A pulsed signal IS an AC signal.

Jeff

[manat...@hotmail.com]

On Jan 30, 3:26 pm, Jeff <j...@jsystems.com> wrote:
> >>> Being really pedantic you could show that it is a combination of an AC
> >>> signal and a DC offset but that is taking things to a ridiculous
> >>> extreme.
>
> >> It is most certainly is a pulsed signal, with pulse width modulation
>
> > DCC is *frequency* modulated.
>
> >> that carries the data. The fact that it is applied to the track in a
> >> differential manner is not really relevant to that fact.
>
> > The fact that it is applied differentially means it is AC as seen by
> > the decoder.
>
> > Are you sure you are not getting confused with the motor drive output
> > of the decoder, which is PWM and is DC, only reversing when the
> > direction of the loco changes.
>
> > MBQ
>
> I think it is you that is confused with the definitions of AC and DC. A
> pulse modulated signal is AC, whether that be pulse width modulated or
> otherwise modulated. DC is an unchanging constant voltage. Perhaps Pulse
> width modulation is not quite the clearest term for the DCC signal as it
> is really a pulse train that caries data, however, the pulses are of

PWM implies a fixed frequency variable mark/space ratio.

DCC is fixed mark/space ratio (discounting stretched zeroes for
driving analogue locos) with variable frequency. A '1' is sent as a
single cycle of approx 8.6KHz nominal, a '0' as a single cycle of 5KHz
nominal.

> different fixed widths for a one and a zero and so it is the width of
> the pulse that encodes the information. There may only be 2 allowable
> widths but that does not alter the fact that it is PWM. It is not

The width of the pulse *and* the width of the idle period are critical
in decoding a DCC bit stream. Both are significant and it is the
overall period that matters. Hence DCC is frequency modulated, not
PWM.

> frequency modulated as such, although the spectral content for a string
> of zeros would be half of that for a series of ones.
>
> Any signal that varies with time is considered to be AC; mathematically
> DC is nothing more than a special case where the time component is zero.
> Even the PWM motor signal is AC as it is a pulse train and it is not
> constant with time.
>
> If you are being really pedantic even the output of a conventional DC
> controller is really AC as soon as you make any change to the voltage!!

This is getting into the real of AC with DC offset, which I did
mention in an earlier post.

MBQ

[Jeff]

>
> PWM implies a fixed frequency variable mark/space ratio.

No, there is no such constraint on PWM, DCC is PWM with 2 discrete
positions. PWM is AC.

>
> DCC is fixed mark/space ratio (discounting stretched zeroes for
> driving analogue locos) with variable frequency. A '1' is sent as a
> single cycle of approx 8.6KHz nominal, a '0' as a single cycle of 5KHz
> nominal.
>> different fixed widths for a one and a zero and so it is the width of
>> the pulse that encodes the information. There may only be 2 allowable
>> widths but that does not alter the fact that it is PWM. It is not
>
> The width of the pulse *and* the width of the idle period are critical
> in decoding a DCC bit stream. Both are significant and it is the
> overall period that matters. Hence DCC is frequency modulated, not
> PWM.

It is merely sent as RTZ with a fixed width for 1 and a different fixed
width for a '0'; that may equate to 5kHz (sic) for a '1' but it is not
frequency modulation. It would be FM if the data stream were used to
modulate a carrier.(Actually '1' is nearer 8.3kHz, 60us per bit 120us
total period).

Any data stream of this type will have frequency components that are
related to the bit widths, that actual spectral components of the DCC
waveform will not just contain 8.33kHz, and its harmonics, due to the
unequal bit periods there will be other components in there as well
depending on the data sent (particularly if there are long highs for non
DCC powering.

Jeff

[manat...@hotmail.com]

On Jan 30, 5:31 pm, Jeff <j...@jsystems.com> wrote:
> > PWM implies a fixed frequency variable mark/space ratio.
>
> No, there is no such constraint on PWM, DCC is PWM with 2 discrete
> positions. PWM is AC.
>
>
>
> > DCC is fixed mark/space ratio (discounting stretched zeroes for
> > driving analogue locos) with variable frequency. A '1' is sent as a
> > single cycle of approx 8.6KHz nominal, a '0' as a single cycle of 5KHz
> > nominal.
> >> different fixed widths for a one and a zero and so it is the width of
> >> the pulse that encodes the information. There may only be 2 allowable
> >> widths but that does not alter the fact that it is PWM. It is not
>
> > The width of the pulse *and* the width of the idle period are critical
> > in decoding a DCC bit stream. Both are significant and it is the
> > overall period that matters. Hence DCC is frequency modulated, not
> > PWM.
>
> It is merely sent as RTZ with a fixed width for 1 and a different fixed
> width for a '0'; that may equate to 5kHz (sic) for a '1' but it is not

No, the period changes since both the mark and space are changed.

> frequency modulation. It would be FM if the data stream were used to
> modulate a carrier.(Actually '1' is nearer 8.3kHz, 60us per bit 120us
> total period).

I don't know where you are getting you information, but I *was*
correct, the nominal bit time, for a '1' is 58us per *half-bit*, 116us
total per bit, or approx 8.6KHz. See http://www.nmra.org/standards/DCC/standards_rps/S-91-2004-07.pdf

Send a stream of 1 bits and you get a square wave at approx 8.6KHz,
send a stream of zeroes and you get 5KHz. That *is* frequency
modulation.

MBQ

[manat...@hotmail.com]

OK, I accept your agreement that DCC is AC, which you seemed to be
disputing.

Now you just have to understande that DCC is FM, not PWM but that's in
the other sub-thread.

MBQ

[None]

On Jan 30, 1:45 pm, "manatba...@hotmail.com" <manatba...@hotmail.com>
wrote:

Maybe it is single side band :>)

[Just zis Guy, you know?]

On Mon, 30 Jan 2012 15:26:04 +0000, Jeff <je...@jsystems.com> wrote:

>I think it is you that is confused with the definitions of AC and DC. A
>pulse modulated signal is AC, whether that be pulse width modulated or
>otherwise modulated.

No it's not. AC has alternating polarity, DC doesn't. Pulse width
modulated DC is still DC, as is DC with a superposed digital signal.

Guy
--
Guy Chapman, http://www.chapmancentral.co.uk
The usenet price promise: all opinions are guaranteed
to be worth at least what you paid for them.

[Jeff]

On 30/01/2012 21:08, Just zis Guy, you know? wrote:
> On Mon, 30 Jan 2012 15:26:04 +0000, Jeff<je...@jsystems.com> wrote:
>
>> I think it is you that is confused with the definitions of AC and DC. A
>> pulse modulated signal is AC, whether that be pulse width modulated or
>> otherwise modulated.
>
> No it's not. AC has alternating polarity, DC doesn't. Pulse width
> modulated DC is still DC, as is DC with a superposed digital signal.
>
> Guy

That is utter rubbish, do an FFT on the DCC signal and see what
components go to make it up!!!!

UJeff

[manat...@hotmail.com]

On Jan 31, 9:20 am, Jeff <j...@jsystems.com> wrote:
> On 30/01/2012 21:08, Just zis Guy, you know? wrote:
>

> > On Mon, 30 Jan 2012 15:26:04 +0000, Jeff<j...@jsystems.com>  wrote:
>
> >> I think it is you that is confused with the definitions of AC and DC. A
> >> pulse modulated signal is AC, whether that be pulse width modulated or
> >> otherwise modulated.
>
> > No it's not. AC has alternating polarity, DC doesn't. Pulse width
> > modulated DC is still DC, as is DC with a superposed digital signal.
>
> > Guy
>
>   That is utter rubbish, do an FFT on the DCC signal and see what
> components go to make it up!!!!
>
> UJeff

So, take an AC signal and full-wave rectify it, connect a load
resistor with no smoothing capacitor. Is the resulting current flow AC
or DC?

Think carefully before answering.

MBQ

[manat...@hotmail.com]

On Jan 31, 9:20 am, Jeff <j...@jsystems.com> wrote:

> On 30/01/2012 21:08, Just zis Guy, you know? wrote:
>

> > On Mon, 30 Jan 2012 15:26:04 +0000, Jeff<j...@jsystems.com>  wrote:
>
> >> I think it is you that is confused with the definitions of AC and DC. A
> >> pulse modulated signal is AC, whether that be pulse width modulated or
> >> otherwise modulated.
>
> > No it's not. AC has alternating polarity, DC doesn't. Pulse width
> > modulated DC is still DC, as is DC with a superposed digital signal.
>
> > Guy
>
>   That is utter rubbish, do an FFT on the DCC signal and see what
> components go to make it up!!!!

Another example. Take a DC power supply with a constant voltage
regulator and a varying current load. Is the load current AC or DC?

Conversely take a constant current supply and connect a varying
resistive load. Is the resulting voltage waveform AC or DC?

Think carefully before you answer.

MBQ

[manat...@hotmail.com]

On Jan 31, 9:20 am, Jeff <j...@jsystems.com> wrote:

> On 30/01/2012 21:08, Just zis Guy, you know? wrote:
>

> > On Mon, 30 Jan 2012 15:26:04 +0000, Jeff<j...@jsystems.com>  wrote:
>
> >> I think it is you that is confused with the definitions of AC and DC. A
> >> pulse modulated signal is AC, whether that be pulse width modulated or
> >> otherwise modulated.
>
> > No it's not. AC has alternating polarity, DC doesn't. Pulse width
> > modulated DC is still DC, as is DC with a superposed digital signal.
>
> > Guy
>
>   That is utter rubbish, do an FFT on the DCC signal and see what
> components go to make it up!!!!

Any time varying waveform has AC *components* as shown by a fourier
analysis. That doesn't mean that the original waveform, complete with
DC bias is alternating.

What do you think the word "alternating" means?

MBQ

[Jeff]

On 30/01/2012 18:43, manat...@hotmail.com wrote:
> On Jan 30, 5:31 pm, Jeff<j...@jsystems.com> wrote:
>>> PWM implies a fixed frequency variable mark/space ratio.
>>
>> No, there is no such constraint on PWM, DCC is PWM with 2 discrete
>> positions. PWM is AC.
>>
>>
>>
>>> DCC is fixed mark/space ratio (discounting stretched zeroes for
>>> driving analogue locos) with variable frequency. A '1' is sent as a
>>> single cycle of approx 8.6KHz nominal, a '0' as a single cycle of 5KHz
>>> nominal.
>>>> different fixed widths for a one and a zero and so it is the width of
>>>> the pulse that encodes the information. There may only be 2 allowable
>>>> widths but that does not alter the fact that it is PWM. It is not
>>
>>> The width of the pulse *and* the width of the idle period are critical
>>> in decoding a DCC bit stream. Both are significant and it is the
>>> overall period that matters. Hence DCC is frequency modulated, not
>>> PWM.
>>
>> It is merely sent as RTZ with a fixed width for 1 and a different fixed
>> width for a '0'; that may equate to 5kHz (sic) for a '1' but it is not
>
> No, the period changes since both the mark and space are changed
>

>> frequency modulation. It would be FM if the data stream were used to
>> modulate a carrier.(Actually '1' is nearer 8.3kHz, 60us per bit 120us
>> total period).
>
> I don't know where you are getting you information, but I *was*
> correct, the nominal bit time, for a '1' is 58us per *half-bit*, 116us
> total per bit, or approx 8.6KHz. See http://www.nmra.org/standards/DCC/standards_rps/S-91-2004-07.pdf

Ok, so from memory I rounded '55 to 61us' to 60us rather than 58us. '0'
bits do not have a fixed duration, they must be in the range 95us to 9.9
seconds (start & stop bits >=100us), the duration is set by the need to
keep the average of 1's an 0's at zero. So there is no nominal frequency
for '0' bits.

> Send a stream of 1 bits and you get a square wave at approx 8.6KHz,
> send a stream of zeroes and you get 5KHz. That *is* frequency
> modulation.

NO. See above: 0's can lie anywhere in the range 95 to 9900us

No it is not FM, it would be frequency modulation if the baseband data
had 1 & 0 of equal width, i.e. a constant data rate, and that data then
caused the modulated data to shift between 2 different frequencies at
the data rate. In the case of DCC the data rate is different between a 1
and a 0 due to the fact that the encoding is contained in the pulse
width, i.e. PWM. If you also add to the mix the stretched 'bits' which
can be up to 12 seconds and form form the start bit of a valid '0' in
arbitrary positions, and the fact that 0 bits are variable length as
well, it is NOT FM.

Jeff

[Jeff]

On 30/01/2012 20:49, None wrote:
> On Jan 30, 1:45 pm, "manatba...@hotmail.com"<manatba...@hotmail.com>
> wrote:
>> On Jan 30, 4:49 pm, Jeff<j...@jsystems.com> wrote:
>>
>>>> A decoder has no explicit reference. All it has is the differential
>>>> signal on the rails, which is AC. Generally the decoder will full wave
>>>> rectify the AC signal to give an on-board DC supply. It is then
>>>> creating it's own virtual reference point at the zero-crossing of the
>>>> input.
>>
>>> I am not disputing that, but it does not get away from the fact that
>>> the AC signal is Pulse width modulated, albeit that there are only 2
>>> positions that the data can sit at. A pulsed signal IS an AC signal.
>>
>> OK, I accept your agreement that DCC is AC, which you seemed to be
>> disputing.

If you read back you will see that I have always stated that anything
other than a pure constant level is some form of AC. I think the problem
is that some people seem to think that AC does not encompass a pulsed
signal.

>> Now you just have to understande that DCC is FM, not PWM but that's in
>> the other sub-thread.
>>
>> MBQ

Unfortunalely is is no FM see the other posts.

> Maybe it is single side band :>)

Due to the differential encoding it is more like DSB.

Jeff

[Jeff]

Switch your dc supply on and off once a day, is that DC or AC?

If you observe it over several days it is certainly AC!!!

Jeff

[manat...@hotmail.com]

On Jan 31, 9:50 am, Jeff <j...@jsystems.com> wrote:

> On 30/01/2012 18:43, manatba...@hotmail.com wrote:
>
>
>
>
>
>
>
>
>
> > On Jan 30, 5:31 pm, Jeff<j...@jsystems.com>  wrote:
> >>> PWM implies a fixed frequency variable mark/space ratio.
>
> >> No, there is no such constraint on PWM, DCC is PWM with 2 discrete
> >> positions. PWM is AC.
>
> >>> DCC is fixed mark/space ratio (discounting stretched zeroes for
> >>> driving analogue locos) with variable frequency. A '1' is sent as a
> >>> single cycle of approx 8.6KHz nominal, a '0' as a single cycle of 5KHz
> >>> nominal.
> >>>> different fixed widths for a one and a zero and so it is the width of
> >>>> the pulse that encodes the information. There may only be 2 allowable
> >>>> widths but that does not alter the fact that it is PWM. It is not
>
> >>> The width of the pulse *and* the width of the idle period are critical
> >>> in decoding a DCC bit stream. Both are significant and it is the
> >>> overall period that matters. Hence DCC is frequency modulated, not
> >>> PWM.
>
> >> It is merely sent as RTZ with a fixed width for 1 and a different fixed
> >> width for a '0'; that may equate to 5kHz (sic) for a '1' but it is not
>
> > No, the period changes since both the mark and space are changed
>
> >> frequency modulation. It would be FM if the data stream were used to
> >> modulate a carrier.(Actually '1' is nearer 8.3kHz, 60us per bit 120us
> >> total period).
>
> > I don't know where you are getting you information, but I *was*
> > correct, the nominal bit time, for a '1' is 58us per *half-bit*, 116us

> > total per bit, or approx 8.6KHz. Seehttp://www.nmra.org/standards/DCC/standards_rps/S-91-2004-07.pdf

>
> Ok, so from memory I rounded '55 to 61us' to 60us rather than 58us. '0'
> bits do not have a fixed duration, they must be in the range 95us to 9.9
> seconds (start & stop bits >=100us), the duration is set by the need to
> keep the average of 1's an 0's at zero. So there is no nominal frequency
> for '0' bits.
>
> > Send a stream of 1 bits and you get a square wave at approx 8.6KHz,
> > send a stream of zeroes and you get 5KHz. That *is* frequency
> > modulation.
>
> NO. See above: 0's can lie anywhere in the range 95 to 9900us

I did say in an earlier post that I was discounting stretched zeroes.
These are only used for driving a single unchipped loco and are not
even supported by all command stations. Use of an unchipped loco on
DCC is deprecated by those who value their locos, due to the risk of
destroying some types of motor.

> No it is not FM, it would be frequency modulation if the baseband data
> had 1 & 0 of equal width, i.e. a constant data rate, and that data then
> caused the modulated data to shift between 2 different frequencies at
> the data rate. In the case of DCC the data rate is different between a 1
> and a 0 due to the fact that the encoding is contained in the pulse
> width, i.e. PWM.

PWM implies a fixed frequency with variable mark space ratio. Perhaps
we'll just have to agree that it's neither.

> If you also add to the mix the stretched 'bits' which
> can be up to 12 seconds

Where do you get 12 seconds from?

> and form form the start bit of a valid '0' in
> arbitrary positions, and the fact that 0 bits are variable length as
> well, it is NOT FM.

As I said, stretched zeroes can be ignored.

MBQ

[manat...@hotmail.com]

On Jan 31, 10:05 am, Jeff <j...@jsystems.com> wrote:

It's DC.

> If you observe it over several days it is certainly AC!!!

It's still DC. How often did the current or voltage reverse?

All time varying signals have an AC component. Not all resulting
dignals are AC.

MBQ

[Jeff]

.
>
> PWM implies a fixed frequency with variable mark space ratio. Perhaps
> we'll just have to agree that it's neither.

It implies nothing of the sort. That may be the case for controlling a
motor, but *not* when encoding data.

>
>> If you also add to the mix the stretched 'bits' which
>> can be up to 12 seconds
>
> Where do you get 12 seconds from?

My apologies, on consulting the spec it should be 10 seconds with a
total of 12s for the complete reversal.

>
>> and form form the start bit of a valid '0' in
>> arbitrary positions, and the fact that 0 bits are variable length as
>> well, it is NOT FM.
>
> As I said, stretched zeroes can be ignored.

No they cannot, as the period of the 0 bits is dependant on the length
of any stretched bits, so as to keep the average at zero.

Jeff

[Jeff]

>>
>> Switch your dc supply on and off once a day, is that DC or AC?
>
> It's DC.
>
>> If you observe it over several days it is certainly AC!!!
>
> It's still DC. How often did the current or voltage reverse?
>
> All time varying signals have an AC component. Not all resulting
> dignals are AC.
>
> MBQ
>

The definition of an AC signal is that it has a time varying component,
you seem to be regarding AC as some sub-set of that which is not correct.

Jeff

[manat...@hotmail.com]

On Jan 31, 11:36 am, Jeff <j...@jsystems.com> wrote:
> .
>
>
>
> > PWM implies a fixed frequency with variable mark space ratio. Perhaps
> > we'll just have to agree that it's neither.
>
> It implies nothing of the sort. That may be the case for controlling a
> motor, but *not* when encoding data.

Every definition of PWM I have seen assumes fixed frequency. I would
be inetrested to see a description of your version of PWM.

> >> If you also add to the mix the stretched 'bits' which
> >> can be up to 12 seconds
>
> > Where do you get 12 seconds from?
>
> My apologies, on consulting the spec it should be 10 seconds with a
> total of 12s for the complete reversal.

Oh FFS. Its 12000 us, or 12 milliseconds.

I am beginning to wonder how much practical, or even theoretical,
knowledge of DCC you actually have.

> >> and form form the start bit of a valid '0' in
> >> arbitrary positions, and the fact that 0 bits are variable length as
> >> well, it is NOT FM.
>
> > As I said, stretched zeroes can be ignored.
>
> No they cannot, as the period of the 0 bits is dependant on the length
> of any stretched bits, so as to keep the average at zero.

Again, you clearly do not understand DCC!!!

Stretched zeroes are ONLY used to give a DC bias to allow unchipped
locos to be driven. This feature should be disabled, and is not
included in some system, due to the potential to damage some motors.
When stretched zeroes are disabled EACH AND EVERY ZERO on the DCC bus
is nominall 200us total (100us each half cycle).

MBQ

[manat...@hotmail.com]

No, the definition of AC is that the current or voltage alternates,
i.e. reverses periodically.

MBQ

[manat...@hotmail.com]

On Jan 31, 1:38 pm, "manatba...@hotmail.com" <manatba...@hotmail.com>
wrote:

Lets see what Cambridge University have to say on the subject
http://www.eng.cam.ac.uk/DesignOffice/mdp/electric_web/AC/AC_1.html

I'm struggling to find a definition that doesn't include reversal of
current or voltage. Would you care to provide one, from an
authoratative source?

MBQ

[Jeff]

So you are trying to say that a pulsed signal is not AC, even though if
it were a perfect square wave it would have components stretching up in
frequency to infinity!!! You have to look more closely to understand
where the periodic signals are, but they are there. A spectrum analyser
would greatly assist you in seeing them.

Jeff

[Jeff]

>
> Lets see what Cambridge University have to say on the subject
> http://www.eng.cam.ac.uk/DesignOffice/mdp/electric_web/AC/AC_1.html
>
> I'm struggling to find a definition that doesn't include reversal of
> current or voltage. Would you care to provide one, from an
> authoratative source?
>
> MBQ

You are missing the point, a pulse wave form is AC, and can be broken
down into its component parts that are sine waves and meet your
definition of AC. Just because at a cursory glance there is no reversal
of current does not make, say a 0-5V pulse train, DC.

View such a waveform on a spectrum analyser, or do an FFT, and you will
not see just a DC component you will see a multitude of lines that make
up the complete waveform, all of which meet you definition of AC.

Jeff

[Jeff]

>
> Oh FFS. Its 12000 us, or 12 milliseconds.

Sorry my brain was working in ms not us.

>>
>>> As I said, stretched zeroes can be ignored.
>>
>> No they cannot, as the period of the 0 bits is dependant on the length
>> of any stretched bits, so as to keep the average at zero.
>
> Again, you clearly do not understand DCC!!!
>
> Stretched zeroes are ONLY used to give a DC bias to allow unchipped
> locos to be driven. This feature should be disabled, and is not
> included in some system, due to the potential to damage some motors.
> When stretched zeroes are disabled EACH AND EVERY ZERO on the DCC bus
> is nominall 200us total (100us each half cycle).

Please read the spec again, there is no requirement for the 0 bit to be
a nominal 200us it may be anything to 12000us. So a nice 100/200us
period fits your view of FM, but the spec allows
100/anything>200<1200us, which does not fit (anything is allowed to
change on a bit to bit basis). In practice this does not usually happen
unless there are stretched bits but it can and a nice 100/200us is just
a sub-set, so you cannot say DCC is FM based on one case when the spec
and stretched bits allow something different.

Jeff

[manat...@hotmail.com]

On Jan 31, 5:14 pm, Jeff <j...@jsystems.com> wrote:
> > Oh FFS. Its 12000 us, or 12 milliseconds.
>
> Sorry my brain was working in ms not us.
>
>
>
> >>> As I said, stretched zeroes can be ignored.
>
> >> No they cannot, as the period of the 0 bits is dependant on the length
> >> of any stretched bits, so as to keep the average at zero.
>
> > Again, you clearly do not understand DCC!!!
>
> > Stretched zeroes are ONLY used to give a DC bias to allow unchipped
> > locos to be driven. This feature should be disabled, and is not
> > included in some system, due to the potential to damage some motors.
> > When stretched zeroes are disabled EACH AND EVERY ZERO on the DCC bus
> > is nominall 200us total (100us each half cycle).
>
> Please read the spec again, there is no requirement for the 0 bit to be
> a nominal 200us it may be anything to 12000us. So a nice 100/200us
> period fits your view of FM, but the spec allows
> 100/anything>200<1200us, which does not fit (anything is allowed to
> change on a bit to bit basis). In practice this does not usually happen

Which is my point.

MBQ

[Jeff]

But you point only works for one specific case where there happens to be
a fixed 100/200us. If there are stretched bits, or for some other reason
the bits periods are not constant, which is allowed by the spec, then
you description fails. You cannot characterise DCC based on a restricted
sub-set of what is allowed.

Jeff

[manat...@hotmail.com]

On Feb 1, 2:53 pm, Jeff <j...@jsystems.com> wrote:

I am basing it on what happens in practice. I was quite clear early in
tne thread about ruling out the use of stretched zeros, as they are
deprecated. I challenge you to find any DCC command station that does
not use fixed duration equal timing when not using stretched zeros to
drive an unchipped loco.

MBQ

[Just zis Guy, you know?]

You said: "A pulse modulated signal is AC", which is simply false. You
can have pulse modulated AC or DC.

And all waveforms can be described by a sum of sine waves, by that
argument DC is AC. I have no opinion on what comes out of my Lenz DCC
system because I still haven't got round to buying that second-hand
oscilloscope I've been promising myself, and I don't care enough to
look it up because it Just Works.

[Just zis Guy, you know?]

And it's fairly common to use +x and -x rails in digital signals, this
is still not described as AC for the simple reason that it's not, in
any meaningful sense.

The question with DCC would be: what does the output waveform look
like with no signal. I don't know the answer to that, I have no way of
checking here at the moment.

[manat...@hotmail.com]

On Feb 1, 7:52 pm, "Just zis Guy, you know?"

<usenet...@chapmancentral.co.uk> wrote:

>
> The question with DCC would be: what does the output waveform look
> like with no signal. I don't know the answer to that, I have no way of
> checking here at the moment.

Not sure what you mean by that. There's never really a situation with
"no signal". If a booster has no input from a command station then
good design would be to shut down the track output.

If there are locos being controlled then the control packets are sent
repeatedly to those locos.

If the command station really is quescent then it outputs a continuous
preamble which is a string of '1' bits which results in a square wave
at 8.6KHz.

MBQ

[Jeff]

>
> You said: "A pulse modulated signal is AC", which is simply false. You
> can have pulse modulated AC or DC.
>
> And all waveforms can be described by a sum of sine waves, by that
> argument DC is AC. I have no opinion on what comes out of my Lenz DCC
> system because I still haven't got round to buying that second-hand
> oscilloscope I've been promising myself, and I don't care enough to
> look it up because it Just Works.
>
> Guy

You can't have it both ways Guy!!! You say that "all waveforms can be
described by a sum of sine waves" and then go on then say that you can
have "pulse modulated DC"!! (DC cannot be described by sine waves unless
you tale the case of zero frequency as a sine wave). If an FFT shows
something other than *only* a DC component (zero frequency) the signal
is AC!!!!
All pulsed waveforms have a time varying component and are therefore AC.

Jeff

[Jeff]

>>
>> But you point only works for one specific case where there happens to be
>> a fixed 100/200us. If there are stretched bits, or for some other reason
>> the bits periods are not constant, which is allowed by the spec, then
>> you description fails. You cannot characterise DCC based on a restricted
>> sub-set of what is allowed.
>
> I am basing it on what happens in practice. I was quite clear early in
> tne thread about ruling out the use of stretched zeros, as they are
> deprecated. I challenge you to find any DCC command station that does
> not use fixed duration equal timing when not using stretched zeros to
> drive an unchipped loco.
>
> MBQ
>

So what do you call it when someone does actually use stretched zeros?
(if was you that introduced the restriction on not considering them). It
is still DCC as per the spec; so does the description of the waveform
suddenly change? No of course not.

(By the way different length 0's are allowed even when stretched bits
are not in use.)

Jeff

[manat...@hotmail.com]

On Jan 31, 4:48 pm, Jeff <j...@jsystems.com> wrote:

> On 31/01/2012 13:38, manatba...@hotmail.com wrote:
>
>
>
>
>
>
>
>
>
> > On Jan 31, 11:47 am, Jeff<j...@jsystems.com>  wrote:
> >>>> Switch your dc supply on and off once a day, is that DC or AC?
>
> >>> It's DC.
>
> >>>> If you observe it over several days it is certainly AC!!!
>
> >>> It's still DC. How often did the current or voltage reverse?
>
> >>> All time varying signals have an AC component. Not all resulting
> >>> dignals are AC.
>
> >>> MBQ
>
> >> The definition of an AC signal is that it has a time varying component,
> >> you seem to be regarding AC as some sub-set of that which is not correct.
>
> > No, the definition of AC is that the current or voltage alternates,
> > i.e. reverses periodically.
>
> > MBQ
>
> So you are trying to say that a pulsed signal is not AC, even though if
> it were a perfect square wave it would have components stretching up in
> frequency to infinity!!!

If it has a large enough DC offset it is not AC.

> You have to look more closely to understand
> where the periodic signals are, but they are there. A spectrum analyser
> would greatly assist you in seeing them.

I know where they are.

The resultant signal is also period.

Being periodic is a neccessary but not sufficient condition for a
signal to be AC.

MBQ

[manat...@hotmail.com]

The components, taken in isolation, are AC. When combined and added to
a DC offset the resulting waveform can be DC. The *definition* of AC
is that the current or voltage reverses or "laternates".

MBQ

[manat...@hotmail.com]

Please provide an authoratative reference for your definition of AC as
anything that varies over time.

MBQ

[Jeff]

>
> The components, taken in isolation, are AC. When combined and added to
> a DC offset the resulting waveform can be DC. The *definition* of AC
> is that the current or voltage reverses or "laternates".
>
> MBQ

The problem is that you are only considering the very simplistic
definition of AC that is taught at school. You have to look at more
advanced text books to see a fuller picture. However, one slightly
fuller definition onthe web is here:

<http://www.animations.physics.unsw.edu.au/jw/AC.html>

Note the section that states:

" As for the computer you're using to read this, its signals are not
ordinary sinusoidal AC, but, thanks to Fourier's theorem, any varying
signal may be analysed in terms of its sinusoidal components. So AC
signals are almost everywhere"

AC signals are *any* signals that vary with time or to put it another
way can be defined as an expansion of a series of signals that may be
described by V*sin(omega*t+phi), where omega is the angular frequency
(2*pi*f), and phi is the phase angle. DC offset may have a term in the
series where the frequency component is zero.

Also have a look at section 2 of:
<http://www.g3ynh.info/zdocs/AC_theory/AC_theory.pdf>

part of which states the following and it goes on to expand further:

"A peculiarity, which is often introduced without comment, is that AC
generators (of the analytical variety) are considered to produce
sinusoidal outputs. Many practical generators (e.g.,mechanical
alternators, radio transmitters) do indeed produce something
approximating a voltage or current sine-wave; but the reason goes
somewhat deeper than that. If we take, for example, a moving-coil
microphone (which is a type of generator which produces electricity from
air-pressure variations), we will find that its output in response to
(say) the sound of the human voice, is extremely complicated. A
technique known as 'Fourier analysis' however, shows that all waveforms
can be built-up by adding-together sinusoidal waves of different
frequencies; and physical investigation shows that these separate
frequency components actually exist."

Jeff

[Jeff]

Let me pose to you a little question.

If you apply a DCC signal or a 0-5V pulse data stream to a suitable
transformer, what to you see on the secondary side??

Answer: the same as you put in, but there is no DC reference, so how in
your view do you classify the output? Is the current reversing? If so in
respect to what? Has it changed a 0-5V pulse that you consider to be DC
to AC?? Of course it hasn't all that has happened is that any DC term if
any has been removed from the mathematical series that described the
waveform previously. AC went in AC came out.

Jeff

[Jeff]

Perhaps "the same" is a slight exaggeration, the waveform will be
distorted slightly depending on the bandwidth and hysteresis of the
transformer.

Jeff

[manat...@hotmail.com]

On Feb 2, 11:11 am, Jeff <j...@jsystems.com> wrote:
> > The components, taken in isolation, are AC. When combined and added to
> > a DC offset the resulting waveform can be DC. The *definition* of AC
> > is that the current or voltage reverses or "laternates".
>
> > MBQ
>
> The problem is that you are only considering the very simplistic
> definition of AC that is taught at school. You have to look at more
> advanced text books to see a fuller picture. However, one slightly
> fuller definition onthe web is here:
>
> <http://www.animations.physics.unsw.edu.au/jw/AC.html>
>
> Note the section that states:
>
> " As for the computer you're using to read this, its signals are not
> ordinary sinusoidal AC, but, thanks to Fourier's theorem, any varying
> signal may be analysed in terms of its sinusoidal components. So AC
> signals are almost everywhere"

Sorry, but that's not good enough. It sounds like the A level physics
teacher who taiught us about various op-amp circuits for adding,
subtracting, etc and then told us that the new Commodore PET computers
that the school had just bought had millions iof these circuits
inside. I had no more reason to believe him than a throw-away comment
in that article. All of the subsequent exammples clearly show
waveforms with zero DC offset.

> AC signals are *any* signals that vary with time or to put it another
> way can be defined as an expansion of a series of signals that may be
> described by V*sin(omega*t+phi), where omega is the angular frequency
> (2*pi*f), and phi is the phase angle. DC offset may have a term in the
> series where the frequency component is zero.
>
> Also have a look at section 2 of:
> <http://www.g3ynh.info/zdocs/AC_theory/AC_theory.pdf>
>
> part of which states the following and it goes on to expand further:
>
> "A peculiarity, which is often introduced without comment, is that AC
> generators (of the analytical variety) are considered to produce
> sinusoidal outputs. Many practical generators (e.g.,mechanical
> alternators, radio transmitters) do indeed produce something
> approximating a voltage or current sine-wave; but the reason goes
> somewhat deeper than that. If we take, for example, a moving-coil
> microphone (which is a type of generator which produces electricity from
> air-pressure variations), we will find that its output in response to
> (say) the sound of the human voice, is extremely complicated. A
> technique known as 'Fourier analysis' however, shows that all waveforms
> can be built-up by adding-together sinusoidal waves of different
> frequencies; and physical investigation shows that these separate
> frequency components actually exist."
>

I don't dispute that you can decompose a signal imto it's components
and use AC analysis on this AC components. The quote above adds
nothing to the debate.

MBQ

[manat...@hotmail.com]

Finger trouble on the previous post.

DCC is AC and you will get an AC output from a suitable transformer
fed with DCC. You can make a very effective block occupancy detector
for DCC using a current transformer.

For the 0 - 5V pulse case, that's more like the old transformer
coupling of audio outputs. The DC offset is blocked turning the DC
input into an AC output.

MBQ

[manat...@hotmail.com]

On Feb 2, 11:36 am, Jeff <j...@jsystems.com> wrote:

Two very, very different scenarios which only goes to confirm your
misunderstanding.

First DCC. We have already established that DCC is *AC*. The current
reverses each half cycle. Put it into a transformer suitable for the
frequency range and you will get an AC signal out. No problem there,
that's how transformers work. You can m

Now a 0 to 5V squarewave.

[manat...@hotmail.com]

On Feb 2, 9:33 am, Jeff <j...@jsystems.com> wrote:
> >> But you point only works for one specific case where there happens to be
> >> a fixed 100/200us. If there are stretched bits, or for some other reason
> >> the bits periods are not constant, which is allowed by the spec, then
> >> you description fails. You cannot characterise DCC based on a restricted
> >> sub-set of what is allowed.
>
> > I am basing it on what happens in practice. I was quite clear early in
> > tne thread about ruling out the use of stretched zeros, as they are
> > deprecated. I challenge you to find any DCC command station that does
> > not use fixed duration equal timing when not using stretched zeros to
> > drive an unchipped loco.
>
> > MBQ
>
> So what do you call it when someone does actually use stretched zeros?

An abortion :-) It's much easier to design for consistent timing.

MBQ

[Jeff]

It is very interesting that you consistently ignore the question when
the result does not fit with your theory.

Jeff

[Jeff]

So is DC if a pulse train has a DC offset and AC it there is not...yeah
right.

Jeff

[Jeff]

Should read "So it is...."

[PV]

"manat...@hotmail.com" <manat...@hotmail.com> writes:
>Not sure what you mean by that. There's never really a situation with
>"no signal". If a booster has no input from a command station then
>good design would be to shut down the track output.

Does any system actually do that? Digitrax certainly doesn't.

>If there are locos being controlled then the control packets are sent
>repeatedly to those locos.

This is flat out wrong. Decoders assume that the previous instructions
stand until they get an update. If dcc commands stop going out on the
rails, the trains will happily keep running unless you kill power. You've
never had any runaway trains? It definitely happens.

>If the command station really is quescent then it outputs a continuous
>preamble which is a string of '1' bits which results in a square wave
>at 8.6KHz.

Not sure what point you're trying to make. This is kind of a dumb message
thread anyway. *
--
* PV Something like badgers, something like lizards, and something
like corkscrews.

[Just zis Guy, you know?]

On Thu, 02 Feb 2012 09:28:28 +0000, Jeff <je...@jsystems.com> wrote:

>
>>
>> You said: "A pulse modulated signal is AC", which is simply false. You
>> can have pulse modulated AC or DC.
>>
>> And all waveforms can be described by a sum of sine waves, by that
>> argument DC is AC. I have no opinion on what comes out of my Lenz DCC
>> system because I still haven't got round to buying that second-hand
>> oscilloscope I've been promising myself, and I don't care enough to
>> look it up because it Just Works.
>>
>> Guy
>
>You can't have it both ways Guy!!! You say that "all waveforms can be
>described by a sum of sine waves" and then go on then say that you can
>have "pulse modulated DC"!! (DC cannot be described by sine waves unless
>you tale the case of zero frequency as a sine wave).

Give the man a coconut, zero frequency is indeed a special case. Now
find a railway power supply that puts out a perfectly smooth DC
waveform.

>If an FFT shows
>something other than *only* a DC component (zero frequency) the signal
>is AC!!!!

Bullshit. If current only flows in one direction out of the supply
then the current does not alternate.

>All pulsed waveforms have a time varying component and are therefore AC.

Bullshit. See above.

[Just zis Guy, you know?]

On Thu, 02 Feb 2012 11:11:12 +0000, Jeff <je...@jsystems.com> wrote:

>The problem is that you are only considering the very simplistic
>definition of AC that is taught at school.

As am I. For values of "school" that include the Faculty of
Engineering and Applied Science at the University of Southampton, from
which I graduated with a B. Eng (hons) in electrical engineering.

Alternating current is current which flows in alternate directions.
Note the word current.

A direct current with a superposed alternating signal is still a
direct current. If current always flows one way from the supply then
to describe it as AC based on the resultant waveform after
superposition of a control signal would be, to use a technical term,
wrong.

Either that or RS Components, Farnell, Maplin and the rest have
mis-labelled every single DC supply they sell.

[Jane Sullivan]

On 02/02/2012 19:44, Just zis Guy, you know? wrote:
> On Thu, 02 Feb 2012 09:28:28 +0000, Jeff<je...@jsystems.com> wrote:
>
>>
>>>
>>> You said: "A pulse modulated signal is AC", which is simply false. You
>>> can have pulse modulated AC or DC.
>>>
>>> And all waveforms can be described by a sum of sine waves, by that
>>> argument DC is AC. I have no opinion on what comes out of my Lenz DCC
>>> system because I still haven't got round to buying that second-hand
>>> oscilloscope I've been promising myself, and I don't care enough to
>>> look it up because it Just Works.
>>>
>>> Guy
>>
>> You can't have it both ways Guy!!! You say that "all waveforms can be
>> described by a sum of sine waves" and then go on then say that you can
>> have "pulse modulated DC"!! (DC cannot be described by sine waves unless
>> you tale the case of zero frequency as a sine wave).
>
> Give the man a coconut, zero frequency is indeed a special case. Now
> find a railway power supply that puts out a perfectly smooth DC
> waveform.

A car battery. Perfectly smooth 12v. DC.

>
>
>> If an FFT shows
>> something other than *only* a DC component (zero frequency) the signal
>> is AC!!!!
>
> Bullshit. If current only flows in one direction out of the supply
> then the current does not alternate.
>
>> All pulsed waveforms have a time varying component and are therefore AC.
>
> Bullshit. See above.
>
> Guy


--

Jane

[Just zis Guy, you know?]

On Thu, 02 Feb 2012 20:26:18 +0000, Jane Sullivan <m...@privacy.net>
wrote:

>> Give the man a coconut, zero frequency is indeed a special case. Now
>> find a railway power supply that puts out a perfectly smooth DC
>> waveform.
>
>A car battery. Perfectly smooth 12v. DC.

I think you'll find it reduces over time, initially in a fairly linear
way and later definitely non-linear.

[Jeff]

>> If an FFT shows
>> something other than *only* a DC component (zero frequency) the signal
>> is AC!!!!
>
> Bullshit. If current only flows in one direction out of the supply
> then the current does not alternate.
>
>> All pulsed waveforms have a time varying component and are therefore AC.
>
> Bullshit. See above.
>
> Guy

Not bullshit at all, the DC component may only cause current to flow in
one direction, but the transitions will cause currents at frequencies
above DC to flow in both directions, they may, or may not, be smaller
than the dc current but they do exist and do flow in both directions.

Jeff

[Just zis Guy, you know?]

Yes, bullshit. Either that or there is effectively no such thing as a
DC power supply.

[Jane Sullivan]

On 03/02/2012 00:22, Just zis Guy, you know? wrote:
> On Thu, 02 Feb 2012 20:26:18 +0000, Jane Sullivan<m...@privacy.net>
> wrote:
>
>>> Give the man a coconut, zero frequency is indeed a special case. Now
>>> find a railway power supply that puts out a perfectly smooth DC
>>> waveform.
>>
>> A car battery. Perfectly smooth 12v. DC.
>
> I think you'll find it reduces over time, initially in a fairly linear
> way and later definitely non-linear.
>
> Guy

Depends if you recharge it between operating sessions or not.

--
Jane

[manat...@hotmail.com]

On Feb 2, 4:07 pm, Jeff <j...@jsystems.com> wrote:

I didn't ignore it. I gave you my honest opinion of stretched zeroes.
They were a useful kludge in the early days for ri=unning an unchipped
loco. Now they have no place in my opinion. There's a danger of
damaging some types of motor and using them unneccasrily just reduces
the data bandwidth available, which is already not great with DCC.

MBQ

[manat...@hotmail.com]

NO. For the purppose of analysis you can split any signal into
components. Those components do not exist in isolation in the
composite signal.

Tell me, how do the electrons and holes in the wire know which
frequency component they are associated with and which way to flow?

MBQ

[manat...@hotmail.com]

On Feb 2, 5:13 pm, pv+use...@pobox.com (PV) wrote:

> "manatba...@hotmail.com" <manatba...@hotmail.com> writes:
> >Not sure what you mean by that. There's never really a situation with
> >"no signal". If a booster has no input from a command station then
> >good design would be to shut down the track output.
>
> Does any system actually do that? Digitrax certainly doesn't.
>
> >If there are locos being controlled then the control packets are sent
> >repeatedly to those locos.
>
> This is flat out wrong. Decoders assume that the previous instructions
> stand until they get an update. If dcc commands stop going out on the
> rails, the trains will happily keep running unless you kill power. You've
> never had any runaway trains? It definitely happens.

Fuck me, where do people get their DCC "knowledge" from? Try reading
the specs. Specifically RP-9.2.4 section C. If you choose cheap
decoders that don't implement the spec correctly, or disable the
timeout then you deserve every run-away you get.

> >If the command station really is quescent then it outputs a continuous
> >preamble which is a string of '1' bits which results in a square wave
> >at 8.6KHz.
>
> Not sure what point you're trying to make.

The point that there is always *something* even if it's just the
preamble on the rails. It is neccessary to continue delivering power.

> This is kind of a dumb message
> thread anyway. *

You didn't raise the level, so why did you bother replying?

MBQ

[Jeff]

On 03/02/2012 09:15, Just zis Guy, you know? wrote:
> On Fri, 03 Feb 2012 07:47:48 +0000, Jeff<je...@jsystems.com> wrote:
>
>>
>>>> If an FFT shows
>>>> something other than *only* a DC component (zero frequency) the signal
>>>> is AC!!!!
>>>
>>> Bullshit. If current only flows in one direction out of the supply
>>> then the current does not alternate.
>>>
>>>> All pulsed waveforms have a time varying component and are therefore AC.
>>>
>>> Bullshit. See above.
>>>
>>> Guy
>>
>> Not bullshit at all, the DC component may only cause current to flow in
>> one direction, but the transitions will cause currents at frequencies
>> above DC to flow in both directions, they may, or may not, be smaller
>> than the dc current but they do exist and do flow in both directions.
>
> Yes, bullshit. Either that or there is effectively no such thing as a
> DC power supply.
>
> Guy

I suppose you could take that view if the DC level changes or you switch
the DC voltage on and off. At the time of the transition there will be
AC components generated. However, it would be silly to classify a DC
power supply as AC merely because of short AC transients caused by
varying the voltage or switching on and off, but in the case of DCC the
AC components are there *all of the time* and make up the
characteristics of the signal, ie the pulses. Without the AC parts the
DCC waveform could only be just a constant DC level.

Jeff

[Chris]

On 30/01/2012 14:01, manat...@hotmail.com wrote:
> On Jan 30, 12:32 pm, Jeff<j...@jsystems.com> wrote:
>>> DCC is not a "pulsed signal" it is AC.
>>
>>> A decoder sees two connections to the track, lets call them A and B.
>>> During one half of the cycle, A is more positive than B and current
>>> flows from A to B. During the other half of the cycle, B is more
>>> positive than A and current flows from B to A. The current flow
>>> reverses twice every cycle. The current alternates. It is an
>>> alternating current.
>>
>>> DCC is AC. It is not a "pulsed signal", nor is it "bipolar DC" as some
>>> would have it.
>>
>>> Contrast this with a real pulsed DC controller where the pulses simply
>>> turn the track on and off. The current only ever flows in one
>>> direction (unless you flip the reversing switch!). That *is* a pulsed
>>> DC signal and not AC [1].
>>
>>> I think the confusion arises due to people looking at the output of a
>>> command station with reference to the local ground. If you measure the
>>> two outputs A and B independently then they do look like pulsed
>>> outputs. The point is that DCC is defined by the two track signals
>>> alone, there is no ground reference. It is the differential signal
>>> between the rails that is DCC and that is AC.
>>
>>> MBQ
>>
>>> Being really pedantic you could show that it is a combination of an AC
>>> signal and a DC offset but that is taking things to a ridiculous
>>> extreme.
>>
>> It is most certainly is a pulsed signal, with pulse width modulation
>
> DCC is *frequency* modulated.
>
>> that carries the data. The fact that it is applied to the track in a
>> differential manner is not really relevant to that fact.
>
> The fact that it is applied differentially means it is AC as seen by
> the decoder.
>
> Are you sure you are not getting confused with the motor drive output
> of the decoder, which is PWM and is DC, only reversing when the
> direction of the loco changes.
>
> MBQ
>
>
Decoders see a DCC signal not an AC signal otherwise they would not be
able to decode the commands. For non DCC locos on a DCC layout they see
PWM by stretching one's or zero's depending on direction to move the
loco. So you could argue its almost DC in that instance and almost AC
with no DC locos.

Note that Hornbys Zero 1 system of the 1980's was very much as you
describe an AC supply with control data overlaid where as DCC is the
supply and signal combined. What this means for the modeller is that if
the supply voltage gets to the decoder so does the commands and
everything works. Whereas the Hornby system was not as reliable as the
supply voltage might arrive but the commands sometimes got lost.

--
Chris

[Just zis Guy, you know?]

On Fri, 03 Feb 2012 09:43:07 +0000, Jane Sullivan <m...@privacy.net>

wrote:

>On 03/02/2012 00:22, Just zis Guy, you know? wrote:
>> On Thu, 02 Feb 2012 20:26:18 +0000, Jane Sullivan<m...@privacy.net>
>> wrote:
>>
>>>> Give the man a coconut, zero frequency is indeed a special case. Now
>>>> find a railway power supply that puts out a perfectly smooth DC
>>>> waveform.
>>>
>>> A car battery. Perfectly smooth 12v. DC.
>>
>> I think you'll find it reduces over time, initially in a fairly linear
>> way and later definitely non-linear.

>Depends if you recharge it between operating sessions or not.

Not really, no. It's a lot smoother than rectified DC but the voltage
will still fluctuate, and will go down as the battery discharges.

[Just zis Guy, you know?]

On Fri, 03 Feb 2012 12:16:59 +0000, Jeff <je...@jsystems.com> wrote:

>On 03/02/2012 09:15, Just zis Guy, you know? wrote:
>> On Fri, 03 Feb 2012 07:47:48 +0000, Jeff<je...@jsystems.com> wrote:
>>
>>>
>>>>> If an FFT shows
>>>>> something other than *only* a DC component (zero frequency) the signal
>>>>> is AC!!!!
>>>>
>>>> Bullshit. If current only flows in one direction out of the supply
>>>> then the current does not alternate.
>>>>
>>>>> All pulsed waveforms have a time varying component and are therefore AC.
>>>>
>>>> Bullshit. See above.
>>>>
>>>> Guy
>>>
>>> Not bullshit at all, the DC component may only cause current to flow in
>>> one direction, but the transitions will cause currents at frequencies
>>> above DC to flow in both directions, they may, or may not, be smaller
>>> than the dc current but they do exist and do flow in both directions.
>>
>> Yes, bullshit. Either that or there is effectively no such thing as a
>> DC power supply.

>I suppose you could take that view if the DC level changes or you switch
>the DC voltage on and off.

You have described AC in a way that will necessarily also encompass
pretty much every DC power supply on the market. This has put you in
your normal position of being the only one marching in step.

Enough of your sophistry.

[Jeff]

On

>
>> I suppose you could take that view if the DC level changes or you switch
>> the DC voltage on and off.
>
> You have described AC in a way that will necessarily also encompass
> pretty much every DC power supply on the market. This has put you in
> your normal position of being the only one marching in step.
>
> Enough of your sophistry.
>
> Guy

You very conveniently snipped the next part of the post that said "it

would be silly to classify a DC power supply as AC merely because of

short AC transients caused by varying the voltage or switching on and off".

With a conventional DC power supply (not DCC or PWM) any AC components
make up a small part of the output. With DCC if you remove the AC
components all that you are left with are 2 DC levels +V and -V which
cancel each other out to zero!! You cannot switch between to 2 DC levels
without the AC components, they are there all of the time and are real
and form the major characteristic of the signal i.e the data.

Jeff

[Just zis Guy, you know?]

On Sun, 05 Feb 2012 10:10:34 +0000, Jeff <je...@jsystems.com> wrote:

>You very conveniently snipped the next part of the post that said "it
>would be silly to classify a DC power supply as AC merely because of
>short AC transients caused by varying the voltage or switching on and off".

Yes, because that was the point I originally made and which you were
apparently disputing. Since you'd accepted my point but were
continuing to argue that you're right and everybody else wrong I did
not see much point continuing.

[Jeff]

On 05/02/2012 11:30, Just zis Guy, you know? wrote:
> On Sun, 05 Feb 2012 10:10:34 +0000, Jeff<je...@jsystems.com> wrote:
>
>> You very conveniently snipped the next part of the post that said "it
>> would be silly to classify a DC power supply as AC merely because of
>> short AC transients caused by varying the voltage or switching on and off".
>
> Yes, because that was the point I originally made and which you were
> apparently disputing. Since you'd accepted my point but were
> continuing to argue that you're right and everybody else wrong I did
> not see much point continuing.
>
> Guy

I think the problem here is one of perception, the problem seem to be
that this is being viewed from the point of view of a motor control
system rather that the general view of how to classify a signal. If the
DCC waveform, or any other pulse waveform for that matter, was being
used in any other electronic application as a means of sending data down
a wire, there is no way that it would be considered as being DC, even
considering the dual purpose of the signal (ie conveying data as well as
a power source after rectification).

Getting back to whether you can classify slowly changing DC or slow on
and off switching of DC as an AC signal, it depends on your point of
view and what information you are trying to recover. If you view it
over a long enough time period then you can recover AC information, the
long integration period being a way to improve the signal to noise ratio
of your measurement. So to one person a slowly changing level or slow
on/off is a DC signal, but to another it is AC data. In a model rail
application it is obviously viewed as DC.

Jeff

[manat...@hotmail.com]

On Feb 4, 8:49 pm, Chris <nos...@nospam.com> wrote:

Have you bothered to read the rest of the thread? We've already
established that DCC *is* AC. The DCC signal is rectified to produce
on board DC power for the electronics and motor. The data is then
decoded by looking at the signal before rectification.

> PWM by stretching one's or zero's depending on direction to move the

Wrong again. only zeroes are stretched. It's still AC,but with a
positive or negative DC offset, depending on the desired direction.

> loco. So you could argue its almost DC in that instance and almost AC
> with no DC locos.

No, it's always AC.

MBQ

[manat...@hotmail.com]

On Feb 5, 12:08 pm, Jeff <j...@jsystems.com> wrote:
> On 05/02/2012 11:30, Just zis Guy, you know? wrote:
>

> > On Sun, 05 Feb 2012 10:10:34 +0000, Jeff<j...@jsystems.com>  wrote:
>
> >> You very conveniently snipped the next part of the post that said "it
> >> would be silly to classify a DC power supply as AC merely because of
> >> short AC transients caused by varying the voltage or switching on and off".
>
> > Yes, because that was the point I originally made and which you were
> > apparently disputing. Since you'd accepted my point but were
> > continuing to argue that you're right and everybody else wrong I did
> > not see much point continuing.
>
> > Guy
>
>   I think the problem here is one of perception, the problem seem to be
> that this is being viewed from the point of view of a motor control
> system rather that the general view of how to classify a signal. If the
> DCC waveform, or any other pulse waveform for that matter, was being
> used in any other electronic application as a means of sending data down

> a wire, there is no way that it would be considered as being DC,]

No one (apart from you, or was that a typo?) is considering DCC to be
DC. DCC is AC.

MBQ

[manat...@hotmail.com]

On Feb 5, 10:10 am, Jeff <j...@jsystems.com> wrote:
> On
>
>
>
> >> I suppose you could take that view if the DC level changes or you switch
> >> the DC voltage on and off.
>
> > You have described AC in a way that will necessarily also encompass
> > pretty much every DC power supply on the market. This has put you in
> > your normal position of being the only one marching in step.
>
> > Enough of your sophistry.
>
> > Guy
>
> You very conveniently snipped the next part of the post that said "it
> would be silly to classify a DC power supply as AC merely because of
> short AC transients caused by varying the voltage or switching on and off".
>
> With a conventional DC power supply (not DCC or PWM) any AC components
> make up a small part of the output. With DCC if you remove the AC
> components all that you are left with are 2 DC levels +V and -V which

If there are no stretched zeroes then DCC is nominally a perfectly
balanced signal and there is no DC component. Remove the AC components
and you are left with nothing.

If stretched zeroes are being used then you will have either a
positive or negative DC component depending on the direction the
unchipped loco is being driven. You clearly can't have both at the
seame time.

MBQ

[simon]

Read the first review of Hornby Zero 1 system in MRC, was well ahead of its
time, quite amazing really.

Cheers,
Simon

[manat...@hotmail.com]

On Feb 7, 9:13 am, Jeff <j...@jsystems.com> wrote:

>
> > If stretched zeroes are being used then you will have either a
> > positive or negative DC component depending on the direction the
> > unchipped loco is being driven. You clearly can't have both at the
> > seame time.
>
> > MBQ
>

> Not correct read the spec. In order to keep the average level at zero,
> if +ve stretched bits are in use then the period of the -ve part of the
> 0 data bits are extended so as to average the DC back to zero between
> stretched bits, that is why 0's are allowed to have different length -ve
> and +ve transitions. However, to the motor the fast -ve bits are are not
> seen as well due to the Low pass characteristics of the motor, so the
> motor sees a net +ve (or everything the other way around if you are
> reversing and using -ve stretched bits).

Totally wrong.

Look at the picture in the spec you seem not to have read for
yourself. The positive and negative portions of the stretched zero are
markedly different. That's how a DC bias in imparted to the DCC
waveform to drive an unchipped loco. A "stretched 0" is one which *by
definition* has unequal mark/space ratio.

MBQ

[Chris]

It thinks its been agreed that it is not AC as most people understand
and for that matter DVM or MC meters, it but modulated DC that is like
AC. Therefore you need meter that reads modulated DC.

Admittedly wrong about the bit stretching, a long time since I looked at
and just thought +ve was 1 and -ve was 0.

From the view point of a DC motor in zero stretching it is enough DC
like to make it work.

Anyway this thread has proved most interesting reading and brought back
a lot of AC theory I studied many years ago at college.



--
Chris

[manat...@hotmail.com]

On Feb 9, 9:57 pm, Chris <nos...@nospam.com> wrote:

No. It's AC, end of.

You need a meter that reads "true RMS" AC instead of a cheap hobbyist
meter that is only calibrated for 50/60Hz sine waves.

MBQ

[Chris]

That would work if the waveform was "true RMS" AC rather than a DCC signal.

--
Chris

[manat...@hotmail.com]

On Feb 10, 6:39 pm, Chris <nos...@nospam.com> wrote:

<sigh>

There's no such thing as "true RMS" AC.

The term "true RMS" is describing the fact that the meter is capable
of measuring the RMS value of any arbitrary waveform, unlike cheap
meters which approximate it but are designed to be reasonably accurate
for a 50/60Hz sine wave.

Since DCC is AC, a true RMS meter will give an accurate voltage
measurement.

MBQ

[Chris]

On 12/02/2012 11:12, Jeff wrote:
>
>>> That would work if the waveform was "true RMS" AC rather than a DCC
>>> signal.
>>
>> <sigh>
>>
>> There's no such thing as "true RMS" AC.
>>
>> The term "true RMS" is describing the fact that the meter is capable
>> of measuring the RMS value of any arbitrary waveform, unlike cheap
>> meters which approximate it but are designed to be reasonably accurate
>> for a 50/60Hz sine wave.
>>
>> Since DCC is AC, a true RMS meter will give an accurate voltage
>> measurement.
>>
>> MBQ
>>
>>
>

> Correct in as much as it will give the RMS value over the sampling
> period of the meter, but if the mark-space ratios of the DCC signals
> vary over time the reading will fluctuate. Also bear in mind that the
> bandwidth of many DVM's is quite limited and will not give an accurate
> reading on a square wave with fast edges, even if the fundamental
> frequency is with its range.
>
> Jeff
>
Which is what I was alluding to, very well put. Probably more true for
moving coil meters.

--
Chris

[bobharvey]

On Monday, January 30, 2012 6:43:38 PM UTC, manat...@hotmail.com wrote:
> Send a stream of 1 bits and you get a square wave at approx 8.6KHz,
> send a stream of zeroes and you get 5KHz. That *is* frequency
> modulation.

No, it is pwm. The apparant frequency variation is an artifact of the encoding mechanism, there is no way to be pwm without that happening.

Put it this way. Transmit 101010 and 010101 and the frequency will be the same. Decoding the frequency will not tell you what the digital content is. If it were frequency modulated you could decode into the frequency domain and know the exact digital content.

[MartinS]

bobharvey <robert...@my-deja.com> wrote:

I had hoped this thread was long dead.

--
Martin S.

[manat...@hotmail.com]

On Mar 22, 3:03 pm, bobharvey <roberthar...@my-deja.com> wrote:
> On Monday, January 30, 2012 6:43:38 PM UTC, manat...@hotmail.com wrote:
> > Send a stream of 1 bits and you get a square wave at approx 8.6KHz,
> > send a stream of zeroes and you get 5KHz. That *is* frequency
> > modulation.
>
> No, it is pwm.  The apparant frequency variation is an artifact of the encoding mechanism, there is no way to be pwm without that happening.

Sorry, but the clasic definition of PWM is a fixed period, i.e., fixed
pulse repetition frequency, with varying mark space ratio. There are
planty pf text books that wll tell you this if you refuse to accept it
from me.

DCC, however, varies the period but keeps the mark space ration
constant, other than when using stretched zeroes to control an
unchipped loco.

MBQ