MOTM

Hi all,

could someone try to explain in layman's terms what it means 
when an input is either AC or DC coupled?

I know it means it assumes either an AC or DC signal. But 
what if I feed it the other type of signal?

And a related question: what is the definitions of an AC and 
DC signal, in modular context? For example, when I an EG 
ouputs a through-zero DC signal, isn't that actually an AC 
signal?

Confused...

- Robert

Here is a simple explanation:  DC (direct curent) means a steady, 
unchanging voltage, which could be any voltage level.  If a circuit is 
DC-coupled it means that the level of an unchanging voltage at the input 
will be passed through.  If a circuit is AC coupled then only changing 
voltages will be passed and there will be a "high pass" function 
applied, i.e. there will be a frequency below which signals won't pass.  
This cutoff frequency is normally about 20 Hz  for audio processing.  If 
you look at a circuit diagram, you will see a capacitor in series with 
the signal path if it is AC coupled.  Both inputs and outputs can be AC 
coupled.

If you feed a DC voltage, say +5V, into an AC coupled input, the output 
will be zero.  You can always feed an AC signal into a DC coupled 
input.  That works fine.  So why not always just use DC coupling?

The benefit of AC coupling is that it removes any DC offset (constant) 
voltage.  For example, the signal input of a VCA is sometimes AC 
coupled, because a DC offset at the signal input can result in a popping 
sound when it is multiplied with the control input.  The control input 
of a VCA has to be DC coupled, or else a prolonged unchanging voltage, 
such as the sustain portion of an envelope, would not produce the 
desired constant output level.

Good question about the EG output.  It is DC coupled, for the reason 
above.  However, an envelope is a type of AC signal, because it changes 
over time.  The interesting thing about an envelope is that it combines 
different frequencies in one cycle.  Attack is typically fast, decay and 
release slower.  An attack time of 20 milliseconds is 1/50 of a second, 
i.e. 50 Hz, up in the AC range for sure.  But a 500 millisecond release 
time is only 2 Hz.  So it is an AC signal with some very low frequency 
parts.

-Richard Brewster


Robert van der Kamp wrote:

>Hi all,
>
>could someone try to explain in layman's terms what it means 
>when an input is either AC or DC coupled?
>
>I know it means it assumes either an AC or DC signal. But 
>what if I feed it the other type of signal?
>
>And a related question: what is the definitions of an AC and 
>DC signal, in modular context? For example, when I an EG 
>ouputs a through-zero DC signal, isn't that actually an AC 
>signal?
>
>Confused...
>
>- Robert
>
>
> 
>
>
>  
>

On 1/9/04, Richard Brewster put forth:
>Good question about the EG output.  It is DC coupled, for the reason
>above.  However, an envelope is a type of AC signal, because it changes
>over time.  The interesting thing about an envelope is that it combines
>different frequencies in one cycle.  Attack is typically fast, decay and
>release slower.  An attack time of 20 milliseconds is 1/50 of a second,
>i.e. 50 Hz, up in the AC range for sure.  But a 500 millisecond release
>time is only 2 Hz.  So it is an AC signal with some very low frequency
>parts.

The output of an MOTM 800 is not an AC signal.  It is a positive 
voltage that never crosses zero.  AC stands for "alternating current" 
-- it "pushes" current one way when the voltage is positive, and 
"draws" current the other way when its voltage is negative.

Most audio inputs are AC coupled.  Their inputs have "blocking 
capacitors" to remove DC offset.  Most CV inputs are DC coupled -- 
they can accept both AC and DC signals.  There is a difference 
between AC and DC coupling when doing ring modulation, oscillator FM, 
etc., but understanding this difference is very important when using 
patching filters and VCA's.  The modulation inputs are DC coupled, 
but the signal inputs are usually AC coupled.

For example, one VCA in each of the MOTM dual VCA's can be used to 
process control voltages (the 110 has a switch allowing you to bypass 
the capacitors) so that they can be used to control modulation depth. 
The other VCA in each dual VCA is AC coupled, so it can only be used 
to process audio signals.  The 820 lag processor functions as 
DC-coupled low-pass filter.  It can be used to process control 
voltages and audio signals.  Afaik, all of the other MOTM filters 
have AC coupled inputs, so they cannot process control voltages.  If 
you try to ignore this, your patch won't "work".

On Friday 09 January 2004 13:01, Richard Brewster wrote:
> Here is a simple explanation:  DC (direct curent) means a
> steady, unchanging voltage, which could be any voltage
> level.  If a circuit is DC-coupled it means that the
> level of an unchanging voltage at the input will be
> passed through.  If a circuit is AC coupled then only
> changing voltages will be passed and there will be a
> "high pass" function applied, i.e. there will be a
> frequency below which signals won't pass. This cutoff
> frequency is normally about 20 Hz  for audio processing. 
> If you look at a circuit diagram, you will see a
> capacitor in series with the signal path if it is AC
> coupled.  Both inputs and outputs can be AC coupled.

I see. Is there also a low pass function applied on AC 
coupled circuits, at say 20 kHz?

> If you feed a DC voltage, say +5V, into an AC coupled
> input, the output will be zero.  You can always feed an
> AC signal into a DC coupled input.  

Ah, this helps! So I can go anywhere on my modular with an 
AC signal, on both AC and DC coupled inputs. Is that 
because a DC coupled input 'sees' the AC signal as a 
quickly changing series of 'unchanging voltage levels'?

> That works fine.  So why not always just use DC coupling?
>
> The benefit of AC coupling is that it removes any DC
> offset (constant) voltage.  For example, the signal input
> of a VCA is sometimes AC coupled, because a DC offset at
> the signal input can result in a popping sound when it is
> multiplied with the control input.

Okay.

>  The control input of
> a VCA has to be DC coupled, or else a prolonged
> unchanging voltage, such as the sustain portion of an
> envelope, would not produce the desired constant output
> level.

Understood.

Thanks Richard, that helped a lot!
- Robert

On Friday 09 January 2004 20:21, Mark wrote:

> The output of an MOTM 800 is not an AC signal.  It is a
> positive voltage that never crosses zero.  AC stands for
> "alternating current" -- it "pushes" current one way when
> the voltage is positive, and "draws" current the other
> way when its voltage is negative.

Ah, I see.

>
> Most audio inputs are AC coupled.  Their inputs have
> "blocking capacitors" to remove DC offset.  Most CV
> inputs are DC coupled -- they can accept both AC and DC
> signals. 

Yes, that's what I just learned from Richards post. Never 
new this, but it's an eye opener for me.

>  There is a difference between AC and DC
> coupling when doing ring modulation, oscillator FM, etc.,
> but understanding this difference is very important when
> using patching filters and VCA's.  The modulation inputs
> are DC coupled, but the signal inputs are usually AC
> coupled.
>
> For example, one VCA in each of the MOTM dual VCA's can
> be used to process control voltages (the 110 has a switch
> allowing you to bypass the capacitors) so that they can
> be used to control modulation depth. The other VCA in
> each dual VCA is AC coupled, so it can only be used to
> process audio signals.  

Aha, that's why Paul recently said that only VCA #2 in the 
190 can be used for DC signals. That input must be DC 
coupled, while input #1 is AC coupled, I assume. So that 
means that input #2 of the 190 cannot handle DC shift and 
could cause pops in the sound. Correct?

> The 820 lag processor functions
> as DC-coupled low-pass filter.  It can be used to process
> control voltages and audio signals.  Afaik, all of the
> other MOTM filters have AC coupled inputs, so they cannot
> process control voltages.  If you try to ignore this,
> your patch won't "work".

I'm starting to understand this now.

I never understood the fact that I can apply an LFO to a DC 
coupled (CV) input. I always assumed that an LFO must be a 
DC signal, since it was used for DC coupled inputs. But an 
LFO is just a slow moving audio signal, very AC. I didn't 
understand. But now that I know that DC coupled inputs also 
accept AC signals, the whole LFO issue is finally 
understood. :)

Thanks,
Robert

--- In motm@yahoogroups.com, Mark <yahoogroups@p...> wrote:
> 
> The output of an MOTM 800 is not an AC signal.  It is a positive 
> voltage that never crosses zero.  AC stands for "alternating
> current" -- it "pushes" current one way when the voltage is
> positive, and "draws" current the other way when its voltage is
> negative.

An EG could be classified as a DC source or an AC source depending 
upon which part of the envelope cycle is active.  In the "off" state 
(at zero) and in the sustain state the EG outputs a constant DC 
voltage.  In the attack, decay, and release state the EG outputs a 
time varying voltage which is essentially an AC waveform.

A signal does not need to "cross zero" to be considered AC.  For 
example, you could take a sine wave (AC) that is 2 volts peak to peak 
and add a 1 volt DC offset.  You still have a sine wave (AC) but 
instead of going from -1 volts to +1 volts, you now have a waveform 
that goes from 0 volts to 2 volts.  (It is still AC but it never 
crosses zero.)

Using that same example, we can think about AC coupled and DC coupled 
inputs.

If an input is AC coupled, it rejects DC offset.  If we feed our 
"normal" -1 to +1 sine wave into an AC coupled input, there is 
essentially no change in the waveform.  (It remains a -1 to +1 sine 
wave.)  On the other hand, if we feed our 0 to +2 sine wave into an AC 
couple input, the DC offset of 1 volt is eliminated by the AC coupling 
and what results is a "normal" -1 to +1 sine wave.

For DC coupled inputs, any DC offset remains in the signal.

On 1/9/04, Robert van der Kamp put forth:
>I see. Is there also a low pass function applied on AC
>coupled circuits, at say 20 kHz?

It is a high pass filter, and the frequency is largely determined by 
the size of the cap.

(However, a capacitor can also be used in a different configuration 
as a low pass filter to reduce RF noise on low level inputs.  If no 
other filters were used, it would still pass DC.)

>Aha, that's why Paul recently said that only VCA #2 in the
>190 can be used for DC signals. That input must be DC
>coupled, while input #1 is AC coupled, I assume. So that
>means that input #2 of the 190 cannot handle DC shift and
>could cause pops in the sound. Correct?

I wouldn't say it couldn't handle an offset signal, but afaik, it's 
not going to remove that offset.

On 1/9/04, strohs56k put forth:
>
>An EG could be classified as a DC source or an AC source depending
>upon which part of the envelope cycle is active.  In the "off" state
>(at zero) and in the sustain state the EG outputs a constant DC
>voltage.  In the attack, decay, and release state the EG outputs a
>time varying voltage which is essentially an AC waveform.

Then you are using a different definition of the term.  The voltage 
between the output and ground does vary, but the current doesn't 
alternate.  I'm not trying to get into a semantic argument, but 
merely offer advice on using the system.

In practice, if I switch the capacitors in series with both inputs of 
the multiplier section of the 110, the output of an 800 does not 
result in a working VCA.  One input must be DC-coupled for what I 
consider a DC signal.  Likewise, I could patch the output of the 800 
to the audio input of a filter and trigger the EG to "ring" that 
filter, but I'm not going to be able to adjust a continuous flowing 
effect that follows the voltage of the EG.

>If an input is AC coupled, it rejects DC offset.  If we feed our
>"normal" -1 to +1 sine wave into an AC coupled input, there is
>essentially no change in the waveform.  (It remains a -1 to +1 sine
>wave.)  On the other hand, if we feed our 0 to +2 sine wave into an AC
>couple input, the DC offset of 1 volt is eliminated by the AC coupling
>and what results is a "normal" -1 to +1 sine wave.
>
>For DC coupled inputs, any DC offset remains in the signal.

Right, we agree on what is happening.  We can look at that as a AC 
signal riding on a DC offset -- that's what you would get if you 
summed a bias voltage and AC signal on DC-coupled mixer.

Two points:

1. We have started discussing AC vs DC signals. 

AC usually implies periodic cycling or repeating of the voltage 
pattern, such as a VCO waveform. However, an envelope generator that 
normally puts out a transient DC voltage can put out an AC voltage if 
being repeatedly triggered from an LFO, for instance.

If a cyclic waveform is symmetrical around zero volts, it is pure AC. 
If not, it is AC with a DC component.

2. Think of DC coupled inputs and outputs as having the ability to 
handle DC and/or AC. They pass everything on exactly as it came in.

An AC coupled input or output will filter out the DC component in the 
signal but pass on the AC component. Apply a pure DC waveform to an 
AC coupled input and it will have no effect, because it is completely 
filtered.

Moe
http://www.hotrodmotm.com