MOTM

I've been trying to determine the proper algorithm for a gate and/or 
trigger delay.  Heres the dilemma.
I assume that the length of the gate should be preserved, so the 
trailing edge of the gate needs to be delayed by an amount equal to 
the leading edge.  Delaying a trigger is easier since its just a 
very short pulse, but if a module is to allow for the delay of both 
gates and triggers, the delay time for the gate and the trigger must 
be the same with reasonable precision.
The first question is what happens if the incoming gate ends before 
the delayed gate has been produced (assume a delay of 3 seconds with 
a gate length of 1 second)  is the delayed gate still produced at 
the output?  or must the incoming gate be present to enable the 
delayed gate to be produced at the output?
This leads to the second question.  If the delayed gate is to be 
produced at the output even if the incoming gate stops, what happens 
if the key (or other event) is pushed a second time before the 
delayed gate is output?  Should the relationship between the leading 
edgeof the first gate, the trailing edge of the first gate and the 
leading edge of the second gate be preserved?
This is important because it seems to me if multiple events are to 
be all delayed and fed through to the delayed output, the circuit 
must be a sort of 1-bit shift register with the delay control either 
selecting a tap or controlling the clock speed.  If multiple events 
are not to be preserved a simple analog circuit involving a 
capacitor for timing and a couple of analog switches will do the job.
I realize this could be implemented using a PIC and in fact I think 
one MOTMer has already done this, but even if a PIC is used the 
question is still valid as the algorithm must still be worked out.
I'd certainly like to hear everyones opinion on this one.

There's no "right answer" to how gate and trigger delays should work; 
everybody has their own opinion. What follows is my opinion on this 
topic.

At 3:56 PM +0000 2003/05/15, paulhaneberg wrote:
>The first question is what happens if the incoming gate ends before
>the delayed gate has been produced (assume a delay of 3 seconds with
>a gate length of 1 second)  is the delayed gate still produced at
>the output?

Yes.

>or must the incoming gate be present to enable the
>delayed gate to be produced at the output?

No.

>This leads to the second question.  If the delayed gate is to be
>produced at the output even if the incoming gate stops, what happens
>if the key (or other event) is pushed a second time before the
>delayed gate is output?

You get a second gate.

>Should the relationship between the leading
>edgeof the first gate, the trailing edge of the first gate and the
>leading edge of the second gate be preserved?

Yes.

>This is important because it seems to me if multiple events are to
>be all delayed and fed through to the delayed output, the circuit
>must be a sort of 1-bit shift register with the delay control either
>selecting a tap or controlling the clock speed.  If multiple events
>are not to be preserved a simple analog circuit involving a
>capacitor for timing and a couple of analog switches will do the job.

Multiple events need to be preserved.

Here's a sneak preview of my logic delay line that also works as a 
gate/trigger delay:

http://www.tellun.com/motm/diy/tln774/TLN-774.html

This panel mockup is already a little out of date, but the thing does 
preserve both edges of a signal and stores multiple events. In my 
opinion, that's the only way to go.

AFAIK, The Old Crow has been working on a Gate/Trigger Delay to be released by
Paul this year or next.  I haven't heard anything on this lately, so this may
no longer be true.  I think this is a VERY important module for the MOTM
line-up.  It was one of the few remaining holes in the short list of
basic/essential modules.  In addition to the gate delay, these include:  triple
pre-amp and envelope follower.

Let's hope this fall and winter provide a bountiful harvest of these
"essential" modules.

And yes, I know that no one will agree with this 100%, but I didn't add my 1/50
Euro a couple months ago when Paul requested.

--Shemp

__________________________________
Do you Yahoo!?
The New Yahoo! Search - Faster. Easier. Bingo.
http://search.yahoo.com

On Friday 16 May 2003 00:19, Tentochi wrote:
> In addition to the gate delay, these include:  triple
> pre-amp and envelope follower.

The triple pre-amp I read about once in a while, what does 
it do?

- Robert

On Friday 16 May 2003 07:57, Paul Schreiber wrote:
> It's a 1U wide module with 3 preamps. 1 is for
> guitar/bass, the other 2 are "line-level". They are all
> DC path (servo'd). It's coming this fall.

Ah, thanks. The guitar/bass input would certainly be useful 
for me. :)

- Robert

It's a 1U wide module with 3 preamps. 1 is for guitar/bass, the other 2 are "line-level". They
are all DC path (servo'd). It's coming this fall.

Paul S.

----- Original Message -----

From: "Robert van der Kamp" <robnet@...>
To: <motm@yahoogroups.com>
Sent: Thursday, May 15, 2003 11:50 PM
Subject: Re: [motm] Gate and Trigger Delays


> On Friday 16 May 2003 00:19, Tentochi wrote:
> > In addition to the gate delay, these include:  triple
> > pre-amp and envelope follower.
>
> The triple pre-amp I read about once in a while, what does
> it do?
>
> - Robert
>
>
>
>
> Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
>
>

Paul Haneberg writes:
>>I've been trying to determine the proper algorithm for a gate and/or 
trigger delay.<<

If I were to do it, this is how.  When the gate delay module receives the start 
of a gate, I'd write a counter into a fifo buffer in memory.  The counter is 
set to whatever delay time the user selects based on some number of clocks per 
second (say 1024 ticks per second).  When the gate ends, I'd write a second 
count representing the length of the gate into a word right after the first 
counter.  You could store hundreds of gates this way, not that you need 
anywhere that many.  When the first counter decrements to zero, you put out a 
gate signal and then start decrementing the second gate-time counter. When it 
hits zero you close down the gate.  Most high-end digital sequencers only have 
a resolution of about 1/1024 of a second, so you would use a time resolution 
like that to decrement counters and store gate times.

BTW, great Jerusalem rendition.

-Elhardt

>There's no "right answer" to how gate and trigger delays should work;
>everybody has their own opinion.

I agree :)

>  >This is important because it seems to me if multiple events are to
>  >be all delayed and fed through to the delayed output, the circuit
>  >must be a sort of 1-bit shift register with the delay control either
>  >selecting a tap or controlling the clock speed.  If multiple events
>  >are not to be preserved a simple analog circuit involving a
>  >capacitor for timing and a couple of analog switches will do the job.

I have a circuit here that uses a flip-flop, two 555 timers, and a
couple of NOR gates.  It will preserve pulse width, if and only if
the pulse widths are less than the delay time and the period between
pulses is more than the delay time.

I do not think it is possible to exceed those limitations without
building anything less complicated than a "real" digital delay.
That's out of my league.  I can only cobble discrete CMOS together.
I don't monkey around with microprocessors.

I also do not think it is possible to do it with only one timer, and
you would need a current source since it's impossible to match (and
in most cases buy) a dual pot.  Imho, a modular approach with
separate timers, logic gates, edge detectors, etc. is much more
flexible.

Yes, you could use the tempo clock to drive shift registers to delay
gates, but only by a number of clock steps.  Which I don't think is
all that useful.

This leads to an important question, what the uses for a gate
delay??  If it is to delay things less than one beat, such as setting
back a snare or delaying  the onset of an envelope or LFO, then the
analogue solution will work.

>Multiple events need to be preserved.
>
>Here's a sneak preview of my logic delay line that also works as a
>gate/trigger delay:
>
>http://www.tellun.com/motm/diy/tln774/TLN-774.html
>
>This panel mockup is already a little out of date, but the thing does
>preserve both edges of a signal and stores multiple events. In my
>opinion, that's the only way to go.

Imho, if you are going to do that, you might as well make a digital
delay that can process control voltages as well as gates.  Let's say
an 8-bit delay instead of a 1-bit delay.  The sampling rate wouldn't
have to be that high for CV.  This way you could delay all the
information required to play a note (not just the gate) and could use
delay effects to modulate that CV.   A DC-coupled digital delay would
be very useful.  I think I saw a mock-up for an MOTM module like that
a long time ago.

>BTW, great Jerusalem rendition.

Definitely :)  Imho, I thought the drumming was no Palmer, but I
thought the singing was better than Lake.   Did you know Blake wrote
the lyrics??

At 11:11 AM -0400 5/16/03, media.nai@... wrote:
>This leads to an important question, what the uses for a gate
>delay??

Of course, that should read "what are the uses for a gate delay??"

> 
> I have a circuit here that uses a flip-flop, two 555 timers, and a
> couple of NOR gates.  It will preserve pulse width, if and only if
> the pulse widths are less than the delay time and the period 
between
> pulses is more than the delay time.
> 
> I do not think it is possible to exceed those limitations without
> building anything less complicated than a "real" digital delay.
> That's out of my league.  I can only cobble discrete CMOS together.
> I don't monkey around with microprocessors.
> 
> I also do not think it is possible to do it with only one timer, 
and
> you would need a current source since it's impossible to match (and
> in most cases buy) a dual pot.  Imho, a modular approach with
> separate timers, logic gates, edge detectors, etc. is much more
> flexible.

The circuit design I had worked out didn't use a current mirror or 
dual pot, it compared the voltage level of a charging cap (actually 
two of them) to a level set by a combination of input voltage and 
pot setting.  The first cap started charging on the gate leading 
edge, the second on the gate trailing edge.  

> 
> Yes, you could use the tempo clock to drive shift registers to 
delay
> gates, but only by a number of clock steps.  Which I don't think is
> all that useful.
> 
> This leads to an important question, what the uses for a gate
> delay??  If it is to delay things less than one beat, such as 
setting
> back a snare or delaying  the onset of an envelope or LFO, then the
> analogue solution will work.
> 
> 
> Imho, if you are going to do that, you might as well make a digital
> delay that can process control voltages as well as gates.  Let's 
say
> an 8-bit delay instead of a 1-bit delay.  The sampling rate 
wouldn't
> have to be that high for CV.  This way you could delay all the
> information required to play a note (not just the gate) and could 
use
> delay effects to modulate that CV.   A DC-coupled digital delay 
would
> be very useful.  I think I saw a mock-up for an MOTM module like 
that
> a long time ago.

I think a 1 bit delay line is in fact the way to go.  I haven't 
decided yet whether to have multiple clock ranges or a long delay 
line with the selection of the tap on a log scale.  I would agree 
that an 8 bit or perhaps 10 or 12 bit delay line clocked at roughly 
1Khz would be a good idea for delaying control voltages.  In fact I 
think there should be a whole family of time delay based modules.  
Perhaps the legendary 200 series?

> 
> >BTW, great Jerusalem rendition.
> 
> Definitely :)  Imho, I thought the drumming was no Palmer, but I
> thought the singing was better than Lake.   Did you know Blake 
wrote
> the lyrics??

I did not know who wrote the lyrics.  The singer I used is in fact 
an incredibly talented vocalist.  He has some operatic training and 
was warming up by switching between singing Merle Haggard and 
something from Wagner (thats the german classical composer, not 
Porter.)   If I recall there is a Monty Python episode where they 
sing Jerusalem while standing in a tub of water.

At 7:20 PM +0000 5/16/03, paulhaneberg wrote:
>
>The circuit design I had worked out didn't use a current mirror or
>dual pot, it compared the voltage level of a charging cap (actually
>two of them) to a level set by a combination of input voltage and
>pot setting.  The first cap started charging on the gate leading
>edge, the second on the gate trailing edge.

I think I understand.  You used two discrete comparators rather than
the built-in comparators of two timer chips.

>I think a 1 bit delay line is in fact the way to go.  I haven't
>decided yet whether to have multiple clock ranges or a long delay
>line with the selection of the tap on a log scale.  I would agree
>that an 8 bit or perhaps 10 or 12 bit delay line clocked at roughly
>1Khz would be a good idea for delaying control voltages.  In fact I
>think there should be a whole family of time delay based modules.

Imho, varying the clock speed would give you a finer range of
adjustment.  I'm sure there are schematics for "real" digital delays
on the web -- you would just have to make it DC-coupled.  You might
be able to rip the blocking caps out of an existing delay.

My thinking is if someone is going go through all of the work of
building a digital delay module, they might as well make it at least
8-bit.  I'm no digital designer, but compared to the cost of the
panel hardware, any increase in cost would be insignificant.

>  Perhaps the legendary 200 series?

Too bad there is a 100, 101, and 110, otherwise he could have named a
digital series 001, 010, 110, etc.  :)

This contest does not include the '900 which is the easiest module overall. 
And this does not include modules that do not come in kit form such as '910 and
'940.

The reigning champion has been consistently the '800.  My nomination is the
'380.  Yes, it does have a slightly higher part count (it's been I while since
I built any '800s), but the parts are VERY straight forward to insert into the
board.  And the parts are all in line, so it is easy to solder.  And the board
is easy to ready.  No coax or cable-ties on the board.  And finally, there are
very few parts to confuse--there are at least 2 of every resistor value, only
one type of small cap (so no magnifying glass needed), only two types of IC,
etc.

In some ways, I think this module deserves a 1/5 difficulty, but this is just
my opinion.  I had a very relaxing time building it.

Cheers!
Shemp

__________________________________
Do you Yahoo!?
The New Yahoo! Search - Faster. Easier. Bingo.
http://search.yahoo.com

I have posted a lot of this in the past, but it was on my mind and I wanted to
share with those who may not have seen this before.

These are just my opinions, so I hope others jump in and add to this or correct
my misconceptions.

If you plan on building a '300, '310 or '320 (I think that is the current
group), make sure you pick up some thermal compound first.  If you don't have
any of these or have not read the instructions (and built) yet, you will not be
aware of this.  Paul is not allowed to ship this item with the kits becuase of
postal regulations.  You pick it up at Radio Shack or from most of the supply
houses.  Here are a few links I found for interesting reading.  I think most
people are buying it for their CPU cooling units.  It can be reallly depressing
to get started and not have everything you need.  Maybe this will get added to
the tools lists out there at some point.

http://www.thetechzone.com/reviews/roundup/thermal_compound/

http://www.viperlair.com/reviews/case_cool/compounds/so_nov02/therm_so.shtml

http://www.heatsink-guide.com/compound.htm

Most axial components (resistors, diodes and ferrite beads) have a lead spacing
of 0.4".  Axial ceramics have a lead spacing of 0.3".  This in NOT documented
in any of the documentation to my knowledge.  It is a pain to bend everything
and then realize the caps don't fit.  Argh!  And your handing lead bending
guide probably only goes down to 0.4" (for 1/4W resistors).  If you use the end
of the guide, it is almost 0.3", but you have to hold the cap very securely. 
Scott Jusciw's PCBs inprove on this by making all of the lead spacings
appropriate to the size of the part.  Is there and "industry standard" for
this?

If you have some MOTM kits and kits (or bare PCBs) from another manufacturer,
build all of the MOTM kits you have first.  You will learn more from Paul's
kits then all of the other kits combined.  Paul's instructions and attention to
detailed are unparalleled without insulting your intelligence.  I am surprised
that other manufacturers haven't followed suit.  But you will become
knowledgable enough to tackle other projects after several MOTM kits--this is
the primary reason I have one of my jobs now.  Thanks Paul!!!

Build the MOTM-900 first.  Follow this by the '800 and then the '380.  This
should get your confidence up to tackle other modules.  Difficulty is related
to part count more than anything else.

Number one MOTM/DIY mantra.  Take your time!  As Paul so famously says, "This
is an accuracy contest, not a speed contest."  This is much more important than
dexterity, experience or knowledge.

Tools are VERY important.  If you buy Radio Shack tools, you will build with
Radio Shack quality and wonder why everything is so difficult and why you are
having so many problems.  Number one is your soldering iron.  Weller WTCPT is
the stock MOTM iron--buy it.  You will NOT need solder since it is included in
all kits.  Next, dykes.  You only need one pair of cutters (not two as
mentioned in some of the literature).  Next, a solder sucker.  Soldapult is
excellent because of its large internal volume.  Next, chain nose pliers.  You
do not need two pairs of pliers.  Next, hollow-shaft nut drivers.  You only
need 2 of these, not a whole set.  One size works for jacks and pots, the other
for switches.  I forget the 2 sizes off the top of my head--I just look at the
color.  Next, allen wrench.  You only need 1 of these, not a whole set.  This
is for tightening the set pin on the knobs.  Again, I forget the size.  For the
allen wrench and nut drivers, make sure you get the proper British size and not
a metric equivalent.  Next, a Lead bending guide.  I probably forgot one or two
things.

I bought a Pana-vise and used it like crazy for my first 20 or so modules. 
When I started doing production soldering work, I used it less and less.  Now,
I rarely use it.  I place a piece of foam under the boards.  When the parts you
are soldering are all at the same height, this makes sure not of the are
inadvertantly popping up.  This does not work well for smaller or light boards,
nor if you don't solder with a light touch.  I suggest getting a Pana-vise.

I don't think most new people need DVMs or oscilloscopes starting out.  These
are luxury items for beginners.  And then you have to know how to use them
properly.  A meter is good for beginners to check a resisor or capacitor value.
 Some may disagree with the necessity of these items.

Finally, the item you don't think you need, but really, really do need--the
heat gun!  You are more likely to mess up your module because you don't have
one than for any other reason.  Using a lighter means a high chance of
accidently melthing the wire casing or possibly a pot casing.  Hair dryers are
not hot enough.  Budget at least $65 for this, not $50.

Keep the new builder's guide that Paul gives you with your first kit close at
hand.  There is a wealth of good information in there!

Use high-quality patch cables.  Cheap cables (like the grey ones from Radio
Shack) are bound to give you a load of problems.  Some will be apparent that it
is the cables fault, and other may seem like bad power or problems with a
module.  If you aren't sure what to get, buy cables from Larry
(http://www.wiseguysynth.com/larry/cable/cable.htm) when he takes orders for
the next batch.  The official cable of MOTM.  They are incredible!

There is nothing sacred about the order Paul has you insert parts onto the
boards.  Always do the water-proof parts first as stated in the documentation. 
A general DIY tip is to start with low profile parts and work your way to
higher parts.  And since MOTM kits generally do not use sockets.  You may want
to solder your static sensitive parts (primarily ICs and transistors) last
before the final board wash.  Paul makes it incredibly easy to do!  Thanks
Paul.  If you are concerned, always follow Paul's instructions exactly.  As you
get some experience, you may find that switching things around a little makes
it easier for you.  For me, I always connect wires to the board before
board-mounted pots.  I feel the pots are someone fragile and bulky, so I like
to do them after the wires, especially after the coax.

That's it for now.  I look forward to hearing other opinions and experiences.

Cheers!
Shemp

__________________________________
Do you Yahoo!?
The New Yahoo! Search - Faster. Easier. Bingo.
http://search.yahoo.com

Easy to build and very cool addition to a rack.  The summed mode 
makes for some interesting evolving timbres when hook to things like 
FREQ modulation or RESONANCE...

--- In motm@yahoogroups.com, Tentochi <tentochi2003@y...> wrote:
> This contest does not include the '900 which is the easiest module 
overall. 
> And this does not include modules that do not come in kit form 
such as '910 and
> '940.
> 
> The reigning champion has been consistently the '800.  My 
nomination is the
> '380.  Yes, it does have a slightly higher part count (it's been I 
while since
> I built any '800s), but the parts are VERY straight forward to 
insert into the
> board.  And the parts are all in line, so it is easy to solder.  
And the board
> is easy to ready.  No coax or cable-ties on the board.  And 
finally, there are
> very few parts to confuse--there are at least 2 of every resistor 
value, only
> one type of small cap (so no magnifying glass needed), only two 
types of IC,
> etc.
> 
> In some ways, I think this module deserves a 1/5 difficulty, but 
this is just

> my opinion.  I had a very relaxing time building it.
> 
> Cheers!
> Shemp
> 
> __________________________________
> Do you Yahoo!?
> The New Yahoo! Search - Faster. Easier. Bingo.
> http://search.yahoo.com

I'm still pondering the idea of whether the gate is preserved if the 
key is lifted before the delay time is up.  
I looked through some of my dog-eared old ElectroNotes.  The 
examples of gate/trigger delays mentioned there all end the delayed 
gate immediately upon key release regardless of the delay time or 
gate length, so gate length is not preserved.
It also appears that my Andromeda with envelopes set to one of the 
normal modes also does not preserve gate length if the key is lifted 
before the delay period has passed, although there are other modes 
with other results.
Maybe I'll go back to my simple R/C timing circuit and save the one-
bit delay line for another module.  Then again maybe not.

Paul, IMO, I think a gate delay module should support both of these modes.
One mode ends the original gate and delayed gate at key release.  The 2nd
mode preserves the original gate length.  I see both seem useful.

Larry H

----- Original Message -----
From: paulhaneberg <phaneber@...>
To: <motm@yahoogroups.com>
Sent: Saturday, May 17, 2003 10:22 PM
Subject: [motm] Re: Gate and Trigger Delays


I'm still pondering the idea of whether the gate is preserved if the
key is lifted before the delay time is up.
I looked through some of my dog-eared old ElectroNotes.  The
examples of gate/trigger delays mentioned there all end the delayed
gate immediately upon key release regardless of the delay time or
gate length, so gate length is not preserved.
It also appears that my Andromeda with envelopes set to one of the
normal modes also does not preserve gate length if the key is lifted
before the delay period has passed, although there are other modes
with other results.
Maybe I'll go back to my simple R/C timing circuit and save the one-
bit delay line for another module.  Then again maybe not.





Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/

At 12:16 AM -0500 5/18/03, J. Larry Hendry wrote:
>Paul, IMO, I think a gate delay module should support both of these modes.
>One mode ends the original gate and delayed gate at key release.  The 2nd
>mode preserves the original gate length.  I see both seem useful.

All you need for that is a switch and an AND gate.