Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss
Arbitrary delay in LTspice
3,115 views
Skip to first unread message
Phil Hobbs
Apr 8, 2015, 4:49:35 AM4/8/15
to
Awhile back we were trying to figure out how to make a voltage-variable
delay line in LTspice. Turns out there is a way to use one signal to
delay another: there's an undocumented function delay(x,y). Unlike the
documented one, delay y can be a simulation variable, like so:
Version 4
SHEET 1 880 680
WIRE -96 112 -112 112
WIRE 96 112 80 112
WIRE 112 112 96 112
WIRE 304 112 288 112
WIRE 80 144 80 112
WIRE 288 144 288 112
WIRE -112 160 -112 112
WIRE -112 256 -112 240
WIRE 80 256 80 224
WIRE 288 256 288 224
FLAG 80 256 0
FLAG 288 256 0
FLAG 96 112 input
FLAG 288 112 delayed
FLAG -112 256 0
FLAG -112 112 delay
SYMBOL voltage 80 128 R0
WINDOW 3 24 98 Left 2
WINDOW 123 0 0 Left 2
WINDOW 39 0 0 Left 2
SYMATTR InstName V1
SYMATTR Value SINE(0 1 0.5)
SYMBOL bv 288 128 R0
SYMATTR InstName B1
SYMATTR Value V=delay(v(input),v(delay))
SYMBOL voltage -112 144 R0
WINDOW 3 -75 149 Left 2
WINDOW 123 0 0 Left 2
WINDOW 39 0 0 Left 2
SYMATTR InstName V2
SYMATTR Value PULSE(0 4 0.2 3 3 1)
TEXT 208 312 Left 2 !.tran 5
TEXT 168 48 Left 2 ;This shows how to make a variable delay.
Cheers
Phil Hobbs
--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC
Optics, Electro-optics, Photonics, Analog Electronics
160 North State Road #203
Briarcliff Manor NY 10510
hobbs at electrooptical dot net
http://electrooptical.net
delay line in LTspice. Turns out there is a way to use one signal to
delay another: there's an undocumented function delay(x,y). Unlike the
documented one, delay y can be a simulation variable, like so:
Version 4
SHEET 1 880 680
WIRE -96 112 -112 112
WIRE 96 112 80 112
WIRE 112 112 96 112
WIRE 304 112 288 112
WIRE 80 144 80 112
WIRE 288 144 288 112
WIRE -112 160 -112 112
WIRE -112 256 -112 240
WIRE 80 256 80 224
WIRE 288 256 288 224
FLAG 80 256 0
FLAG 288 256 0
FLAG 96 112 input
FLAG 288 112 delayed
FLAG -112 256 0
FLAG -112 112 delay
SYMBOL voltage 80 128 R0
WINDOW 3 24 98 Left 2
WINDOW 123 0 0 Left 2
WINDOW 39 0 0 Left 2
SYMATTR InstName V1
SYMATTR Value SINE(0 1 0.5)
SYMBOL bv 288 128 R0
SYMATTR InstName B1
SYMATTR Value V=delay(v(input),v(delay))
SYMBOL voltage -112 144 R0
WINDOW 3 -75 149 Left 2
WINDOW 123 0 0 Left 2
WINDOW 39 0 0 Left 2
SYMATTR InstName V2
SYMATTR Value PULSE(0 4 0.2 3 3 1)
TEXT 208 312 Left 2 !.tran 5
TEXT 168 48 Left 2 ;This shows how to make a variable delay.
Cheers
Phil Hobbs
--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC
Optics, Electro-optics, Photonics, Analog Electronics
160 North State Road #203
Briarcliff Manor NY 10510
hobbs at electrooptical dot net
http://electrooptical.net
John Larkin
Apr 8, 2015, 11:55:29 AM4/8/15
to
On Wed, 08 Apr 2015 04:49:30 -0400, Phil Hobbs
<pcdhSpamM...@electrooptical.net> wrote:
>Awhile back we were trying to figure out how to make a voltage-variable
>delay line in LTspice. Turns out there is a way to use one signal to
>delay another: there's an undocumented function delay(x,y). Unlike the
>documented one, delay y can be a simulation variable, like so:
>
It avoids complications by refusing to do negative delay.
<pcdhSpamM...@electrooptical.net> wrote:
>Awhile back we were trying to figure out how to make a voltage-variable
>delay line in LTspice. Turns out there is a way to use one signal to
>delay another: there's an undocumented function delay(x,y). Unlike the
>documented one, delay y can be a simulation variable, like so:
>
We're about to ship the first unit of an 8-channel waveform playback
box. It stores big waveform files on an SD card and each channel has a
DDS-clocked FIFO and a DAC. We added a programmable time-shift
parameter per channel, that allows us to tweak time alignment, up to
+-999.999 samples, fractional part interpolated, with obvious
causality concerns. We pad the waveform data with 1024 zero samples,
when we open the file and again at EOF.
How did you find the delay() function? I bet LT Spice has all sorts of
undocumented tricks. Even the documented features aren't usually very
well documented.
--
John Larkin Highland Technology, Inc
picosecond timing laser drivers and controllers
jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com
Jim Thompson
Apr 8, 2015, 12:00:05 PM4/8/15
to
On Wed, 08 Apr 2015 08:55:28 -0700, John Larkin
<jla...@highlandtechnology.com> wrote:
>On Wed, 08 Apr 2015 04:49:30 -0400, Phil Hobbs
><pcdhSpamM...@electrooptical.net> wrote:
>
>>Awhile back we were trying to figure out how to make a voltage-variable
>>delay line in LTspice. Turns out there is a way to use one signal to
>>delay another: there's an undocumented function delay(x,y). Unlike the
>>documented one, delay y can be a simulation variable, like so:
>>
>
>
>It avoids complications by refusing to do negative delay.
>
>We're about to ship the first unit of an 8-channel waveform playback
>box. It stores big waveform files on an SD card and each channel has a
>DDS-clocked FIFO and a DAC. We added a programmable time-shift
>parameter per channel, that allows us to tweak time alignment, up to
>+-999.999 samples, fractional part interpolated, with obvious
>causality concerns. We pad the waveform data with 1024 zero samples,
>when we open the file and again at EOF.
>
>How did you find the delay() function? I bet LT Spice has all sorts of
>undocumented tricks. Even the documented features aren't usually very
>well documented.
I have a collection of manuals, tutorials and undocumented features in
<jla...@highlandtechnology.com> wrote:
>On Wed, 08 Apr 2015 04:49:30 -0400, Phil Hobbs
><pcdhSpamM...@electrooptical.net> wrote:
>
>>Awhile back we were trying to figure out how to make a voltage-variable
>>delay line in LTspice. Turns out there is a way to use one signal to
>>delay another: there's an undocumented function delay(x,y). Unlike the
>>documented one, delay y can be a simulation variable, like so:
>>
>
>
>It avoids complications by refusing to do negative delay.
>
>We're about to ship the first unit of an 8-channel waveform playback
>box. It stores big waveform files on an SD card and each channel has a
>DDS-clocked FIFO and a DAC. We added a programmable time-shift
>parameter per channel, that allows us to tweak time alignment, up to
>+-999.999 samples, fractional part interpolated, with obvious
>causality concerns. We pad the waveform data with 1024 zero samples,
>when we open the file and again at EOF.
>
>How did you find the delay() function? I bet LT Spice has all sorts of
>undocumented tricks. Even the documented features aren't usually very
>well documented.
LTspiceTutorials.zip on the Simulation Tools & Macros Page of my
website.
...Jim Thompson
--
| James E.Thompson | mens |
| Analog Innovations | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| San Tan Valley, AZ 85142 Skype: skypeanalog | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |
I love to cook with wine. Sometimes I even put it in the food.
Phil Hobbs
Apr 8, 2015, 12:13:44 PM4/8/15
to
On 04/08/2015 11:55 AM, John Larkin wrote:
> On Wed, 08 Apr 2015 04:49:30 -0400, Phil Hobbs
> <pcdhSpamM...@electrooptical.net> wrote:
>
>> Awhile back we were trying to figure out how to make a voltage-variable
>> delay line in LTspice. Turns out there is a way to use one signal to
>> delay another: there's an undocumented function delay(x,y). Unlike the
>> documented one, delay y can be a simulation variable, like so:
>>
>
>
> It avoids complications by refusing to do negative delay.
>
> We're about to ship the first unit of an 8-channel waveform playback
> box. It stores big waveform files on an SD card and each channel has a
> DDS-clocked FIFO and a DAC. We added a programmable time-shift
> parameter per channel, that allows us to tweak time alignment, up to
> +-999.999 samples, fractional part interpolated, with obvious
> causality concerns. We pad the waveform data with 1024 zero samples,
> when we open the file and again at EOF.
>
> How did you find the delay() function? I bet LT Spice has all sorts of
> undocumented tricks. Even the documented features aren't usually very
> well documented.
>
>
It was in the undocumented features file on the LTspice Yahoo group,
> On Wed, 08 Apr 2015 04:49:30 -0400, Phil Hobbs
> <pcdhSpamM...@electrooptical.net> wrote:
>
>> Awhile back we were trying to figure out how to make a voltage-variable
>> delay line in LTspice. Turns out there is a way to use one signal to
>> delay another: there's an undocumented function delay(x,y). Unlike the
>> documented one, delay y can be a simulation variable, like so:
>>
>
>
> It avoids complications by refusing to do negative delay.
>
> We're about to ship the first unit of an 8-channel waveform playback
> box. It stores big waveform files on an SD card and each channel has a
> DDS-clocked FIFO and a DAC. We added a programmable time-shift
> parameter per channel, that allows us to tweak time alignment, up to
> +-999.999 samples, fractional part interpolated, with obvious
> causality concerns. We pad the waveform data with 1024 zero samples,
> when we open the file and again at EOF.
>
> How did you find the delay() function? I bet LT Spice has all sorts of
> undocumented tricks. Even the documented features aren't usually very
> well documented.
>
>
under files/FAQ.
http://tinyurl.com/kxrzmup
Jim Thompson
Apr 8, 2015, 1:30:03 PM4/8/15
to
>Awhile back we were trying to figure out how to make a voltage-variable
>delay line in LTspice. Turns out there is a way to use one signal to
>delay another: there's an undocumented function delay(x,y). Unlike the
>documented one, delay y can be a simulation variable, like so:
>
>Version 4
>SHEET 1 880 680
>WIRE -96 112 -112 112
>WIRE 96 112 80 112
>WIRE 112 112 96 112
[snip]
>delay line in LTspice. Turns out there is a way to use one signal to
>delay another: there's an undocumented function delay(x,y). Unlike the
>documented one, delay y can be a simulation variable, like so:
>
>Version 4
>SHEET 1 880 680
>WIRE -96 112 -112 112
>WIRE 96 112 80 112
>WIRE 112 112 96 112
>SYMATTR Value PULSE(0 4 0.2 3 3 1)
>TEXT 208 312 Left 2 !.tran 5
>TEXT 168 48 Left 2 ;This shows how to make a variable delay.
>
>
>Cheers
>
>Phil Hobbs
What's that supposed to do? When I run your file I get bizarre
>TEXT 208 312 Left 2 !.tran 5
>TEXT 168 48 Left 2 ;This shows how to make a variable delay.
>
>
>Cheers
>
>Phil Hobbs
results...
<http://www.analog-innovations.com/SED/HobbsDelay_2015-04-08_10-25-36.jpg>
almost like an FM function ??
Jim Thompson
Apr 8, 2015, 1:37:48 PM4/8/15
to
On Wed, 08 Apr 2015 10:29:58 -0700, Jim Thompson
<To-Email-Use-Th...@On-My-Web-Site.com> wrote:
>On Wed, 08 Apr 2015 04:49:30 -0400, Phil Hobbs
><pcdhSpamM...@electrooptical.net> wrote:
>
>>Awhile back we were trying to figure out how to make a voltage-variable
>>delay line in LTspice. Turns out there is a way to use one signal to
>>delay another: there's an undocumented function delay(x,y). Unlike the
>>documented one, delay y can be a simulation variable, like so:
>>
>>Version 4
>>SHEET 1 880 680
>>WIRE -96 112 -112 112
>>WIRE 96 112 80 112
>>WIRE 112 112 96 112
>[snip]
>>SYMATTR Value PULSE(0 4 0.2 3 3 1)
>>TEXT 208 312 Left 2 !.tran 5
>>TEXT 168 48 Left 2 ;This shows how to make a variable delay.
>>
>>
>>Cheers
>>
>>Phil Hobbs
>
>What's that supposed to do? When I run your file I get bizarre
>results...
>
><http://www.analog-innovations.com/SED/HobbsDelay_2015-04-08_10-25-36.jpg>
>
>almost like an FM function ??
>
> ...Jim Thompson
Works OK when you don't slew the delay so fast relative to the
<To-Email-Use-Th...@On-My-Web-Site.com> wrote:
>On Wed, 08 Apr 2015 04:49:30 -0400, Phil Hobbs
><pcdhSpamM...@electrooptical.net> wrote:
>
>>Awhile back we were trying to figure out how to make a voltage-variable
>>delay line in LTspice. Turns out there is a way to use one signal to
>>delay another: there's an undocumented function delay(x,y). Unlike the
>>documented one, delay y can be a simulation variable, like so:
>>
>>Version 4
>>SHEET 1 880 680
>>WIRE -96 112 -112 112
>>WIRE 96 112 80 112
>>WIRE 112 112 96 112
>[snip]
>>SYMATTR Value PULSE(0 4 0.2 3 3 1)
>>TEXT 208 312 Left 2 !.tran 5
>>TEXT 168 48 Left 2 ;This shows how to make a variable delay.
>>
>>
>>Cheers
>>
>>Phil Hobbs
>
>What's that supposed to do? When I run your file I get bizarre
>results...
>
><http://www.analog-innovations.com/SED/HobbsDelay_2015-04-08_10-25-36.jpg>
>
>almost like an FM function ??
>
> ...Jim Thompson
incoming signal... then you can see it working properly.
Phil Hobbs
Apr 8, 2015, 2:29:13 PM4/8/15
to
On 04/08/2015 01:29 PM, Jim Thompson wrote:
> On Wed, 08 Apr 2015 04:49:30 -0400, Phil Hobbs
> <pcdhSpamM...@electrooptical.net> wrote:
>
>> Awhile back we were trying to figure out how to make a voltage-variable
>> delay line in LTspice. Turns out there is a way to use one signal to
>> delay another: there's an undocumented function delay(x,y). Unlike the
>> documented one, delay y can be a simulation variable, like so:
>>
>> Version 4
>> SHEET 1 880 680
>> WIRE -96 112 -112 112
>> WIRE 96 112 80 112
>> WIRE 112 112 96 112
> [snip]
>> SYMATTR Value PULSE(0 4 0.2 3 3 1)
>> TEXT 208 312 Left 2 !.tran 5
>> TEXT 168 48 Left 2 ;This shows how to make a variable delay.
>>
>>
>> Cheers
>>
>> Phil Hobbs
>
> What's that supposed to do? When I run your file I get bizarre
> results...
>
> <http://www.analog-innovations.com/SED/HobbsDelay_2015-04-08_10-25-36.jpg>
>
> almost like an FM function ??
>
> ...Jim Thompson
>
Right. It has a sine wave source, a ramp, and a behavioural voltage
> On Wed, 08 Apr 2015 04:49:30 -0400, Phil Hobbs
> <pcdhSpamM...@electrooptical.net> wrote:
>
>> Awhile back we were trying to figure out how to make a voltage-variable
>> delay line in LTspice. Turns out there is a way to use one signal to
>> delay another: there's an undocumented function delay(x,y). Unlike the
>> documented one, delay y can be a simulation variable, like so:
>>
>> Version 4
>> SHEET 1 880 680
>> WIRE -96 112 -112 112
>> WIRE 96 112 80 112
>> WIRE 112 112 96 112
> [snip]
>> SYMATTR Value PULSE(0 4 0.2 3 3 1)
>> TEXT 208 312 Left 2 !.tran 5
>> TEXT 168 48 Left 2 ;This shows how to make a variable delay.
>>
>>
>> Cheers
>>
>> Phil Hobbs
>
> What's that supposed to do? When I run your file I get bizarre
> results...
>
> <http://www.analog-innovations.com/SED/HobbsDelay_2015-04-08_10-25-36.jpg>
>
> almost like an FM function ??
>
> ...Jim Thompson
>
source which is the sine wave delayed by the value of the ramp voltage
in volts. The flat spot in the middle is where there's no data for the
BV to use, but it catches up again at the end.
Jim Thompson
Apr 11, 2015, 11:11:19 AM4/11/15
to
On Wed, 08 Apr 2015 14:29:08 -0400, Phil Hobbs
Pondering my navel, I was wondering how to implement such a creature
in PSpice.
I can't.
Mikey can because he writes the executable and can make a device model
that looks back into the data stream.
But I can make a voltage-controlled all-pass that will be quite useful
for simulating DLL's (delay-locked loops).
Watch for it on the Device Models & Subcircuits Page of my website.
in PSpice.
I can't.
Mikey can because he writes the executable and can make a device model
that looks back into the data stream.
But I can make a voltage-controlled all-pass that will be quite useful
for simulating DLL's (delay-locked loops).
Watch for it on the Device Models & Subcircuits Page of my website.
0 new messages