Delay
Dominique Jouniot
Hello,
Has someone an idea , how i can build an analogic delay (aroud 10 to 30
ms) for a audio signal.
I know how to do it in digital electronic, but i dont know if it is
possible with only analogic aproach.
Thanks
Best regards
Dominique
stanle...@cheerful.com.nospam
【 在 sci.electronics.design 見到 Dominique Jouniot <d...@worldnet.fr>
君於 Wed, 03 Jun 1998 15:39:02 +0200 所寫的大作中提到:】
You can do it with LC network. Same way as the digital (active) delay
lines but without the input and output inverter.
David Sherman
Dominique Jouniot wrote:
>
> Hello,
>
> Has someone an idea , how i can build an analogic delay (aroud 10 to 30
> ms) for a audio signal.
>
> I know how to do it in digital electronic, but i dont know if it is
> possible with only analogic aproach.
>
> Thanks
>
> Best regards
>
> Dominique
of approximation (like everything analog). Another name for what you are
looking for is an "all-pass" filter. I wish I could easily draw you a
schematic, but email doesn't do that. On the S-plane, you want zeros in
the right half-plane which are the mirror image of poles in the left
half-plane. Thus you can see that as you travel along the jw axis, the
amplitude doesn't change but the phase does.
The one-pole, one-zero version is pretty simple and easy to understand.
Start with a standard inverting opamp circuit with both resistors the
same value. Disconnect the + input from ground and insert a capacitor
between that input and ground. Connect a resistor from the + input to
the inverting configuration's input. Now it is intuitively obvious that
when the input frequency is very low, the circuit input voltage appears
at the + opamp input, hence there is no current through the resistor
going to the - opamp input, and the overall circuit has a gain of 1 and
no phase shift. At very high frequencies, the capacitor holds the +
input to ground and the circuit has a gain of 1 with 180 degrees phase
shift (inversion). In between, the gain stays constant (not intuitively
obvious, but write out the equations in S) but the phase varies with
frequency. A phase angle which increases linearly with frequency is a
time delay. This circuit is only approximately linear over a restricted
frequency range. You add more pole-zero pairs to get the linearity to be
better over a wider range.
"Operational Amplifiers: Theory and Practice" by Roberge includes a good
practical discussion of the Pade approximation to a time delay including
how to optimise the design with more pole-zero pairs.
You can also do a "pseudo-analog" approach and use a "bucket brigade
device" this is a MOS IC which contains a large number of small
capacitors (e.g. 1024 of them) with FET switches in between. The
switches are clocked in such a way that with each clock cycle, the
charge on one capacitor ("bucket") gets transferred to the next one. An
analog voltage fed into the input of the device appears at the output
some time later, the time being determined by the clock frequency and
the number of buckets. For audio, you must clock fast enough to preserve
the high frequency information, say 40 kHz min. Then you must have
enough buckets to provide your time delay (1024 stages would give you
about 26 mS). Good easy-to-use ones are made by EG&G Reticon. They may
be hard to find and run about $10 in small quantity. I think the
audiophile would complain that they "chop up" the sound or some such.
The truest analog delay is a delay line made of coaxial cable, but you'd
need about 6 miles for 30 mS and its hard to change the delay.
Good luck.
---David Sherman
David Sherman
stanle...@cheerful.com.nospam wrote:
>
> 【 在 sci.electronics.design 見到 Dominique Jouniot <d...@worldnet.fr>
> 君於 Wed, 03 Jun 1998 15:39:02 +0200 所寫的大作中提到:】
>
> >
> >Has someone an idea , how i can build an analogic delay (aroud 10 to 30
> >ms) for a audio signal.
> >
> >I know how to do it in digital electronic, but i dont know if it is
> >possible with only analogic aproach.
> >
> >Thanks
> >
> >Best regards
> >
> >Dominique
> >
>
> lines but without the input and output inverter.
The LC network is an approximation, placing zeroes in the right half of
the S plane to match poles on the left half. The more LC sections you
have, the better the approximation (that is, the wider the range of
frequencies over which the circuit approximates a constant time delay)
In the limit, where you have an infinite number of LC sections, you have
a perfect phase shifter, a resistive impedance looking into the circuit,
and you have a transmission line. Coax, twin-lead, etc can all be
considered to be a constant time delay. Coax is obviously better for
coiling up since the E-field stays inside. All told, an LC network with
lots of sections is likely to be difficult (i.e. expensive) for such a
long time delay. Iron-core coils are needed, which introduce their own
frequency-response problems, etc. The LC design theory is nonetheless
valid and can be translated into active circuitry by well-understood
methods.
---David Sherman
Sam Goldwasser
You can do it with a CCD but that is only a semi-analog approach :-).
--- sam : Sci.Electronics.Repair FAQ: http://www.repairfaq.org/
Latest Sam stuff: http://www.repairfaq.org/sam/
Lasers: http://www.repairfaq.org/sam/lasersam.htm
http://www.misty.com/~don/lasersam.html
Sam Goldwasser
But you will need order of hundreds of LCs to be able to delay an audio
signal with a 20 KHz bandwidth by 30 ms!
--- sam : Sci.Electronics.Repair FAQ: http://www.repairfaq.org/
Latest Sam stuff: http://www.repairfaq.org/sam/
Lasers: http://www.repairfaq.org/sam/lasersam.htm
http://www.misty.com/~don/lasersam.html
In article <35756de1...@news.hknet.com> stanle...@cheerful.com.nospam writes:
>Hello,
>
>Has someone an idea , how i can build an analogic delay (aroud 10 to 30
>ms) for a audio signal.
>
>I know how to do it in digital electronic, but i dont know if it is
>possible with only analogic aproach.
>
>Thanks
>
>Best regards
>
>Dominique
>
You can do it with LC network. Same way as the digital (active) delay
Sam Goldwasser
You are going to need a lot of poles and zeros to handle a 20 KHz audio
signal with a delay of 30 ms! Of course, if his bandwidth is limited to
20 Hz, it won't be so bad :-).
--- sam : Sci.Electronics.Repair FAQ: http://www.repairfaq.org/
Latest Sam stuff: http://www.repairfaq.org/sam/
Lasers: http://www.repairfaq.org/sam/lasersam.htm
http://www.misty.com/~don/lasersam.html
In article <357571EA...@virtual-cafe.com> David Sherman <dav...@virtual-cafe.com> writes:
Dominique Jouniot wrote:
>
> Hello,
>
> Has someone an idea , how i can build an analogic delay (aroud 10 to 30
> ms) for a audio signal.
>
> I know how to do it in digital electronic, but i dont know if it is
> possible with only analogic aproach.
>
> Thanks
>
> Best regards
>
> Dominique
Charlie Edmondson
The easiest analog ways to do this are the old fashioned ways, distance and
tape delays. Distance says you put a speaker and a microphone 10 to 30 ms
apart! Kinda lossy though... 8-)
Tape delays can be pretty precise, but I haven't seen one for a while. It
was simply a reel to reel tape deck using a long loop of tape. There was
an erase head, a record head, then one or more play heads. (Most had
multiple play heads for various delays and for multiple echo effects, some
had a single moveable record head for variable delay.) No recording studio
was complete without one!
Charlie
David Sherman
Okay, two messages in one here.
First in response to Sam Goldwasser:
>You are going to need a lot of poles and zeros to handle a 20 KHz audio
>signal with a delay of 30 ms! Of course, if his bandwidth is limited to
>20 Hz, it won't be so bad :-).
Maybe a ridiculous number of poles and zeroes. If I didn't have anything
else to do, maybe I'd calculate it. Long ago, when I played drums in a
garage band, our lead guitar player bought a "phlanger" which was a
little sound-effects box, the core of which was an analog time delay. I
traced out the schematic of it (all opamps and discretes back in 1976)
and it was a 5 stage (5 poles and 5 zeroes) approximation to a linear
time delay. Obviously the bandwidth of an electric guitar is something
less than half that required for the complete audio band, and distortion
is more allowable in a "special effects" box, and I don't know what the
overall time delay was, but in all it worked quite well.
Second, thinking about that "phlanger" reminded me of another analog
delay line which I saw in a Marshall tube guitar amp. It was basically a
long, fairly floppy, coiled spring attached at one end to a small
loudspeaker and at the other end to microphone. The delay was simply the
length of the spring divided by the speed of sound in the steel. It also
worked okay, but was subject to vibration, and the little speaker used
to drive it wasn't very good.
So, there's lots of ways to make an analog time delay.
---David Sherman
Eric Inazaki
Two ways I know:
1. A hell of a lot of coax.
2. A bucket brigade device.
In article <357551F6...@worldnet.fr>, Dominique Jouniot
Boris Mohar
On Wed, 03 Jun 1998 15:39:02 +0200, Dominique Jouniot <d...@worldnet.fr>
wrote:
>Hello,
>
>Has someone an idea , how i can build an analogic delay (aroud 10 to 30
>ms) for a audio signal.
>
>I know how to do it in digital electronic, but i dont know if it is
>possible with only analogic aproach.
>
>Thanks
>
>Best regards
>
>Dominique
Length of hose of hose would do.
Regards
Boris Mohar
VIATRACK printed circuit designs
Saleh
Dominique Jouniot wrote in message <357551F6...@worldnet.fr>...
>Hello,
>
>Has someone an idea , how i can build an analogic delay (aroud 10 to 30
>ms) for a audio signal.
>
>I know how to do it in digital electronic, but i dont know if it is
>possible with only analogic aproach.
>
>Thanks
>
>Best regards
>
>Dominique
>
There is another method: Convert your analog signal into modulated pulse
signal (for example, pulses with modulated width - of course, depending on
your analog bandwidth you have to calculate a frequency of pulses), send
these pulses to a delay circuit, convert back the sequence of pulses into
analog signal.
Saleh. sig...@hotmail.com
mes...@my-dejanews.com
In article <6l5fuf$ggg$1...@news01.iafrica.com>,
I couldn't get ahold of the original post on this subject so I will
post from here. I hope this hasen't been said already.
There is a device, made by Panasonic, called a "Bucket Brigade Device".
There is a variety of different types from 64 to 4096 stages. If you
look at http://www.digikey.com, on page 175 of their catalog, you can
get a better idea of what is avaliable. These devices are very simple
to use and have a variety of applications as fixed and variable delay
lines. They are really just analog shift registers that pass stored
charges from capacitor to capacitor, all controlled by a 2 phase
clock.
If you don't have access to this information, feel free to E-Mail
me, and I will send it to you.
Martin Meserve
Engineer, Program/Project Specialist
CAD/CAE Manager
Lockheed Martin - AZ
-----== Posted via Deja News, The Leader in Internet Discussion ==-----
http://www.dejanews.com/ Now offering spam-free web-based newsreading
stanle...@cheerful.com.nospam
On Thu, 4 Jun 1998 08:48:22 +0200 "Saleh" <sig...@hotmail.com> wrote:
>
>Dominique Jouniot wrote in message <357551F6...@worldnet.fr>...
>>Hello,
>>
>>Has someone an idea , how i can build an analogic delay (aroud 10 to 30
>>ms) for a audio signal.
>>
>>I know how to do it in digital electronic, but i dont know if it is
>>possible with only analogic aproach.
>>
>>Thanks
>>
>>Best regards
>>
>>Dominique
>>
>
>There is another method: Convert your analog signal into modulated pulse
>signal (for example, pulses with modulated width - of course, depending on
>your analog bandwidth you have to calculate a frequency of pulses), send
>these pulses to a delay circuit, convert back the sequence of pulses into
>analog signal.
>
>Saleh. sig...@hotmail.com
>
I think it is a good idea, but it will be very complicate and you have
to write your own program.
One of the simple method to implement an delay line for audio signal
is use mechanical delay line instead of electrical one. Connect two
audio tranducers to a spring of suitable length (remark: the length is
depend on the audio wave travelling speed over the steel times the
delay you need times 2*pi*radius of the spring).
Lasse Langwadt Christensen
In article <3575BC37...@irvine.orcad.com>,
Charlie Edmondson <Charles....@irvine.orcad.com> writes:
> The easiest analog ways to do this are the old fashioned ways, distance and
> tape delays. Distance says you put a speaker and a microphone 10 to 30 ms
> apart! Kinda lossy though... 8-)
>
> Tape delays can be pretty precise, but I haven't seen one for a while. It
> was simply a reel to reel tape deck using a long loop of tape. There was
> an erase head, a record head, then one or more play heads. (Most had
> multiple play heads for various delays and for multiple echo effects, some
> had a single moveable record head for variable delay.) No recording studio
> was complete without one!
>
> Charlie
U could just adjust the pseed of the tape to get different
delays. I actually think theres still people using them
because they can be pretty good if they get a new piece of
tape one in a while. It will probably be quite trendy, just
like dragging a big old leslie around :)
-- Lasse
----------------------------------------------------------
Lasse Langwadt Christensen, M.Sc. EE (to be in 1999)
Aalborg University, Department of communication technology
Applied Signal Processing and Implementation (ASPI)
e-mail: F...@control.auc.dk, http://www.control.auc.dk/~fuz
snail-mail: Lobovej 20 - 9230 Svenstrup - Denmark
Phone: +45 4026 5572 ICQ# 13068090
John Fields
> >Dominique Jouniot wrote:
> >>Hello,
> >>
> >>Has someone an idea , how i can build an analogic delay (aroud
> 10 to 30
> >>ms) for a audio signal.
> >>
--
1. Point a directional microphone at a speaker located 10 - 30 feet
away from the mic.
2. Do the same thing but couple the mic and the speaker with a piece
of pipe.
John Fields, Austin Instruments, Inc.
El Presidente Research, Design, and Development
"I speak for the company" Austin, Republic of Texas