Simple time delay circuit?

[acehobojoe]

What is the most simple circuit I can create to delay the time of a signal by a few milliseconds? Could it be a pt2399, or some sort of BBD? In the studio there's an effect where you delay one of the stereo channels by a little bit and it makes it wider, anyone thought of doing this?

[aron]

Yes, any simple delay circuit will do it. I used to do this live with my memory man. The first time we did it, it was incredible!

[Transmogrifox]

PT2399 coupled with the typical application circuit is probably your best bet.  Messing with stereo delays produces some very incredible sounds.  You can create some really neat subtle effects -- especially when you mix filtered outputs (different bands with different delays sent selectively to left or right).  I have played with this a lot in the digital realm (Linux/jack and C programming), but never tried to implement a dedicated "box" for this.  It would be useful to have several dedicated delay lines with an adjustable delay (like a PT2399 with a pot to turn delay time), and then something like a mixer to left and a mixer to right.  On each mixer input would be an EQ to set how each delayed tap is filtered.

What you are asking to do is a really simple sub-set of this.  You could even do something really cool with a BMP style tone stack, a single delay feeding to one of the BMP tonestack taps (feed to HP or LP).  Then you have 3 outputs:
High pass
Low pass
Delay

Some kind of combination of those could be panned between left and/or right, or mix something to center and put delay to left or right, etc.  Just a good experimentation platform.

This looks like a promising chip if coupled with an ADC/DAC:
http://www.ti.com/lit/ds/symlink/tpa5052.pdf

Not sure your level of skill whether you could devise a circuit to combine a codec with something like the TI chip and get it to work.

I think "simple" is the PT2399 because it has built-in ADC/DAC.

[Mark Hammer]

Are we sure a 2399 can do "a few milliseconds"?

[acehobojoe]

Now that's my question. I'll have to read those datasheets alot before I'm comfortable with them.

[blackieNYC]

I think what you are looking to do would be at the very bottom of the time range with a PT2399.  The lowest resistor you can put there is a 1k on pin 6.  I'm not sure what that delay is - 50ms?  50 ms is very cool for stereo.  I think even fast playing would remain fairly clear.  Build one anyway, they're great. know anyone with a rackmount digital delay with a readout?  Maybe you could find the delay that you like.
Then again, if ever there is a case for a "full fidelity, full bandwidth" delay, this may be it.  Do you want that 2nd channel to be warm but a bit dark? Or do you want the most accurate reproduction of the first channel?  If you want channel two to be 100% wet, I myself would be curious to hear how a nice digital delay sounds, compared to a 2399 circuit meant to have "analog warmth" and the filtering required to roll off noise. Know what I mean?  But build a rebote or some 2399 circuit, skip the modulation, and try that.  A must build.

[blackieNYC]

Merlin's valve wizard Small Time delay is a pt2399, and he says the 1k resistor gives you 35 ms.  I would think that's a good minimum for stereo separation. (Not short enough for flanging apparently, and barely enough for chorus IMO) 35ms up to maybe as much as 100ms would be very cool sounding.

[merlinb]

Merlin's valve wizard Small Time delay is a pt2399, and he says the 1k resistor gives you 35 ms.

Yes, the PT2399 will only go down to 35ms reliably. You can get down to 25ms (ish) with a 100R delay resistor, but you need latch up protection if you want to do that. A cheat would be to use a PT2399 in both channels, one set to 35ms (say) and the other for 40ms (say), so the difference is only 5ms. But then you're degrading the audio quality in both channels instead of just one...  A BBD can do shorter delays than the PT2399.

[mth5044]

There was an ADT circuit posted back when this forum did the FX-X contest. It was a BBD based circuit that use like 10 parts for a few ms delay. Have a look for it.

[samhay]

What is the most simple circuit I can create to delay the time of a signal by a few milliseconds?

For short delay times, this is about as simple as PT2399-based circuit can get:
http://sound.westhost.com/project26a.htm (Fig 3)

[midwayfair]

I don't see how "a few" (like 3-5mS) is worth building an effect, even if you could find something that can do a delay that small, which I don't think you can. If it's in the studio, you can just nudge one channel a little. If you're live and not micing the amps, put the louder amp 5-10 feet (1m = 3mS) to the rear. If you're micing the amps, close mic one and mic the other from a few feet away.

[Mark Hammer]

I have a few Omnisonic "stereo image enhancers" from the early 80's.  Radio Shack had one built into a 10-band EQ, and Carver had their "sonic hologram".  Radio-Electronics had a project that I built (and still have the board for if you've got a spare SAD1024!).

These created a much larger sound field by artificially creating a "sound shadow", using BBDs.  When a sound source is not directly on-axis (straight ahead) the sound arriving at each ear is slightly different.  What arrives at the closest ear is a few milliseconds ahead, and has wider bandwidth than the "version" arriving at the farther ear.  So what all of these electronic devices (and later, in software, with the Hughes SRS found in Windows Media Player) do is to identify whatever is unique to one side/channel (left side minus right, right side minus left), and cross feed it to the other channel, with a tiny bit of delay, some lowpass filtering, and mix it in with the other channel at a lower level.

You don't notice the impact right away, but when you turn it off, you wonder where the music went.  That small addition creates a much wider stereo soundfield, and allows the listener to spatially isolate instruments.  I also find that it reduces listener fatigue dramatically.  You can listen on the worst sound system for a long time, simply because the spatial decoding workload for the brain is reduced.  It's like the difference between reading a crisp 12pt laser-printed Arial-font text, and reading a bad photocopy of a 9-pin dot-matrix text; the one just zips along more smoothly than the other.

I imagine there is no end of software plug-ins that can accomplish this with greater fidelity than achieved by my old kit.  But if one wants to do it in real-time in the analog mode, I guess you'd need a couple of BBDs to get the sort of very-short delays required for the job.

[Processaurus]

Take a look at the PT2396.  The lesser known sibling of the PT2399.  It can do short, low millisecond delays, and has an "easy mode" (outlined on page 5 of the data sheet).  Easy mode makes it simple to choose from a short list of static delay times without using a digital control.  12ms is the shortest it does in easy mode.  Next is 24ms, and 36ms, (and steps up to 200ms) are available.  If you could live with the limited steps great, but if you needed shorter, or finer control there is the microprocessor controlled mode, but would be a programming project (and thus no longer "easy") to send the correct serial data to the PT chip, to access all of the delay times in between.  I wonder what would happen by "overclocking" the chip with a faster crystal?

Datasheet

[armdnrdy]

You can get .16ms - 6.4ms from a MN3003,
.32 - 6.4ms from a MN3006 or MN3206,
.64 - 12.8ms from a MN3009 or MN3209,


Although not very common, (you won't find them at Small Bear) they are still available if you search the internet.

[Ice-9]

What is the most simple circuit I can create to delay the time of a signal by a few milliseconds? Could it be a pt2399, or some sort of BBD? In the studio there's an effect where you delay one of the stereo channels by a little bit and it makes it wider, anyone thought of doing this?

That is what is done in the recording studio and is referred to as double tracking, many guitar players used to do this by recording two guitar tracks exactly the same and the slight differences produced this effect. One guitar player that did this to perfection was Randy Rhodes.

It could be emulated with a chorus or flanger pedal if you can disable the modulation.

1ms delay can be made easily with the FV-1 chip and it isn't as complex as some think.

[acehobojoe]

DSP?

So far there are great ideas listed. I just need to decide on one.

[Ice-9]

DSP?

So far there are great ideas listed. I just need to decide on one.

The FV-1 I mentioned above is DSP, and easy to implement a very short delay , you could even make it triple track. 

[snap]

more about BBD from almost 30 years ago: http://www.cjlectronics.com/tech_info/bucket_brigade_info

[merlinb]

Depending on what it is you're trying to acheive you might also want to look into all pass filters, or even the Blumlein shuffler:
http://www.audiosignal.co.uk/Resources/Stereo_shuffling_A4.pdf

[deadastronaut]

many guitar players used to do this by recording two guitar tracks exactly the same and the slight differences produced this effect.

i  do this on home recordings , it gives a great full wide sound..especially on heavier stuff,

i even record 4 tracks of riffing, pan 1+2 extreme, and 3+4 half left/right/

those little nuances in the riffing makes it really pop out, and gives a 'live' feel to it....