Measuring Delay Time

[Paul Marossy]

How does one measure delay time? I have a DIY pedal for which I would like to determine the max. delay time but I am clueless on how you would do that. Anyone know how that would be done?

[GibsonGM]

Do you have access to an O'scope, Paul? 

[Paul Marossy]

Do you have access to an O'scope, Paul?

Yes. I figured that might be involved.

[GibsonGM]

I could see using a storage scope to 'track' the repeats after an initial input signal...

If you (like me) only have an old CRO, you could use a repeating input (LFO) to get a handle on 'how many squares away' the delay signal shows up, if that makes sense...

In practice I think that once you have dialed in the scope to the right range, this should be pretty easy.  You know, just counting the squares between signal and repeat and multiplying by the 'per division' unit...

[Paul Marossy]

I could see using a storage scope to 'track' the repeats after an initial input signal...

If you (like me) only have an old CRO, you could use a repeating input (LFO) to get a handle on 'how many squares away' the delay signal shows up, if that makes sense...

In practice I think that once you have dialed in the scope to the right range, this should be pretty easy.  You know, just counting the squares between signal and repeat and multiplying by the 'per division' unit...

Hmm... not as easy as I was hoping. I do have an old Tek 453, it's all analog. I also just made a "DSO Shell" DIY kit (handheld digital scope) but I am not sure that would do the job or not.

[mth5044]

I put it into my DAW, hit record and made a quick strike on the guitar to produce a quick blip and had it repeat a few times, then used the time marker on top as a ruler.

[Paul Marossy]

I put it into my DAW, hit record and made a quick strike on the guitar to produce a quick blip and had it repeat a few times, then used the time marker on top as a ruler.

That's clever

[anotherjim]

A recording program audio editor would be my first choice too. Scopes get creaky working timings beyond milliseconds. Freeware Audacity would do.

If using a scope, and you know the delay technology, it's easier to calculate from the delay clock period since that is fast & easy to measure. We know how many stages a BBD chip or PT2399 has, so it's that many multiples of the clock period.

[Ice-9]

I put it into my DAW, hit record and made a quick strike on the guitar to produce a quick blip and had it repeat a few times, then used the time marker on top as a ruler.

+1, exactly this.  The DAW can switch between ruler and time so if you select the time you can get almost the exact delay timing (zoom in for accuracy).

[Ice-9]

Double post

[Paul Marossy]

I might have to try the DAW thing, sounds like the easiest way to do it. But I do wonder how a big company like Roland-Boss would do it? I guess they would know that because they are programming the microprocessors, huh? 

[Josh?]

If you don't want to use a DAW, there are a ton of free apps with tap tempo metronomes, and it's pretty easy to tap along with something accurately enough for the metronome to tell you what tempo it is.

As for the big companies, they hire wizards. It's all magic

(On a more serious note, I'd like to learn how they program that stuff too)

[PRR]

Get long guitar cord. Guitar and amp and big space. Tink and hear the sound come back, from 5 foot, 50 feet, 500 feet. Your ear can learn to estimate short delay times. 1 foot air = 1mS for estimation purpose.

[Ice-9]

Get long guitar cord. Guitar and amp and big space. Tink and hear the sound come back, from 5 foot, 50 feet, 500 feet. Your ear can learn to estimate short delay times. 1 foot air = 1mS for estimation purpose.

a = sqrt [g * R * T]

[Rob Strand]

One way is to use a pulse input signal then just measure the distance between peaks.

The other way is simply to measure the clock frequency and calculate the delay.

If you have a modulation effect then measuring the delay is trickier.   A simple way is to measure min and max voltage the LFO waveform (after any filtering) you then feed a those DC levels into the VCO, effectively freezing it in time, and you can measure the fixed delay.

Other ways are to read the waveform into a PC and do correlations using software like SciLab/Matlab/Octave. If you want more accuracy an easy way out is to upsample the signals to get around the errors from discrete time sampling resolution.