Mimicking Analog Delay in a Digital Delay Pedal

[holmburgers]

Hi All,

I wanted to ask about a mod that I saw written about on a different thread, which sounds rather interesting. Essentially adding some components to the circuitry of a digital delay to emulate the decaying repeats of an analog delay.

Specifically the effect in question is the digital delay section of a Boss BE-5 (schematic here).

So, user codamedia wrote, "I just did another MOD to my BE5 this morning. This is a delay "high cut mod" which makes the delay sound more like an analog or tape delay.

This is what I did, but of course you can alter the cap values to get exactly what you want. The larger the cap value, the more high end that you cut.

1: On the EFFECT LEVEL control for the delay, I soldered in a 223 (22nF) cap between the wiper and ground. This gives the "initial repeat" a slight drop in high end so the 1st repeated note does not interfere with the original note. If you don't proceed with step two then the high end cut provided here will stay consistant with all additional repeats.
2: On the REPEAT control for the delay, I soldered in a 503 (50nF) cap between the wiper and ground. With this cap, every repeat that occurs will have less high end which is very much like the workings of a "tape echo"."

This seems like a great idea and I am planning to try it out on my own BE-5, though I would just implement both mods into a single position on a 2-pole switch, meaning it'd be the normal digital delay OR the full "analog" emulation effect described here.

First off, I'm just curious if anyone else has done this and can comment on the sound, if it really approaches an analog delay sound, and secondly, what are the mechanisms at work here to reduce the high-end of the repeats in the circuit? (warning - I am still pretty green to electronics...)

Lastly, would this work on other digital delays, specifically the Ibanez PT3? (schematic available here)

I love the squishy and rounded sound of analog delay repeats, and would love to have this feature on the above-mentioned multi effects, which I'm currently fiddling around with.

Cheers all!

[And a shout-out to codamedia, whose website is a great resource!]

[ElectricDruid]

Yes, it'll probably sound a lot more like analog delay. The most obvious characteristic of those old analog delays is the limited bandwidth, so adding heavier filtering to a digital delay gets you much closer. This is true of any digital delay.

The added caps acts as lowpass RC filters and gently remove high frequencies. Especially on the repeats control, you don't *need* a huge amount of cut since the signals go round and round and get more highs removed each time.

Give it a try and see what you think!

[holmburgers]

Thanks!

So in the schematic it's labeled slightly different, but I assume VR011 "Delay" is actually effect level (and "Feedback" is of course repeats), but in a simple block diagram kind of way, what exactly is happening with the signal?

Is it that... the dry signal is going to the delay circuit, the "E. Level" pot controls the volume/gain initially going to the delay, so with a capacitor the treble is getting rolled off to begin with, and then the feedback loop from the output of the delay and back in again (for repeats) is attenuated by the "F.Back" pot, and thus adding a cap here gives the additional low-pass filtration each time it goes by...?

I might have this a bit off, but a simple explanation of the signal path would be mucho appreciated!

[Mark Hammer]

Been doing that mod for close to 40 years.  The key to a pleasing version of it is to make the treble-cut in the repeats shallow, and pick your corner frequency right.  What you want is for there to be a little bit of high-end left, such that there is something to filter out on successive repeats. 

This has several benefits:

• It moves the repeats to the perceptual background, so that you can have many repeats without sounding "cluttered"
• It more closely mimics the manner in which physically reflected sound loses bandwidth from non-reflective surfaces
• It attenuates any accumulating audio grime from multiple passes through any delay circuitry, digital or analog

[holmburgers]

Hi Mark,

Thanks for the input!

So I assume then that you have done this on a variety of different delay pedals, meaning it can essentially work with many different circuits?

As for the component values... could you recommend a range in which I could experiment?

I'm really weak on my electronic theory, so I'm not confident in saying which capacitor value range I should play around in, but I could certainly see spending a few bucks on a handful of capacitors and experimenting with what sounds best.

I'll of course use codamedia's values as a starting point, but how low or high can I go on each of the two capacitor values in the example given?

Many thanks in advance

[ElectricDruid]

So in the schematic it's labeled slightly different, but I assume VR011 "Delay" is actually effect level (and "Feedback" is of course repeats), but in a simple block diagram kind of way, what exactly is happening with the signal?

Ok, you asked for it!

Here's the schematic with some scribbles on it:



The signal comes in in the top-left, down the red wire. 18b looks like pre-emphasis, and then the three transistor stages are filters to remove any harmonics that would cause alising when sampled by the delay. IC17a is a compressor to boost quiet signals and improve the Signal/noise ratio. From there it goes off down to the delay processor/RAM at the bottom. The delayed signal comes out of that processor (green line) and goes back up through another couple of transistor filters, this time to remove digital images caused by the sampling process, to IC17b which is a expander, designed to make quiet signals quiet again like they were before IC17a messed with them. The side-effect of this is that noise gets squashed as well when you make the quiet bits quieter. From the expander, the signal goes right to the Feedback and Delay pots. The Delay pot sets the level of the signal going to the output mixer - pink path, to 16a. The Feedback pot sets the level of signal returned to the compresor to be sent back into the delay - blue path. The output mixer IC16a is a simple inverting op-amp mixer. The dry signal comes from after the pre-emphasis stage to the mixer down the other pink path.

Hope that's clear, ask if not!

[holmburgers]

Hi Druid,

Sorry for the slow reply, but thanks so much for posting this! That's extremely helpful for a newbie like me.

This definitely clarifies a lot about the basic workings of a DD circuit for me.

So when it comes to adding an RC filter to either the Delay/Effect level pot or the Feedback pot, to make the cutoff frequency calculation, do I simply use the potentiometer value + the capacitor?

For instance, in codamedia's original recommendations, he puts a 50nF cap on the 10k feedback pot, which according to the calculator I found online would give ~318Hz as the -3dB cutoff frequency.

And he says to put a 22nF cap on the 50k on the effect level pot, which gives a cutoff frequency of ~144Hz.

Is this right, or am I barking up the wrong tree here??

[ElectricDruid]

I'm not sure. I thought I knew, but then I did a quick simulation to check, and got a result I wasn't expecting.

I'd have said that the cutoff would be set by the top half of the pot down to the wiper, so there'd be no filtering at maximum feedback, down to the 318Hz you mentioned at the very bottom of the pot with almost no feedback.

However, it looks from the sim like the cutoff is several KHz across the range, and while it varies a bit, it doesn't go in only one direction. This implies (to me, at least) that the bottom of the pot must interact with it too. The cutoff comes down like I'd expect at first, but then seems to go back up again once we get past halfway.

My guess is that the relevant "R" for the RC is the two halves of the pot in parallel, since that would explain what I'm seeing in the sim. That gives a small value R value at the top and bottom, maxing out at half the pot's value in the center. Perhaps someone who has the benefit of a technical education in this stuff can enlighten us as to whether that's correct and why that is?