KDLA
A Modulated Delay with Adjustable LFO Shape and Extended Delay Time
Overview
Does the world need one more delay pedal? Why, yes, yes it does, and the KDLA is it. So what sets it apart? The modulation can be adjusted from subtle to extreme, mimicking warped vinyl, tape "wow", and other, more subtle, effects. In addition, the delay time can be stretched to over 1500 ms, though it does degrade the delayed signal significantly. It's perfect for that old-school, lo-fi, tape-style delay.
The KDLA is really born out of an amalgamation of the Stalker and the Wobble Box. It utilizes the delay topology from the Stalker (with added feedback control) and a modified version of the Wobble Box LFO that allows for adjustable shape. The plus side of this circuit is that the ES56033 allows for more usable delay times, both shorter and longer, than the ubiquitous PT2399. Aside from that, the overall circuit topology is pretty straightforward.
Additionally, since I have plenty of ES56033's kicking around from previous project development, I decided we may as well continue rounding out the stable of effects that can be made with them. The KDLA brings us to a double tracker, chorus/vibrato, and delay. Other projects may well follow.
How It Works
The KDLA is a rather straightforward delay circuit that utilizes a digital delay chip. Aside from slightly different time control topology, the rest of the circuit is very similar to your typical PT2399 delay.
The KDLA uses the same input buffer that many of my circuits use. It works well, so let's keep on using it.
KDLA Input Buffer
Next up is the delay stage. The multifeedback filters for input and output are the same as the Stalker, as are the modulator/demodulator smoothing caps. However, this stage includes a time control as well as feedback path to provide the typical delay controls. Also note the connection of the point labeled “CS”. This stands for current sink. Much like the Wobble Box, KDLA uses a current sink driven by an LFO to modulate the delay time, providing the modulated audio signal.
KDLA Delay Stage
The delay time is controlled by two potentiometers wired in series. Time 1 provides the shorter delay times, while Time 2 allows for longer delay times. Maximum delay time is achieved with both pots maxed out. According to my measurements, the delay time is ~1560 ms on my prototype. This is a pretty straightforward arrangement meant for simplicity. Of course, with the simplicity comes the ability to create the same delay time with different pot combinations. Hey, you can't win 'em all, right?
KDLA Time Control
Now that we have delay, we need to create our modulation. This starts with an LFO. The topology for this LFO is the same as what I used on the Wobble Box.
Note that the speed control is a reverse log pot. This is to give more resolution at the faster speeds. With a linear pot, there hits a point where the speed shoots off and it can make that range very hard to dial in. So don't sub this for a linear if you don't have to.
One other thing to note is that the output level of the LFO is fixed. I found that adjusting this doesn't have much effect on the sound because of how we are introducing the modulation.
KDLA LFO
With the LFO signal now generated, we are going to a little trick to mess with the current that the pin 2 capacitor can sink. Because the ES56033 (appears to) use an internally compensated VCO for delay time setting, we can mess with the delay time just a little until the internal compensation catches up. This allows us to create modulation while not actually changing delay time. To change the current that the VCO is sinking, we use a current sink. This is the same idea as the EchoWreck, except instead of using a smoothed PWM signal to change delay time, we are always changing the amount of current that the current sink's BJT lets through, so that the VCO's internal compensation is constantly on it heels. This creates the wobble sound.
The current sink uses the BJT as a variable "resistor" (how open/closed the collector/emitter connection is) to change the current allowed through. We also have a series resistor from emitter to set the maximum current to sink. With none, we don't have any modulation. Raising this resistor more makes the minimum depth setting more drastic. Setting it lower makes the minimum depth setting less noticeable. I found 68 Ohms to have a minimum depth right on the verge of being unnoticeable.
The depth pot is setting a limit as to the minimum amount of current that the chip will sink. I found that it is a very small resistance range (note the parallel 120 Ohm resistor!), and I also found that a logarithmic pot works best. With linear, the lower settings on the depth control go from mild to very noticeable over a very narrow range, so with the log pot, you get a much more even sounding distribution of the depth.
KDLA Current Sink
The remainder of the signal path is the summing of dry and delayed signal and output buffering. This is, again, the same output buffer that I have used several times now. The delayed signal is summed using the output of the mix pot.
KDLA Output Buffer
Finally, we have the power section. This is pretty basic, but please note the use of the LM7805 instead of 78L05. This is done because it is pulling about 60 mA off the 9V line, which, with a 4V drop is nearly 1/4W of power dissipation, which gets that little TO-92 case of the 78L05 really hot to the touch. I prefer the larger, TO-220 packaged LM7805 for these ES56033 circuits just because they don't really even get warm. They are bigger, and you could probably just use a 78L05, but I'm just playing it safe. Who knows what could happen when it gets hot, maybe it can't supply enough current or something. I don't know.
KDLA Power Section
Well, there you have it! Another entry in the ES56033 series is in the books. Fancy making one yourself? You'll find a full size schematic PDF, build documentation, and Gerber files for the board here. Let me know what you think!