Monthly Archives: May 2014

Building a PT2399 based delay pedal

The populated board, close-up

The populated board, close-up

My friend at DIYEffects was kind enough to send me a PCB he had constructed for a delay pedal. His is based on a PT2399, which is a very popular (and extremely cheap) integrated circuit that will take an audio signal and produce an echo of up to 500-600 milliseconds. It does all this in 44k of digital memory. That’s nothing at all. Oh and they cost about $2 at the time of writing. Seems too good to be true?

Maybe. There are compromises of course. Mainly in what amounts to fidelity of the repeating signal. With only 44k of memory there’s no way this will be a sparklingly clean 16bit/44 kHz digital delay that you might expect from Lexicon. As is fairly typical in the world of guitar effects, it turns out nobody actually wants a crystal clear delayed signal. If you ever get to try it out you’ll find that the repeats tend to drown out the dry signal as a hi fidelity ‘wet’ signal will have all the same frequencies as the dry signal so it kind of stomps all over it. The ‘low fidelity’ wet signal of the 2399 has somewhat muted high frequencies, so it sits under the dry signal quite nicely.


The biggest issue I had with this circuit was that it would only work intermittently. In fact, it worked perfectly on first power-up, and I was able to play it for an hour or so with no issues. But when I came back to it this year ( I had populated the board and got it running a couple of years ago) and fired it up, there were no echoes. None at all.

I always blame power first, so I did some measurements to ensure that the op-amp had some power, and the 2399 chip itself had things in the right place. It looked like the 5v regulator wasn’t quite doing what it should, so I replaced that and made sure the pinouts were going to the right sockets. This helped as now I had a steady 4.8v at pin 1 (Vcc) of the 2399.

Still no echoes though. I read a post (here) that talked about “latching problems”:

First, I’ve had the PT2399 ‘lock up’ on me two or three times now.  When this happens the circuit passes signal but you get no delay.  The regulator inside also gets hot (and could cut out I guess).  The effect can come back to life with the same PT2399 after disconnecting it from the power supply for a spell.

This sounds exactly like the issue I’m having. So this post lead to one from Merlin himself:

Some devices appear to latch up or fail to function properly if digital ground (pin 4) is left unconnected. This pin should usually be connected to analog ground (pin 3).

So I made a simple jumper from some scrap bits of stranded copper wire, and soldered it across pin 3 and 4. Problem solved.


Tagged , , , , ,

A real effects pedal power-supply


Over my playing years (more than 30 now) I’ve owned many effects pedals, spent a ton on 9V batteries and even tried many different power supply devices. I had one of those VisualSound “1-Spot” daisy chain things, which worked quite well. And recently I was given a Furman SPB-8 which appears to be the mother of all pedal boards. If you’ve been to this site before you know I’ve built an amp, and a whole slew of different pedals. All of these things I mention have one thing in common: there’s a lot of attention paid to converting alternating current from the wall outlet into direct current to power the circuit.


After looking at many different power supply circuits, I realised they looked very simple, but I did not know the first thing about why each component was there and what they did.

My cousin Mike was in need of a power supply for his pedal board that could handle anything that he threw at it (i.e. multiple voltages and a useful current handling capability). Naturally there are products one can buy, such as the Voodoo PedalPower or the MXR Brick, but these can be quite expensive. And lets face it, I wanted to build one, so I could find out how they work.


So, what does this thing need to do? Basically the following is desirable:

  • Provide a steady 9 volts.
  • Do not introduce noise into the circuit.
  • Handle at least ten pedals’ loads, which might approach 1 Amp.
  • Be small enough to fit under Mike’s pedal board.

Transform, rectify, smooth and load

From my work with tube amps and reading everything I could find in print or online, it became clear that, at a fundamental level, there are only a few discrete steps to accomplish:

  1. Transform the outlet supply from 115V to something closer to the desired voltage.
  2. Use a rectifier to convert the AC supply to a DC supply.
  3. Use filters to smooth out the now rippling DC.
  4. Load the circuit with whatever you wish to actually provide power. Whatever that is, it require voltage, will draw current and therefore has resistance.

A transformer is easy to get hold of. Even Radio Shack sells a 9V AC ‘wall wart’ that will handle 1 amp. A rectifier is easy too; just a bunch of diodes arranged in a certain way.

The filtering though was a mystery. How do you know what values to use? After gathering quite a few different circuits it seemed everyone was using different values, so who had it right?

The idea is this: use the characteristics of a capacitor to ‘fill in the gaps’ of a rippling DC supply. Remember, we rectified AC, which gave us rippling DC. That ripple will be at a frequency that can be heard by us humans, so we need to reduce that ripple to be inaudible.

Here’s a graphic showing a number of things. Firstly, this is a time-series showing voltage over time for the circuit at the top of the bitmap. Its a simple circuit showing an AC source going through a rectifier. Then the AC signal flows across C1, a capacitor. This graphic is to explain how to calculate the value of C1, and the influence it will have on the green waveform, which is the rippling DC source (post-rectification).

Ripple smoothing calculation

Ripple smoothing calculation

So imagine it like this:

1) The rippling DC source voltage rises to reach about 32 volts. During this time C1 is charging.

2) as the DC source starts to come back down, C1 starts to discharge. This serves to “fill the gap” between the ripples.

The chart is showing, with the different coloured lines, the influence of different values of C1 on the rippling DC.

Here’s a great article on this subject.

A more modern approach.

It seems to me that it is rather than painstaking to install a network of capacitors to perfectly smooth out the rippling DC. Also, transformers are far from perfect, so under load they sag, in terms of voltage. Even if you got all your calculations right, under load it will behave differently.

It is far easier, but not necessarily more efficient (in terms of trees and polar bears), to use a voltage regulator. The 78XX series works very well. You give it more voltage than you require, say 12V, and it does the work to provide you with an almost entirely ripple-free steady 9V. Any excess voltage is ‘given up’ as heat. This is why you might need a heat-sink (and why dolphins and rainbows are destroyed). Here’s a typical circuit using regulators:

AMZ power supply

A typical regulated power supply for guitar pedals

The end result

I came up with a simple circuit that used part of the above circuit to provide 9V and 18V for about 10 pedals. My initial circuit was flawed in that I’d ordered the wrong voltage regulators. They could only handle 100mA. Once we’d connected up all of Mike’s pedals, it would literally cut out. I even connected each one individually to my bench supply so I could measure each ones current draw.

Investigation showed that the regulator was extremely hot, so it became clear that the thermal protection was being tripped. I replaced the 7809V regulators with higher current handling ones and all was well.

Here’s some photographs of that build:

The finished 9V/18V power supply

A very simple decal makes the enclosure look quite spiffy, I think.

Gut shot of 9v/18v power supply

The circuit mounted inside, with all the DC connectors

The final schematic

The final schematic

And the final schematic, with values underneath:

  • D1 – 1N5401
  • C1 – 470uF
  • C8 – 0.1uF
  • D2 – 1N4148
  • IC1 – L7809CP
  • C3 – 10uF
  • R1 – 100R 1Watt
  • C2 – 100uF


Tagged , , , ,