Modifying An ISD1820 Voice Recorder Board
breaking out the controls and making a lo-fi sampler
The ISD1820 is an audio recording IC that'll record up to 20 seconds of audio that you can trigger back. The recordings are also non-volatile, so they won't disappear when you power the chip down. Now, while 20 seconds sounds great, as with most things - there's a catch.

If you want the full 20 seconds of audio then your sampling frequency drops to just 3.2KHz (giving about 1.3KHz of bandwidth). At the other end of the scale you can get 8 seconds at a much more respectible 8KHz sampling frequency - absolutely not audiophile teritory, but certainly more than usable.

The sampling (and playback) frequency is controlled by a resistor, so changing the resistor value gives us different sampling / playback frequencies and associated times. Here's a quick table showing them -

Resistor Sample Time Sampling Frequency Bandwidth
80K Ohm 8 seconds 8.0KHz 3.4KHz
100K Ohm 10 Seconds 6.4KHz 2.6KHz
120K Ohm 12 seconds 5.3KHz 2.3KHz
160K Ohm 16 seconds 4.0KHz 1.7KHz
200K Ohm 20 seconds 3.2KHz 1.3KHz
Fig. 1 - Table of resistance vs sampling time / frequency (from the datasheet, linked below)

This relationship between sampling frequency and resistance means that if we use a potentiometer instead of a fixed resitory we have pitch control. In this guide I'm going to show you what I did to a board based on the ISD1820 to make my own lo-fi sampler with pitch control.


Fig. 2 - The finished item!

Picture above is my finished lo-fi sampler. It's got 1/4" input and outputs, two buttons - red for record and green for playback, and a knob for pitch control. And an external battery pack because I underestimated how much space I'd need.

I'm not going to show you the exact nuts-and-bolts of how I built this, because - being frank - I pretty much made it up as I went along. I'll show your what I did to modify the board itself, but the case is up to you to sort out - I can barely drill two holes in the same alignment... Read it through first so you don't make the same mistakes I did.

Disclaimer: I am absolutely a beginner at this whole electronics / soldering / making stuff malarkey, so most of what is below is be figuring stuff out as I go along, watching Youtube videos and reading things on that there inernet - so take everything I say within that context. Also, I'm no great shakes when it comes to soldering so some of the joints you'll see below are truly gruesome and for that I apologise to you all.

Let's get started!

Ok, so first off, here's the board we'll be modifying. These are incredibly common on all the usual online tat bazaars. I picked mine up from Amazon in the UK for about 3.50 (US$5-ish) which included a 40mm speaker and connecting wire.


Fig. 3 - The board in question

The quick rundown of what's on the board -
  • Top left - the on-board mic
  • Left hand side - Dupont headers to break out the various functions, including the power-in (3v)
  • Centre - the ISD1820 chip
  • Right hand side - the white connector is for the audio out. Not a Dupont, but can be used with them
  • Along the bottom - function buttons

  • A quick description of what the function buttons do -
  • REC - While held this will record through the mic
  • PLAYE - Press this once and the sample will play to the end
  • PLAYL - While held this will play the current sample, stopping when released

  • The Dupont headers on the left hand side are the main interface points of the board. You'll notice that there's two rows, this lets us break out the functionality without having to unsolder the function buttons.

    From the top we've got -
  • VCC - positive voltage in (left), positive voltage out (right)
  • GND - ground
  • FT - pass the input of the mic to the speaker (referred to in the datasheet as 'Feed Through')
  • P-L - same as pressing PLAYL
  • P-E - same as pressing PLAYE
  • REC - same as pressing REC

  • You'll notice that there are a couple of jumpers already in place and also two labels just to the right marked 'FT' and 'P-E'. In their current places the jumpers don't do anything. However, when placed across the jumpers marked as 'FT' or 'P-E' then they will either 'feed through' or 'play endlessly' respectively.

    NOTE: We have two things near each other with the same 'P-E' name. They are different, and in this build we'll just be using one of them. When use the shorthand name of 'P-E' below then I'm referring to the pins on the right-hand side next to the label, the left-hand pin header will be referred to as 'PLAYE'.

    If you wanted to, you could break out buttons for each function - a record, three playbacks, and a feed-through. But, this board has a bit of duplication and we can utilise this to cut down the amount of work we need to do.

    I'm not going to bother with the feed through functionality for this build - I have no use for it - so that's a button gone right off the bat.

    The three playbacks - PLAYE, PLAYL and P-E - are all actually just variations of each other. While I was testing this board I discovered that if it shorted the two P-E headers really briefly then I got the functionality of the PLAYE button. If I help it longer then then it would loop (aka Play Endlessly). Even better, was that if I did let it loop and then broke the connection it'd play until the end of the sample. If that isn't what you're after - and you want it to stop as soon as it is released, then you might want to wire up the PLAYL jumper as another button.

    Right, a quick recap on what we're going to be doing -

  • Removing the mic and replacing it with a 1/4" socket
  • Replacing the speaker with a 1/4" socket
  • Having a separate button for recording
  • Having a separate button for the P-E (Play Endlessly)
  • Adding a potentiometer to change the sampling frequency

  • The first job will be to remove the mic off the board and solder in a couple of wires in its place. The mic is through-hole so will come out with your favoured way of removing components - solder, heat, wicking tape, stern looks, etc. If you don't break it then you've got a mic to use with your next project.

    To replace the speaker you can either just use the cable supplied (which does fit properly) and solder your jack to that (doesn't even need to be 1/4", if you're a Euroracker then you can use 3.5mm instead); or you can use a couple of Dupont connectors. I went for the Dupont connectors as the cable just wasn't quite long enough for my small box. With hindsight I would just have extended the cables. Ah well, you live and you learn.

    The record and play buttons were next. These were momentary push-buttons (rather than toggles) and connect to the headers so required Dupont connections. I removed the wires they came with and replaced them with some cut-down wires with Dupont connectors. If you've got some way of creating Dupont connectors then you could probably just crimp them onto the existing wires. I'll tell you where these connect a little bit later on.


    Fig. 4 - The record and play endlessley buttons with new Dupont connectors (colours chosen at random)

    Right, now we're up to the fun bit - and also nearly at the end. We're going to be replacing the R4 resistor with a potentiometer to give us some pitch control.

    The R4 resister is at the top right of the board and is a SMD item. A bit of heat and you'll be able to remove it.


    Fig. 5 - The R4 resistor has been removed

    The datasheet for the chip says that 80K ohms is the lowest resistance you can have for R4. However, in my testing I found that I could playback at much lower resistance - and therefore a higher pitch; but couldn't record - it'd just give me a short sample of unhappy noise. With this in mind I decided that I should probably put some kind of static resister in line with the potentiometer to shift the range a bit.

    For reasons that were nothing more than "it was all I had", I went for putting a 10K resister on the pot. I was using a 200K pot so this would give us an effective range of 10K to 210K. If I had one, I would have liked to have a go with a 20K resister and 250K pot to go really lo-fi (if such components exist - see my disclaimer above about being new to this). Anyhow, I only had a 10K resister and a 200K potentiometer, so that's what I used. I soldered the 10K to the left leg of the potentiometer. Ignore that this is a stereo pot - again, it was all I had to hand - a normal 3-leg pot will be just fine (and take up a bit less space).


    Fig. 6 - A 10K resistor (badly) solder to the left leg of the 200K pot

    Here's the challenging bit - soldering a bit of wire to an SMD pad. The way I did it was to put some additional solder on the pad - making sure that I didn't accidentally bridge the two - so that I'd have something for the wire to grip to. With a bit of blind luck I managed to get them soldered on - if a little bit closer to each other than I'd like.


    Fig. 7 - Wires soldered to the SMD pad for R4. I later glued the cables to the board for extra security

    The red wire - from the upper pad - was soldered to the 10K resistor on the left leg; the yellow wire - from the lower pad - to the centre leg. This gave me the lowest resistance - and therefore highest sampling frequency - at fully anti-clockwise. If you want it to go the other way - with lowest resistance at fully clockwise - then the red wire (and it's associated resistor) will need to go to the right leg instead.


    Fig. 8 - The board with a 1/4" socket for the mic input and the resistor/pot combo for pitch

    Nearly there - just got to plug it all in now!

    Apart from the speaker connection - which goes into the white connector on the right-hand side of the board - everyting else goes into the header pins on the left-hand side of the board. Here's where they go -

    What Where
    Power (Positive) - 3v VCC - Left
    Power (Negative) GND - Left
    Record Button (1) VCC - Right
    Record Button (2) REC - Left
    Play Button (1) P-E - Right
    Play Button (2) REC - Right

    To connect the Play button connections, the jumper that is already covering the 'P-E' and 'P-L' right-hand pins needs to be removed.

    From the datasheet, if you were wiring up button the the PLAYL (P-L) and PLAYE (P-E) pins on the left-hand side of the header matrix, then these all work when the pins go HIGH. This means they - like the REC pin above - would also need to have a connection to the 'VCC - Right' pin.

    Here's a (terrible) picture of my finished contraption in a tiny box, complete with a disconnected speaker cable. It's not great, but hopefully you should be able to see where all the cables go. It's not got the power connected at this point - which in some ways makes it easier to see where things go.


    Fig. 9 - The finished wiring (bar the power) in the box

    Links:
  • ISD1820 Datasheet
  • Lonershy's video about building his version (though his is slightly different and he does mistakenly say the buttons should be connected to GND and not to VCC)

  • All work © Darren Shaw 2020 (except where noted)
    darren@anexium.com

    Back to the index