Picture of Squawkbox - Your Personal Vox
This project started as an instrument for ['http://dorkbot.org/dorkbotcolumbus/Sched.htm#musicevent Sonic Tooth], our local Dorkbot music event. I was looking for something where you could put any sound into the mix.

As an instrument, it's basically a mash-up of an answering machine and a megaphone. It ends up being a small package with a loud punch, so you could use it to play the same sample over and over (and over...). It uses a solid-state recorder with a 120 second capacity (Winbond's ISD25120)

The really cool potential that I didn't explore in this project is the ability to control the recorder via a microcontroller. It looks like the recorder's memory is addressable via a 10-bit address setting. With that, you could have the resulting playback be driven by any number of environmental stimuli. I was looking at using an Amtel ATiny13 to drive it, because I only really needed a 4-bit address space, leaving 4 ADC lines for input. But, that's all to come still.

Step 1: The Design

Picture of The Design
You could think of this thing as the guts of a digital answering machine hooked to a portable PA system. However, instead of being limited to the outgoing/incoming messages, we have full access to the capabilities of the chip.

The chip is a Winbond ISD 25120p. It records up to 120 seconds of audio at a 4KHz sample rate. It's about the quality of an answering machine (surprise!), but coupled with the amp in the megaphone, it produces a very textured sound.

The design here is the reference design shown on page 33 of the chip's datasheet The only things I really added were a couple of status LED and a 7805 voltage regulator. I drive it off a 9v battery, but you might get longer life out of 4-6 AA cells. YMMV.

The capacitors and resistors above the chip are filters for the mic and speaker, as well as a gain-preset.

I soldered this design together, more as an excuse to practice using a new soldering iron. If you plan on making more than 2 of these things, plan on etching a board. This is a very simple design, and you could probably make a single-sided board. If anyone wants to design one, I'll post it here for you....
jackillac926 years ago
Good progect but use a different case there only like 5 bucks at radio shack, or just buy it online and suffer shipping it because radioshack is a ripoff.
Cool but i'd add a 3.5 mm mic jack or something OR a 2.5mm one and buy one of those old bluetooth adapters and speak through bluetooth to it having it stuck on your car roof lol
I'm also currently hacking 'round a ISD25XX, but, unfortunately, Winbond consider the 25XX series to be obsolete... :-(

S, tig
Yeah, I saw that. I was looking around and the ISD4002 series looks like a replacement (ISD4002-120 has 120s recording at 8KHz). Even cooler, it support a 4-wire SPI interface to be driven by a microcontroller. However, it doesn't look like it supports a standalone mode. I need to dig into the app notes more. What type of applications are you looking at building for? M
I'm building a stomp-box looper. I've done a proof-of-concept (lovely, beautiful, lo-fidelity sound), but I'm yet to do the final build. (BTW, my approach was slightly different from yours in that I took a an existing PCB/kit and modified that.)<br/>As far as possible alternatives to the ISD25XX, I was looking at the <a rel="nofollow" href="http://www.winbond-usa.com/en/content/view/160/290/">ISD17XX</a>. Interestingly, unlike the 25XX, the internal clock on the 17XX is set by an external resistor which might means realtime pitch/time stretching by wiring a pot up to the chip.<br/><br/>tig<br/>
otakutheiii8 years ago
i like this, would i be able to just attach a 1/4in. socket instead of the the mic?
matterantimatter (author)  otakutheiii8 years ago
Yep, it's just a question of playing with the input filter. For a mic, the data sheet calls for an external capacitor/resistor filter (which I think gives you the ability to limit the frequency sensitivty ). That's what's in this design.

However, it also says:
If the desired input is derived from a source other than a microphone, the signal can be fed, capacitively coupled, in the ANA IN pin directly.

Also, if you need more capacity, you can daisy-chain multiple chips together and they will automatically cascade over to each other.