This is great! Are you looking to sell the MIDI PCBs or provide the Eagle files? I'll bet a number of people would find it useful.

This version adds the AC coupling capacitor to the input of the on-board audio amp that was missing from the initial version. It also brings out the I2S clocks and data input to a connector making it possible to add an ADC for stereo audio input, although there is currently no firmware support for this. And it also improves the thermal dissipation around the 5V regulator.

This board ships with firmware v1.00, which is available on the download page. None of the hardware changes affect the firmware, so the latest firmware will still run on the previous board version. The product description has simply been updated to reflect the current firmware capabilities.

You can easily set the WAV Trigger up to automatically loop a track on power up, without any jumpers or trigger connections. However the WAV Trigger can only loop on increments of it's audio buffer, which is 128 samples long. Unless your wave file exactly fills the last buffer, there will always be some number of samples of silence at the loop point. It will be very small, as in a couple of msecs, but very likely not 0.

I always suggest using your sample editor to seamlessly loop your file many times and save it as a really long track - there's plenty of space on the flash card. That way, the short gap will occur very infrequently.

Those comments apply only to the original hardware version (WIG-12000) and not to the current version (WIG-12897) of the WAV Trigger, wherein the missing cap has since been added to the design. (Mentioned in the comments just above this one.) I will try to make this more clear.

How are you changing banks? By changing the MIDI Channel on your keyboard, or with a trigger input using the BankUp/BankDn function? Firmware v1.04 and above added support for using MIDI Program Change messages to change banks. Is it possible that you're sending Program Change messages? In any event, I'll have a look at what else might have changed, although I've not heard any other reports of problems.

If it's easy, can you try the versions between v1.02 and v1.10 and see which one breaks it?

By the way, I just posted v1.20, which increases the number of tracks from 999 to 2048. Filenames can now have 3 or 4-digit track numbers, up to "2048xxx.wav". This version also maps all 128 MIDI notes to a bank and supports 16 banks (up from 10), so the MIDI to track number mapping has changed. If you're using multiple banks, you'll have to rename files with this version. See the release notes on the download page.

Need more info: What firmware version are you using? How are you triggering the tracks? With trigger inputs or serial control? Is it consistent? Does a track that seems to play too slow always play slow, and fast tracks always play fast? How did you create the tracks? What's hooked up to the WAV Trigger?

SD Card and track length should not matter. One test would be to remove all tracks from the card except one that plays too fast, remove all the connections to the WAV Trigger and play the track using the on-board button (which plays the lowest number track it finds.)

I must admit that cold is not something I'm used to worrying about. It's usually heat. That doesn't sound too bad, but I certainly have not conducted cold tests. The only thing I could suggest would be to look at the individual datasheets for the major components and review specified operational temp ranges. Not conclusive, but at least a start. Condensation would of course be bad.

If you record the output using an audio editor like Audition or Audacity, you can zoom in on any anomalies. With sine waves anything wrong should be readily apparent - both audibly and visibly. I've spent a lot of time looking at the output and haven't seen or heard of this, so I'd be very interested in any data to the contrary.

The introduction of pitch bend / sample-rate control in firmware v1.05 did introduce some distortion (bad samples) when actually pitch bending, but this was fixed in v1.08 and later. Firmware v1.10 is on the download page now.

It's not supposed to do anything if you request a track that doesn't exist. Assuming you're using the serial control port, can you post your code? Are you using the Arduino library or did you implement your own serial port code? Which command are you using?

Sorry you are having problems. A volume of +10 (dB of gain) is pretty loud if your wav file is already hot.

When did you purchase your board? I ask because the first version was missing a capacitor and needed this mod when using the on-board amp. The latest version (WIG-12897) has the cap and does not require the mod.

I assume you have tried another speaker? I would agree that it's more likely that you would have fried the 0.25 Watt speaker. The stereo jack output still works, right?

I'm glad I read this. I was struggling to get my files to play at all and it turns out they were mono. I guess I was interpreting the instructions to mean the trigger supports UP TO 16-bit, 44.1kHz stereo (assuming that anything less would also work). I could then get the sound files to play, but they were 4x faster than they should have been. In Audacity I needed to set the "Project Rate" and the sample rate to 44100. For some reason, initially the project rate was 11025. The wav sounded normal from within Audacity, but 4x faster from the trigger. Once they were both set to 44100, all is fine and dandy.

Thanks for the feedback. I've tried to be clear that the WAV Trigger supports 16-bit, 44.1kHz stereo wav files. Mono support would not increase polyphony because it's the number of files, not the number of channels, that's important. In fact, mono would actually slow things down because it would involve extra buffer copying. I made a short video on using Audacity to convert audio files for use with the WAV Trigger.

Not sure where you got that impression, but the WAV Trigger is a play-back only device. It reads FAT16 or FAT32 formatted microSD cards which are written using a computer and a microSD "reader". The WAV Trigger serial interface is just that, not a USB device connection, although you can use a USB-to-serial converter like the FTDI Basic to create a VCP connection to your computer. But that won't support a class-compliant mass storage device connection. Hope this helps.

Well, no. Not directly. USB-MIDI requires an on-board USB device controller and the appropriate descriptor tables for a class compliant MIDI device, which the WAV Trigger does not have. The easiest USB connection is using the FTDI Basic, but that's just a USB-to-serial converter, and looks like a COM port on the host computer, not a MIDI port.

One way to get true USB MIDI would be to use a Teensy with its class-compliant MIDI library (I've never used it myself) and then serial (MIDI) to the WAV Trigger.

Yes, if you're using MIDI, as opposed to the serial control protocol, then your MIDI controller needs to supply pitch bend messages. You could implement your own pitch wheel, or use one from another midi controller, and then merge the MIDI streams from that and your PIC controller, but that would be a bunch of hardware.

Are you using the Arduino WAV Trigger library or your own code? Can you replicate the problem just using a single track and the steps above? We should probably spare the product page the back-and-forth. If you email me at info(at)robertsonics(dot)com, I can follow up with specific questions if necessary.

LM4990 datasheet indicates 1.25W into 8ohm, and 2W into 4ohm. Triggers can be programmed to be master volume up/down with 0.5dB steps. If you want a continuous controller (pot) for volume you'll need to use the serial control protocol - the WAV Trigger does not have an analog input. The serial control protocol does indeed have a master volume command, which is also supported in the The WAV Trigger's Arduino serial library. MIDI Controller #7 also does master volume (In MIDI mode).

If you're planning to drive the speaker at high volumes, I would recommend limiting your supply voltage to 9V - 6V would be even better - to prevent the 5V regulator from getting overly warm.

What did you have in mind? The only external communications path to the WAV Trigger is the serial port, which is limited to standard serial baud rates. Even if the Wav Trigger supported writing files to the microSD card, which is currently does not, it would take over half an hour (best case) to update 20MB of audio - presumably longer than opening the box and changing the card.

Are you thinking that you want to update remotely, without physically going to each location?

I think you're talking about a different, and possibly more expensive, product. How much would you be willing to spend for something like the WAV Trigger but geared more toward commercial use cases such as yours?

I did it this way for two reasons: First, the current track numbering system only allows for 999 tracks - not enough for 16 channels of 128 notes. Second, I was trying to keep the MIDI note number to track number translation simple. I'll point out that there's no reason you can't use all 128 notes in a given bank - that will work, it's just that note numbers 100 - 127 will overlap with the next channel/bank.

The right way to do this is to increase the number of tracks to 2048 (16 * 128), and to make the track number = (Channel * 128) + Note. But in order to get the latency down, I have to store track file info in RAM, and I'm not sure I have enough RAM to hold that many tracks. I'll have to see...

Thanks for the input. I only added MIDI support after I realized that the performance I was able to achieve allowed the WAV Trigger to be used as an instrument. It's a work in progress!

(PS - Thanks for the kind words.)

This is not normal behavior. Did it work before you soldered to it? Where is the noise coming from if you don't have the audio output plugged into anything? If you suspect that the board is malfunctioning, best bet is to contact Sparkfun tech support.

At the moment, no. The only way I can think of to control individual track volumes via MIDI would be as you suggest: Using channel volume to change the volume of all the notes (tracks) triggered on a given MIDI channel. But there's no support for this in the code today. All I can suggest is to use an Arduino to receive MIDI and use the WAV Trigger's serial control protocol to do all of the audio control. That would give you ultimate flexibility to control the volume of any track, along with looping and sample-synchronous playback.

My thinking has always been to focus the WAV Trigger on being the best audio mixing engine it can be, and to use a separate micro for more complex and custom control I/O functions and logic, and the serial control protocol to control the WAV Trigger audio playback.

Piezo triggers present several challenges, the first being that they produce very short spikes of high voltage - as much as 20V or more. You can't simply connect a piezo trigger to a GPIO of any micro without risking damage to the input. Each piezo will need an interface circuit to prevent the input from seeing more than 5V. Secondly, if you want it to be velocity sensitive, you have to sample into an analog input at a pretty high frequency so as not to miss the spikes. It's no longer a digital input.

I'd suggest using something like a Teensy to interface the piezo into an analog input and sending serial message to the WAV Trigger. You can google "Arduino piezo drum trigger" and find suggestions of how to connect piezo drum triggers to an Arduino input.

Not yet. Am planning to add this capability to the serial protocol, but haven't decided how to implement. Would you be satisfied with short filenames, i.e. 8.3 DOS format? Supporting arbitrarily long filenames will be a little challenging.

The WAV Trigger plays 16-bit, stereo, 44.1kHz .wav files, with no meta-data. It's likely that the wav files you're trying to play are not in that format. You can use the free audio editor Audacity to easily convert from any format, including mp3. Here's a short video showing how.

I've used this one, as shown in the (out-dated) video below. Mine came as a kit, and I only populated the parts required for input. Worked perfectly. What's the other one?

Yup. I've actually already done this with the violins, choir, cello, flute and full orchestra - Just haven't had a chance to post all the files. I've also added support for switching banks with the MIDI Program Change message, and am making a short video showing how all this works with a RockBand 3 keyboard, which provides Program Change capability from it's controller buttons. Hope to get it all posted this weekend.

It's a little hard to say what's going on from the description. I'd suggest trying to simplify. Can you duplicate the problem if you take the Arduino completely out of the equation and just use the WAV Trigger's button to trigger a track with one of the cards that exhibits the problem?

Well, that kinda sounds like a bug to me. I'll have a look and either fix it or explain it better. You may need to give me until the weekend. What version of firmware are you using?

You don't need any other board or software if your switches or buttons are either contact closures or logic level (3.3V or 5V). In fact, I show how to do this in the video link below (The video is out of date with the latest firmware features, but it demonstrates the concept.)

However, piezos generate large voltages for a very short period of time, and if you hook them directly to any processor's GPIO logic input you risk damage. You will have to experiment with at least a zener diode to clamp the voltage to 5V or less and perhaps a R/C filter to lengthen the pulse.