I’ve had Kaseta on my wish list for a long time, it seems. Then it came up on Thonk new products page. I’ll describe the DIY build and then talk about using it.

Building Kaseta

Kaseta is designed around a Daisy Patch Submodule, which came programmed with the latest firmware. One PC board holds all the panel controls. The Daisy Patch plugs into the back. It’s a very streamlined design and easy to build.

Twenty-three pots mount on this panel, four of which are slide pots. There is one button and nine LEDs. After soldering headers on the back, all of the panel parts are inserted, then the panel placed on top, parts secured and everything soldered. It’s a bit of a juggle, but not too bad. I had trouble only with one slide pot that had a slightly bent pin that was obstinate about going into its hole. Two kinds of pots are used, 14 tall trimmer types and 5 right angle with Davies clone knobs. The manual says not to solder the pot frame tabs. I obeyed this for the slide pots and right angle ones, but I did solder the tabs on the tall trimmers, because otherwise the only support is the signal pins.

When I was cleaning a bit of flux off, I smudged the serial number. It’s 229 (also marked on the kit bag, and PCB photos, above).

The panel is aluminum. Very nice looking.

Trying it out

This delay is mono in, stereo out. It has four virtual read heads, each with buffer position slider, volume in the mix, feedback level, and pan pot. The four CV inputs can be mapped to any pot and can also be mapped to a clock input.

The flow diagram is essential to knowing how to use Kaseta. The input section includes preamp and saturation in series, with four pots to control the level, drive and bias of the saturation, and a Dry/Wet knob for the saturation. It’s not a Dry/Wet for the delay!

Notice that the Feedback level is independent of the output level for each read head. You often want to feed back from later areas on the tape, but without mixing that into the output.

The filter and wow/flutter can be placed in different locations in the flow.

There are three possible ranges for the Speed. By default, this length is between 5 minutes and 10 ms. Imagine my mystification when my first attempts took up to five minutes to clear the buffer! Setting this to between 8 seconds and 10 ms recovered my sanity.

Here is a short demonstration that starts out with a ping-pong delay and then plays around with feedback and different speeds. For input I used a drum from Starling Via Meta, using Meta’s trigger button to make a single hit. I changed the pitch occasionally to demarcate hits.