Once we have a non functional prototype we only have to add the electronics to make it functional. First we have to insulate all the different parts within the shells (iron wire). Once this is done we can add the prints. We also have to connect the two sides with each other. This goes through the back.
Another thing we didn’t mention yet was a new connection point. This point is used as ground. This point is on the earlobe. This is no problem because there is already a connection on the earlobe now we only make it a conductive point. We do this by making a hole in the mousse and putting a piece of steel-wool inside it. To this piece we connect a conductive cable and connect this cable to the print.
In this final part I added the tentacles to the headset. First I had to drill several holes (for each tentacle one) at the correct places. Then I inserted all the tentacles. I fixed them using the iron wire and the back plug.
Then we close everything up again and the prototype (without electrics is ready).
We will make 2 applications: one working on a computer and one as a show model on a smartphone. The show model will have the look and design of what we want the app to look like. For the opening screen we want to make sure we nail the look and feel we want to project to potential customers. So we made multiple options. But the one we chose was this one:
In this second part I will explain how all the parts are assembled.
We started at the cover plate. At the back of the cover plate we need to build a case for the band to fit in. We glue these parts together because they will never have to open again. These cases exist out of 4 parts: the plate, the two side parts and a top part. In the picture you can see it without the top part. The band can’t leave this case because of the vertical part which fits in a slit in the two side parts. This part is glued on the band. Once this all is done the top part is glued upon the whole.
This part can be screwed in the shell part. In the shell there are compartments printed where plugs can be put in. The holes in the cover plate will perfectly align with these holes.
Now we only need to attach the speakers to these parts. This is done in a few steps. The speaker is attached to a plexi part. The mousse has to be folded over this plexi part. The speaker will be directly glued on the cover plate and on top of that the plexi part will be glued. This provides a small gap between the cover plate and the plexi part. Because of this gap the mousse can be attached to this plexi part.
Now we only need to attach the tentacles and make from this normal functional headphone a emotion reading headphone.
In this post I’ll explain the proces of making all the different parts.
The shells covering the ears are 3D printed. First we surface modeled these parts and then we printed them:
The closing plates which cover the inside of these shells we made in two steps. First we lasercutted these parts out of plexiglass. Once these parts were cut we heated them and pressed them against a mold so they have the right curve to fit into the shell. There are also some other parts lasercut. These parts are the parts that are used to hold the band at the back. These parts don’t need any extra work after they are cut.
For the part at the back we use a steel part. This part is formed into a curve using the metal bender at school. The radius has to be slightly smaller as the radius of a head. Once this part is formed there is added a layer of mousse at each side. These three layers (mousse/steel/mousse) is also covered in leather to give it a clean look.
We reuse the speakers out of an old headphone but we make new mousse parts because we want them bigger so they cover the entire ear.
In a previous post we explained how the tentacles will be made but we changed it slightly. We don’t use foam on the end but a sort of thermoplast clay. This way the end will be strong and the connection between the mousse and the electrode will be assured.
There will still be an iron wire include so the user can form it according to his own head. This iron wire and the cable will be encased in a mousse so it is soft to touch and this whole will be covered by leather so it will have a clean look. This leather will be sown so the sew will be invisible.
For the connector to the skin we chose steel wool because of its conductivity and its soft texture.
When the second PCB was designed, we have tested the PCB last week. In the measurement set- up, the 2 PCB’s ( one for every channel ) was tested. But we didn’t know which test set- up was the best so we tested different methods.
Here below, you see the scope figures of all the test set- ups. We will go through them in this post. The main difference between them is the material that we have used to connect electrode and head or how we connect the ground of the PCB with a certain point on our body.
Here you see our first measurements. Here we used a mousse material to connect elektrode and head. the problem was that the measurement wasn’t consistent because the connection wasn’t always very great, this was a good measurement, but when we had a bad connection we measured 50 Hz noise.
Because the measurements weren’t consistent, we have searched for a solution, so we used steel wool, because that material conduct very good and we connected the ground to different places on the body like ear lobe, center head to look for all the best results.
Here, you see a measurement with steel wool and the ground connected to the center of the head. Here we connected the reference to measure the signal with the ear lobe. We have seen less 50Hz noise, but the measurement was not always good.
Then we used the ear lobe as a ground connection and the center of the head as a reference connection, this can we see in the next scope figure. We have seen that the measurements were much more reliable and we think that this will be be the best measuement method for our application.
In the last weeks, we have designed a new PCB. This PCB has some improvements since the last PCB we have created.
On this PCB, we have our instrumentation amplifier to measure the difference between the active and the reference elektrode. Then, we have our high pass filter to block the DC- signals and our low pass filter to block all the frequencies higher then 14 Hz. We also added a notch- filter to extra block the 50 Hz- noise in the environnement like from the electrical grid.
We amplify the measured signal ca. 200 000 times to read it in the µC. We do this in 2 stages, 100 times with the instrumentation amplifier and 20 000 times with a non- inverting amplifier.
Here below, you find some pictures of the PCB and the first measuresement set- up. We discuss the measurement in a next post.
Because the electric cables are 7-8 mm thick, we decided to use other materials instead of 3D print material. Also because we think it’s not strong/flexibel enough.
We make a had for the connectors so it will stay on his place. This has to be a light but strong material. In the back there’s going to be 2 holes, 1 for the electric cables and one for the iron wire that will support the stability of the connector on the head.
(the cilinder in foam is the connector)
Next step: we put foam around the wire. We kept the half circle design so we put a piece of cardboard under the wire.
Because we want that the headset looks expensive, we finish it of with leather.
This week we presented how far we are right now.
We did a few tests to read in a good signal. We did this in different set-ups. (KENNY?)
The results of this are:
We also worked further on the design of the headset and made a few renders and a video: