The time it takes for one marble to drop from here to here is constant. It doesn’t matter how fast I turn the crank, this time will always be exactly the same But if I turn the crank slowly. The crank handle doesn’t travel so far during the time it takes for the marble to reach the kick-drum. So that’s the 90 bpm. But if we then increased to 140 bpm the crank handle travels much longer during one marble fall. So the kick drum will be late. I have one contact microphone in the kick drum and have another contact microphone down here. And the one down here will react exactly to my crank because this note of the rhythm machine has the same rhythm as the crank. On the bottom here is the kick drum let’s listen to it. So you hear that the kick drum and the rhythm machine is not playing together. We need to fix this and depending on the tempo of the song the kick drum and the rhythm machine will be in different distances from each other timewise. We need something indexable to offset this latency for every specific song for every specific tempo We need the musical timing clutch. So this music timing clutch is designed by Will from Machine Thinking in collaboration with me but Will came up with most of this cleverness with this function and here’s also machined the pieces. And he has made an amazing video about how he made this masterpiece. Machine thinking is a wonderful youtube channel I can easily recommend you to go over there and hit subscribe because, he does fantastic stuff. I have set up a test on the Marble Machine X to show you exactly what this musical indexable clutch do and how it works, but first I want to show you how it felt to unpack this masterpiece and I also needed to cut some key ways to mount it properly to the crankshaft. I have set up a test to show you exactly how this thing is working. So the music timing clutch is here. I made a program with a note for every fourth bar. One two, three, four, one. And I can see with my color-coded registrators, I know that this is the first channel of the kick drum. You see they keep from moving there. The black and white registrators are cnc’d from black and white POM (Polyoxymethylene), also known as delrin, which is a great bushing material. And I’m now going to perform a timing test to see if we can get the kick drum, to play tight with the rhythm machine. So returning the machine in 90 bpm. And now I’m gonna activate the kick drum. On the top here, is the temporary rhythm machine sound from the whole note and on the bottom here, is the kick drum. Let’s listen to it. We can immediately notice, that the kick drum is not playing at the same time as every fourth stroke. We can although notice, how consistently it is off. Let’s now check how much it is off. 210. So it’s 210 milliseconds off and the kick drum is late. So now we’re gonna try to take this kick drum stroke and move it to hit that rhythm machine stroke. Which means, that we want the kick drum to play earlier. Which means, that we will offset the programming wheel earlier towards the crank with the indexable clutch. On the Marble Machine X the relationship between the crank and the programming wheel is sacred. It’s 1 to 65 gear ratio, and they are really, really connected exactly to make all this line up, musically. But with this timing clutch, we are able to break this gear train. And adjust the relationship between the crankshaft and the programming wheel, we wanted the programming wheel earlier. I’m gonna move ten steps. Like that. It’s much closer. Listen to the previous example. And the new one. Tackle! The previous value was 200 milliseconds off and now we’re 100 milliseconds off. So let’s go to the indexable clutch and put ten more steps. Two, three, four. Haha! This is working really awesome. Oh, no, there’s some issues with this cable. I have to redo this test. That’s a shame because it’s actually just nailed. This street here… marks the smallest possible adjustment and now I wanna go the opposite direction. I wanna go one line like that. Test four I’m going to give you a close-up of the marble gate. This pulls this back, which closes the path for the marble and not until the fork has returned, the marble will be let back. Okay, that was a marble gate action. Flywheel is still going over there. Stop. Okay. We have good signal for this recording. Haha, look at these transients, it’s absolutely unbelievable. Finally I build something that works. Now I’m gonna head over to the machine and perform the exact same test, but this time, in 140 bpm. And you can play a game and guess: What’s going to happen with the kick drum? Will it be tight with the rhythm machine? Will it be earlier than the rhythm machine, or will it be later than the rhythm machine? So, have you placed your bids yet? My guess is that it’s going to be late. Yes. And how late? 970 to 041. So. 70 milliseconds late. In 140 bpm. And that is why we need: This beauty. Now we were 70 milliseconds late in 140 bpm. Which means: We have to, again, make the programming drum go before the crank. We have to advance it this direction and we know that ten lines is about hundred milliseconds. So. Let’s do seven lines. One, two, three, four, five, six. For reference, let’s listen to the first test in 140 bpm. You can clearly hear the kick being late and here is the new test with the first adjustment of the timing clutch. So the kick is late, which means that we have to forward the programming wheel that’s this direction one, two. For this test, I want you to notice how the Fish Stair and the conveyor belt moves in rhythm with the music. Thanks to these Pistons, that are connected to the crankshaft in one to two gear reduction, they move halftime compared to the crank. So this actually moves on the offbeat. Sound test. Listen to the conveyor belt. Let’s check the results. Pretty awesome. Yeah, 10 milliseconds behind again. One more click on the indexable clutch. Look at that. Do you see that transient? I, can’t tell you how happy I am when I see those kinds of results. This is what I’ve been fighting for. If you’re a musician. You might have thought about already that if I play the Machine without a computer click in my ears the rhythm machine will tell me exactly what bpm I should play in, so I will play and play and play and when I start to get a little bit slow, the marbles and the rhythm machine are slowly gonna go out of sync. So it will tell me that I need to hurry up a little bit. The latency of the marble drop will actually give me a help, to keep a certain bpm. So I will feel the sweet spot and I will then maintain a bpm throughout the whole song. That’s the indexable clutch, everyone. By the way, you’re watching Wintergatan Wednesdays. I want to say a personal thank you to everyone who are following along in this long serious, this long journey I see more and more people getting invested in the destiny of the marble machine X. So, thank you so much for watching. And I want to say a special thanks to everyone who are supporting this project through patreon or through YouTube channel memberships. Thanks to your continuing support, I’m able to get more help from people around me and I can focus on finishing the Marble Machine X and I’m also able to say no to sponsorships money. And I can’t tell you how proud I am, that we can keep the integrity of this little corner of the internet that we have together. I am so thankful for that. And lastly Will from machine thinking thank you for your magical contribution to the Marble Machine X. I met will in Alabama and on the trip over I have made these sketches of the timing clutch. I showed them to Will and we hit it off immediately solving the problem together there and then in like two minutes and that was the start of the collab. Head over to machine thinking to check out his video on how he built the first brass piece ever mounted on the Marble Machine X. For being an afterthought, it looks like the machine was designed to have this part from the beginning, which is a sign of great engineering. So, thank you Will and thank you for watching.