 and I think we are ready to go and it's open source organ construction I'm really excited about it all all things that you can build yourself has is being open sourced and prototypes prototypers are conquering everything and I'm very glad Benjamin Wandes here to talk about organ construction with 3d printers and electric tinkering he brought along yannick yannick Bayer stadt is assisting with the mechatronic parts and communication technology parts and I'm very excited what the update is on this topic thank you very much yeah hello thank you for coming on the last day that you may find your way here most people probably think an organ looks like this but in hacker spaces I wouldn't recommend building a big one may or more like this one maybe you could wheels on it maybe I should say that I'm not planning to build a big organ but if a hacker space wants to build one I can really support you I'd really like to do that and maybe trying to define what an organ really is their data has to go in a compressed air has to go somewhere and data and air have to convert it to structured air and that goes into the pipes and sound will come out there are some gray areas maybe is an harmonium is that an organ but I don't have to answer that it's just a gray area and you could also say that you can make a wider definition of an organ and that's perhaps something like a machine that that makes continuous sounds that are not percussive sounds and then maybe a Hammond organ is also an organ then in that definition I wouldn't really require a manual or a keyboard because little automated automated things that also produce tones like that may also be organs and here are some pipes pictures of some pipes typical organ pipes until number 14 they are lip pipes they work like a recorder and there are read pipes they more work they work more like a clarinet and there are some miscellaneous pipes other pipes who likes experimental things who can they can they can search for Vox Maris that's another another pipe type with very high pressures and very high volumes except some very few exceptions a wind instrument always needs a generator and a resonator and the resonator is the pipe and the generator is what makes it resonate what makes it oscillate with a brass instrument you use the the lips and maybe there's a reed in some instruments and some resonators and generators have different sounds and that's why there are different types of organ pipes so they all sound a little bit different my own designs I always thought that you put in a pipe you and you are restricted to cylindrical pipes then all the pipes are parametric in in free cat or open-ass cat and you have to put in the value and fall for diameters and such into the file and then export as stl and then you can put it into a 3d printer in in free cat it's a spreadsheet there's a spreadsheet and for entering variables and there's one one area to to enter your values you see here some schematics of metal organ organ pipe on the left and a recorder and some of the things that I do on the right there are many reasons for design decisions for details for organ pipes they're all quite different but roughly they 99% of the reasons why they look the way they do roughly are questions of practicability of building the things these metal organ pipes are from from sheet metal soldered together and that's why they have thin walls everywhere and there that's why there's air in the bottom here because it's the easiest way to build it on the other hand a recorder you just drill a hole into into a piece of wood and then you make the lip and then you put in the block there and that's why there's the big black block there that's caused by the way it is built and if you use a 3d printer there completely different set of rules I I push in the block of wood this isn't relevant here but what's really interesting here is that on the bottom it is it is hollow below the below the lip and I didn't think of that immediately but at first I looked I printed pipes like recorders and and yes that works but the the generator hasn't doesn't have to be at the end of the pipe so if you look at a transverse flute and the generator isn't at the end of the pipe and I really didn't look at it but Helmholtz said the generators should be at the seven eighth point so if you don't know where to start maybe that maybe that's a good starting point at the seven eighth of the length I have made several flu pipes from the thingy verse and I wasn't really excited by any of them these organ pipe shape depends on on the way it is it is machined after building with with machining tools and with organ pipes with metal pipes that works very well and it doesn't work very well with plastic from a 3d printer and you don't really have to imitate the shape of an of a normal organ pipe if you really want to have something like that and I wanted to say something about the design programs I started working with free cat and then went over to open scat and they are very different approaches but in practical handling I'd say if you need to do something very quickly just use free cat and you can just just roughly make something and if you think it's going to be a real big pro project and it has to scale then maybe it's a good idea to to code it again in scat because then it can really version can be versioned and it doesn't crash quite as often one of the reasons why it's such a problem with free cat is if you enter the parameters in the spreadsheet then after every entry the the model will be recalculated and maybe there are mathematically impossible models and then it'll just hang and crash you can try to to enter the parameters in the correct order but it's quite unnerving and takes a long time and it's annoying it and it's no fun at all and a problem with open scat that I found is that the files can only work with specific versions and someone gives me an open scat file and then I then it doesn't work and you have to talk about program versions and it can be fixed often but it's still very very not quite ready yet maybe they really have to think about what they really want to achieve so what about plastic doesn't they don't they break quickly well organ builders really claim that they their organs last for centuries and that's a big thing to to claim because they are moving parts and an organ always comes along with a maintenance doesn't come along with a maintenance contract and they sometimes break and then they are half broken or something and I would suggest that be realistic about that and make some estimates how long the parts will last and you could perhaps consider that one version lasts for 100 years and one last for 50 years but only costs 20 percent and you can weigh your priorities and I wouldn't really print something that will should last from PLA that's nonsense but you can also print nylon and you can also come and look at this I have something printed from nylon and that's almost serious really okay on the Easter hack I had just finished design with a 45 degree lip and I couldn't play it because there was something a problem with a 3d printer on Easter hack but it's done now and I can demonstrate it now it's what it looks like I did that because I always printed them at 45 degrees angle and all the the flu pipes had problems with the support materials and if I make the lab I'm at 45 degrees then I can print them upright and it'll stick to the building plate and there are really no problems with that it makes sounds just like other flu pipes and that's all fine I couldn't find any difference really if you build it 45 degrees or straight up and spring I had done experiments with the reed the regenerator but I didn't follow up on that and I really wanted to to mention it because I still have the project to build a prototype with real dynamics like with a piano and so keyboard dynamics I mean if you if you touch more if you hit it harder then it's going to be louder and if you maybe something like this so this is a piano and for all non-musicians and if you make the same thing with a Hammond organ it's always exactly the same volume and I would introduce the spectrogram spectrogram and I did that rather lazily on the Easter egg rather sloppily and the lower the lower bars are always what you hear consciously and and everything above that are harmonics and whistling has hardly any harmonics but but singing does and can you can you see the stripes there and I think that's really a very interesting toy if you if you build instruments as a hobby musical instruments as a hobby and it's a very nice toy and then you can you really have a visual feedback for for what you are doing there and where's the where's the pipe here okay here's the pipe so this is a pipe that I had a moment ago one problem with the dynamics is there there are several problems but one of the problems is that the flow pipes the pitch changes when you blow hard and if you and they will overblow if you blow harder they overblow and go into the upper octave so just change the the airflow is quite limited and I have this I have an inverted read generator here and you can see in the spectrum that it gets more overtones more harmonics but it doesn't so you can see it gets more harmonics but it it the pitch doesn't change and they don't overblow so if you do something with dynamics then that works so it's a potential candidate for high pressure register a high pressure stop and most flow pipes that I have made they work with 40 and 50 between 40 and 50 millimeters of water column and the read pipes more or less at 100 so talking about millimeters of water column this is there's this toy here this is an experiment making a very simple intonation work and building an organ they have different jobs some people build the pipes some people tune the pipes and tuning the pipes isn't done during the concert or in the organ but in the in the workshop and you can try them here you can why don't you just blow in with your mouth because they contain lead but now I can't open the bottle right otherwise we'll have to take water that's not how it because organ pipes are specified for a certain air pressure and this specified air pressure isn't working when you're blowing with your mouth I'm the most of the time where there were organ where we build organs people couldn't read and write but there's no problem there's so what I'm doing with this is I'm filling water inside this columns oh it's been a bit of much water where is my cup it's a little bit of experimental setup could you actually see this it's a bit shitty that's this colored water okay at least it's now a good amount of water can you see it on the video yes it seems so here you have it open you have it right because then you can actually see it so that's going to be out of the setup so this is also print from a 3d printer so just a little bit I suggest to fill it into the half so that's been meant and this this thing here okay I'll hold it up so that you can see it and then I hold it in a camera so this was my first try to make a nightingale I put inside their real water and what we are trying to find out experimentally is Janik please hold the nightingale so that it doesn't turn over and I make a lot of mess so I'm blowing inside the the chamber and we don't hear anything by now so there are there are keys like a gravure and now they're coming tones and the question is at which pressure it started trilling because that's what it should do so we take a different one because it sounded shitty why is it not working but here is the valve broken that's the thing when you build something on your own okay okay oh there's not enough water inside one can't see it from the outside how much water there is in them let's try it again yeah right so now there's tone and on the difference of the both water columns you can see what pressure there is try it again okay this is what I meant with the starting right and this is how you can attach 3d printed organ pipes to this and test them that's the goal of the exercise here okay as I said I wanted to do something with dynamic and one could now different take different input vectors for the dynamic and if you measure the velocity your turn pull you can do it like a digital piano and then convert that to the to the volume to the amplitude and then you can also measure how how strong you press the key down with what force that's really something you use for real in real organs and I wouldn't call it dynamic but expressive play because with the following reason with half pressed keys no one can play that means if you take the height of the key as an input vector you wouldn't influence the loudness of the key but the start of the swinging of the pipe and this is of course interesting but it's not the dynamic as I thought it should be what you also could take is the pressure on the key and I have a pressure sensor here which I take for as a sensor for this and what you see is just read in with an Arduino this is uh this here is what I think it should be and you can come forward to me after talk if you want to see it now you would need not just a sensor but also an actuator and not just one but many of them these are the wolves as they are inside the intonation needle and there is there's a block and a spring inside I decided for this form because I could 3d printed it very easily I don't know I wanted to do this and Janik wanted to work with actuators for a long time and the state was for a long time on maybe but as he's seen this it was going forward so we swap seats all right okay the problem was we need some sort of valves that are cheap because we need a lot of them and otherwise it's going to get expensive very quickly and the valves have to be driven somehow and I started with something like um like a cone pushing a cone into another cone and something like that and then Benjamin came up with something for the for the keys there and that's where that was my starting point and for driving it we need some position positioning drive and it's uh it has to know the position or I need to know the position when I want to drive it and so I can do it by knowing the position from from the driving itself some people will know what I'm talking about but I'll explain it in more detail or I'll have a positioning coder and they are linear or or for for rotational movements but I'm then I need a control feedback loop and I have one measure one value that I measure and I know where I want to go and I have to put the control to such that the difference goes to zero and for this feedback control loop I need real time controlling features and if I do that with one microcontroller with maybe 25 drives there that's rather challenging perhaps and a lot of needs a lot of computational power and the hardware is quite expensive then so what possibilities do I have there there's this kind of drives that the actuators that Benjamin showed us from from auto hoist making parts for organs and selling them and these actuators can also regulate and change the position I want to have it 100% open or 50% open but they are quite expensive and then there are two two coils working against each other and and it maybe uses a lot of power I had tried to do that with some solenoid valves um so what what a loud speaker makes what a loud speaker uses and but that's quite expensive it doesn't really work very well and there are these RC model servos they are rather cheap but when they when they are cheap then they are they're expensive no no sorry then they're if they're cheap they're loud and also quite slow but controlling them is easy because the position control mechanism is inside the servo and there's something that that measures the position and it's all automatic and the professional version is this is a servo actor what's used in industry in or in the powered steering in a car there's a there's a motor and a positioning coder this is only the encoder you see and with with good quality quality that's quite big and and expensive so yeah and many of you probably know from from 3d printers no stepping motors and by controlling them it is known where it is at which position it is because for one controlled step there's a certain rotation standard stepping motors probably usually have 200 steps per revolution so 1.8 degrees you can do some tricks to to increase resolution but normally you know the position so what are the advantages and disadvantages of stepping motors the advantage is you know the position unless you put too much force on it too much talk on it and it then it may skip a step and then you have a problem and if someone with a 3d printer knows the problem maybe one one shift one layer is shifted then but position control is is relatively easy because the controller controlling is you just have to count the steps in each direction and just calculate plus and minus and an addition subtraction and you know where it is and the position resolution is quite good it's quite high it's precise enough for what we need to do the first estimate is that you perhaps you need about 20 steps 20 volume steps to to use to make a useful discrimination so that you don't really hear steps in the volume of the instrument so yes that was my my estimation about 20 perhaps and the disadvantage is that normally if you have small stepping motors that are cheap then they have a very low low force and low torque and the valve design that we currently plan to build and that's not a problem i'll show you in a moment and they are sometimes loud and it depends on the way you control them for example in a 3d printer maybe you know that they make funny noises if they drive back and forth and there are several special driver chips that can do that better and are almost noiseless very quiet and then about valves it's nice if you just have this the key at the bottom and just have to push it down and then i why does it have to be a linear movement you can also do it in a in a different way okay so i can it also simply build in a way that the motor is put directly to the valve because then you put this you but this is not good because if you put you need to put the spindle there and then you have this nut which is going forward and backwards which is pretty pretty expensive in terms of parts and building but you also could put their stepper motor to one attach one stepper motor to one of these tubes which then turns around the tube about 90 degrees and you can think about how far you might go but you need to take care of that you can switch fast enough because the motor needs some time so for the controlling that's pretty straightforward you use a microcontroller which is a processor for professor for multiple in and output per stepper motor you need three in and outputs because the most easy interface for a stepper motor you have a step direction interface so you have one pin you go from zero to one and each each time you change the the amplitude you go forward or backward and then you can say that it should go one step this this you're doing over the direction pin also you would need an end stop switch so that you know that you are being on your final position or position zero and the valve is completely open or closed so you need for for contacts and there are controllers which are which can handle this I have a Tiva bot which has about 100 GPIOs which you can use so this is a good thing you can use for control organ pipes you can imagine one octave has 12 point tones and if you want to have two two tones you have 24 if you would have one tone for the third 25 and therefore you could use with my device control with my device about 30 pipes so now we have a valve we have a controller we have a driver for the controller we can attach it to the microcontroller and what we are still missing is the data from the input device it might be a keyboard with force sensors or you also could use punch cards from turning liars you would you would get rid of one digital layer by using this mechanical way to control it and it would be a good idea to use something like this for a generic storage way for what someone put into a keyboard and then bring it to a hardware device so our current idea is to use MIDI which is a serial protocol over the UART port going to the different devices if you don't know what this meaning it's universal asynchronous receive and transmit so universal is asynchronous reception and sending the asynchrone means that is just so on the microcontroller programmed that is not so important but it is just not so easy to use a very simple so medic can even more first bite says what what what it is what needs to be done and then spun for not on and not off and there's various different things and the second is the pitch of the tone of the note and the third is velocity that's usually used for for loudness for for the the amplitude so what really should be better in the future is a protocol called open sound controls called OSE that is based on network based you have ethernet cables and switches and they talk to each other using UDP packets this it's possible to if you don't build just a small device with but have many pipes then you need more controllers on the output side because sometimes you don't have enough pins on the controller and maybe you really want to have different input devices and on an ethernet switch that's very easy to to connect them rather than with MIDI cables and loop through and things like that and it has some more features for example to change the volume on a note that's already playing and it's quite easy to build a network like that and you could think how to configure that but you'll have one unit with one organ voice and it says you have c3 up to c5 and for example then it could react to the network inputs and not to any others okay now we perhaps have a small demo if it works perhaps hopefully it's all just during the last days we covered that together and i would have liked that it sounds like something yeah i wanted to try this with our our motor to do a tones but i left the air pump back at home so you can just see the motors can you see this yes maybe no put it a bit more over to me you might see now eventually a big heap of cables like what as it is today you use a breadboard and put in a lot of stuff there but later on we wouldn't use different boards with chips on it but in a way cheaper but as said the problem was to print out all the walls but if we would put pressure here this key you would see that the motors would turn around and in theory it would play a free software song but not here maybe in the other demo i could turn on the free software song and it says they are pretty fast if you just go to 90 degrees so that's why we choose them and if someone is interested you can come and ask me and this is what in general should be enough to play fast with an organ organ music is in general pretty slow what some of it is slow but the question is how fast you can play the tones it's depending on how fast the pipes are going to respond and that's more like the problem you know from here i don't hear anything i was i had my problems with stepper motors they might be loud but i was wrong the other one i did with the coil i pushed it on the on on a wooden part and you can't hear it with a wooden part so they are pretty quiet so now so we came to our end and everything else in a q and a for details or other questions or ideas what you we can improve thank you there's a call for the mic there's one mic one so hello what's speaking against the mechanical control like mechanical swing and then you put down electromagnet yeah that works if all you want is turning on and off and i really want to make a dynamic organ but with dynamical pipe the freed pipes this wouldn't flu pipes flu pipes yeah okay so i have to i have to go back a little more there are lots of pianists you couldn't play a caviar chord and of course there are people that can play the caviar chord and of course you couldn't play on that and sometimes i sometimes we go i started playing caviar chord and the first two months it sounded horrible and you really have to if you if you really have such a sensitive actuator which changes the the pitch that doesn't mean you can't use it it's something that you have to learn or what the the musicians have to learn to to that the pressure is is measured and the and that the pipes work differently than otherwise and you can perhaps with that not all the literature that that exists but i think that even with flu pipes and dynamics you can do something it's not a big dynamic range but i think it's a big gain in expressivity for which it would work and i think that if you do it with dynamics with flu pipes then maybe you end up with literature like for the caviar chord with only two or three voices because you can't control more than that so if you can do it that would be great but um a control board would be nice yeah yeah yeah if you if you could have a swell that would be nice but okay hi so thanks for your talk in the beginning so um a question to microcontrollers and ios have you thought about what would happen if you multiplex the exits the inputs from the motors so of course the steps the step time there would be a question how it works but in general you can multiplex it yeah the direction and end stop you can multiplex those certainly but at some point in time you'd always arrive at the point where you need more than one and in the architectural design you probably have to to allow for getting it together and that point alone that you want to separate input and output device then the midi would be enough for that but um okay the the step the the step input i wouldn't like to multiplex really thanks we have another question here have you thought about what's the um what's the size of the diameter of the valves and uh how the air is flowing you think the rotational valves we haven't done that yet but i haven't really done a lot of fine tuning so far at all with the cross sections for the for the pipes for the hoses that's a factor for organ pipes really but i really would would couple that to that seeing what kinds of hoses i can get which which cross sections uh works with which pipe and i would just uh try that with these rotational valves but i haven't done any calculations yet yet so i don't know but it's i think it's one of the more charming aspects of the the rotational valves that you can easily build them in different sizes and different thicknesses and you don't really need two magnets for big pipes and things like that or do you have do you have an idea or suggestion so the second thing would be if you would build a triangle structure you could have an um with rounds with rounds forms you could have a better reaction time yes we're not there yet but but we thought about it yes thank you you already said that you left the pressure air pressure motor back at home but how do you um how do you generate the pressurized air because compressors in general are pretty loud so what kind of construction do you use therefore so yes compressor is quite loud if you want to especially if you want to have a big one so for big big organs it's put in the in the basement and is damped on a damped set up and if you don't need the huge pipes even for for small organs um i did some experiments but i didn't bring it along because i don't have a car so a construction such as in so different different bellows um small construction but if you really want to have a serious and stable um air supply that's why the folding bellows was invented that goes up and down as a whole um which is more complicated than a small bellows but if you have a bellows that only does does it like this it depends on which angle you have it and if it has different weights but these days you can you know that the bellows knows how much air it is in it because you have a sensor and and the other bellows that are really pumping then they change their speed so that you um so it doesn't go up and down too badly but keeps at the same level more or less from what they are made are they made it's also a 3d printed stuff it's um the the bearings of skateboards because they are cheap and so and these things that the motor is really the loudest part and but i haven't really tried optimizing that making the motor quieter but maybe that's not it's not a stupid idea in a maker's place it's not a bad idea to make such construction let such as the the um the on the fairground organ um an automatic one to to make a construction like that with two pumping bellows and one one storage bellows okay so you have controlled the pipes separately have you um thought about controlling one register with this dynamic similar to a synthesizer where you have after touch you could also just control the whole keyboard with one signal not that you do this mechanically but if you have all pipes in one register open the whole all the waves to register okay then you get something the same as if you use a swell you have a dynamic factor that works so the same for for a whole voice and i really think it depends on what kind of music you want to play with that there's a um something you can use that for but i really want per key dynamics but it's but that too is interesting for certain types of music yes of course maybe another thing it's not there's not any valuation that's not but some people i want the key by key dynamics others will others really want dynamics for the whole voice and but you can't play a fugue with that for example but there are so many beautiful things on this world okay with this which software do you make the spectrum of the overtones so there was ff mpeg and the exact command line command is on the on the far plan page on the schedule page on the office talk and if you try that on the terminal on the command line interface that should work otherwise you have to ask nerd how how to debug that okay yeah one could build an organ at every time arbitrary difficult and you had the the main point the controlling but have you have an idea when you would use uh the different registered register i would really like to make an actuator below every pipe so there's the the question how you can you make it cheaper so no but we haven't really done that i'd like to build a prototype so you can at least start start playing music so i can show you what what what it's about so two actives and two registered and one flu flu pipes and one read read register and i would make an actuator under every pipe and if someone is then going to build an organ an organ with multiple registers then they can of course decide how much money to throw at it but i if i don't really intend to build a huge organ beyond the prototype among other things things because i just don't have the room for it so is that is that an answer okay thank you yeah also the call for document what you're doing make a documentation of what you're doing so that others can build this too and choose from what you're doing to compose it to a new organ yes right um i also have most of the time the same deck number so at the next event just call me so thank you everyone and thank you for all the questions and if you build an organ bring them here to congress