Making a PCB Electromagnet to control Ferrofluid

do you remember when i tried to control ferrofluid  with my pcbcoils yeah it didn't work that well   this idea actually kicked off when the applied  procrastination guys reached out to me to try and   remove all the wires in their ferrofluid display  but i didn't think it was possible because my pcb   coils were too weak but a few months ago i created  a pcb electromagnet which was way more efficient   so i shipped some samples to Simen to test it  with the ferrofluid display and it didn't work   time to elevate things my first idea was to swap  things around and have a magnet sandwiched between   the ferrofluid and the pcb there would be  a coil array that could move the magnet   in the x and y direction which as a result  moves the ferrofluid there was one problem   only one magnet can be used because otherwise the  magnets will start interacting with each other   Simen wasn't very happy with this because it would  limit the resolution of his display so we reverted   back to the idea of attracting ferrofluid with  the pcb coil hey carl your flexible pcb coils   are really thin so it's actually possible to stack  them on top of each other to increase the number   of turns wouldn't that increase the magnetic field  strength this is a question i get asked a lot   can you stack two pcb coils to create a stronger  magnetic field to test this out i connected two   coils in parallel and the magnetic field does  improve but the temperatures becomes way too hot   this has to do with the fact that the coils have  a series resistance so connecting two coils in   parallel would result in a drop in resistance  which means more current pass through the coil   increasing the magnetic field but also raising the  temperature connecting them in series would result   in a higher resistance which lowers the current  and thus lowering the magnetic field strength   theoretically by adding more turns the B should  increase but in practice this also increase the   series resistance which doesn't allow that to  happen now there is another way how we could add   more turns without affecting the series resistance  and that is by increasing the copper thickness i   have never experimented with using the care copper  layers flexar currently uses 0.5oz thick copper   so imagine how much more turns you could fit  with like 5oz it was very important that the   area of the coil doesn't get larger than what  it is right now so that if it works it doesn't   affect the resolution of the display this means  that any new turns we're going to fit needs to   be added to new layers i asked my pcb manufacturer  more details about copper thickness and the main   limitation here is that the thicker the copper  the wider the track has to be so given that my   flexar actuator has only 4mil track width  and clearance it can only be manufactured   with 1oz thick copper something like 2oz  would require a width and clearance of 8mil   i ran some simulations to estimate how much more  turns could fit for the same resistance value and   these are the results i got to keep the same area  we need to fit these new turns in more layers so   i'm going to design two new pcbs and one of them  is going to be my first 12 layer pcb now i wasn't   completely confident with my track resistance  simulation so for the four layer coil i made this   test board that have multiple coils with different  turns it is divided into four coil groups that   have different variables like copper fills and  a steel stiffener i took a similar approach for   the 12 layer coil design making two coils one  with polygon pours and one without i also left   the bottom vias untented which would allow me to  bypass layers in case the resistance is higher   both of these pcbs were designed using the latest  version of altium designer in my opinion this is   the most polished and high-end pcb design software  it is what all the big tech giants use to design   electronics and although it's a professional  tool it still manages to make it really easy   for everyone to learn how to develop their  own circuits click the link in my description   to get access for a free trial of altium  designer and get 30 percent off your license   thank you pcbway for sponsoring my new pcbs  let's start by testing the 12 layer coil   this board was manufactured with a two millimeter  thickness the coil's inductance was increased to   two five five micro henrys because of the new  extra turns but its resistance was a little bit   smaller than my initial 20 ohms target so at  a constant 5 volts the coils was drawing more   current i was afraid that this would make it  get too hot but it still managed to fit with   my current flexar's temperature range as you can  see in this test the exposed copper sample managed   to keep it a little cooler without there being  any effects to its magnetic field strength this   means that the 12 layer coil is 5 times stronger  than my 2 layer coil which makes it the strongest   pcb actuator i ever made and remember that this  is still without steel adding this back core   increases the magnetic field to 16 milli tesla  which means that this could work much better with   fleurofluid but before we jump into ferrofluid  testing let's test the four layer coils i split   the board into two so that first i can measure the  parameters of the non-stiffener area and started   eliminating the ones that got too hot and the  samples got within range compared to my two-layer   coil their magnetic field was around twice as  strong but the stiffener samples didn't improve   so i added my thicker steel plate and made it go  up to seven milli tesla i'm very impressed with   the performance of this four layer coil because  i think it can improve a lot of my projects but   when it comes to ferrofluid control it didn't work  that well the best result i was getting was with   the 12 layer coil but the ferrofluid in this jar  wasn't that liquidy and the glass also had some   thicker areas so i ordered some better quality  ferrofluid and tested it in a thin petri dish my new pcb electromagnet managed to  attract a small bubble of ferrofluid   but here's when Simen suggested  something that changed everything the way this works is by biasing the core  with a magnet i'm not sure if this considered   cheating but it worked very well with my setup  so let's forget about this magnet biasing idea   and see how this would compare to an off-the-shelf  electromagnet well the electromagnets create a   much larger bubble at five volts they generate  a magnetic field of around 20 milli tesla   which to be fair isn't that larger than  my 12-layer pcb electromagnet but these   guys can be powered with 12 volts which means  that they are much more powerful 12 volts now okay going back to five volts i tried to power my coil with 12  volts just to see what happens   and the magnetic field was pumped up to 30  milli tesla which created a much stronger bubble but as you're probably guessing  the heat is a problem here   now this doesn't mean that you cannot power to  12 volts Simen created this music visualizer   that supplies short period of 12 volts pulses  which still kept the coil at low temperatures   these are some 12 volts pwm tests i've  made which exclude the magnet's biasing so can we use these pcb electromagnets  to make a full ferrofluid display   to answer this question we need to use a  much more buoyant ferrofluid tank which   the applied procrastination guys have so go  over to their channel and watch their video

Channel: Carl Bugeja

Views: 45,091

Keywords: arduino, diy, how to make, raspberry pi 4, flexi rigid, pcb design, altium, design, engineer, malta, simple, easy, science, technology, art, 3d printing, fun, pcb, arduino projects, carl bugeja, electronics, electronics design, circuit, micro, tiny electronics, microcontroller, robot, flexible actuator, circuit design, flexar, open source, flexpcb, ferrofluid, electromagnet, pcb electromagnet, ferrofluid control, ferrofluid display, display, ferrofluid clock, magnets, coil, diy coil, diy electromagnet

Id: Ce5tP98XgxY

Length: 8min 12sec (492 seconds)

Published: Sun Nov 07 2021