We used to use wire EDM to cut ceramics we made a lot of custom parts for our edms such as bearings and contacts as far as to get the power to the wire used a similar looking device basically a bearing where the carbide ring around it sealed with o-rings filled with Mercury.
Nice work, glad you got it sorted. I expect you have but i could not see from the video but remember to keep the electronics outside the "loop" of the feed and return to your pulse generator. It is so easy to think i will make this tidy and forget that anything inside the "loop " is effectivly coupled by a 1 turn transformer.
Also make sure you ground your pcb. Seems like it's floating now? Many times properly grounding a pcb will already solve many EMC issues. The many screws used to mount a mainboard into a PC definitely serve a second purpose. Well done!
At around 11:30, you can see the bolt of the top left roller oscillating a bit. Trying to determine chicken from egg, is it part of the cause of the issues from runout or over zealous mesh engagement on the gear set, or is it just a reponse to the torque changes and the fact that from memory that particular axle is mounted on the eccentric bearing with the lever under spring tension? Either way, would it make sense to use a camming closure mechanism to constrain it more, versus just leaving it to the spring tension? This a fascinating channel, thanks so much for the effort you have put into all this and for sharing it!
So I´m thinking the load cell could be moved inside the control box and connected to a roller wheel via a pushrod. The pushrod could be grounded where it enters the box. That way you wouldn´t have to worry about EMC and filters and copper tape.
Also if the wire was wrapped almost one full turn around a driven wire tension wheel the other spring loaded wheel probably could just be an idler. Removing the need for expensive gears. With one spring loaded idler from each side the side load on the driven wheel goes down and the driven wheel could be put directly on the motor. The cheap geared DC motor you used before with great results.
Hi, I am not familiar with EDM, however it seems that wire should pass at the same speed through both feed and pull rollers, correct? in this case, why not have just one servo controlling both feed and pull through mechanical coupling (gear), and provide tension control with a mechanical (spring) tensioner? sorry if i am out of the loop on this one though.
Just came across your videos. Have you thought about a different solution to the tention system? We use to make this a bit different for our maschines we build for textile and wire winding. We use an angle sensor for it. The guide bearing/roller is attached to a swing arm with counter weight (adjustable), pretty fast way to correct the feed speed and that way the wire tention. Basic solution was just feeding the 0-10V signal of the sensor into a controler adjusting the motorspeed or the more sofisticated variant is feeding an angle value from a rotary encoder into a sps to then adjust the speeds, the encoder variant gives you more options in setting the boundries of your setup and pick a resolution (even adjustable by gear reduction for example) And damn you are making me wanne build one of these things.
My gut feeling says it will, granted with lower performance, now that both motors are driven with velocity/position control instead of torque-control. Stepper-motors would also reduce the entry-barrier for people not ready to go for ClearPath servos just yet. Then again, in the big picture, the servos aren't too expensive.
From a control point of view, steppers would work. However, I strongly recommend not to use them. The clearpath motors are really strong, and Im driving them up to 70% of their max RMS power. A stepper that can deliver that much torque would be really big and if you would change the gear ratio for steppers, you cant reach higher wire speeds. Another disadvantage is that steppers inject lots of nasty vibrations into your precision mechanics, you really dont want that.
I am all for Clearpath servos as they are an excellent product so it’s nice to hear their purchase was not for nothing! However, I was also hoping there was a less expensive alternative. In any case, happy to see forward progress!
I have a feeling that many of those problems could be fixed by an oversized motor with some gearing. It will take more space and power. But it might be worth looking into if you decide to make the whole machine into a product.
I looked at the HX711 before I made my own amplifier and did not like it. The sample rate is really low. The control loop runs at 1000Hz, the HX711 can only do 80Hz, which will serverly hamper control performance.
Hey Mike, would you mind answering a question from a noob? From what I've been able to research, the general goal of an EDM arc generator is to deliver a fixed amount of energy per spark. And to do so, the current and duration are controlled after the spark ignites. But the circuitry to do that is complex and expensive. So why not simply use a transistor to apply a voltage to an inductor, wait until the current stabilizes so you know exactly how much energy is stored in the magnetic field, and then switch the transistor off to cut the circuit? That should generate whatever voltage is necessary to ignite the spark, and deliver a fixed amount of energy. And all you need is a coil, a MOSFET, and a timer to switch it on and off. And some circuitry to sense the distance between wire and workpiece so the motion control knows whether to speed up/slow down/back up, but that should be easy enough. I should really just build the sucker and see what happens, but surely everyone would be doing it this way if it were really so simple.
Im not going to dive into lots of details, but that won't work for multiple reasons. Here are just a few: *for a long arc on time, you would need a gigantic inductor *there is a random ignition delay between applying a voltage and getting an arc *you need good control of the open gap arc voltage *the frequency is high, while you are busy charging the inductor, the next arc is already supposed to happen. With your idea you will definitely get arcs, but the performance will be dissapointing. If tried similar concepts years ago. If you would like to make your own design, I recommend to perform a literature study first. A lot of research has been performed in this field. Watch out for this rabbit hole, once you go in, it will consume a huge amount of time. I know a few guys who have been at it for years without good results.
@@baxedm9806 Ah, so even with constant energy per spark, the uncontrolled rate of flow causes a large difference in how much material is removed? Thanks for the response.
It would be interesting to know how Immersion worked around the problem with nonlinearity in the touch-sensitive gamepad buttons they developed for Sony in the mid-90s.
I learn something every episode. Thanks for all your effort! Incredible.
We used to use wire EDM to cut ceramics we made a lot of custom parts for our edms such as bearings and contacts as far as to get the power to the wire used a similar looking device basically a bearing where the carbide ring around it sealed with o-rings filled with Mercury.
Nice work, glad you got it sorted. I expect you have but i could not see from the video but remember to keep the electronics outside the "loop" of the feed and return to your pulse generator. It is so easy to think i will make this tidy and forget that anything inside the "loop " is effectivly coupled by a 1 turn transformer.
Also make sure you ground your pcb. Seems like it's floating now? Many times properly grounding a pcb will already solve many EMC issues. The many screws used to mount a mainboard into a PC definitely serve a second purpose. Well done!
Quite interesting thanks for the series.
Also couldn't you have moved the tension sensor to the new wire feed in side to bypass the spring ?
ude, it’s so cool, everyone likes CNC milling machines, they don’t f**king understand how much EDM are cool. good luck, from Russia with love_)))
Very well done. Not an easy nut to crack. Thank you
At around 11:30, you can see the bolt of the top left roller oscillating a bit. Trying to determine chicken from egg, is it part of the cause of the issues from runout or over zealous mesh engagement on the gear set, or is it just a reponse to the torque changes and the fact that from memory that particular axle is mounted on the eccentric bearing with the lever under spring tension?
Either way, would it make sense to use a camming closure mechanism to constrain it more, versus just leaving it to the spring tension?
This a fascinating channel, thanks so much for the effort you have put into all this and for sharing it!
So I´m thinking the load cell could be moved inside the control box and connected to a roller wheel via a pushrod. The pushrod could be grounded where it enters the box. That way you wouldn´t have to worry about EMC and filters and copper tape.
Also if the wire was wrapped almost one full turn around a driven wire tension wheel the other spring loaded wheel probably could just be an idler. Removing the need for expensive gears. With one spring loaded idler from each side the side load on the driven wheel goes down and the driven wheel could be put directly on the motor. The cheap geared DC motor you used before with great results.
Hi, I am not familiar with EDM, however it seems that wire should pass at the same speed through both feed and pull rollers, correct? in this case, why not have just one servo controlling both feed and pull through mechanical coupling (gear), and provide tension control with a mechanical (spring) tensioner? sorry if i am out of the loop on this one though.
Great video!
Just came across your videos. Have you thought about a different solution to the tention system? We use to make this a bit different for our maschines we build for textile and wire winding.
We use an angle sensor for it. The guide bearing/roller is attached to a swing arm with counter weight (adjustable), pretty fast way to correct the feed speed and that way the wire tention. Basic solution was just feeding the 0-10V signal of the sensor into a controler adjusting the motorspeed or the more sofisticated variant is feeding an angle value from a rotary encoder into a sps to then adjust the speeds, the encoder variant gives you more options in setting the boundries of your setup and pick a resolution (even adjustable by gear reduction for example)
And damn you are making me wanne build one of these things.
Excellent news! Glad you got it all working! Will this load sensor and circuitry work using stepper motors?
My gut feeling says it will, granted with lower performance, now that both motors are driven with velocity/position control instead of torque-control. Stepper-motors would also reduce the entry-barrier for people not ready to go for ClearPath servos just yet. Then again, in the big picture, the servos aren't too expensive.
From a control point of view, steppers would work. However, I strongly recommend not to use them. The clearpath motors are really strong, and Im driving them up to 70% of their max RMS power. A stepper that can deliver that much torque would be really big and if you would change the gear ratio for steppers, you cant reach higher wire speeds. Another disadvantage is that steppers inject lots of nasty vibrations into your precision mechanics, you really dont want that.
I am all for Clearpath servos as they are an excellent product so it’s nice to hear their purchase was not for nothing! However, I was also hoping there was a less expensive alternative. In any case, happy to see forward progress!
@@ScotY808 The JMC servos might be a cheaper alternative, search for "JMC iHSV57-30-18-36" on Aliexpress, costs about 1/3rd of a ClearPath.
I have a feeling that many of those problems could be fixed by an oversized motor with some gearing. It will take more space and power. But it might be worth looking into if you decide to make the whole machine into a product.
Think about using the hx711 very effective with the load celles and less sensitive.
I looked at the HX711 before I made my own amplifier and did not like it. The sample rate is really low. The control loop runs at 1000Hz, the HX711 can only do 80Hz, which will serverly hamper control performance.
Hey Mike, would you mind answering a question from a noob? From what I've been able to research, the general goal of an EDM arc generator is to deliver a fixed amount of energy per spark. And to do so, the current and duration are controlled after the spark ignites. But the circuitry to do that is complex and expensive.
So why not simply use a transistor to apply a voltage to an inductor, wait until the current stabilizes so you know exactly how much energy is stored in the magnetic field, and then switch the transistor off to cut the circuit? That should generate whatever voltage is necessary to ignite the spark, and deliver a fixed amount of energy. And all you need is a coil, a MOSFET, and a timer to switch it on and off. And some circuitry to sense the distance between wire and workpiece so the motion control knows whether to speed up/slow down/back up, but that should be easy enough.
I should really just build the sucker and see what happens, but surely everyone would be doing it this way if it were really so simple.
Im not going to dive into lots of details, but that won't work for multiple reasons. Here are just a few: *for a long arc on time, you would need a gigantic inductor *there is a random ignition delay between applying a voltage and getting an arc *you need good control of the open gap arc voltage *the frequency is high, while you are busy charging the inductor, the next arc is already supposed to happen. With your idea you will definitely get arcs, but the performance will be dissapointing. If tried similar concepts years ago.
If you would like to make your own design, I recommend to perform a literature study first. A lot of research has been performed in this field. Watch out for this rabbit hole, once you go in, it will consume a huge amount of time. I know a few guys who have been at it for years without good results.
@@baxedm9806 Ah, so even with constant energy per spark, the uncontrolled rate of flow causes a large difference in how much material is removed?
Thanks for the response.
hay Hunny its the guy with the dutch accent
Excellent!
Oh, at first I thought it's This Old Tony's video, but sometimes it looks like hand expressions a little bit unemotional, otherwise - perfect!
good job!!
It would be interesting to know how Immersion worked around the problem with nonlinearity in the touch-sensitive gamepad buttons they developed for Sony in the mid-90s.
what is the purpose of the copper tape ?
ffs try watching the video!
@@Andrew_Fernie it was in the next video but i got my answer