I am currently working on a project to build a vibratory bowl feeder and am looking for some way to the vary the frequency of AC powering it. First I will give some background as to why I think this is necessary (because I know it is quite hard to do).
The basic principle of a bowl feeder is a bowl with a ramp sits on 3 (or more) springs set about 60degrees from the horizontal. On the bottom of the bowl is a steel armature, which is just above an electromagnet mounted on the base with an air gap between the two. When the electromagnet is driven with AC the armature above it vibrates up and down and the springs translate some of the at movement into a rotational vibration around the vertical axis. his serves to move the parts in the bowl around and up the ramp.
The issue is that the bowl vibrates at 50hz (driven by single phase mains AC), to get good movement of the parts in the bowl the driving frequency has to be reasonably close to the natural frequency of the bowl and its contents. This either means tuning the bowl to have a particular frequency (hard and time consuming, probably beyond my ability) or changing the driving frequency (also hard but a more versatile solution).
So now you know why I want to build a variable frequency drive (buying one is too expensive), I know it is hard to do but I am just exploring my options, it needs to output close to 240v and a reasonably high current to give the electromagnet enough power. I know I need a rectifier to convert the AC to DC, then a large capacitor to smooth the DC and then hopefully some solid state (eg TRIAC thyristor or similar) driven by pwm from an arduinoto change back to AC of a desired frequency.
Thanks for reading all this, I appreciate all feedback and suggestions
Producing a variable AC frequency at 240V is not a trivial task, and it depends a bit on whether the solenoid
needs a Sinusoidal waveform or whether it will work with a square wave.
Either way , its not a task for a beginner.
I would expect the actuator to be similar in principle to a loudspeaker driver and 240 V at a "reasonably high current" is a lot of power. It would send those little screws flying all over the place.
Perhaps you could use a bass driver with the coil connected mechanically to the bowl? 24 V supply and an H-bridge to give 48 V p-p to an 8 ohm speaker would give you 72 W safely (144 W with a 4 ohm driver). I would guess that would be enough to move small screws if you hit the resonant frequency of the bowl.
Tesla uses a giant voice coil (about 4 to 6 ft diameter ) to drive their shaker table for vibration testing of various components. To drive it you need a very high power audio (AC) signal generator and amplifier which delivers a sinewave to the voicecoil.
A 300W home stereo (used) with a large speaker coil would probably work. You can
breadboard the signal generator with an arduino and pipe the line level (1V p-p) output into your stereo Line In jack.
Archibald:
Use a motor to shake the bowl instead of an electromagnet. Vary the speed of the motor.
This was my original thought on how to do it, and how I thought these feeder bowls actually worked. However an unbalanced motor is vibrating in 2 dimensions which is difficult to deal with. Also how well does an unbalanced motor deal with sustained running? Does it damage itself if providing large vibrations for a long time?
mauried:
Producing a variable AC frequency at 240V is not a trivial task, and it depends a bit on whether the solenoid
needs a Sinusoidal waveform or whether it will work with a square wave.
Either way , its not a task for a beginner.
Is it not trivial because it is conceptually hard or because it is potentially dangerous and you have to deal with lots of heat dissipation etc? I have done a lot of reading around the use of electromagnets in bowl feeders and as far as I know the shape of the wave form doesn't really matter so long as it is alternating (though a soft sinusoid would provide quieter part movement)
russellz:
Do you have the actuator? Can you post the specs?
I would expect the actuator to be similar in principle to a loudspeaker driver and 240 V at a "reasonably high current" is a lot of power. It would send those little screws flying all over the place.
Perhaps you could use a bass driver with the coil connected mechanically to the bowl? 24 V supply and an H-bridge to give 48 V p-p to an 8 ohm speaker would give you 72 W safely (144 W with a 4 ohm driver). I would guess that would be enough to move small screws if you hit the resonant frequency of the bowl.
I don't have the actuator all fixed up yet, I have modified a microwave oven transformer to be an electromagnet and yes it is quite powerful but I am pretty sure it needs to be.
As for your idea to use a bass driver, that is a pretty awesome idea to be honest and I am definitely going to look into that. Can you give some examples of a bass driver you think would work well?
A high powered speaker system might be the way to go, it wasn't something I had even considered. I will be looking into getting some used 300W speakers asap to try it out. To be honest this sounds a lot more doable than trying to build a variable frequency drive
Function generator IC's are available but if you use the arduino to generate the AC
(using a DAC or the ToneAC LIBRARY then all you need to do is make convert the ac signal to 1V p-p to be compatible with your stereo Line In jack.
Archibald:
Use a motor to shake the bowl instead of an electromagnet. Vary the speed of the motor.
Haldus:
This was my original thought on how to do it, and how I thought these feeder bowls actually worked. However an unbalanced motor is vibrating in 2 dimensions which is difficult to deal with. Also how well does an unbalanced motor deal with sustained running? Does it damage itself if providing large vibrations for a long time?
I was thinking of the motor being solidly fixed but connected to a shaft with slight eccentricity machined at the end to drive the bowl. That's how orbital sanders and most delta sanders work. My gut feeling is that tiny orbital motion in two dimensions will work better than the motion your electromagnet gives. I am not sure which would be the quietest.
An option could be to drive your existing electromagnet from a powerful audio power amplifier via a step-up transformer but I would advise doing measurements and calculations first (an electrical engineering qualification would help).
I know it is hard to do but I am just exploring my options, it needs to output close to 240v and a reasonably high current to give the electromagnet enough power. I know I need a rectifier to convert the AC to DC, then a large capacitor to smooth the DC and then hopefully some solid state (eg TRIAC thyristor or similar) driven by pwm from an arduinoto change back to AC of a desired frequency.
I agree that a DIY VFD is unrealistic. I'm not even sure a EE would attempt it but
it is beyond the ability of any beginners.
An option could be to drive your existing electromagnet from a powerful audio power amplifier via a step-up transformer but I would advise doing measurements and calculations first (an electrical engineering qualification would help).
I think that is going to be an impedance matching nightmare. If you had enough data on the electromagnetic it would be possible to calculate the transformer ratio but obtaining a core with the right characteristics and winding it is another thing all together. They do that sort of thing here at work but they have $40,000 simulation software to crunch the numbers and spit out the formula . They order the cores and I wind it per their specs. The cores are usually in two halves and I'm given an air-gap spec to follow so the transformer (or inductor) will have the correct inductance (which I measure with a $5000 LCR meter) and won't saturate below the rated current. Once the prototype is built , we run it through tests using expensive 1 kW load resistors and $20000 (or more) Tektronix O-scopes and $1500 Diff probes. So yeah, it takes a little more than an idea and the desire to do it to make it happen.
CrossRoads' transducer idea looks good. They seem to be loudspeakers without cones so in principle should work well. You don't need a sine wave, a square wave is fine and easy to generate a variable frequency with the Arduino. Driving the transducer with an h-bridge will be the simplest way to get the power into it - no need for a high power audio amplifier. You will need to vary the frequency to find the best performance.
If you can hit resonance for your system then the power requirement should not be great.
When designing a system driver the FIRST thing you need to consider is the system parameters of the device you are trying to vibrate, none of which you have provided.
Where did you get the idea of using 50 Hz from? I have seen these devices in operation in a number of factories and I seem to remember that the vibration was definitely sub-sonic. Perhaps around 10 Hz.
As engineering trivia goes, I wouldn't put vibratory bowl feeder frequency very high on the list of things to remember. Now if you forgot how Line to Neutral voltage relates to Line to Line voltage for a 3-phase system then I would be concerned.
[VL-N = VL-L/SQRT(3)]
