# DC to 3 phase inverter for hard drive motor

Discussion in 'The Projects Forum' started by lostowl05661, Apr 18, 2008.

1. ### lostowl05661 Thread Starter Member

Apr 18, 2008
28
0
This probably seems to simple to the rest of you, but...

I have removed the spindle motor from a couple of hard drives. I think these are three phase motors. I want to run them off of a simple battery. I think what I need is a circuit that inverts DC to 3 phase to run this, with some kind of speed controlling variable resistor.

I did a few searches on the Internaughty and in this forum, but everyone seems to be in love with stepper motors, which I always thought was the motor that controlled the heads.

So, anyway, I have two types of motors. One with three leads and one with four. By playing around with wires and a battery. I can get the motor to jump back and forth by jumping any two of the connectors, one after another. This is why I think it's a 3 phase.

I am working with the Elentco 500-in-1 electronics kit for components and breadboard. Can someone help me with a simple circuit diagram or other information to get me going?

Appreciate it.

2. ### SgtWookie Expert

Jul 17, 2007
22,201
1,809
Well, do you have a Schmitt-input IC that has at least 3 inverters? Something like a CMOS 4093, 40106?

If they are indeed 3-phase motors (I don't know, never took one apart) the one with three leads is probably a Delta wound motor; the windings are connected like an equilateral triangle with the three leads coming from the "points" of the triangle. Using an Ohmmeter, the resistance between one pair of leads will be approximately the same as between any other pair of leads.

If the 4-lead motor is 3-phase, it's likely wound in a wye ("Y")configuration. In a wye, there is one phase connected to each "tip" of the "Y", and neutral is connected to the center. Perhaps not so obviously, the neutral line is the one that has the least resistance to the other three phases. It won't be zero Ohms, though.

3. ### SgtWookie Expert

Jul 17, 2007
22,201
1,809
I've done a bit of searching, but can't find a list of what parts are included in your 500-in-1 kit. If you post it, that would help a great deal in figuring out what might be built from what you have.

4. ### lostowl05661 Thread Starter Member

Apr 18, 2008
28
0
I don't know what a Schmitt-input IC is. Sorry, I am on lab 38 out of 500. Oscillators. I looked on Wikipedia and saw the circuit diagrams for a 3 phase inverter ( but I don't understand why none of the transister base's are connected). I was hoping this was a standard kind of thing where a few transisters, resistors and capacitors would get me spinning.

For theory, I understand the 3 phase uses three single phase AC currents offset by a certain amount, and that to do this you need to use transistors and diodes to create the AC. I am hoping to learn a bit more with this side project.

The 500-in-1 kit has the usual smattering of resisters, capacitors, transistors, diodes etc.

For IC, it has the following:
High Speed CMOS IC
74HC76
74HC00
74HC191
74HC4028
74HC4511
74HC02

Timer IC 555

Audio Power Amplifier 546

Operational Amplifier 324

5. ### lostowl05661 Thread Starter Member

Apr 18, 2008
28
0
that is what I thought as well. i'll check with the the ohm meter as you suggested.

6. ### lostowl05661 Thread Starter Member

Apr 18, 2008
28
0
Yep, one is delta and one is star. The three prong one has a resistance of 2.8 ohms between the prongs. The four lead one has 1.3 between the fat leads and .8 to between the skinny lead and any of the fat ones. So, that means the skinny is the neutral, right? Interestingly (to me) the delta motor has tactile cogging and the star doesn't.

Thanks for your prompt reply. I am a computer administrator, etc, and I know how annoying it is when people ask questions without RTFMing, and I try not to do it. However, I am kind of reaching a mental block about reading any more about 3 phase, single phase, AC/DC, induction brushes and toilet paper. I learn best by doing, so I'd like to cobble up a circuit and try to figure it out after wards.

BTW, the IC idea, is the circuit too complicate to do with transistors, capacitors and resistors? I'd help me understand it better if I can see the circuit that way. From what I've read though, it's starting to seem like what I want to do isn't that simple at all....

7. ### SgtWookie Expert

Jul 17, 2007
22,201
1,809
Well, sure you can do it without IC's - but the IC's have built-in functions that remove the necessity of having to do a lot of wiring.

OK, let's go with the 324 opamp.

What kind of transistors do you have, and in what quantity?

8. ### SgtWookie Expert

Jul 17, 2007
22,201
1,809
Wait a minute!
Look in your manual, in book 2
Experiment #320, "Three Phase Oscillator", page 37?

Scan and post it.

Might as well start with something that you KNOW you have parts for.

Basically, from that circuit it will be a matter of using three transistors driven by each of the three phase outputs. You might be able to use the wye-wound version with V+ being fed to the neutral, the collectors of three NPN transistors tied to one each of the phases, their emitters tied to ground, and bases connected to the 3-phase outputs via resistors, say 20k each. It depends upon your transistors, and how the circuit is configured that is generating the three phases.

9. ### SgtWookie Expert

Jul 17, 2007
22,201
1,809
OK, threw together a 3-phase generator using an LM324. See the attached.

It's output frequency is about 19.2kHz, which is way too fast for your motors. But I did it that way so that it would be easier for you to see the pseudo-sine wave output.

If R5 through R7 are removed, the outputs become square waves and the frequency drops to around 250Hz, which should work fine with your motors. However, the opamp alone won't likely have enough current to drive the motors.

Note that R4 sets the voltage level for the + inputs of all three opamps. This is to allow adjustment for the duty cycle of the outputs; I wanted 50%. In this simulation, I wasn't worried about getting it exact; ballpark was enough. It really needs a 0.1uF cap to ground to eliminate noise.

Note that there should be a 0.1uF (100nF) bypass cap across the supply pins. This goes for just about any IC.

File size:
77.2 KB
Views:
1,035
10. ### lostowl05661 Thread Starter Member

Apr 18, 2008
28
0
Sarge, you rawkith mightily. I looked through the book for inverters in the table of contents, I totally blocked on the oscillator doing essentially the same thing. (or exactly the same thing).

Thanks for the second post. I can't wait to study it.

here is the page you quoted from book two.

http://www.viciousbunny.net/images/3phaseOscillator.jpg I think you are right. I'm going to skip to this experiment for now, and also examine your diagram. Thanks for you help.

11. ### Papabravo Expert

Feb 24, 2006
12,280
2,723
I'm sorry to throw cold water on this effort, but it is extremely unlikely that the spindle motor in a hard drive is an AC motor. Precise speed control is required of a spindle motor or the bit rate of the bits coming off the drive would be changing by an unacceptable amount.

What you have is called a Brushless DC Motor. It is operated by sequencing DC currents into the coils at the precise times when maximum torque is exerted on the rotor. As the rotor turns a sensor switches one coil off and switches the next coil on. I've seen both Hall effect sensors and back emf sensors used for this purpose.

Microchip has a pretty good application note on brushless DC motors. Here is a link.

Here is another one from Microchip on choosing drive transistors

This is just a guess but you might be biting off more than you can chew at this point. But hey, I've always believed that "man's reach should exceed his grasp".

12. ### SgtWookie Expert

Jul 17, 2007
22,201
1,809
OK, now we're on the same page (literally!)

I modified the circuit I put up earlier to (mostly) agree with what they have in the schematic. Differences to note:
1) Instead of a fixed R12/R13, I'm using a pot (R4) (*See below)
2) Instead of a fixed R1/R2, I'm using a pot (R11)
3) I'm using 10v instead of the 6v shown. I'll change that later.
4) A minor difference is for C5 I'm using 1uF instead of 10uF. Little consequence.

The big difference between this circuit and the one I put up before is that the previous one was designed primarily to produce square waves 120° apart, whereas this design is for producing sine-like waves 120° apart, although there is significant clipping going on.

If you look at the simulated O-scope trace below, you'll note that the frequency of the output will be roughly 0.518Hz - very slow indeed! That's OK though - it can be sped up later. Your mission, Mr. Phelps, is to get the thing working.

(*) On that very subject, I suggest that instead of the fixed R12/R13 combination, that you use a potentiometer as I did in 1) above. Initially, set it for equal resistance from the wiper to both 'ends'. You can use a 10K, 25K, 50K pot (pretty much whatever you have) but it should be a linear pot, not an audio or log taper pot. This will enable you to adjust the symetry of the waveform, but more importantly will also provide an adjustment in case the oscillator won't start on it's own - although if you've assembled things properly, it should.

File size:
75.9 KB
Views:
946
13. ### SgtWookie Expert

Jul 17, 2007
22,201
1,809
OK, hadn't seen Papabravo's post before I made mine. As I said before, I don't really know for certain what kind of motors are used for hard drive spindles, except that steppers would be highly unlikely (limited speed, limited torque at speed).

If you're really new at electronics, just getting a 3-phase oscillator working satisfactorily will be challenging enough - and you'll learn a good bit in the process. If you feel you're getting frustrated, take a break. Don't be afraid to ask questions.

14. ### SgtWookie Expert

Jul 17, 2007
22,201
1,809
OK, here's the rest of it with a modified H-bridge transistor driver circuit.

Since the program I used for the simulation is basically a "student demo" version, it's limited in the number of components that a circuit can consist of. I hit that limit, so a portion of it is rather crudely drawn.

The original 1K resistors to the LEDs stay where they are, and a pair of NPN/PNP transistors are arranged with their emitters tied to a common point, and their collectors to +6v/ground respectively. Both of the paired bases are controlled by the current supplied via the 1k resistors. If you don't have 2N4401/2N4403 transistors, you could use 2N2222/2N2907 or 2N3904/2N3906 pairs, which are all NPN/PNP respectively.

With this arrangement, it is impossible to have both an upper(NPN) and a lower(PNP) transistor turned on at the same time, which would short out the power supply in a condition commonly known as "shoot through".

Unfortunately, it is not as efficient as if the collectors were tied to a common point, with the emitters to ground/6V - but we're looking for simplicity here.

Coming down from the emitter common tie points, you'll notice three wires with some kinks in them. Those kinks are supposed to represent the windings in your delta-wound motor (it may very well be a wye-wound, don't know at the moment). BUT - each lead should have a current limiting resistor wired in series with it (between the emitters common tie points and each motor lead), otherwise your little batteries will get drained in a really big hurry, along with overheating the transistors.

For these three series resistors which are not shown, use 33 Ohm resistors (Orange Orange Black Gold) that are rated 1/4 Watt or higher.

Even if it IS a BLDC (Brushless DC) motor, at the extremely low frequency this circuit is generating, it should still spin - at about 30 RPM. At higher frequencies, all bets are off!

File size:
75.7 KB
Views:
548
15. ### Papabravo Expert

Feb 24, 2006
12,280
2,723
In my experience, hard drive spindle motors rotate at speeds between 4000 rpm and 7200 rpm. They may go faster than that, but I'm not aware of those designs.

The last time I remember an AC motor in a disc drive, it was an 8" Floppy Disc Drive that would hold about 240,000 bytes.

16. ### SgtWookie Expert

Jul 17, 2007
22,201
1,809
Agreed! Some of them are mighty zippy indeed. Seems that most of them are 5400 or 7200 RPM, but I've seen a few at 9600 RPM. Can't help but wonder about platter distortion at those kinds of speeds.

If these motors he has are indeed BLDCs, he won't be able to spin them up very fast at all unless using a VFD (variable frequency drive) at the minimum. But the 3-phase oscillator circuit as built isn't capable of that. As it is, trying to spin it much faster will probably just wind up with the motor oscillating back and fourth.

I waited to go "high-tech" - I picked up a trio of BASF 5-1/4" SSDD FDD's capable of stepping 40 tracks at a 12mS rate; 240K formatted capacity. My, they were quick on my Model I TRS-80 that I'd tricked out with a Lobo Drives expansion interface and a whopping 48K of RAM, complete with dead-bug lowercase mod, hand-assembled keyboard debounce and lowercase driver routines, along with a JKL screen print utility that would actually reproduce the graphics character blocks on my Epson MX-80 F/T w/Graftrax. The FDD's sure beat loading programs from the cassette deck, which was initally a lot better than typing them in by hand, but later about as exciting as watching paint dry.

Oh well, off topic...

17. ### Papabravo Expert

Feb 24, 2006
12,280
2,723
At least the edge of the platter isn't going supersonic and creating shock waves.

18. ### lostowl05661 Thread Starter Member

Apr 18, 2008
28
0
I have seen a lot of references to the BLDC, and the VFD products on the Internet, and that may be what I need. However, I'll give it a whorl. I am somewhat new to circuits. Brief bio (if you care)

Used to play with Radio Shack Science Fair lab kits when I was 9 - 12. Love them. Been programming computers for fun and money since, now a Cisco engineer. Lately, I have been drawn to wind and solar energy. i want to experiment with various things like that. My end goal is to build my own solar array and suntracking drive, as well as two productive wind generators.

To that end, I bough the 500-in-1 kit to re-learn electronics all over again. I also am taking a welding course, and hope to progress to making my own Gingery series machine tools. Today, in fact, for practice, I welded up a dodecahedron out of three inch base pentagons.

So, for fun, I am slaughtering about 40 hard drives that my employer needs destroyed for security reasons. I intend to meld down all the aluminum into ingots. I also have become enamored with the pretty aluminum machined bits and platters and motors kind of got me starry eyed. It's so cool it HAS to be good for something!!!!!!!!

Any way, that's the story. I turn 38 next month and I hope to have made significant progress on my goals in within the next five years.

Phew!!! Thank you mister exposition. Anyway, thanks for the advice. I'll post any info as I try this. Papabravo is probably right, and these motors are pretty, but pretty useless without a VFD. If so, most will go into the foundry.

But I can dream....

19. ### SgtWookie Expert

Jul 17, 2007
22,201
1,809
Nice bio
My folks gave me a Norelco electronics kit when I was a kid.

They were made in Holland, and had Euro-type transistors. The projects were schematics printed on glossy-faced cardboard with holes that you'd punch out, and lay on top of a pre-drilled Masonite board (like pegboard) - you'd insert springs in the board holes, clip components between the coils of the springs, run buss wire around... it was great fun. One quirk about that kit was it's heavy use of PNP transistors. As a result, I decided that I didn't "like" NPN transistors for a few years. Funny!

Never had one of those Radio Shack Science Fair kits; they were probably early Elencos. Picked up a R.S. Electronics Learning Lab years ago that was designed by Forrest M. Mims III, who is actually a member of this site - although he posts very infrequently. R.S. still sells the Electronics Learning Lab. Mr. Mims did a really nice job on it. The pair of manuals that come with it are easy to read.

I noticed the other day that my local RS store started selling a Sensor Lab, which was also designed by Forrest. Since you seem interested in that sort of thing (wind turbine, solar array), you might consider picking up one of them.
http://rsk.imageg.net/graphics/uc/rsk/Support/ProductManuals/2800278_PM_EN.pdf

Welding is fun besides being really useful. Are you doing gas, electric, tig/mig or a combination?

I know what you mean about getting starry-eyed over such nicely machined bits. Seems a crime to not re-use them somehow! Perhaps use the platters for brake discs for a racing kart your dog can drive (big evil grin)

But hey - if you can even get the motors turning at any rate of speed, then you will have accomplished your immediate goal, right?

The motors were designed to do one thing - get a lightweight aluminum platter up to a fixed high speed, and maintain that speed in an efficient manner. You won't get much torque at all from them. But, it's the challenge of getting some inanimate object to do your bidding

Thanks for the Gingery reference. I've found a number of foundry and DIY machine tool sites, but Gingery's series of books progress in a logical manner, from foundry to lathe to further machine tools. He left a great legacy for generations of budding machinists to come.

I've been intrigued with the idea of building a small CNC mill for awhile, and have accumulated much of the supplies I need to build one - but the design concept keeps changing as I explore various ideas. One of these days...

20. ### lostowl05661 Thread Starter Member

Apr 18, 2008
28
0
The class covers all types. We are on stick arc welding. We have done gas cutting and plasma cutting, and gas brazing. Next is mig, then tig, then actual gas welding.

And, yes, it's about refusing to be beaten by an inanimate object.