dekengineering U2 resonance problems

[vacpress]

Hey! I know I could probably figure this out by searching the forums, and I have read some stuff, but I also want to ask advice..

I am building a set of U2 stepper drivers per http://www.dakeng.com/u2.html

I had it on a breadboard and added alot to the stock circuit suggested to hopefully 'take care of the problems', which were wildly triggering inputs, possibly from some RF or EMF or whatever off the breadboard?


So, I added some 10k resistors to ground... This seems to be working fine.

The problem is that there are lots of resonant areas... the steppers i am using didnt have these troubles with a plain logic IC full step setup... Incidently, in the frequencies which are good, the U2 is much smoother and quiet than a full step logic drive.. the same motors does not have these resonant dead zones with a xylotex driver..(i blew up 4 of them! i suck

i am using a 4mhz ceramic oscillator and the modified code provided on the U2 site.. If I can get these working well, I plan to convert the code

anyhow, see my schematic. please advise!

www.robertguyser.com/u2.jpg

Thanks in advance! (Pnmino, i am looking at you!)

and now, off to drum lessons!

[kreutz]

Tell us about your motors. What current and voltage are they? Your limit resistors seem to be a very low value for L/R driver (1 ohm), you might be running them with a phase current way over the specs.

[vacpress]

my motors are the 'laserjet 2 motors'

these are 1.4 amp/phase, 5.4v.... i am running them with 16v power supply for now... it is a switching power supply...

at worst, the resonance problem is so bad it trips the breaker in the power supply...

i have a 24v 3amp transformer.. i just am too lazy to hunt up the proper rectifier...

also, I am using TIP102 transistors.. not Tip122.. Proteus has a limited number of the most usefull parts... and infinite ceramic capactors.... one very disconcerting thing is that you dont actually wire a oscillator.. but, that is another topic alltogether!

[kreutz]

Originally Posted by vacpress my motors are the 'laserjet 2 motors' these are 1.4 amp/phase, 5.4v.... i am running them with 16v power supply for now... it is a switching power supply... at worst, the resonance problem is so bad it trips the breaker in the power supply... i have a 24v 3amp transformer.. i just am too lazy to hunt up the proper rectifier... also, I am using TIP102 transistors.. not Tip122.. Proteus has a limited number of the most usefull parts... and infinite ceramic capactors.... one very disconcerting thing is that you dont actually wire a oscillator.. but, that is another topic alltogether!

You are driving those motors with almost 4 Amps per phase!!! that is why you are tripping the breaker.

In order to limit the current you need to replace the 1 ohm resistors by 5.0 ohm resistors (4.7 ohm are ok) 25 Watt each. (power dissipation is approx 12.5 Watt), and when you are driving half step or full step two phases on you will be over the maximum current your power supply will deliver safely. I recommend a 5A power supply for one motor. (all those values are for a 16 Volts power supply)

[vacpress]

ah. that seems to make sense.. my unipolar board has... a bunch of piggybacked ceramic resistors...

the effect of these 4 amps is... heat and magnetic resonance\feedback in the stepper body? less efficiency on account of heat loss?

let me ask, what sort of motor could i use with what ive got? it is wired up and pretty! last time i put that much solder on $30 of components without figuring out at least the important math! these are supposed to hold me over untill i can afford more xylotex boards or geckos or something better... i like the electronics at least.. it isnt a total loss..

one thing i had tried was 1amp, 60v schottky diodes across the transistor outputs... these got very hot and melted my breadboard.. i suppose that was the 4 amps on em...

ohms law! hah!

[kreutz]

Originally Posted by vacpress ah. that seems to make sense.. my unipolar board has... a bunch of piggybacked ceramic resistors... the effect of these 4 amps is... heat and magnetic resonance\feedback in the stepper body? less efficiency on account of heat loss? let me ask, what sort of motor could i use with what ive got? it is wired up and pretty! last time i put that much solder on $30 of components without figuring out at least the important math! these are supposed to hold me over untill i can afford more xylotex boards or geckos or something better... i like the electronics at least.. it isnt a total loss.. one thing i had tried was 1amp, 60v schottky diodes across the transistor outputs... these got very hot and melted my breadboard.. i suppose that was the 4 amps on em... ohms law! hah!

What is producing your resonance is current waveform distorsion due to magnetic saturation. If you continue using the motors without current limit you could demagnetize the permanent magnets.

Try to get the resistors, it will be fine after replacing them. Don't forget to mount a heat sink on the transistors.

[vacpress]

Originally Posted by kreutz What is producing your resonance is current waveform distorsion due to magnetic saturation. If you continue using the motors without current limit you could demagnetize the permanent magnets. Try to get the resistors, it will be fine after replacing them. Don't forget to mount a heat sink on the transistors.

Thanks for the advice! I will share the code when done. I meant to say I am converting it to PicBasic at the top of the thread. This will make it easier for those of us who use some variation of pic basic to work with the code..

I dont know much ASM, but the U2 code is simple, and I am looking forward to converting it over...

Now if only proteus supported PicBasicPro which i purchased not even knowing about: Protron, which seems superior and cheaper, and supported by proteus!

[H500]

You need fast recovery diodes across the transistors or they could get damaged.

[vacpress]

So, I am working on this again, right now. I do not have any 5k 25watt resistors. I have now got 11ohms, through a 1 and 10, 10watt resistors in series.. This seems to run smoother, no more power supply freak outs.

I do not know exactly how to do the diodes. I had 60v Schottky diodes across the transistor outputs. They got hot and melted! The drives didnt work at all untill i cut them out of the circuit... What kind of diodes?

Is there a good example of a well engineered unipolar output section that you can direct me to? A nice appropriate output design i can tack onto my pic indexer microcontroller?

In single step mode it is currently nice and smooth sounding. In half step mode it is smoother sounding but it seems to be hiccuping every so often regardless of speed... I think there may be lots of 'back emf' or similar from the coils switching.. How do I smooth this out?

Thanks.

Thanks.

[kreutz]

Originally Posted by vacpress I do not know exactly how to do the diodes. I had 60v Schottky diodes across the transistor outputs. They got hot and melted! The drives didnt work at all untill i cut them out of the circuit... What kind of diodes? Thanks.

The diodes should be ultrafast diodes (less than 150 nSec reverse recovery time) 5Amperes minimum, connected across the C-E of the TIP122 with the cathodes connected to the collectors. Vclamping (peak reverse voltage) about 80-90 volt.

The TIP102 already has a diode inside, so you don't need any external diode unless the speed of the internal diode is too slow. You will need an oscilloscope measuring VCE during switching in order to find out if you need the external diodes or not.

In oder to suppress instabilities due to ground noise connect a 1 uF /250 volt plastic film capacitor (not polarized) across Vmotor and Ground as near as possible to the board.