[mr.eemay]

Greetings, may I ask your help and experience?

I have designed and built a test board that uses TB6560AHQ to drive 2 phase, 8 wire superior electric slo-syn stepper motors with supply voltages of 24VDC and 5VDC for the logic. (Photos of a fried one attached)

There is a problem with the board. When connected to the motor via series (or parallel, it doesn't make any difference whatsoever, but gives less torque) connection and driven, the motor moves inconsistently, skiping steps, changing direction without input and making a lot of noise. When I looked at the output with osciloscope, there were a lot of uneccessary spikes and plunges, generated by the chip that (I think) forced the motor to behave like that. Playing with the settings of torque, decay or driving mode improved the situation a bit, but the problem still existed.

As I have read more and more about this particular IC on forums, it became clear, that I have made a mistake by underestimating it's sensitivity to noise and inductive resistance, so I used wirewound ceramic resistors for the "sensing" resistors. Also, the circuit may have poor mass and inductivity because of poorly designed PCB layout.

So, here are the questions:

1. Can all those problems be caused by the inductivity of resistors and PCB layout?
2. If yes, can I use a few parallel resistors instead of one for the "sensing" resistors? (0.5Ohm non-inductive resistors are hard to find, you know )
3. I have seen people using low ESR capacitors for the "filtration" of motor side power supply. May that help?

Thanks,
Mykolas J.

P.S. My english isn't top-noch, hope you'll be able to understand everything clearly enough

[jalessi]

Mr.Eemay,

Good source for resistors:

http://tinyurl.com/2dp2ywh

This will help,

Jeff...

[amishx64]

mr.eemay,
You have a SEVERE lack of components on your board. Do you have eagle for PCBs? If so, I will send you a schematic of what you should have in a driver like that. Otherwise, I'll just post an image. I can also tell what you need component-by-component, but that might take a while.
Two musts are:
Rectifier diodes (for back-emf)
Large Electrolyte capacitors

Maybe you have these off-board tho?

- amishx64

[amishx64]

Sorry 2x post...

You should also definitely include a heatsink on that driver IC.

I've looked (briefly) at low ESR Caps and I don't think they are necessary for CNC applications, especially when driving motors, since there is no through current (for one terminal thru to the other) powering anything with a load. Low ERS's are used on motherboards, and I would imagine devices sensitive to slight temperature and voltage changes, such as precision electrical instruments.

[mr.eemay]

Thanks for the link, Jeff, I somehow forgot to look for those in american ebay.

Amishx64, I was almost sure that I will have to scrap this board, so I tried to make it as basic as posible (by the schematic in the datasheet with slight changes) and then try changing, adding components, but I have ran out of ideas whats wrong, as the board came out to be very sensitive to the slightest changes, which, well, produced just a bit less or a bit more noise. The heatsink is removed by the way

mr.eemay,
You have a SEVERE lack of components on your board. Do you have eagle for PCBs? If so, I will send you a schematic of what you should have in a driver like that. Otherwise, I'll just post an image. I can also tell what you need component-by-component, but that might take a while.
Two musts are:
Rectifier diodes (for back-emf)
Large Electrolyte capacitors

Maybe you have these off-board tho?

- amishx64

Though I design in (a very outdated) accel EDA suite, I have access to eagle, so if you would do that, I would appreciate that.

By the rectifier diodes do you mean putting them on the motor ouptut? I thought the IC has them built in?
Concerning the large electrolytes, it depends where they should be. If you mean on the power circuit, then yes, I have them off-board by the power supply.

Thanks for your help, guys,
Mykolas J.

[amishx64]

I will send you my eagle diagram as soon as I find it.

By the rectifier diodes do you mean putting them on the motor ouptut?

Kinda. The output from the micro goes to the stepper wires and a diode also connects from ground to the same pins. Two per stepper coil, so one each for A+,A-,B+,B- or however you like to call the coils.

I thought the IC has them built in?

Not that I know of, and if they did, they would insufficient for CNC purposes anyway. The diodes would be way too small anyway you look at it.

Concerning the large electrolytes, it depends where they should be. If you mean on the power circuit, then yes, I have them off-board by the power supply.

All power supplies need large caps nearby, as you have. But they also need placed very near to the driver IC chip that does the switching because things can get very 'noisy' near the chip with all the current switching and all.

Schematic is coming...

- amishx64

[mr.eemay]

This is the schematic I have done so far, is that the diode layout you've had in mind? Am I missing anything more? Please, let me know,

Thanks,
Mykolas J.

[amishx64]

It needs a lot of work. You would've blown a lot of electronics.

- amishx64



Edit: The CNCzone website reduced the quality. I'll have to host it elsewhere also.

[amishx64]

Here's the hosted version:
http://img62.imageshack.us/img62/365...errevisedc.jpg

[mr.eemay]

Thanks for the advice, amishx64.

I have to admit, my circuitry skills are a bit rusty, the last time I've designed a circuit more complicated than a voltage regulator was like ten years ago

I guess that the notes in blue are not critical. So, I know that pull-down's are built in, but on my test board, there was quite a bit of noise there and adding a 1k pull-down resistor eliminated that. I might have to increase the value, but it seemed to work just fine then. Resistors are cheap, you know, so I've put them there, just in case. The resistor for the LED has just a generic value, I'm not sure what LED's I will get from the store, so I will adjust the value when I get them. I totally agree with all the other notes in blue.

Regarding the notes in red, well, for the voltage vs. GND issue, this software is pretty dumb, and it goes haywire when you put ground or vcc symbols on it, so what I do, is instead of using symbols, I use screw terminals for representing the power. And it's easier to distinguish between grounds for me just by writing the -5VDC and -24VDC. (sorry for the whole offtopic whining about bad sofrware ). The electrolytic decoupling capacitor is there (C2 actually), I have just represented it further away from the IC and missed the value quite badly . You have made me scratch my head with that diode part there to be honest. I have seen a few circuits that have diodes arranged like that and I think that by putting them there would let current flow one direction only trough the coil of the motor. Does the IC reverse the direction of current while in operation or something?

EDIT: Oh, now I have just realised that by doing that, I have basicaly made a short-circuit between the outputs, is that right? Hmm, I now wonder why on earth other circuits had those there then

Thanks for the help,
Mykolas J.

[amishx64]

You are quite welcome!

That's alright, we all need to (re?)learn sometime! I'm glad to help.

Your assumption that the blue isn't critical is correct. The red is most important by far. Extra components, especially something as cheap as resistors, are always encouraged if it makes the circuit more stable / reliable. If they reduce noise, I'd add them. I am not 100% sure about the 100K vs the 1K value. I saw 100K on the datasheet, but I don't think the difference will be overly critical (hence blue ). I figured you knew what you were doing with the led and resistor there, but wanted to point it out just in case.

LOL dumb software. Always a plus! I figured you meant GND, but I mentioned because the error could be fatal to the chip. In electronics (you probably already know this) -24V and +24V gives you a 48V equivalent voltage. [I use screw terminals to represent power too, dunno any better way.]

I understand you confusion with the seemingly backwards diodes. One way current won't flow (GND to volts? Come on! That can't happen! Or can it?), the other way motor current short-circuits into GND! Well one way must work, and short circuiting and blowing everything up here just isn't an option. You definitely don't want to do the way I crossed out with current short-circuiting into GND! The other way (that seems like they'd do nothing) actually allows for the induced back-EMF generated by the coils when the stepper moves and coils are disengaged / engaged to get fed right back into the coils to kill off the excess current with hefty motor coils. This noise would normally go tight to GND and make a bunch of noise, disturbing and resetting your whole circuit constantly. That's why these diodes are so important. If you don't have them, current can be generated with a higher voltage than you actually source (>24V) and flow in reverse through your driver IC. To my understanding, this usually is what would kill the chip. The noise just screws everything up by shifting the voltage levels around.
I hope this is a good enough explanation. I've never had it explained to me professionally, but it seems to work for me and adheres to what I've read on the subject

To answer:

Does the IC reverse the direction of current while in operation or something?

The answer is yes, but not for that reason. It just switches the current back an forth in order to reverse the coil polarity and make the stepper 'step'.

Oh, now I have just realised that by doing that, I have basicaly made a short-circuit between the outputs, is that right? Hmm, I now wonder why on earth other circuits had those there then

Yes, that would be correct. I hope you read the other circuits wrong, because if that is true, they certainty won't last for long! (About as long as it takes for the IC to turn on the motor coils, which is about a second or two after logic power up I believe.)

BTW, a higher voltage results in a lower induced-EMF. However, there are obvious disadvantages to running everything at higher voltages.

If you have any more Q's, feel free to ask away.

Hope this helps,
- amishx64

[mr.eemay]

Thanks guys, hardware works fine now. I have some problems still, but only with software and I can ask a local friend to come and see what's wrong, as he is familiar with it

[eSilviu]

You are quite welcome!

That's alright, we all need to (re?)learn sometime! I'm glad to help.

That's why these diodes are so important. If you don't have them, current can be generated with a higher voltage than you actually source (>24V) and flow in reverse through your driver IC. To my understanding, this usually is what would kill the chip.
BTW, a higher voltage results in a lower induced-EMF. However, there are obvious disadvantages to running everything at higher voltages.

- amishx64

If I understand you well, you are saying that:
1. the 4 diodes between motor coils and V+ should not be there
2. higher voltage results in a lower induced-EMF

If so,you are totally wrong:
1. Scope of those diodes: If the voltage at that point is bigger than (V+ + 0.6V) because of the back EMF, because you slow-down the motor or because the driver is not powered and somebody rotate the motor by hand!!!! the current will flow thru those diodes to the V+. Since V+ is a voltage source it should absorb the energy without rising the value of the V+. This is the reason for which a BIG capacitor is put between V+ and GND.
so those 8 fast diodes are necessary - ALL of them, unless you want to use 4 external diodes, and 4 internals diodes - not a good idea

from my experience, stalling a Nema23 stepper from 1000RPM could produce even 100V back emf for about half a second on a bad power supply.

2. back-EMF voltage is a function of V+, how fast the transistors in the H bridge commute, and how quick are you slowing down the motor (kinetic energy will be transformed to electricity, that should be absorbed by the voltage source V+ without overheating or rising V+ considerably)

3. you write in the picture "do not connect!!! 0 logic is not the same as 0V stepper power". Probably you based your assumption on the last schematic in the datasheet. Well, this is ABSOLUTLY WRONG!!! those two grounds should be connected together!!! in a single point - at the power supply, or at the terminals of the filterring capacitor. And by the way, 470uF/35V is much too low for this application. I would say at least 3300uF/50V!