challenge : Protect stepper drivers from getting fried

Hi there all you electronics developers

It seems very common for someone to mention they have fried their stepper driver of extrucer controller - probably because of connecting/disconnecting their motor while power is being supplied to the board.

It would be worth the effort to develop a methadology to ensure the board does not get fried due to this type of action - is that even possible?

I would imagine that if you clamped the outputs to 12V and ground with 8 Schottky diodes it would stop it being fried by a live disconnection.

---

[www.hydraraptor.blogspot.com]

That should work, but fitting 32 diodes sounds like a lot of work if the only cause is user error, which could be more easily prevented by writing "switch off before (dis)connecting" next to the connectors. Have there been cases of the wire breaking off on its own, or other faults where the diodes would have helped but a warning wouldn't?

If one only wants to use the RepRap to print in plastic, an easier hardware solution is to solder the wires to the board, but if one wants to be able to swap the extruder for a paste extruder/milling head/etc, that wire at least needs to have a plug.

Hi Rebecca

In fact my blown stepper was due to a wire that came loose while I was testing.

The fault was probably mine in not ensuring the wires were securely attached, but there you go - I did not intentionally disconnect...

Probably a mission connecting the 32 diodes as you say, but then the number of times I have seen someone mention a blown stepper is quite high, and for those people a driver that was resistant to this error would have been great.

Cheers

simply using a "real" stepper driver instead of the horrific kludge on the extruder board sorts it for most people. Pololu's A4983 breakout board is very popular for this

---

-----------------------------------------------
Wooden Mendel
Teacup Firmware

True, but they still blow. Most of them are not that well protected.

I've had one blow from what appeared to be a loose cable fault.

Yes it seems to be a problem with all the Allegro chips. The internal diodes aren't fast enough to protect against sparks.

Having said that I have had loose connections which have latched up the driver but have never destroyed one. I think it is perhaps because my 5V feed is provided by low current regulator. I have noticed, when powering an A3977 from a bench supply, that when it latches up it tries to take a lot of current from the 5V rail. If that is current limited perhaps that is enough to stop them frying.

---

[www.hydraraptor.blogspot.com]

Hi Nop

I have just recieved 2 A4983 boards (the standard ones) along with a host of capacitors and resistors.

I will be using 1 to replace my burnt out G3 stepper and the other to drive my extruder - I can see that whine the G3 Extruder generates in the extruder motor from becomming verry irritating very quickely.

Anyway, can you maybe tell me how one would limit the current of a power rail? If it is complicated perhaps direct me to a web page that I could learn more?

Cheers

Simply use a 100ma 5V regulator e.g. 78L05 rather than the bigger one normally used.

---

[www.hydraraptor.blogspot.com]

nophead Wrote:
-------------------------------------------------------
> Simply use a 100ma 5V regulator e.g. 78L05 rather
> than the bigger one normally used.


eh? what? huh?

I was planning on not doing much with my stepper motor board other than adding the 47uF and 10uF decoupling capacitors, a 10 pin IDC connector for communication with the motherboard, a 12 and 5v supply from my ATX PSU (using a molex connector), a 4 way screw connector to connect the stepper motor and a 3 pin header (connecting directly to the 10 pin IDC) for the min opto-switch. and of course the Pololu stepper driver... oh, and a 3 pin dip switch to manually set the microstepping level.

no thoughts about a 5V voltage regulator... ??

Or is that used and only supply 12v from the ATX, generating the 5v using the regulator and 12v supply rather than a 5v supply from the ATX?

I think I probably will need to post my finalised design here for sanity checking before I wire it up. - just when I think I am starting to understand.... I see or hear something that shows me I still have no clue.

For example, I thought the 47uF decoupling capacitor you mention here was in series on the 12v supply to the pin, but looking at tbuser's pololu extruder I see he has a 100uF capacitor that connects the 12v and ground rails, so I assume that is what you mean by a decoupling capacitor?

man this stuff is confusing..

Edited 3 time(s). Last edit at 11/26/2010 12:11PM by AgeingHippy.

AgeingHippy Wrote:
-------------------------------------------------------
> nophead Wrote:
> --------------------------------------------------
> -----
> > Simply use a 100ma 5V regulator e.g. 78L05
> rather
> > than the bigger one normally used.
>
>
> eh? what? huh?
>
> I was planning on not doing much with my stepper
> motor board other than adding the 47uF and 10uF
> decoupling capacitors, a 10 pin IDC connector for
> communication with the motherboard, a 12 and 5v
> supply from my ATX PSU (using a molex connector),
> a 4 way screw connector to connect the stepper
> motor and a 3 pin header (connecting directly to
> the 10 pin IDC) for the min opto-switch. and of
> course the Pololu stepper driver... oh, and a 3
> pin dip switch to manually set the microstepping
> level.
>
> no thoughts about a 5V voltage regulator... ??
>
> Or is that used and only supply 12v from the ATX,
> generating the 5v using the regulator and 12v
> supply rather than a 5v supply from the ATX?

Yes if you use 5V from the ATX PSU then there is virtually no limit to the current the chip can draw from the 5V rail in a fault condition. If you regulate 12V down to 5 with a 78L05 then it is limited to a little over 100ma. I am not sure that stops it getting fried when you disconnect the outputs but I think it may do. I have a persistent problem with connectors coming loose but none of my stepper drives have fried yet. Sometimes the LED on the 5V rail goes dim if the machine gets an ESD strike so I know the regulator has shutdown. The V3 boards I got from Zach have a 78L05 but the V2.3 boards have a 1A regulator. Perhaps that makes the difference.

>
> I think I probably will need to post my finalised
> design here for sanity checking before I wire it
> up. - just when I think I am starting to
> understand.... I see or hear something that shows
> me I still have no clue.
>
> For example, I thought the 47uF decoupling
> capacitor you mention here was in series on the
> 12v supply to the pin, but looking at tbuser's
> pololu extruder I see he has a 100uF capacitor
> that connects the 12v and ground rails, so I
> assume that is what you mean by a decoupling
> capacitor?

Yes the word "on" implies in parallel whereas I would have said "in" for in series. Putting a capacitor in series with a DC power rail makes no sense.


>
> man this stuff is confusing..

---

[www.hydraraptor.blogspot.com]

Hey guys

The following concerns the design of a stripboard design for a pololu stepper driver to replace my blown G3 stepper.

Looking at tbuser's extruder stepper design and the comments added by Antonio Santos there is a mention of needing to connect the GND of the 5v and 12v supply if you use separate supplies. Now if I use my ATX to provide both 12v and 5v to my stepper driver board do I need to connect the ground somewhere, and how would I connect them? Do I connect using a capacitor, resistor or simply a bit of wire?

Further, the IDC connection coming from the motherboard also carries a GND... do I also need to connect this GND to the GNDs coming from my ATX supply?

Since the 12v and 5v provided to my motherboard is from the same PSU supplying this stepper driver will they already be connected?

Finally, Triffid Hunter mentions keeping the 5v and 12v GND separate except for one point due to the noise on the 12v rail.

Any advice would be greatly appreciated.

Thanks.

the 5v and 12v grounds are connected inside the psu. connect all external logic grounds (eg motherboard) to your 5v ground

---

-----------------------------------------------
Wooden Mendel
Teacup Firmware

alternatively instead of changing the 7805 regulator, using a resistor right after its 5v out, for the entire board, would that work? and what would be the value 10k, 100k? how much logic current a chip should use overall?

Yes if you get the value right and put the decoupling after the resistor it may work. The value would probably be in the 10 - 100 ohm range. I have used that technique with another chip to stop it latching up with ESD. It is easier and less components to just fit the smaller regulator in this case though.

---

[www.hydraraptor.blogspot.com]

Wouldn't a fuse work? Or would it not react quickly enough to save the driver chip? Has anyone consider using the STM Stepper drivers? I was looking at the L6208 for my driver, but then I saw the L6470 which as a 1/128 step size...seems like that would give you much better positional control and it also has a lot of features integrated into it that could offload the uC.

I was also considering placing the drivers on the same board as the uC, but this thread makes me concerned about that. If these drivers blow that easily, then you don't want them on the uC board.

Where did you find the L6470? It seems to be hard to get in small quantities: Mouser has a minimum order of hundreds, RS and Farnell don't sell it at all.

The L6208 is more readily available and does have built-in protection diodes, but requires several other external components, so would be about the same price and more work than adding protection diodes to a Polulu; it also can't do less than half steps without an analog input.

They claim that Digikey will be getting them in stock next week. Hopefully the other suppliers will be getting them soon. I have a design with the L6208 using the analog inputs to achieve the microstepping functionality. But I then discovered the L6470 and have converted my design to that. My original app required four stepper motors, so I integrated four drivers on a single board with a Mega 1280. BTW the L6470 looks like it would offload a lot of the workload from the uC, so I could possibly go with a smaller uC. While I was researching my current app, I ran across the mendel bot stuff and it peaked my interest and i thought I could probably support both apps with a single board. Do you think the Mendel community would be interested in a single board solution?

Since we are actually using 3 (actually 4) stepper driver boards, wouldnt make more sense to have a common rail line dedicated with a small addition of features of protection like over-voltage (/-some bemf at some extent); even maybe one diode for reverse polarity, etc?
Can anybody with some inspiration come up with something like that?
Further on, would be enough having just one 78L05 for 3-4 steppers or should that remain on stepper board? And eventually how many cm can decoupling caps be set from the driver boards for that purpose and still do their job (like 5cm is ok, and 10 cm is too much?). I have also seen some fuses on mains, on some semi-commercial boards, but i dont know how usefull those would be.

Edited 1 time(s). Last edit at 12/15/2010 08:34PM by NoobMan.