[bo.gardmark]

I did purchase 3 pcs of Leadshine DM556 stepper drivers. I did open them to check the PCB and the internal layout. I also made similar measurements as I have done earlier on the HY-DIV268N-5A drivers.

Info about the DM556 can be found here: Leadshine Technology Co., Ltd.
The DM556 is a quality product. I did no modification to the driver, the driver worked perfectly right out of the box. Good documentation and even a test/tuning program was included.

I was able to run the stepper at up to 4000 mm/minute without losing steps. They have an auto configuring feature that seems to work well. The automatic idle current reduction did also function well. I can recommend this product. However, it is important to know that it must be Leadshine original product, there are copies out there called DM556 that don’t have this performance.

Regards

Bo, SM6FIE

[jbase]

Hello.

I am new here and this is my first post.

Just a little follow up on these DIV268N stepper drivers. I was, should I say that, unfortunate to purchase 3 of them on Aliexpress for my DIY CNC router (I never built one before, but I have good experience with 3D printers). My PCB schematics matches Lucas' drawing 100%. Lucas, great job on that schematics, thanks, man!!

Now, concerning the TQ (torque) signal. I have managed to make the thing work w/o any modification to the PCBs. My steps per mm is 640 (5 mm pitch ball screws, 1.8 degree steppers, x16 microstepping), controller is GRBL 0.9j running on a 2560 Mega (BTW Mega is now fully supported). What I did was set DIV268N current to the lowest level (to not bother the motors) and then tried sending move commands with different feedrate to see how the TQ schematic would react. What I found out was that at feedrates above 1500 mm/min approx, the schematics would actually trigger and set TQ pin of TB6600 to high level. At 1000 mm/min the output was low, at 1250 it was at "uncertain" level. In my setup the steppers run fine at 500 mm/min, however at 1000 mm/min there is not enough torque and they start to skip. So, I checked GRBL settings and found out the step pulse width was set to 10 usec. I changed the setting to 20 usec and voila!! The TQ was working as desired. It is now activated at 750 mm/min approx and motors run w/o skipping at 1000 mm/min. I even wired a separate LED between GND and TQ signal (with 470 ohms resistor in series) to indicate when high torque mode was activated. I checked all 3 drivers that I had and they all function in the same fashion.

So, it looks like the torque schematic IS working. However, the design flaw is that it will only work in certain conditions. For example, if you use x8 or x4 microstepping, it won't work the same way due to lower step pulse frequency: you would need 40 usec and 80 usec step pulse to make them work, respectively. Also, it depends heavily on your setup: screw pitch, etc etc if this setting would fit your needs.

I personally think that these drivers are usable. They are really cheap, they DO require some modification, but that's sort of fun, don't you think? And of course, they do feature wrong microstep and current table like it was mentioned before. Someone here mentioned poor durability of these drivers because of schematic design flaws. I personally disagree. This schematic is built AROUND the TB6600 chip. TB6600 is a quality product from Toshiba. It really can't go wrong provided you don't hit its limits with higth voltage/current and that it cools properly. It's also a good idea to check if enough thermal paste was used (mine had enough). It also, probably, needs the optocoupler power supply capacitor modification mentioned before (I have not done that yet). Then, they are usable.

[iokiok]

hey,all!
what could cause this transistor to blow up? could it be caused by the interconnected grounds from the two separate power supply modules? what other components could be affected? (is this the one which connects to pin #1 of the driver? 3906?)

[ingovogel]

Hello together,
i have also byed 4 of them ....

now spend 2 hours over the pcb ...

microstep dip switches are changed ... and low is on and high is off position ... if you take a look at the tb6600 datasheet ...

SW4 is to M1, SW5 is connected with M2 and SW 6 is to M3 ...

i think i will prog a ATtiny45 to set the Vref and to go in low torque after 2 sec ...

if someone is interested ... i will give infos ...

Greetings Ingo

[Ginjaian]

Hi!

I'm building a CNC router at the moment, and have reached the perennial problem of what stepper driver to use. I have a 24V 10A PSU, with pretty standard 4A NEMA 23 stepper motors. I'm leaning towards the Massmind or Lucas THB6064-based boards, which sound well designed, but there is also the PiBot Stepper Motor Driver board, that uses a TB6600. Someone did mention this earlier in the thread, and I've seen others ask about it on CNCZone, but there's little feedback. I believe it's quite often used on Shapoko machines. Does anyone have any feedback on how well this one works?

If Lucas happens to read this thread... Do you have any plans to do a surface mount version of your board(s)? I only ask because most of my soldering nowadays is SMT, rather than through hole. And it would be good to get a stepper driver design for mass manufacture, that uses the Toshiba chips, that actually works!

Ian

[James Newton]

...it would be good to get a stepper driver design for mass manufacture, that uses the Toshiba chips, that actually works!
Ian

You are going to mass manufacture such a design? Who is going to do that? I'm interested, but I can't cost justify it vs the cheap garbage China puts out. E.g. people will buy the cheap one and deal with the failures instead of buying one I finance the mfgr of, because they don't see the difference. If you want to help do a mass run, let me know. I've seriously thought about it, but the cost won't be low. I don't think we can get below $30 actual sale price. China can sell for nothing because they use knockoff parts that don't work right. And the buyers never know what they are missing. Ask me how I know:
https://www.youtube.com/watch?v=3Zva2QvXRFA

Right now, people who like to build kits buy the drivers. They know they can repair them cheaply (although, based on my poor sales of the replacement driver chips, that doesn't happen often) and they have pride in having made it.

[Ginjaian]

Hi James

Thanks for the reply! Don't get me wrong, I've had my fair share of rubbish; I had one of the early TB6560 3-axis boards. I got this working, just, before starting to understand why it didn't work very well. Which is why I'm researching more carefully this time around! I just thought the time might be right for a more integrated, quality solution, that doesn't have the cost attached of, say, a Geckodrive.

It's all very well producing a DIY kit, but realistically, this limits your market to people who want to build EVERYTHING themselves. A set of modular, pre-built electronics, something like PiBot are doing, seems like the middle way between inferior far East products and expensive Western electronics; ie Eastern manufacture to Western design and quality control. It's pretty much what's happened in the 3D printer market (of which I'm a veteran!); I'm not sure many people are soldering their own Sanguinololu or RAMPS boards any more (as they were a few years ago), because there are plenty of usable, reasonably priced (though integrated) driver boards around. And the CNC world will develop that way too. But to move forward, a surface-mount design is needed, which is cheaper to produce than a through hole design. And yes, I would consider helping development of this, with my time and money (if need be) - though an actual production run could be financed by running a Kickstarter.

I was part of the team that developed a through-hole Toshiba TB6560-based set of electronics for 3D printers: Phoenix Electronics - RepRapWiki
I know you talk about the 'step elongation' built into your board, but it's not really necessary; you just need to increase the length of the step pulse in the firmware. We did this 4 years ago; this is the original commit: https://github.com/buserror/Marlin/c...1ec6fec33d1818
Marlin has changed rather a lot over the last few years, but the switch to enable a longer step time is still there! See line 296 of https://github.com/MarlinFirmware/Ma...Conditionals.h
I haven't looked at Marlin for a while, but I think this part now controls the step pulse (from line 476): https://github.com/MarlinFirmware/Ma...in/stepper.cpp

Ian

[James Newton]

Thank Ian,

I'm very aware of the limitations of the current kit... LOL... my sales figures can attest to it. The issue is that I can't compete with China with a ready made solution. Or I don't think I can, do you? I can beat them senseless on quality, but that doesn't matter when it comes to selling to people who aren't electronics savvy. My current customers "get it". The average spindle jock who just "wants an electronics package" won't be able to grok the difference between my quality unit at $30 and the crap China is pushing at $25. If he buys it at $25 from China #1, and it dies, he will buy it again at $25 from China #2, and so on. At some point, he will switch to a different type of driver "because those 6064 drivers just aren't reliable". See? I'm already getting that. "Why would I buy your kit for more than the cost of a fully assembled unit from China? And anyway, those 6064 drivers just don't last / move well / etc..." MY 6064 drivers run like a dream. But if you see the difference in the components I'm selling, you won't mind soldering it up.

On the pulse stretching thing. Yes, but there are a number of people who are comfortable doing electronics and NOT comfortable doing firmware. Not having to change the firmware is an advantage.

[crisis]

Hi all yet another newbie . quick background I'm more from engineering background than electronics so yep i bought a kit from EBay which included the HD-DIV 268N-5A (yes I now know it's not great)
overall my build went well and now configuring the system and reading everything I can thats when I found that even the chart for stepping printed on it and documentation were wrong I've
seen lots of threads with revised charts half of them conflicting with each other so I thought I would see if i could resolve this for my self and if right help others.

So just reverse engineered the problem if I screwed up please let me know

first info I read was higher the step rate shorter the distance traveled IE 1 step= 1mm 1/4 step = 0.25mm
so I just jogged my x axis to travel a set distance of 100mm and read the distance that Mach3 thought it had traveled it didn't matter if it was accurate just needed a base line
next I went through all of the dip switch settings and repeating the test and getting a distance traveled for each setting
and yes I got a different distance for each setting so working from this if I'm right the greatest distance traveled should equal 1/1 and the shortest distance equal 1/16.
So this gave me a chart that had to be correct. Or did I get something wrong with the premise I wont pint the chart here no point if its wrong but if you guys can say that I'm not barking up the wrong tree ( local saying) I'll post my results.

By the way great forum

[Jon.N.CNC]

Hi!

I'm building a CNC router at the moment, and have reached the perennial problem of what stepper driver to use. I have a 24V 10A PSU, with pretty standard 4A NEMA 23 stepper motors. I'm leaning towards the Massmind or Lucas THB6064-based boards, which sound well designed, but there is also the PiBot Stepper Motor Driver board, that uses a TB6600. Someone did mention this earlier in the thread, and I've seen others ask about it on CNCZone, but there's little feedback. I believe it's quite often used on Shapoko machines. Does anyone have any feedback on how well this one works?

If Lucas happens to read this thread... Do you have any plans to do a surface mount version of your board(s)? I only ask because most of my soldering nowadays is SMT, rather than through hole. And it would be good to get a stepper driver design for mass manufacture, that uses the Toshiba chips, that actually works!

Ian

You don't get 4a repeatble current out of any of the toshiba chips. You need chopper type mosfet drivers to power those motors correctly. Toshiba drivers are a waste of time for cnc routing.

3D printing nema 17 tech being tweaked to appear suitable for nema 23's with bogus amp ratings.

Pibot are awful, most I could get out of them was 0.9a repeatble cranked right up and the 3a motors sounded loud and clunky and lost steps.

[lucas]

China can sell for nothing because they use knockoff parts that don't work right. And the buyers never know what they are missing. Ask me how I know

Those could be remarked THB6064H's, these have seperate OSC input's for each channel, one of the decay mode pins is used for an osc input. This could explain why the motor turns strange, actually on a single channel

If Lucas happens to read this thread... Do you have any plans to do a surface mount version of your board(s)? I only ask because most of my soldering nowadays is SMT, rather than through hole. And it would be good to get a stepper driver design for mass manufacture, that uses the Toshiba chips, that actually works!

I agree with James, designing is no problem but mass production requires a heavy investment and people are still guided by the cheapest price, why are there still crapp chinese TB6560 boards sold?
BTW: the chip is not to blame, it's the design. I have 2 machines running with TB6560 without problems.

So this gave me a chart that had to be correct.

I never made a correct table, there were 2 colums swapped if memory serves me well, yours will probably be correct and should match some mentioned here.

You don't get 4a repeatble current out of any of the toshiba chips. You need chopper type mosfet drivers to power those motors correctly. Toshiba drivers are a waste of time for cnc routing.
Pibot are awful, most I could get out of them was 0.9a repeatble cranked right up and the 3a motors sounded loud and clunky and lost steps.

I can't comment on the Pibot, never used it.
The TB6600 runs a bit rougher (even if well designed) than the THB6064AH.
I get the current out of all the different Toshiba chip's into the motor, no problem. If you can't then you either have a bad drive or use a wrong measuring setup.

[Jon.N.CNC]

Those could be remarked THB6064H's, these have seperate OSC input's for each channel, one of the decay mode pins is used for an osc input. This could explain why the motor turns strange, actually on a single channel



I agree with James, designing is no problem but mass production requires a heavy investment and people are still guided by the cheapest price, why are there still crapp chinese TB6560 boards sold?
BTW: the chip is not to blame, it's the design. I have 2 machines running with TB6560 without problems.



I never made a correct table, there were 2 colums swapped if memory serves me well, yours will probably be correct and should match some mentioned here.



I can't comment on the Pibot, never used it.
The TB6600 runs a bit rougher (even if well designed) than the THB6064AH.
I get the current out of all the different Toshiba chip's into the motor, no problem. If you can't then you either have a bad drive or use a wrong measuring setup.

What do you mean you get the current? half an amp will turn a 4a stepper motor. Or if your measuring current then you need to be measuring repeatable current not peak.

Tested various tb drivers, for one thing it's a gamble being that the market is flooded with counterfeit tb chip drivers which doesn't help, but none got anywhere close to thier rated amps. Motors run quieter when correctly powered, a single step being barely audible, often see in tb related threads people saying they have turned the amps down because the motors were too noisy. Lol

Extra £10 per driver you can get chopper type mosfet driver that runs cool so don't suffer thermal issues and thier rated current is repeatable and 100% smoother motion than tb.

Go for something like a cw5045 or even the cheaper slightly less quality wantai drivers.

Toshiba recommended for anything more than 3D printing is the doing of complete novices or eBay sellers

I sent an email to Pibot about the amp ratings being nowhere near the measured amps, the response I got was "current cannot be measured, the rated amps is based on simulated figures" pmsl that email got around loads.

Measuring current I just use vref math over a shunt resistor with my oscilloscope, it's not as accurate as expensive probes but accurate enough to see that all the tb drivers were outputting under half current when continuous power was drawn.

But all you really need to do it try both types, and there's simply no argument.

[Ginjaian]

Hi Jon.N.CNC

Thanks for your input, but I have to disagree with you about the Toshiba drivers; I don't think the chip itself is the problem (so long as it's genuine), just how the cheap versions of the PCB have been implemented. Unfortunately, the poor design of the board has become an epidemic, and for some reason, despite the negative reports about these, the Far Eastern manufacturers still keep producing the same thing! I have seen exactly the same problem with other budget electronics; despite clear problems, they just never bother fixing the underlying issues, so they keep selling these the same poor quality boards to unsuspecting people, who haven't done adequate research. They work up to a point, because, as you point out, you can run a 4A motor on 0.5A, you just won't have much torque!

I suspect the PiBot stepper driver has the same issue as the other Chinese boards; that the stepper driver is put into idle mode far too quickly after the last step, so it appears that it has virtually no torque. As Lucas and James have pointed out, it's a problem with the board design; their boards don't do this. If perhaps Lucas could take a look at the schematic, we might know more? The schematic is here, but I don't know what version it is: PiBot TB6600 Stepper Driver - RepRapWiki
The reason I asked for feedback on the Pibot board was to see if it was an improvement on the old boards.

Thanks for the information on the CW5045 driver, it looks interesting. Unfortunately, there isn't much feedback/reviews on it (apart from your recommendations on other threads on cnczone), and I don't know enough about the different types of stepper driver to know if this is a good choice or not. Anyone else got any thoughts on these? I'm tempted to get one of each (I already have a TB6560 board, so James's/Lucas's TB6064 board, a Pibot TB6600 and a CW5054), set them up as similarly as possible, and run a back-to-back test. I'm going to order board kits from Lucas anyway - sorry James, but I'm in the UK so Luc's closer, I prefer the opto-isolated board, and you don't seem to have much stock according to the massmind website!

Ian

[crisis]

Ok this is the results I've got. Can only say this is for the HY-DIV268N-5A which seems the most common one off ebay so it's going to be the one that a lot of newbies get stuck with.

as I said in other post these results were from measuring actual distance traveled on my cnc

Switches

4 5 6 Distance traveled Miro step

OFF ON ON 100mm 1/1
ON OFF ON 50mm 1/2
OFF OFF ON 50mm 1/2b
ON ON OFF 25mm 1/4
OFF ON OFF 12.5mm 1/8
ON OFF OFF 6.25mm 1/16


So after all this it seems that what is printed on the unit is correct.to what results I'm getting. so is it just another one of those statements that gets posted everywhere that has no basis in fact
or am I still missing something lol

[Ginjaian]

Hi crisis

I think that just means that the drivers you have, have the correct microstepping chart on them. Unfortunately, the other problems with these drivers are unlikely to have been corrected, ie noisy motors and lack of torque. However, they may still be sufficient for your application.

Ian

[crisis]

Hi Ian

yep I suppose thats the big problem for builders like me making a cnc in a workshop/garage/shed when you are fist starting out you have nothing to compare to. I can watch hundreds of youtube vids but they will never give you a real idea of when a cnc system is optimal as apposed to just working unless you can see one in the flesh and hear what it should sound like, the power/torque that should be expected. I think thats called experience...lol. but a big thanks to James for posting vid that did help a lot.

[rprilutsky]

4 5 6 Distance traveled Miro step

OFF ON ON 100mm 1/1
ON OFF ON 50mm 1/2
OFF OFF ON 50mm 1/2b
ON ON OFF 25mm 1/4
OFF ON OFF 12.5mm 1/8
ON OFF OFF 6.25mm 1/16

Hi I am new to this so i apologize in advance if my question sounds stupid. Why with increase of the number of microsteps the distance traveled by the axis decreases? If each pulse on PUL/CLICK wire results in one microstep with the higher number of mictosteps per step I would expect the speed to decrease rather then the distance, The distance should stay the same? What am i missing.

[_ID_]

With microstep settings you divide the main step into smaller sections of it.

1:1 means that one pulse would move the stepper 1 step ahead
1:2 means that one pulse would move the stepper 1/2 of a step ahead (you need 2 pulses for one full step of the stepper)
1:4 means that one pulse would move the stepper 1/4 of a step ahead(you need 4 pulses for one full step of the stepper)
1:8 means that one pulse would move the stepper 1/8 of a step ahead(you need 8 pulses for one full step of the stepper)

With microstep you gain resolution of the stepper motor. For instance: if the stepper is making 1mm of linear path with 1 full step it will make 0.25mm with 1/4 step. But there is a catch. The higher you go with resolution the less torque you have. In real world 1:16 is more than enough. You can make a calculation how much is 1:16 of 2mm(linear distance of 1 full step).

[fungunner2]

Hi All
Sorry my bad English ...
The Stanby Current never works for me and my Steppers are fast hot...
I know this China Import can never run as a Leadshine ,but a one Minute Operation and my Steppers says Thank you
I work with Linuxcnc and there is a Way to controll the standby Current .
For this I separate Pin 3 from the TB 6600 and wire this Pin eg. to Pin 14 from Parrallel port.
Now a little bit configuration in linuxcnc and the current sinking works for me: No more Problems with fast Stepper Heat.

[yaidelcnc]

I will build a pcb with the scheme corrected lonewolf55 (I assume what is the latest and has corrected the greatest number of defects, please correct me if I'm wrong) trying to keep as many components of the original plates (whenever possible ) and upload it.
regards