[A_Camera]

Hi,

just one more user of the eBay tb6560, and I would like to share my experience and my measurements. Going through this very long thread, one of the main themes is about the shortcomings of the design with respect to the timings and noise of the pulses after the optocouplers.
My motors tend to loose steps at more than 600 mm/s, and indeed, the oscilloscope shows a rather sloppy rise time in the step input of the tb6560. The levels seem quite OK to me:

Attachment 304904

As a test, I tried following someone's advice to connect a 1K pull-up resistor to each optocoupler, and the waveform becomes much, much better, with a reasonable rise time:

Attachment 304916

I did this for all the dir and step optocouplers, which is a pretty simple operation, and I noticed that the motors sound somehow smoother and the machine can run up to 1500 mm/s. I'll keep them at a safe 1000 mm/s, which is more than enough for my needs.

Seems a big improvement, and in the next weeks (eh, free time is what it is...) I'll do more testing after installing the new spindle, as I already destroyed two dremel-like spindles by doing relatively intense carvings...

The other big improvement that I had experienced some time ago was to wire only half of the motors' coils, to lower the inductance. At the time the max speed I could achieve was ~200 mm/s, which was quite disappointing.
My motors are marked 57BYGH603B, and seem to be quite common in cheap eBay kits. They have a very large inductance, which may explain the difficulties in driving them.

I hope that my experience is useful to others, and I wish to thank all those who posted the very useful info and explorations in this forum.

Here's a photo of the mod with the 1K resistors, I'll be happy to check if other people with the same issues as me find any analogous improvement by mounting them.

Attachment 304932

Cheers
fbx

Those optocouplers are of no use, they give you a false sense of safety. Remove them all together and your performance will increase more. Another thing is that I don't think you run at 1500mm/s, possibly 1500mm/min... and if that is really the case then you are really lucky. Even 1500mm/min is a very high speed for those drivers, assuming you are using a ball screw with 5mm pitch.

[Ladi268]

Hello,

@petsat: thanks a lot for the quick reply. The power on circuit will be one of the next things I will implement. I hope the Toshibas will hold on until then.

Modifications so far as already mentioned: jumpered optos, grounded heat sink and removed resistor in current regulator circuit. Tomorrow I'll get 150pf caps (radial none smd but I don't think that this will cause any trouble).
The odd thing so far is that the hissing noise is more obvious than before (which doesn't really bother me, just wanted to mention it). But really weird is one thing: if I jog the axis from the keyboard and let go the key the axis continues to travel for another second or so. This was not the case before the changes. Maybe somebody knows this behaviour.
One more thing I wanted to ask you: the Linuxcnc Wiki also suggests to replace the 16 diodes against super fast UF5408 - what do you guys think of that mod?

Cheers from Germany,
Juergen

[A_Camera]

Hello,

@petsat: thanks a lot for the quick reply. The power on circuit will be one of the next things I will implement. I hope the Toshibas will hold on until then.

Modifications so far as already mentioned: jumpered optos, grounded heat sink and removed resistor in current regulator circuit. Tomorrow I'll get 150pf caps (radial none smd but I don't think that this will cause any trouble).
The odd thing so far is that the hissing noise is more obvious than before (which doesn't really bother me, just wanted to mention it). But really weird is one thing: if I jog the axis from the keyboard and let go the key the axis continues to travel for another second or so. This was not the case before the changes. Maybe somebody knows this behaviour.
One more thing I wanted to ask you: the Linuxcnc Wiki also suggests to replace the 16 diodes against super fast UF5408 - what do you guys think of that mod?

Cheers from Germany,
Juergen

Remove the diodes. They have no purpose since the 6560 chip already has those built inside.
The hissing noise is really annoying and you can't really get rid of that. You can improve but I don't remember how.

The thing is, you are wasting time and money. Get three DQ542MA drivers and you'll never look back. It's like day and night.

[petsat]

The thing is, you are wasting time and money. Get three DQ542MA drivers and you'll never look back. It's like day and night.

Wasting time YES! (although I believe that there is no such thing as wasting time in making something better... its EDUCATIONAL TIME)
Money? ...well the fixes we all suggest here cost a few euros. Even the power up and down mod could be replaced by two switches that you turn on and off at the right sequence.
On the other hand I agree that DQ542MA are a much better controller but it costs a few more euros.
If one go for the DQ542MA and throw away the TB6560 he will never get the feeling that he "made something work as it should". This is one of the few happy moments we hobbyists have

The decision is his anyway.

Petsat

[A_Camera]

Wasting time YES! (although I believe that there is no such thing as wasting time in making something better... its EDUCATIONAL TIME)
Money? ...well the fixes we all suggest here cost a few euros. Even the power up and down mod could be replaced by two switches that you turn on and off at the right sequence.
On the other hand I agree that DQ542MA are a much better controller but it costs a few more euros.
If one go for the DQ542MA and throw away the TB6560 he will never get the feeling that he "made something work as it should". This is one of the few happy moments we hobbyists have

The decision is his anyway.

Petsat

...but in the end the card will end it's life in blue smoke. So, indeed that's educational. What I learned was that I should have made my move earlier. I personally didn't learn more than it was waste of time and money, it delayed my CNC project time considerably.

BTW, I used two switches as well, but a beautiful day my card went up in smoke after a sudden and short main power outage. Power was lost for a few seconds and when it came back the X was not working.

[fbx]

Those optocouplers are of no use, they give you a false sense of safety. Remove them all together and your performance will increase more. Another thing is that I don't think you run at 1500mm/s, possibly 1500mm/min... and if that is really the case then you are really lucky. Even 1500mm/min is a very high speed for those drivers, assuming you are using a ball screw with 5mm pitch.

Hi,
indeed it's mm/min, sorry about that. And yes, my machine has acme screws.

I agree with you about the optocouplers. From my point of view, putting the pull up resistors is way quicker than desoldering stuff, so I stuck to that. Still the inputs will be somehow decoupled, for what it counts.

After all this time I can still confirm my speed numbers, the machine works very well.

f

Sent from my ONE A2003 using Tapatalk

[A_Camera]

Hi,
indeed it's mm/min, sorry about that. And yes, my machine has acme screws.

I agree with you about the optocouplers. From my point of view, putting the pull up resistors is way quicker than desoldering stuff, so I stuck to that. Still the inputs will be somehow decoupled, for what it counts.

My opto couplers were in sockets, so desoldering was not necessary, just removed those and shorted the inputs with the outputs of them. Takes five minutes to do it, less than adding the resistors, which in my opinion is not a good idea at all.

After all this time I can still confirm my speed numbers, the machine works very well.

Very good. As long as it works...

[stragenmitsuko]

Throw a few 100$ at it and you'll get something better . Sure
Throw in a few 1000$ an and you'll get something superb . Offcourse
Now ad some more 10.000$ and you can buy your own ready made cnc . Why bother to build one yourself

See what I'm trying to say .
Any fool withg a lot of $$$$ can accomplish anything

The real challenge is getting somthing to work or improve it , not throwing big bucks at it .

My 2 ct's .


Oh and btw my modified tb6560 3axis card has worked for a couple of years now without smoke or whatever .

[petsat]

Throw a few 100$ at it and you'll get something better . Sure
Throw in a few 1000$ an and you'll get something superb . Offcourse
Now ad some more 10.000$ and you can buy your own ready made cnc . Why bother to build one yourself

See what I'm trying to say .
Any fool withg a lot of $$$$ can accomplish anything

The real challenge is getting somthing to work or improve it , not throwing big bucks at it .

My 2 ct's .


Oh and btw my modified tb6560 3axis card has worked for a couple of years now without smoke or whatever .

Couldn't agree more

petsat

[aarggh]

Any fool withg a lot of $$$$ can accomplish anything

And then of course there are those who keep on buying the same product over and over to replace failed products in the quest to get it "working", in the expectation that repeating the same thing over and over will somehow produce a different outcome.

cheers, Ian

[A_Camera]

Throw a few 100$ at it and you'll get something better . Sure
Throw in a few 1000$ an and you'll get something superb . Offcourse
Now ad some more 10.000$ and you can buy your own ready made cnc . Why bother to build one yourself

See what I'm trying to say .
Any fool withg a lot of $$$$ can accomplish anything

Not really true. $$$$$$$ doesn't give you the knowledge to accomplish anything at all.

When I bough mine I knew I was going to have to fix it but I also knew I had more than enough knowledge to do it, had all the instruments I needed and also all the material necessary. Without that background information I wouldn't have started and wouldn't have bought that crap.

The real challenge is getting somthing to work or improve it , not throwing big bucks at it .

My 2 ct's .

I fixed mine as well, worked excellent for a while, at least I THOUGHT they worked excellent, until I had something else to compare with... after that... that's a different story.

Oh and btw my modified tb6560 3axis card has worked for a couple of years now without smoke or whatever .

Good for you, or should I say you are one of the few lucky ones. I keep hearing that now and then, but of course, there is no way to check it out but there is also no reason to doubt it. Never the less, facts are facts, those cards are just crap and there are MUCH better alternatives, . Dealers don't really tell you to buy those cards ONLY if you know a lot of electronics and not afraid of working with the fixes needed. I mean, do you also buy brand new cars and think it is OK that it is not according to the specifications, needs modifications and fixes just to be able to drive it? I don't. If I buy something which supposed to work of course I am disappointed if I have to fix it, regardless how much money it costs or if I have the knowledge to do the job or not.

BTW, already your first line is nonsense. You don't need to "throw a few hundred $$$$$" out to get something better than this card. Here is a quick link:

De Free SHIP to EU 3Axis Wantai Stepper Motor Driver DQ542MA 50V 4 2A 128MICROS | eBay

From Germany to Sweden, 112$ three drivers and a BoB from a seller with 100% rate. That is really cheap, and I haven't even looked for cheaper when I searched eBay a minute ago. If I wanted the cheapest I could probably get it for less than 100$ if I was ready to wait for shipment from China.

In other words, all your arguments are failed.

My

[Ladi268]

Hello again guys,

right now I am quite (no, not quite, really) happy with the results. After all the mods I am running my machine with 600 mm/s with 1.25 mm pitch lead screws (regular M8 stainless steel screws) in half step mode.
Maybe I'll give full step a try tomorrow - since I have only 1.25 mm pitch the accuracy in theory is still better then I'll expect in reality. My machine is a proof of concept for me and is completely made of wood with drawer guides on each axis and still it was accurate to 0.15 mm on a test run this evening. That was more then I expected. OK, for some 0.15 mm might be far off, but regarding woodworking 0.15 is nothing.

@petsat: if I understood your power up/down circuit correctly you are giving the board approx. 12 V for 2 seconds and then you go up to 24 V. What would be the difference if I adjust my lab PSU down to 12 V and then turn it up to 24 V after some seconds? Since it is only me working on that "Proof of concept" I can live with that as well - during the last few days I always turned it down to 11 or 12 V before shutdown and the next time when I turn the PSU on it starts again with the 12 V. Is that safe as well in your opinion or should I get my soldering iron out again :-).

Cheers and again thank you very much guys,
Juergen

[petsat]

Hello again guys,

right now I am quite (no, not quite, really) happy with the results. After all the mods I am running my machine with 600 mm/s with 1.25 mm pitch lead screws (regular M8 stainless steel screws) in half step mode.
Maybe I'll give full step a try tomorrow - since I have only 1.25 mm pitch the accuracy in theory is still better then I'll expect in reality. My machine is a proof of concept for me and is completely made of wood with drawer guides on each axis and still it was accurate to 0.15 mm on a test run this evening. That was more then I expected. OK, for some 0.15 mm might be far off, but regarding woodworking 0.15 is nothing.

@petsat: if I understood your power up/down circuit correctly you are giving the board approx. 12 V for 2 seconds and then you go up to 24 V. What would be the difference if I adjust my lab PSU down to 12 V and then turn it up to 24 V after some seconds? Since it is only me working on that "Proof of concept" I can live with that as well - during the last few days I always turned it down to 11 or 12 V before shutdown and the next time when I turn the PSU on it starts again with the 12 V. Is that safe as well in your opinion or should I get my soldering iron out again :-).

Cheers and again thank you very much guys,
Juergen

Well, the concept of my design is that I give to the circuit 12v for the "on-delay" (555 ic) and, at the same time, through the 7805 regulator I supply the 5 volts to the 6560 chip. Only when these two conditions met ("and" gate does that) the 6560 ic gets 30v for the motors. This ensures that the 6560 chip is on and stable WITHOUT any motor power and then the motor power (30v) is on.



Power down reverses that: It cuts instantly 30v power and gives 2 seconds time to the 6560 ic to get stable without any high voltages before shutting it off.

If I understood well your thoughts you intent to supply 12v initially and then power it up to 24v. That violates the manufacturers suggestion for powerup and down the IC

Attachment 313872

petsat

[petsat]

If you want to simplify the power up and power down procedure you need two power supplies DC 5v & DC 24v and 2 switches.
You remove 7805 Voltage regulator from the board, you connect 5v supply through switch1 to the board (see photo) and connect 24v to the suggested connector (those 2 supplies are grounded to the same point as well).
The sequence is:
You turn on switch 1 (5v) and give power to 6560 and then after 1-2 seconds you turn on switch 2 (24V) to give power to motors.
Then you turn on your computer and start mach3.

to power off you reverse the procedure. Computer OFF, 24V OFF, after 2 seconds 5V off.

Of course if you make a mistake to the sequence you QUARANTEE FRYING THE 6560 chips

That is the reason I designed the circuit!!!! I'm prone to mistakes...

Attachment 313886

petsat

For as long as I was using the 2 switches I had switch 1 always on an 6560 chips always getting 5 volts! This isn't something I suggest though!!!

[kilbold]

too thansk ...

[ehtisham-syed]

what about shutdown due to power loss, both inputs will go out at same time will it damage the chip

[petsat]

what about shutdown due to power loss, both inputs will go out at same time will it damage the chip

The problem will be when the power will come back! Not at power loss there will be no power present! Both switches will be on simultaneously though and the powerup sequence will be violated. For that reason I designed the circuit above!

[KH0UJ]

I got 2 of these in the shop, funny mine is working perfectly bolt-on directly last two years ago and it`s up and working until now, with this as a spare, so what`s the difference about my board and the OP`s board? it worked perfectly on a 24VDC power supply, one thing for sure LM78XX IC`s are rated 32VDC max not 36VDC.

[petsat]

I got 2 of these in the shop, funny mine is working perfectly bolt-on directly last two years ago and it`s up and working until now, with this as a spare, so what`s the difference about my board and the OP`s board? it worked perfectly on a 24VDC power supply, one thing for sure LM78XX IC`s are rated 32VDC max not 36VDC.

You are in LUCK!
The whole issue of smoked ICs comes from the fact that the designers of the board did not follow the recommendations set by Toshiba for the powerup and power down sequence.
There are other minor design faults but this is the one that leads to 6560 failure.

[ncysys]

Hi Guys. Here is some tips for users of the TB6560 stepper motor driver:

Make sure your pulse signal duration is around 10% of the time between pulses. If you use a 555 timer or a signal generator with 50% duration and you notice the motors do not have good torque or limited speed the pulse duration is the problem. Mach 3 normally outputs a narrow pulse.

Also when powering up the driver, make sure the disable signal is on before the power is applied to the driver. This is the power-on sequence recommendation from Toshiba.

If you want to optimize the performance of your driver, consider at least the schmitt trigger between the opto-coupler and the driver clock input and if there is the torque reduction circuit, it should be removed.