rajshinai, Thank you, also! It's good to know we are on the right track.
rajshinai, Thank you, also! It's good to know we are on the right track.
Hi all,
I've been reading this thread since some time and decided also to buy one of those boards from Ebay. They are cheap and I don't fear soldering
I also got the newest 3-axis version in combination with a power supply and three nema 23 steppers (1A). I didn't make any modifications and all axis seem to work. I will try to replace C25/C26/C27 as soon I get the parts. What's the point of jumpering the opto couplers if the board works with them untouched?
For those who have problems setting up the board with Mach3, take a look at the guide I got with my board. Fig.5 and Fig.6 on page 7 should help. And if you are a Mach3 beginner like me: Read the Mach3 manual! Especially the installation guide.
Jumpering Optocouplers does indeed not improve functionallity. I ended up doing so though because I used the Optocouplers from the board to properly isolate the input (PC) from the Stepper driver. The board has one common ground accross the whole of the board which is really bad in case something goes wrong, because your PC could very well get damaged. I built myself a small extra PCB that properly implements an optical isolation with the optocouplers from the board (easy if your using an arduino with grbl which is what I'm doing, a bit harder if you use an LPT cable and a PC). The way the optocouplers are used on the board simply makes them useless.
Also remove the resistor i marked in one of my earlier posts, this reduces noise a great deal. I made some videos from before/after each mod, I'll be uploading those to youtube as soon as i've got some time.
This is in my opinion all the modification for blue board.
Please comment if see something wrong .
Attachment 251264
I removed the R55, R60, R61 and replaced C25, C26, C27 by 270pF as Joh-nny recommended. My board is now working with much more steps per second than before (about 2.5x on 1/8 step mode)!
Thank you!
Hi guys,
Firstly, thanks for taking the time to post your findings on the problems with this board, and for sharing the information on how to fix them.
I have a board which looks slightly different to the pictures on this thread, but it seems to be suffering with the same problem, missing steps when using mach3, even when the speed is set to just 50%. If i move more than 1 axis at a time it also happens. The DRO might say i've moved 10mm when I might have only moved 2mm.
My setup is a cheap little X2 mill, pc running mach3, 6560 board and some arcade buttons and levers in a custom built touch screen display console with a jpac USB arcade controller to manually control each axis.
By looking at the pictures i've attached, can anyone please tell me if it's possible to fix my board with the same components (resistors, capacitors etc) and where to solder the parts?
Thanks, Alan
Attachment 265178
Attachment 265180
Attachment 265182
Attachment 265184
There are some users have the knowledge and desire to experiment and tweak this board and be happy with it. However, Those in the know understand how poorly designed this board is. I had this board and threw it away after trying to give it away.. Simply because even though I have the ability to make the mod there are too many other negative variables with this board that will make it no fun to operate.
One could spend hours with it dreaming of a very low cost means to get 5 axis. The reality is this board is junk. The driver chips themselves are old school and have known faults that newer systems do not worry about. You can be running along fine and suddenly another problem rears its head. This may have been a good deal 20 yrs ago. But today I feel one would be better served purchasing a hobbyCNC kit with motors at a very low cost and be done with it.
Hi Alan,
Your pics don't work, at least for me. Yes there are problems with this board, but it should work at low speeds. I can't help you much with the electronic stuff. Are you sure your settings in Mach3 are correct? Check if your steps/mm setting suits your hardware. Also check your dip switches on the top of your board. It can be set to full-step, halve-step and so on. If you have chosen 1/16-step there may be too much steps per second, which can't be handled.
Thanks Curtis,
I'm aware that buying cheap usually means buying twice, but thought this fix would make my board a lot better. If I bought a Gecko 540 controller, would this solve my issues?
Felix,
the board is fine below 65% feed rate, it's just when I push it above this each axis stutters and freezes. I'm sure all the settings are correct as it works ok slowly, but thanks for the info on that, i'll check the switches and see how that goes. I'll try and get some pictures up.
Thanks guys!
You might simply be running into the limitations of your system. What is the inductance of your motors and what power supply voltage are you using?
My power supply is 24v and each motor inductance is 3.6 mH.
cheers, a
With such a high inductance, you are simply running into the limits of the power supply. The voltage should be around 60v for maximum performance.
Also, 24v is likely the highest that the 6560 can take without self destructing. You need a high end product, like the ones from Geckodrive, to get the maximum performance from those motors.
Hi all,
The first 3 axes 6560 board I bought from ebay some time ago (www.hyu68.com) blew up after connecting it to the power (12v). The supplier replaced it instantly without asking any questions (??) for a new one.
This made me suspicious and when the new one arrived I never connected it to the power supply but I seek information on those boards.
This search lead me here where after reading all the posts (THANK YOU ALL) I have made the following mods:
I CHANGED THE OSCILLATOR FREQUENCY
I have replaced capacitors on pin7 (silkscreened 102, i.e. 1000pF) with 3 smd capacitors salvaged from an old PC motherboard marked KE2 (i.e 150 pF)
Attachment 275602
CURRENT REDUCTION CIRCUIT REMOVAL
Like many people here I have also disabled the "clever current reduction circuit" (that apparently doesn't work as it should) The mod was simple I have just removed a resistor a transistor and the capacitor that form the circuit.
Attachment 275604
GROUNDED HEAT SINK
Heat sink is now Connected to pin 15 through the screw. A lead from one screw of every driver connected to the corresponding ground Pin 10 or/and 15 but NOT pin 6 (this is the 5V-GND - it should not be disturbed by additional driver currents!)
DESIGNED AN MADE A POWER DELAY CIRCUIT TO POWER ON & OFF THE BOARD
The main problem I had was the power sequence that was not right according to Toshoba's datasheet. In order to care for it I designed a small circuit to power up and down the board in a safe way.
To do that I removed the two voltage regulators (7812 and 7805) from the board and supplied the 5 and 12 volts from the new power supply circuit. This power supply circuit also supplies 30 volts for the motors and operates as follows:
When S1 is turned to ON position, it gives a 2 sec delay between 5V and 30 V power to the chip and when S1 is switched OFF it turns 30V off instantly and turns off 5V after 2 seconds.
I have tested the circuit and it works (so far) without any problems.
If anyone spots any flaws on that I will be happy to discuss and modify it.
Petros
I am unsure, what it is you want to do.
If you read through this forum thread, you will find a lot of tips about modding a TB6560-based driver card.
Some tips I personally consider valuable, some others bogus or even wrong (like changing the oscillator frequency to values outside specification for the chip).
The typical chinese cards with TB6560 chips are pretty badly designed and laid out: there is no separate power planes for the H-Bridges and the control circuits for example and other weird design decisions such as the "clever" power saving circuit or using optocouples as analogue delay lines to be somewhat compatible to Mach3 pulse timings make these boards a more or less bad choice.
Out-of-the-box they will work somehow and (strictly speaking) need no modification.
But you posted here, so you want to modify the board?
All of the blue-chinese boards I have seen already included the 7414 Schmitt Trigger, either on the board or on the "breakout board" that shipped with the board.
The diagram you posted is pretty straightforward: every Schmitt-Trigger has a pullup resistor on input and an output.
I have NO idea, where the input pullups may be necessary: PC parallel ports have push-pull output stages.
Modern LPT implementations use 3.3V supply and therefore only provide a 3.3V high level on output. If they are standards-compliant (and most are), they accept up to 5V on the inputs.
If you have such a port - a multimeter will quickly tell you, just measure a high output - and your board does not work with that (because it is designed for 5V levels), you want a level shifter.
The 7414 circuit that you posted may do the trick. Maybe even without hte pullup resistors.
Just connect the LPT outputs to the inputs of teh Schmitt Triggers and the trigger outputs to the inputs of the driver board.
If you lack documentation about the driver board - and most of us do, I have not seen any chinese supplied board to come with any documentation - then you will need to reverse-engineer the pinout (like most of us did :-).
Grab the datasheets for
- the driver chip
- the optocoupler
- the two SO ICs on your board
Find an input on the Toshiba chip, find the opticoupler its connected to, find the parport pin that the input of that optocoupler is connected to, write it down, repeat for the other relevant inputs.
Some boards I have come across have separate enable signals for the driver stages (which is essentially a good thing).
But LinuxCNC only has one driver enable output signal.
So I usually wire this output to three inputs of the 7414 (or actually to inputs of a 4049 which I usually use as level shifter and driver) and the three outputs are my three "separate" enable signals.
BTW: TTL technologies 101:
74xx: stone-age ancestor of the TTL series
74Sxx: S as ins Schottky: faster variant, even more power hungry
74LSxx: Low-Power-Schottky: approx. as fast as 74xx but much less power consuming
74ALSxx: Advanced low power Schottky: faster variant, slightly more power than LS
74HCxx: Highspeed CMOS reimplementation. speed approx between LS and ALS, *much* less power consumption but the minimum HIGH voltage on inputs is raised from 2.4V to 3V (if memory servers, no bashing please)
74HCTxx: T as in TTL-compatible: HC variant but accepting anything above 2.4V as "high", so fully TTL compatible. Excellent replacement for LS in most circuits
Does that matter to you?
Probably not: the signals we use in drivinf steppers are so slow, any TTL technology is plenty fast enough.
But always beware about driving power: not all TTL I are rated to drive >5mA as is required for driving the LED of an optocoupler. The HC14 may be able to drive that much.
But you *should* grab the datasheet yourself and check the absolute maximum ratings for the chip.
Just to be sure and just to learn a little today :-)
All set?
Just hook up the board and post your results.
Hartmut "hase" Semken, [email protected]
As your board looks different from the old ones I suspect you must read many pages of this forum to verify it is the same design on a different PCB. I wiil suggest starting from page 55 and read #656 by Neon22 before going any further to try and workout the connections.
People made different approaches to the subject as in #91, #204, #283 but you have to verify your board is designed in the same principle as the earlier ones.
Petros
Hi harmut,
Thanks for your reply,
Let me try to explain myself.
I had the old blue chinnese board working for about 1 week until exploted. (no idea why this happened).
so i bought two of these again.(one for replacement)
After reinstalled and being working for about 5 minutes it exploted again. (this time all 3 TB6560 chip exploted)
Before this happened you cuold hear a very loud motor noise, so i start wondering what was wrong....
Today a have a third board running, i have set current to 50%, and its been working fine, but the noise keeps on all of the time.... im waiting until it will explode again....
But in the meantime the chinese seller offered a new one for free, so i would like to try modifying this one.
Thanks for your ideas i will try then i'll get back with the results...
Thanks Petsat,
i'll take a look to those pages!
Durango I would suggest the following ESSENTIAL MODIFICATIONS LIST:
1. NEVER power up the unit without motors connected to the outputs and do not unplug a motor when the board is powered up. The poor grounding (note2) may destroy the 6560 chips.
2. Ground the heat sink for each TB6560 through its mounting screws. In the current design, there is a danger of blowing all chips if one goes bad. Adding the grounds helps protect them all.
3. Disconnect the current reduction feature. This generates extra pulses and is not useful.
4. Pin 7 on driver chip has a cap that sets the internal frequency. Replace the 1000pf capacitor which is installed with a smaller value capacitor (150 to 330 pF).
5. Clean up the pulse to Clock (Step) and Direction signals. The pulse coming from the computer through the optocouplers is very slow on rise time (because of the optocouplers). The chip misinterprets the signals and misses steps. Bypass optocouplers.
6. The way the ICs are getting power violates the manufacturers guidelines for power up and power down sequence. Do not use a single switch on the power line coming from the power supply. Either power the ICs on/off by a separate power supply or use a dedicated powerup/down circuit to deal with the sequence matter.