How I fixed my Chinese TB6560 controller (updated)

Ok, started mods on my spare board, first removed the auto current reduction circuit, found that the 50% decay mode seems to have improved the issues around mid-band resonance, but the steppers are noticably noisier ( hiss ), this is probably due to the low current switching frequency, looking at the TB6560 timing diagrams the internal osc looks like it needs to be substantially higher than the maximium step frequency, should also eradicate the audio hiss, I'm looking at using 330pf to up the frequency to about 130kHz, back later.
.
.
.
Later.
====

Ok, tried that, much much quieter or at least I can't hear it ( cats may be able to though ), I'll setup two dial gauges tomorrow and try running a long program ( 1 hr ) just to make sure that I am not loosing any steps, I have not yet tried the 7414 buffer fix yet to square up the step signal, will only do this if needed to correct for missing steps.

Quote:

---

Originally Posted by davidmb

Ok, started mods on my spare board, first removed the auto current reduction circuit, found that the 50% decay mode seems to have improved the issues around mid-band resonance, but the steppers are noticably noisier ( hiss ), this is probably due to the low current switching frequency, looking at the TB6560 timing diagrams the internal osc looks like it needs to be substantially higher than the maximium step frequency, should also eradicate the audio hiss, I'm looking at using 330pf to up the frequency to about 130kHz, back later.
.
.
.
Later.
====

Ok, tried that, much much quieter or at least I can't hear it ( cats may be able to though ), I'll setup two dial gauges tomorrow and try running a long program ( 1 hr ) just to make sure that I am not loosing any steps, I have not yet tried the 7414 buffer fix yet to square up the step signal, will only do this if needed to correct for missing steps.

---

What was original value for the capacitor...100pf?
I'm also having audible hissing noise. Thinking to try your 330pf fix.

Hi, the original capacitor was 1000pf, putting the chopper frequency into the 40kHz region, I still have some 100pF caps to try, may get around to it at the weekend. Be aware that not all the noise goes away but it is very much attenuated, or mostly out of hearing range.

Quote:

---

Originally Posted by davidmb

Hi, the original capacitor was 1000pf, putting the chopper frequency into the 40kHz region, I still have some 100pF caps to try, may get around to it at the weekend. Be aware that not all the noise goes away but it is very much attenuated, or mostly out of hearing range.

---

Hi...do not forget to share your findings.

Thanx

Ok, ran a 1 hr test last night, the system was loosing steps, so after looking at the board layout, I have removed the three optocouplers that are used for the clock/step signals these have now been wired so that the output of the onboard 7414 drives directly into the clock pin of the TB6560, re-running the test right now, it's not clear from the data sheet about the rise time of the clock signal, but it does indicate that the chip uses the positive/ rising edge of the clock signal.
---------
Ok, test re-ran without loosing any steps, now running a more advanced test 3hrs, max speed is currently 2m/min ( 78ipm ) with 4mm acme leadscrews ( so hitting 500 rpm on steppers without any problems ), will report back later.
---------
Next night. Just finished running a 12,600 line of gcode feed range 500 mm/h to 1100 mm/h rapids at 2036mm/h, runtime of approx 3 hours, no steps lost.

1 I did not bypass the opto's, just swapped the legs around so the correct side was down to ground.
2 changed the caps on pin7 and Freq changed from 44 to 150. Decided to order the 100pF for them now.
3 Also added in the 7414 but am not sure If I should keep this right now

Edit: 1 was done based on post one

Quote:

---

Driver IC ENABLE input improperly driven: The opto-coupler for the ENABLE input is wired incorrectly having the npn emitter connected to supply. Although the circuit does function, the improper wiring of the npn will cause very low transistor gain opening the possibility of system noise corrupting the ENABLE signal. Solution: Rewire the npn of the opto-coupler with collector to supply and emitter to output.

---

2 was done basted on the post just above and the data sheet

Quote:

---

PWM Oscillator Frequency (External Condenser Setting)
An external condenser connected to the OSC pin is used to internally generate a saw tooth waveform. PWM is controlled using this frequency. Toshiba recommends 100 to 3300 pF for the capacitance, taking variations between ICs into consideration. Approximation: fosc = 1/(Cosc × 1.5 × (10/Cosc + 1)/66) × 1000 kHz

---

3 was done based on post one and readings I obtained myself with my board and scope

Heat sink was also Grounded, and a separate supply used for the Fan (added a 7812 for the fan)

Groot

Hi,

looking at the specs for the PC817 opto couplers on my board, datasheet gives response times as 18uS, so you probably need to keep the 7414 buffer after the opto to square up the signal, I have some high speed 10MHz opto isolators with scmitt logic outputs, but that would be like building a new board, perhaps this should be considered, not sure what best price is for TB6560 devices, biggest negative on these devices as I can see is the low motor voltage 34v is just not high enough for very high performance stepper motors.

For My application 34V is already almost to hight (I actually run on 28V) and are currently happy with speed and performance I get.
I use this for small work and with the current set-up I can do 1m x 1m by 300mm and this can be expanded.

After the above mods it runs 90%better than originally, however I am still going to replace those caps with 100pF's

For me this was based on Budget 1st, In hind sight I could have bought something nicer, but then I was going to miss out on all the FUN fixing this

I am still going to try the other kits of James Newton and compare it to this one.

Like the 1st post here I will still list and show all my final changes and post a few pictures.

Groot

I have abandoned this project.
It took too much of my time!
Gone to a proper Breakout Board and Stepper drivers (Chinese too).
The Emco5 Steppers are now working as should be, no more loss of steps.

However, using the TB6560 board gave me a lot of insight of Mach3.

Jos Raven, Holland

Hi just finished running a 4 hour gcode file, no problems except the pc shutdown during the roughing algorithm ( I turned on the big vac 2kW to clean up and the pc shut down - argh!! ), reset and restarted the code so total cut time is somewhere in the region of 5.5 hours, no issues with the driver at all, I will now adapt my original board so they are both the same.

So these boards whilst not being a 'Gecko' are cheap and relatively easy to make work properly, at least mine have been.;)

This is what I have done.
1) Disconnect current save feature ( remove three 10k resistors )
2) Replace oscilator caps with 330pf
3) bypass or replace the three opto isolators that pass the step signals

my Board in the end.
Fan runs on its own 7812
Main Heat-sink is grounded.
Caps on pin 7 replaced by 100PF
7414 added
Pin 25 trimmed back
All changes are based on tests and information from this thread and the Data-sheet of the TB6560

Some soldering needs still to be cleaned up, but all in all it works for me now. Next I will try a home-kit based also on the tb6560

Groot

Hi,

Any news on the effect of 100pf caps for the on-board oscilator, from the datasheets it means that we should be able to use a 10uS step signal (20uS per step ), which should get the max step rate up to 50khz may be able to use the higher microstep settings with this and still maintain reasonable speeds.

has anyone done a dummys quide to modding this board.
i have just finished building my cnc machine with one of these boards.
it misses a few steps now and again. i thought it was something i had done but after reading this post i think it may be down to the board.
if their is a step by step quide could someone point me in the right direction.
many thanks

Look at my post #91 (I see the text is bad cause of size limit, I might redo the picture)it is more or less step by step, but before you do anything, lets just Check all your settings....
Please tell me what software you are running, and then what is the setting you have it at. Lets work from there, but I am also fairly sure following what I have done will fix most, if not all problems.
Also remember this is cheap and from china ....

thanks GrootWitBaas for your reply.
im using mach 3 on a desktop running windows 7.
the dip settings for all three axes are
1: off
2: on
3: off
4: 0n
5: on
6: on
the dr pulse and step pulse are all set to 5
is their any thing else you need to know.
many thanks.

Yes I need to know what is your "Kernel Speed" set at.
Also try turning Dip 3 to on.
How is your soldering skills, and have you ordered some 100pF caps already ?

Edit: also Motor Size and PSU size/type, just to confirm, but that should be ok :D

Stepper Motors are 32 57BYGH78-401A 18.9Kgf.cm
power supply 24 volt it came with the kit.
kernel speed is 35000hz
my soldering skills should be ok
the 100pf capasitors do they need to be electrolytic caps, also what voltage do they need to be.

sorry the motors arnt nema 32 they are nema 23

Quote:

---

Originally Posted by mark.a

the 100pf capasitors do they need to be electrolytic caps, also what voltage do they need to be.

---

100 pf is a very small value of capacitance. You should not be looking for an electrolytic capacitor with that value.

You will probably easily find ceramic disc or monolithic capacitors of that value with a voltage rating of 50 volts (but you'll probably be fine with an even lower voltage rating).

For example, see DC100: VARIOUS: Passive Components

(I'm sure that you can find a nearby UK supplier as well)

I haven't had time to play with my TB6560 board yet, and so I don't know whether or not the layout will permit you to use tiny SMD surface mount "chip" capacitors, but if so, they would be a good choice in this application (but would be more difficult to solder into place).

The caps you need is smd 100Pf, I have send you a pm with the link where I got mine in Germany.
once they are changed then "kernel speed is 35000hz" can be changed up to the next level (or even more) and this should most probably stop missed steps. Hissing of the steppers is also reduced by this change. 330pF can also be used, but I went with 100pF. Looking in the store doorknob linked this will do (you are replacing smd caps with smd caps)
Personally I use EMC2 not Mach3 (I don't do Windowzzz) but I know from feedback from other members that this does help.

O and I am known to be wrong at times :D