[H500]

The max speed depends largely on the power supply voltage unless your system happens to suffer from resonance. I'm not using one of these boards, but for comparison, with a 34v supply and 280 oz-in steppers on a SX3 mill, with preloaded ball screws, I can reliably move at 150 ipm on the x and y axes and 60ipm on the Z. I plan to use 45v, so in theory I should be able to do 200 ipm unless I need to tighten up the gibs for more rigidity.

[texaspyro]

I am using them on a MAXNC15 pile 'o crap mill. It has 1/4-20 acme lead screws and I can drive it at a whopping 12 (count 'em - 12) inches per minute with the stock steppers on 28V. With some other steppers (half the size of the stock steppers) I get 24 IPM. I am going to try some real steppers next...

[lodberg]

I'm using the TB6560 and I'm only able achieve 700-800mm/min (27.5-31.5ipm). My settings are: current 25% (off,off), fast decay (on,on), 1/8 micro step (off,off), 24V 15A power supply, 5mm Ball screw. I have tried with 28V, but that don't give any significant improvement.

Could it be the stepper motors I'm using that is preventing more speed? I'm using them in unipolar configurstion. However, I don't have the datasheet of the steppers: Nema 23, 1.8°/step
Type: 23LM-C005-29V
MINEBEA CO. LTD
Made In Thailand

I hope you can help me.

[doorknob]

I'm using the TB6560 and I'm only able achieve 700-800mm/min (27.5-31.5ipm). My settings are: current 25% (off,off), fast decay (on,on), 1/8 micro step (off,off), 24V 15A power supply, 5mm Ball screw. I have tried with 28V, but that don't give any significant improvement.

Could it be the stepper motors I'm using that is preventing more speed? I'm using them in unipolar configurstion. However, I don't have the datasheet of the steppers: Nema 23, 1.8°/step
Type: 23LM-C005-29V
MINEBEA CO. LTD
Made In Thailand

I hope you can help me.

I don't have any information about those stepper motors. Do they have six wires?

I'm not sure what you mean by "I'm using them in unipolar configuration" because the TB6560 is designed to drive motors using a bipolar configuration. So you must be using only four of the motor's six wires. That means that you must choose which of the wires to hook up, and which ones to leave disconnected. If the motor has a high winding inductance, then you may get improved speed if you choose a "half-winding" hookup instead of a "full-winding" hookup.

Reportedly, some of the TB6560 boards are miswired so that when they are set for 1/8 microstepping they are instead using 1/16 microstepping (and vice versa) - if that is the case, does it improve the speed if you change to "off,on" or have you already taken that into account? Are you using Mach3? What is your setting for steps per unit (mm/inch)? Have you verified that your axes move the expected distance for a specific number of steps?

Have you tried the 50% current setting? Does it make any difference in the operation of the motor?

[H500]

Check your gibs. They should not be binding. With the screws removed, I can move my 90lb table if I push hard enough. It slides with a solid but smooth motion.

[lodberg]

Thanks a lot for your kind help

@doorknob
Yes, my stepper motor have 6 wires and I’m only using 4 of them in a unipolar configuration (sorry it might be that it is called bipolar half-coil / single winding?) Maybe I should check the wires ones again and try to change the wiring. But I’m sure that using both the coils in bipolar serial connection will not give me more speed or should I give it a try and increase the voltage to 28-30V?

I understand that some of the TB6560 boards are swapping 1/16 and 1/8 micro stepping, but since I have the “steps per unit” set to 320 steps/mm – it must be correct with 1/8 at (off, off). I have also tried to give 50% current, but without any luck. The stepper motors get a little hot, but that is about it – no more speed.

I’m using Mach3. I have tried to use the auto calibration of x,y,z. This gives 320 steps/mm. By the way I don’t have any delay in steps and dir puls and I think that I’m using 25kHz kernel speed.

@H500
What are gibs? J My support rails... ? Currently my CNC machine is made of 20mm (0.787inch) MDF and is therefore a bit heavy (I know that is good for stability.. vibrations). I'm working on a “version 2” made of 15mm (0.59inch) plywood and it will less heavy than my MDF CNC machine. This should of course give more speed, but I’m not sure how much it will help.

[doorknob]

@doorknob
Yes, my stepper motor have 6 wires and I’m only using 4 of them in a unipolar configuration (sorry it might be that it is called bipolar half-coil / single winding?) Maybe I should check the wires ones again and try to change the wiring. But I’m sure that using both the coils in bipolar serial connection will not give me more speed or should I give it a try and increase the voltage to 28-30V?

I understand that some of the TB6560 boards are swapping 1/16 and 1/8 micro stepping, but since I have the “steps per unit” set to 320 steps/mm – it must be correct with 1/8 at (off, off). I have also tried to give 50% current, but without any luck. The stepper motors get a little hot, but that is about it – no more speed.

OK, if by "unipolar configuration" you meant "half coil configuration" then I would not change that, other than double-checking that you are actually using half of each winding instead of the full windings. No, you do not want to use the bipolar series hookup.

If your steps per unit gives you the axis travel distance that you expected, then it would seem that your 1/8 DIP switch setting is correct.

So, maybe your motors are simply a poor match for the TB6560 and its maximum voltage rating. Or maybe your Mach3 motor tuning settings can be improved.

[lodberg]

Thanks a lot for the very fast reply! I will check the wires again and the Mach3 settings. If I can't find anything I guess I need to find better stepper motors. Which stepper motors (Nema23) do you recommend as a perfect match for the TB6560?

[Larken]

To get the most speed, you should raise the PS voltage to the max 36 volts.

Saying your drive has a max current of 3 amps, you will get more torque at high rpm if you were to run 5 amp/phase rated stepper @3 amps then a 3amp/phase rated stepper @ 3 amps. (You would get less low speed power but more highspeed power. )

But running under 30 volts, high speed is a term that doesn't apply !

If speed is what you want, you'l never get it from that drive. You would be better to keep you steppers and dump the drive for at least a 60 volt drive and PS.

[lodberg]

I'm sure going much over 32V will destroy the TB6560 chips according to previous replies in this thread.

I'm just trying to figure out how I can use the TB6560 board at best configuration. I have found the following on YouTube where they are using the board at a very nice speed (1800-3360mm/min):
[ame=http://www.youtube.com/watch?v=WML5mECPGGs]CNC TB6560 3axis_RED. cncVulcano - YouTube[/ame]
[ame=http://www.youtube.com/watch?v=Ym8X82npqw8]Home made CNC router DVD Laser Fixed tb6560 driver MOD - YouTube[/ame]

Btw I have been searching for any information on the steppers I'm using: 23LM-C005-29, 1,8°/step, 0.44A/Phase. Therefore, I might be looking for other step motors - any good suggestions....? I mean some of you are using the board with a stepper that gives you good speed (above 1000mm/min) ?? Thanks a lot.

[texaspyro]

the steppers I'm using: 23LM-C005-29, 1,8°/step, 0.44A/Phase.

0.44A/phase is VERY low (like suitable for toys). You are probably not going to get much power/speed out of those. Most serious (but small) steppers are more likely in the 2.5A-3.5A range.

[Marc N Fournier]

Well I did it again, sent the BBFH TB6560 chips up in smoke and sparks. This time I know exactly what I did wrong.

I was lowering the 5V source to see if this might be the cause of an issue as I was reading 6V on the high end once in a while.

Bottom line, I went down too low and the Z immediately went up in smoke followed by X,Y,A. My error: I had NOT un-set the "enable" pins.

---

As I said previously, I removed both the 12V and the 5V regulator and wired up connectors to supply the 12V and 5V from the PC power supply this was all going to reside inside of.

During my testing this weekend I had a 12V and 5V adjustable power source setup plus the motor power supply. So everyone was on there own power source.

Mach3.

Nema 23 motors from Keling 270 Oz/in, 3.2V/2.8A, 3.6mH on X and Y.
Nema 23 motor from Keling 495 Oz/in, 5.1V/3A, 7mH for the Z.

I had X and Y on half step at 2000 steps/ps, velocity of 150, acc at 10 on both.

Z was full step, 1000 step/per sec, 85 velocity, 25 acc.

Card switch settings are
1/2= OFF, 3/4= ON, 5=ON, 6=OFF for XYA.
1/2=OFF, 3/4=OFF, 5=ON, 6=ON for Z.

So that comes out to 25% current for all, full decay for all, Half step for XYA and Full step for Z.

The card was not missing steps at a feed rate of 60. It all ran the Roadrunner test flawlessly, not missing steps or stalling. Did this several times over the course of the week end, setting up the motors at a known position to start and checking the position at end of run for any changes. No missed steps.

Pushing the feedrate higher (85) would stall or drop steps. So I was scoping trying to find the source of the glitch.

----

For those encountering smoked TB6560 chips, from my experience.

On un-modified boards running the 12V and 5V regulators off the motor power supply, anything that affects the power supply, 12V and 5V regulator can/will cause the TB6560 chips to smoke while enabled.

Anything that causes the 12V regulator to shut off momentarely will blow the TB6560 chips while enabled.

The Un-protected spindle relay (no back-EMF diode) can cause the 12V regulator to hickup, that will also affect the 5V regulator and blow the TB6560 chips while enabled.

Not respecting the power up/power down sequence can/will blow the TB6560 chips.

---

I assume the card needs a fuse of an appropriate value to protect things. Since there is only the single power connection for the motor power supply, what value fuse is recommended ? One with the value of the highest Amp draw motor or onw with the lowest amp draw motor value ? A value matching the total AMP draw of the motors ?

---

Modification to this board were removing the regulators, replace the 1000 pFcap with a 100pF cap and by-passing the Optos. But not the spindle relay opto, trying to by-pass that one will cause it to not work, period, as it MUST go to ground.

Someone had asked how to get the spindle relay to work. Found out. In Mach3 you set the ports and pins; output. Output #1 for the spindle relay output. In the case of the 4 axis card you set output #1 for port 1, pin 2.

Then again in ports and pins, "spindle Setup", remove the tick mark from Relay control, change the clockwise (M3) Output # to a "1" so this matches the output #1 setting we did earlier. Now the relay will activate at the start of a run.

---

So I have to order 4 more TB6560 chips from China.

Until my own mistake, the card was running very well for my use. Bettre than tha it ever had.

Marc

[doorknob]

Well I did it again, sent the BBFH TB6560 chips up in smoke and sparks. This time I know exactly what I did wrong.

While I'm sorry to hear about your smoke, your experiments are quite interesting and they seem to underscore the concern about proper power sequencing being required to keep the chip healthy.

I'm not sure that a fuse is the way to go, though. A TB6560 will probably blow more quickly than a fuse will blow (and even if the fuse did blow first, the action of abruptly cutting off the power to the motor might cause an inductive voltage spike that would blow the chip).

[H500]

What are gibs? J My support rails... ? Currently my CNC machine is made of 20mm (0.787inch) MDF and is therefore a bit heavy (I know that is good for stability.. vibrations). I'm working on a “version 2” made of 15mm (0.59inch) plywood and it will less heavy than my MDF CNC machine. This should of course give more speed, but I’m not sure how much it will help.

The gibs are part of the slide mechanism on a dovetail type machine. Since your machine is made of wood, it doesn't have any.

Are you using threaded rods or ballscrews? Ballscrews will give you 2 to 4 times the performance.

[HankMcSpank]

Bought one of these about a year ago - & wouldn't you know it.... pretty much within 2 hours of first useage the Y axis TB6560 IC blew ....I then put the board to one side, where it has been ever since. Just pressed it into play (after instaling a replacement TB6560 IC on the Y axis)...& had no end of troubles with missing steps & lumpy jogging

My biggest problem (to date!) was that the DIL switch is junk ...ie whilst my Y axis DIL switches we visually set - they weren't electrically! This is because the switch's plastic surround was slightly loose & therefore the white selectors had nothing to push downward against 9to make the bras contacts touch) ...this meant that while I thought my current was set to 100% it was actually only set for 25%!! (which would account for why my Y axis was running dreadfully). The solution - in my case - was to hard wire the bottom of the DIL switch with little jumer wire for the premanent setting I want (I don't think most people would change them anyway - ie once they're happy with their own setup). I've set mine for 100% current, fast decay and 1/8 microstepping.

While I was on, I changed out the Osc 1000pf cap for a 100pf cap on all axises - aah bliss, that awful sound has disappeared.

Tomrrow I intend 'jumping' the opto isolators (pin 2 to pin 4)....I don't have any buffers & I don't intend throwing any more money into this POS.

I've read that many have also disabled some clever current feedback circuit (which can't be that clever, as apparently it doesn't work well), but I can't actually find the mod - can anyone link me to it please?

[doorknob]

I've read that many have also disabled some clever current feedback circuit (which can't be that clever, as apparently it doesn't work well), but I can't actually find the mod - can anyone link me to it please?

I believe that the mod involves cutting the trace (where it says "cut" shown in red) between the transistor base and the 10K resistor shown in this diagram (from post #19 of this thread):

http://www.cnczone.com/forums/attach...5&d=1283911876

[Marc N Fournier]

Blown my 3rd JPG to about 200% and you will see the red ovals. You remove a resistor and a transistor. Resistor is silkscreened 4.7K. Transistor is further, on same trace.

There is no need to cut any traces if you are by passing the optos using the PC817 sockets as you are not going to install the third 78hc17 chip everyone is fitting. You will be getting the same results with out it.

You can leave or remove the "ENABLE" optos as they have no effect on performance but are very usefull for chip life (power on/off sequence). The important ones are the step and direction.

Leave the opto for the spindle relay as is otherwise it will not work, if you intend on using it. it is the last one in the series.

Since the optos are totally useless in this particular inception, you can even get rid of the all EXCEPT the last one (spindle relay)


Marc N Fournier

[Marc N Fournier]

[QUOTE=HankMcSpank;1066532]

Tomrrow I intend 'jumping' the opto isolators (pin 2 to pin 4)....I don't have any buffers & I don't intend throwing any more money into this POS.

QUOTE]

Heads up here. It is not 2 to 4 in all cases... Step and direction should be 2 to 4 effectively. If you plan to remove the enable optos then it is 2 to 3 for the jumper.

I used cut down staples to short pin to pin. How is that for mizerly.

Wonder why you would run at 100%. If your steppers are rate to close to the limit of the card, does no seem like a good idea to me. Correct me if I am wrong but current here will not give you speed. Unless you are pulling one heck of a load 2A wold be good for anything.

The G540 is only 3.5A but can accept 50V. That si were the speed will come from, voltage. So do not expect to get fast speeds out of the 34V limit.


Marc N Fournier

[sr43]

What size did you use for the decoupling cap
thanks

[HankMcSpank]

Oh dear...my X axis isn't working now & it looks like it's the TB6560 IC itself (pin 3 - the step pulse/clock input into the TB6560 appears to be loading down the pulse as supplied by the opto...when i desolder/lift pin 3 on the TB6560, the opto output signal is fine....but not there when i solder the TB6560's pin 3 back to the board)...of course nobody sells these ICs in the UK, which means a 14 day wait on a slow boat from China!

Seriously...these boards suck.