Haoyu TB6600

I recently acquired a Haoyu TB6600 drive and thought some might find the following write-up of interest.

Looks can be deceiving but this little drive does get a number of things right.

For a start the board is well assembled, only its hand soldered connectors and opto isolators are covered in messy flux residue.

Unlike the TB6560 boards, this implementation gets the opto isolation right...well almost.
The step and dir pins use fast 6N137 opto isolators with a slow PC817 on the enable pin.
The error is that the drive defaults to always enabled at power-up. A current must pass through
the enable opto isolator to shut the drive off. I'd rather the reverse situation apply.
There are two spare inverters on the board and one could be used to correct the phase of the enable signal.

The motor step occurs at the very instance when current is made to stop flowing in the step opto isiolator.

The tab of the TB6600 is isolated from 0V allowing the chip/heatsink to be bolted directly onto the case chassis or
more TB6600 attached to the same heatsink without fear of motor currents flowing in the chassis.

The connectors are the weakest link IMO. I'd much rather a simple screw terminal than incur the extra contact surfaces.
The weakest is the 2 pin power connection. I would not wiggle the 4 pin motor connection when operating, for that is
sure to destroy the drive.

There are very short and direct PCB traces from the drive to the A+/A- and B+/B- motor phases made possible
by the engineers at Toshiba adopting an easier to route device pinout.

The sense resistors do not look to be the low inductive current sensing type but there are 3 of them per phase.
What ever the value, the total inductance it is at least a third that of any one resistor.

I very much doubt if the bulk capacitance on board is the low esr type. This could be a point
of failure when the electrolytic dries out, but is easily rectified by replacing.

There's no decoupling caps close to the +Vm pins of the TB6600 but it only took a few minutes to get some fitted.

When mounting the heatsink, one of the holes is blocked by the connector.

The PCB mounting holes are symmetrical and the board can be rotated 180 deg to make air flow/wiring easier.

The switching frequency was measured to be 40kHz.

In the absence of step pulses the phase current is reduced.

I kept motor wiring neat and tidy, and took full advantage of the opto isolation to keep motor currents isolated from the chassis and
opto drive signals. There was 7ft of 16AWG/4 between the drive and motor and 10ft of CAT5 cable from the
drive to the BOB, the drive was set for half stepping, the motor was in-situ on the mill and operating at 2.5A, +Vm was 27V.
There was no missing steps nor any motor hissing or squeals. The motor had plenty of torque.

Given the choice I much rather this drive over ANY TB6560 design.

Congrats, very nice and detailed evaluation.
I did look at the schematic but didn't notice the lack of decoupling caps on the VM supply. (chair)

Could you please check one more signal if possible, apparently you have the equipment and the knowledge for it :) ?
How long does it take for the TQ input to drop low after a step pulse is received?

I think that it will be short and will influence torque at low speed and this could create problems during acceleration.
Some machine setups might suffer from this, others might not.

Glad you asked. I'll measure some actual TQ values.

I'm not a big fan of modulating the TQ in this way. I'd rather 100% current all the time. On a ball screw setup the possibility
of back driving the motors is high, it might be Ok on a gib setup with more friction.
On the mill with TQ current reduction active, there was a 1/50 division (i.e. 360/50) that was easily twistable by hand.

Totally agree re: acceleration. Control software setting of acceleration could slow down the step pulse rate to let the drive recover
from this powered down state.

There is one little ceramic cap across the bulk capacitance but its 3 or more inches away from where it should be placed.
There's two +Vm pins on the TB6600 and this was the only possibility of getting a pair of 150nF caps on top of the Vm rails.

Some pictures of the Haoyu TB6600 board.
The underside of the PCB is very messy and its worth while taking a look for solder splashes before first powering up.

C8 & C7 provide the decoupling for the motor supply.
The trouble is, C7 is far from the power pins of the TB6600 to have any meaningful impact at high frequencies.

The last pic shows my mod to get two 0.15uF caps directly across the A and B phase power supply pins.

On these drives the phase current is reduced to 30% when the step pulses slow to a halt.
This is was problematic on my machine as the holding torque in that state allowed a 1.8deg movement in the motor coupling.

There's more than one way to disable this 'feature' but I ended up lifting pin 9 and attaching it to ground pin 7 -> see pic.
Once tweaked, the drive supplies 100% of its target set current.

Quote:

---

Originally Posted by bobgrant

On these drives the phase current is reduced to 30% when the step pulses slow to a halt.
This is was problematic on my machine as the holding torque in that state allowed a 1.8deg movement in the motor coupling.

There's more than one way to disable this 'feature' but I ended up lifting pin 9 and attaching it to ground pin 7 -> see pic.
Once tweaked, the drive supplies 100% of its target set current.

---

Bob,

Thank you for these very useful posts.

Out of curiosity, were you having the same problem with the torque being reduced to 30% almost instantly, after receiving the step pulse, as was identified by lucas in the HY-DIV268N-5A driver board?

The reason I ask is that a friend and I tried emulating the torque reduction circuit performance of the HY-DIV268N-5A driver, in Spice. With the present values, Spice predicted the same performance as lucas noted in his scope photos. We are going to revisit the situation with a larger cap at C2, in hopes that this will give us an acceptably longer period of time before current reduction kicks in, thus fixing the situation of reduced idle current kicking in too soon.

Since the torque reduction circuit in the HY-DIV268N-5A driver appears to be very similar to the one in the Haoyu driver, our work may be of interest to you.

Best,

-gnu_B

No, no such problems here.

Check with the schematics, if you use an active low rather than active high (check polarity)
clock signal the effect of the torque control can be rendered ineffective "IF" the clock pulse width is short (tens of microseconds or less).

FWIW, for $22, the Haoyu drives are quiet and create textbook micro-step current waveforms.

Bob Grant,

Do you have any video of the stepper drives running your machine?

How do you like them after using them for a few months?

Where did you purchase them?

Your review and modification instructions are awesome!

Schmeckola

Mistake, the post above is the correct one! This one can be deleted.

I am new here, and my native language is German. I am a electronic engineer and i have all the equipment to make proper measurements and tests.

I also had intensive look at my Haoyu TB6600.

The defiantly have some Problem but the drivers are working without major problems.
My stepper motors are 4 Wantai Stepper Motor Model: 57BYGH633B 3.0A
Power supply is a 25 Volt 15 Amp Switching power supply.
Breakoutboard is:
3rd Gen 5Axis CNC Router Breakout Board Set Display/Control Panel Gcode Recorded
3rd Gen 5AXIS CNC Router Breakout Board Set Display Control Panel Gcode Recorded | eBay
The Haoyu TB660.
TB6600 1 Axis Stepper Motor 2 Phase Drive Board for CNC Router Beter with TB6560 | eBay

1. The Haoyu TB660 board is constructed with torque reduction to 30% if clock signal is in low state without pulses, once pulses are present the board switches to 100% of torque set up by the trim potentiometer R7. The actual low limit of switching between 100% and 30% Power consumption limit is at 1% of maximum speed with approximately 50Hz Clock frequency on pin CLK. Torque is still strong enough at this low speed even the current is limited to the lower 30% current level. I use 1/8 Step mode. I used mach3 to connect via a breakout board to the driver. The width and frequency of the pulses is sufficient that the TQ pin is set correctly when pules are present. My maximum feed rate is 4600mm per min on X and Y.
2. To have 100% Torque all the time even if the CLK Frequency is below 50Hz one can invert the CLK pulses in mach3.
3. Torque adjusting: Turning the current pot lets easily adjust the torque to the maximum possible, the maximum is reached when the WARN led gets on. The average current measured is well below 1 Ampere with 25V supply. The TB 6600 is suitably only for 28 Watts not more see datasheet, and even gets lower if the environment temperature gets higher as 25° Celsius with heat sink ( 15W at 80° with heat sink). Peek current of 5 Ampere is only for a very shot time 100 msec so for periods when the motors speed up or reduce speed the current gets to this limits. This limits are nearly reached as soon as the WARN Led is on. Corrections about this peek current is only possible by changing the tree 2 x 3 parallel 0,680 Ohm resistors = 0,2267 Ohm. Datsheet: 0,2 Ohm for maximum current. Impediance of this Risistors is not a problem.
4. The position of the pot depends a lot oft the type of motor and the mass that has to be moved.
5. The absolute peak maximum of power supply voltage for the TB6600 is 50 Volts but this does not mean that one can supply 45V because the EMK produced by the Motor cols is much more as the Voltage supplied, peak EMK is easily doubled Supply voltage, the power supply in use must be able to coup with this circumstances as well since there are no additional Voltage limiter diodes on the board. Be aware the TB6600 withstand only 50 Volts even it has internal diodes on the output.

So far all is usable.

1. Not correct is that the internal Vreg voltage is loaded with to much load so the 5V Vreg is actually only 4 Volts.
2. To correct this i removed the LEDSs for PWR and RUN. And changed the serial resistor for WARN LED to 20k value. Maximum sink current on WARN pin is 1 mA not 5 mA. And i added an additional Tantal Capacitor for the Vreg with 47uF. Resulting in a stable 4,8 Volts without spikes on Vreg. The Datasheet explains that one can connect the WARN LED via a 33k resistor to Vreg, but there is no mentioning that one can use this Vreg voltage for more as 2 mAmp. So using this pin for the HC14 and for the tree 1k resistors to the opto couplers is surly beyond the limits. But there are no detail information how much current can be realy drawn from this Vreg pin. One also could use ab external 5 V supply by connecting all 5V connections to the external 5V. Then the Vreg pin only has a 100nF capasitor to GND.
3. The datasheet also states that the board is always save in all condition even when the state of the input pins are in a unpredictable state.
4. There is also no problem with the ENABLE High on power on. The chip protection is always active.
5. Still if possible it would be best to switch on the PC with the Mach3 or other software including the breakout board and switch to DISABLE (emergency stop) before powering the TB6600 on or off. (Preferable would be that all mode switches would be connected to the software via break-outboard, so software could control standby turning on and of the TB6600 and also the micro-stepping division could be controlled via software.)
6. I also added the two extra 100nF to VCCA as shown here in the forum (did not make a difference, but sure the are better as without more and better decoupling is always worth thinking about.)
7. Of major importance is a neat short wiring between break-outboard and TB6600 Flat cable with 6 wires every second wire is GND. Shelded twested pair also works well.
8. A short wire linke between GND brakoutboard and GND 24V powerconnection of the board is also needed to reduce spikes. All 4 driverboards and the breakoutboard are mounted on a aluminium metal farme. Resulting in absolutely silent motors in standby or DISABLED (emergency stop) and quieter running at all speeds. Motors and mechanic still has some resonance peeks at some speeds.
9. Link to a schematic from this forum with minor errors: (on cap is labelled wrong C3 is the cap on Vreg not C, RESET is connected to Vreg, some more 100nF Capacitors near the active components are on the board.) http://www.cnczone.com/forums/attach...0&d=1372455055
10. Datascheet link: http://www.haoyuelectronics.com/Atta..._Datasheet.pdf

So i just got to this thread today again and i just want to mention that i use the drivers up to this day, no problem at all. I also use the machine quite a lot.