[kt88]

1. I will contact the seller and ask. Plan B would be to simply solder everything myself (I just don't really have the time).
2. About the linear PS, it seems to make sense actually. I guess it could easily be done in the SMPS's as well - simply add a significant delay (say an LPF with 10mS or so time constant) to the reference voltage of the chip. If the controller has a build in reference generator, the voltage divider at the output can be made to operate in a way that it's first set to ~7V (to allow the 5V voltage to build up), and than ramp up to 24V. It should be quite easy (one extra resistor + relay + LPF).
3. Obviously I'll put the caps as close as possible to the driver board, or it would simply have no benefit

[kt88]

I went a head and drew the schematic for a simple start up circuit. The time constant of the LPF at the output of the 7805 will determine the delay at start-up, and the diode+capacitor in-front of the 7805 will enable it to stay alive a bit after the 24V supply is shut down.

It is used as a replacement for the on-board 5V regulator. The supply should be connected to the '+24V Supply' terminal, the '+5V driver' is connected to the 5V pad on the board (again, instead of the 5V regulator that's already there), and the '+24V Driver' is connected to the 24V terminal on the driver board.

The exact values of components will depend on the relay in use, and the amount of delay you'd like to get from this circuit. The 7805 will probably need a HS as it drops about 18-19V over it, so it'll depend on the current the driver will demand from the 5V rail. I didn't add the capacitors at the input and output, but obviously if you don't have these on the board already you should add them as well.

I haven't tested it yet, just drew it so I won't forget it for when I'll need it (a few weeks I guess, I don't have as much time to put into this project as I'd like). If anyone notices something wrong with it please let me know so I'll fix it.

[kt88]

Well seems like I have one more update today. I've built a quick proto-type of the circuit I posted earlier. If anyone will be interested in the values I've used I can post them, but they aren't of much importance. What is important are the results. I will say that I've inserted some dummy load for both rails.

Please find attached 2 images.
The first image shows what happens when the power is connected - this was done by physically connecting the cable after the supply is already steady at 24V (to mimic what would happen with a SMPS). As you can see, the 5V is powered up first, than ~220mS (which is exactly the time constant of the LPF) the relay clicks and 24V is allowed to go through.

The second image shows power down sequence. This time the 24V rail dies out completely ~150mS before the 5V rail starts dropping. This delay time will change depending on the load connected on the 5V rail - the heavier the load, the shorter the delay. However, C1 can be increased in value to compensate. I've used a 470uF cap for C1, and a 150ohm resistor as load for the 5V rail (~33mA of current draw). If you are using a higher supply voltage (like 28V or 30V) this delay time will be a bit longer.

[boldford]

Well seems like I have one more update today. I've built a quick proto-type of the circuit I posted earlier. If anyone will be interested in the values I've used I can post them, but they aren't of much importance. What is important are the results. I will say that I've inserted some dummy load for both rails.

Please find attached 2 images.
The first image shows what happens when the power is connected - this was done by physically connecting the cable after the supply is already steady at 24V (to mimic what would happen with a SMPS). As you can see, the 5V is powered up first, than ~220mS (which is exactly the time constant of the LPF) the relay clicks and 24V is allowed to go through.

The second image shows power down sequence. This time the 24V rail dies out completely ~150mS before the 5V rail starts dropping. This delay time will change depending on the load connected on the 5V rail - the heavier the load, the shorter the delay. However, C1 can be increased in value to compensate. I've used a 470uF cap for C1, and a 150ohm resistor as load for the 5V rail (~33mA of current draw). If you are using a higher supply voltage (like 28V or 30V) this delay time will be a bit longer.

That's exactly the sort of thing I had in mind.
I do think lucas's theory has more than a little merit and kt88's mod should overcome the issues IMHO.

[James Newton]

kt88, what happens when you shut off the power and K1 opens? Where does the back EMF from the magnetic field around the motor coil go?

While I do think it is possible to work with the TB6560 chip by being very careful about the power on sequence, and I do think it is /probable/ that Lucas idea will /decrease/ the chance of frying the chip, I really question if all the hassel is worth it. For the extra cost of that Relay, and the pain of dealing with the possibility of fried boards, why not just buy a Gecko and enjoy the magic?

Now... if there were another chip... a chip /like/ the TB6065, but which would NOT fry on power up without logic supply, and which actually manages more current... ,o) :cough: THB6064AH :cough:

[kt88]

A. I've noticed I didn't add the diode in parallel with the relay coil in the schematic I've published a few posts back. Obviously every relay should have that.
B. Gecko is really not an option. It would set me back well over 300$ (international shipping + tax included). This one would be around 60$ for 3 channels. For these who can afford it, there's really no alternative (at least from what I read in the forum), but for me it is way too expensive. BTW, I know that in most thread that are related to these cheap boards, the thread always gets onto the issue of "Gecko/other expensive driver is much better, just buy it". While I understand where its coming from, I would like to try and keep this thread on the issue of getting these cheap boards running good enough for a basic machine. I don't expect anything more out of it.
C. About the back EMF issue, do you have any suggestions? I would assume that since there's a large cap after the relay, the 24V supply won't drop instantly, and will allow some roll-off and will limit the back EMF issue. Obviously, I don't have much experience in this area so other ideas/suggestions are welcome.

[bobgrant]

I'd bought a 3 axis TB6560 board to test and used a scope to watch the current in one of the phases. The board was set in the finest micro step mode and I was expecting to see a voltage waveform resembling a rectified sine wave on the scope by examining the voltage across the sense resistor.

What resulted was definitely not sinusoidal, its was full of spikes and when calculated, the peak current was not even close to the desired level so the motor was definitely under powered.

[kt88]

Finally some updates. While some of the parts seem be on their way for over a month, the TB6560 driver board arrived

I've connected it, and set it to 1.2A - hooked up to the first stepper that I found in the drawer. It actually a 3.6A 2V stepper, but I've started at 1.2A for now.

First I've hooked it to a 555 for pulses, and a 24V lab PS just to make sure it works. I was glad to see it works fine (took me a minute to figure out the inputs are inverted, so EN should be GND for it to run). Next I've hooked a 1R metal-oxide resistor in series with one of the windings, and measured the current to see what it really is.
I was very pleased to see it works pretty good with the current constantly around 1.2A as it should be. Please see the attached screenshot from my scope for more details.

It seems to work well with somewhat higher currents as well, and with 1,1/2,1/8 and 1/16 stepping to. I'm yet to fully test it, this was just a quick test to see if I even works as it should.

BTW, I was surprised by the small size of the PCB, I was expecting about twice the size judging from the pictures in the eBay add.

[boldford]

The advert for this far eastern variant of these boards at TB6560 3A stepper motor drive single-axis stepper motor driver board controller | eBay seems to be a little more realistic on the capabilities of the TB6560.

[kt88]

Seems to be almost 100% identical to what I've ordered:
1PCS, CNC Router Single 1 Axis Controller Stepper Motor Drivers TB6560 3A | eBay
(except I've paid 13.8$ shipped, for some reason he pumped the price up by 3$)

BTW, I took out the 5V and inserted a current probe in series with the regulator and measured the current at just 13.3mA maximum from the 5V rail. So I a TO-220 Package with a basic HS (or even no HS at all) will be sufficient to run 3 of these boards. So I'll use just a single start-up/turn-off circuit for all 3 boards (I've posted the schematic earlier).