retrofit on EMCO PC turn 55

jrkeat wrote:

How would the D525MW work? Chuck says the job requires a DB25 for the turret, and either the Gecko or the 7i39 requires another. So two DB25 parallel ports, which seems like its one more than the D525MW can be set up for.

You may need to add a PCI parallel port card in addition to the onboard one. As I said, you will need to count pins to work out what is required. The Gecko is wired to use a complete parport, but individual drives can be wired individually.
The parport has either 12 outputs and 5 inputs or 4 outputs and 13 inputs.
4 outputs is not enough to drive s axes and the turret, so the question is whether 5 inputs is enough to handle the axis limits and the turret. You need 2 lines for the X limit and Z limit (though it is actually possible to share them, if you really need to). That leaves 3 lines for the turret. I don't know how many it needs.

The 7i39 actually uses a connector on a Mesa PCI card, so is a completely different thing. (It is a brushless motor drive for connection to the Mesa 5i20, 5i22, 5i23 or 7i43 cards)

If you actually meant the 5i25 then the point is that 5i25 connector _looks_ like a DB25, but is actually rather different and configurable. You can connect (for example) a 7i76 with 5 channels of step/dir stepper control, 48 lines of digital IO, and a +/-10V line for the spindle VFD. Then you can connect another 48 IO card to the expansion port on the 7i76, and if you really want to get silly, add another 8 x 48 channel IO cards to the expansion header on the 5i25.
Or the pinout of that DB25 can be identical to a Parport and drive the Gecko G540 drive mentioned earlier, but you still get the option of adding another 8 cards to the expansion header, or a second Gecko G540. It's all down to the magic of firmware.

TBH I don't see much point adding extra parport cards when the $80 5i25 offers so much more in the same footprint. But also bear in mind that many PCI parport cards are dual-headed.

Also I want to make it clear that the 7i76 board does not drive stepper motors directly, it is more of a smart breakout board to which stepper drives such as Gecko / Leadshine / TB6540 / Original Emco drives are connected.

I think you are going to need to spend a minimum of $120 to get the machine working. You can spend a lot more, for an incrementally better result. I have been trying to describe the least-cost method.

The G540 has 4 inputs. I used 1 for Z axis home switch and 1 for X axis home switch.

Originally I used the other 2 inputs for the spindle encoder which worked fine. However the interface was a bit ackward because the G540 inputs are internally pulled up to +12 and the EMCO optical interrupters are 5 volt parts. This means I needed to use an open collector buffer to shift from +5 to +12.

When I started looking at the tool turret, it required 2 more inputs so I was forced to add another parallel port. Since the EMCO optical encoders are already buffered and debounced and 5 volt parts, I connected both the spindle encoder and turret encoder directly to the second parallel port without a break out board. The optical encoders already provided the necessary isolation. So in summary the second parallel port is only used for 4 inputs. (2 spindle encoder and 2 turret encoder). The home switches are still connected to the G540. I still have 2 inputs left of the G540 for things like a touch probe if I ever choose to add something.

The G540 has 2 outputs. One is used to drive a relay that turns on the VFD. The other is used to move the tool turret using the existing tool turret control board.

The G540 has a VFD speed interface. One DB25 pin has a PWM signal that is converted to 0-10 volts that control the VFD speed.

All in all, the G540 is a fairly complete package. Yes there are 2 drives in it that are not used and that wastes 4 output pins but I didn't need them anyway. I looked at alternative solutions but I could not find anything that was cheaper. Keling Inc sells the G540 for $240.

cs

andypugh wrote:

As I said, you can run a 5-wire motor on any drive. On a bipolar drive you ignore the centre tap wire and leave it unconnected. On a unipolar drive, it is the 0v return wire.

Isn't that going to double the impedance of the coil (two coils in series, rather than one)? Wouldn't you have to increase the voltage by sqrt(2) to get the same power out of it?

I seems like in order to have it behave according to the original specs, you would do just what you said, but ground the center tap.

jrkeat wrote:

On a bipolar drive you ignore the centre tap

Isn't that going to double the impedance of the coil (two coils in series, rather than one)? Wouldn't you have to increase the voltage by sqrt(2) to get the same power out of it?

Yes, and possibly no.
You have two coils in series, so the impedence (and inductance) is higher. But you are energising both sides of the winding, not just half of it, so the magnetic flux is doubled.
Quite how that translates to torque and power for a particular motor I am not entirely sure. Looking at this datasheet for reconfigurable stepper:
www.slidesandballscrews.com/pdf/stepperm...SY60STH88-3008BF.pdf
It seems that, compared to unipolar use, the holding torque is higher wired bipolar series (you can't do bipolar parallel with a 5-wire) and the inductance is higher. The increased induction might limit the motor speed at a given voltage, if the machine does not hit physical limits first.

HOWEVER! I have just noticed that, contrary to the article I first linked to, it is not possible to run a 5-wire motor with a Gecko G540.
There is more information here, including suggestions that it might be a simple modification to allow the motors to run with Geckos.
(That will teach me not to allow something I read on the Interwebs to over-ride my original opinion)

This only matters if you decide to swap the drives, though.

andypugh wrote:

HOWEVER! I have just noticed that, contrary to the article I first linked to, it is not possible to run a 5-wire motor with a Gecko G540.
There is more information here, including suggestions that it might be a simple modification to allow the motors to run with Geckos.
(That will teach me not to allow something I read on the Interwebs to over-ride my original opinion)

This only matters if you decide to swap the drives, though.

Where did you read this? If I were to construct a driver in the simplest way I can imagine, it would also drive a 5-wire motor.

jrkeat wrote:

Where did you read this?

It isn't in the Rev 8 manual (interestingly)
But: www.geckodrive.com/gecko/images/cms_file...%20MANUAL%20REV1.pdf

"1) MOTORS: One to four step motors are required. All NEMA-17, many NEMA-23 and a few NEMA-34 motors are
acceptable. The motors preferably should be square in cross-section, not round. The motors can be 4, 6 or 8-wire
motors. 5-wire motors cannot be used with the G540. Choose a motor that has a rated current of 3.5A or less.
Choose a motor that has a rated winding inductance of 2.5mH to 3mH if maximum power output (>100W mechanical)
is a requirement. Never use a power supply voltage greater than 32 times the square-root of the motor inductance
expressed in milli-Henries (mH)."

It might be something that would need to be checked with Gecko, as they have changed the drives since that document was written.