7i76 + 8i20 for BLDC motor control

In the process of converting my PM25MV to CNC. I got the X,Y and Z going now I was working on the spindle control. I was probing the Potentiometer leads to see what the BLDC driver was working with and dropped one of my alligator clips and shorted the thing right out. I emailed PM to get a cost on a new one but in the Meantime I'm looking at other options.

I was looking for BLDC drives but simple drives with just hall effect inputs and can output 750w are few and expensive. Some servo drives can do it but still expensive. I'm not a big fan of VFD driven AC motors on mills this small. Poking around it looks like the 8i20 was meant for this sort of thing. It's listed for an "AC servo" but I'm pretty sure an AC servo is the same thing as a BLDC motor with an encoder for positioning rather than hall sensors for commutation.

The plan is to split an ethernet cable and wire it to the RS-422 on the 7i76 and run it back to the 7i80. There is no namelate on the BLDC motor, all I know is that it is "750w". I'm going to assume the voltage from the drive is on the upper end since there isn't any transformers on the drive. I'd like to build an 80v or so toroidial linear supply and power the 7i80 spindle motor as well as my Z stepper driver so I can get it off my 48v switching supply which is getting drawn down to ~36v with all 3 axis on it.

So my questions are:
Is it safe to assume I can run the motor off of around 80v and still get it upto speed without an issue? I assume that the BLDC component will use field inputs for the hall effect sensors, each should only be at around 50hz @3000rpm. The 8i20 has a brake resistor output, but I assume it's not necessary since the BLDC rotor is pretty light. The thing stops immediately on the stock drive. ANy thing else I'm missing?

www.aliexpress.com/item/6-5E-36v-72v-700...ard/32796616039.html

Doesn't look like something I'd drive a spindle with

Just a entry point for a search when you look for a cheep solution. And by the way. When the spindle turns with usual force, and nothing gets to hot, what's the problem.?
M5C

Mike

blazini36 wrote: Is it safe to assume I can run the motor off of around 80v and still get it upto speed without an issue? I assume that the BLDC component will use field inputs for the hall effect sensors, each should only be at around 50hz @3000rpm. The 8i20 has a brake resistor output, but I assume it's not necessary since the BLDC rotor is pretty light. The thing stops immediately on the stock drive. ANy thing else I'm missing?

You can probably get an idea of the rated motor voltage by spinning it in reverse and measuring the AC voltage generated.

The 8i20 will probably do what you want. I have used them in all my machines. It would be better to also fit a spindle encoder, just because that gives the drive a cleaner position input. (it also means you can ridid tap, hob gears, and other spindle-synched things).

The STMBL drive is another option, but good luck finding one (you can build one, though, I have PCBs)

Another good option for your motor is probably the Pico brushless motor drive: www.pico-systems.com/acservo.html

Technically, a BLDC motor wants to be driven with flat DC current or (PWM DC). There are 6 different states for the 3 phase connections that are stepped through, triggered by positional info from the Hall sensors. If you spin one of these moors you should see a sort of flat topped waveform. Cogging torque and noise is difficult to optimise with them. Look for "6 step control" or "trapezoidal control".

An "AC servo" or BLAC motor generally requires a PWM sine wave - and will generate such a waveform if you spin it. The drive waveform can be either be a "dumb" sinewave from an EEPROM or (better) generated with vector control (this requires phase current sensors in the drive). Ideally you'd have decent position resolution, so a high count encoder or a resolver is best.

The motors themselves, though, are pretty much identical. It's more about how they are driven.

andypugh wrote:

blazini36 wrote: Is it safe to assume I can run the motor off of around 80v and still get it upto speed without an issue? I assume that the BLDC component will use field inputs for the hall effect sensors, each should only be at around 50hz @3000rpm. The 8i20 has a brake resistor output, but I assume it's not necessary since the BLDC rotor is pretty light. The thing stops immediately on the stock drive. ANy thing else I'm missing?

You can probably get an idea of the rated motor voltage by spinning it in reverse and measuring the AC voltage generated.

The 8i20 will probably do what you want. I have used them in all my machines. It would be better to also fit a spindle encoder, just because that gives the drive a cleaner position input. (it also means you can ridid tap, hob gears, and other spindle-synched things).

The STMBL drive is another option, but good luck finding one (you can build one, though, I have PCBs)

Another good option for your motor is probably the Pico brushless motor drive: www.pico-systems.com/acservo.html

I was planning on adding an encoder of some sort. For now I was going to repurpose the optical break sensor that the mill came with for the tach. It's 4PPR but it will at least give me speed feedback on the spindle itself.

The replacement from PM was the cheapest and easiest option to get it running again so I wound up ordering one. Not sure whats up with this one but it's not working either. This one is rated for higher speed and a 1.5hp motor. Tested the pot and the 5v control wires and that's all fine. The board is fully potted so that's about all you can do. I don't trust this thing much anyway, I get some pretty wild AC voltages riding on the DC control wires when I was testing it. I'm gonna see if I can get a refund on the replacement, at this point I think I'd rather go with another BLDC drive.

I spun the motor by driving the drawbar hex with a drill. The drill is rated for 1300RPM, on the 1:1 pulley I got 35VAC between phases. On the 2:1 pulley which should be about 2600RPMs at the motor, I got 66VAC. I didn't realize you said to spin it in reverse so I think this was in forward.

I have a 35v+35v toroid that I'm hoping to use to run this. So it looks like I'll either be running it at about 50VDC or 100VDC after it's rectified. I assume with the stock drive it's something like a 70V motor, Is it safe to assume that If you limit the speed in software, the PWM generated at the amplifier would average out the voltage to the lower voltage? In other words, is it safe to supply the drive with 100VDC and expect the motor to handle it if limited to 2500rpms?

I was looking at the Pico drive but it requires PWM rather than analog. Other than adding more hardware I'd have to sacrifice a stepgen and write a bitfile for the 7i76e to output PWM. I'm not sure about building the STMBL drive right now because I have enough headaches with getting the rest of the mill straightened out, but I might want to pick up a PCB from you for down the road if you can spare one. You said you use the 8i20 on all your machines, what types of motors do you run?

blazini36 wrote: I was planning on adding an encoder of some sort. For now I was going to repurpose the optical break sensor that the mill came with for the tach. It's 4PPR but it will at least give me speed feedback on the spindle itself.

You get more info than that from the Hall sensors.

You said you use the 8i20 on all your machines, what types of motors do you run?

I have a range of motors, 3 x 0.54kW Lenze motors on the Mill, and then 900W and 1.2kW random servos on the lathe.

All these motors use resolvers. One of them had an absolute encoder, but I couldn't get it to work so swapped it with a resolver I had.

andypugh wrote:

blazini36 wrote: I was planning on adding an encoder of some sort. For now I was going to repurpose the optical break sensor that the mill came with for the tach. It's 4PPR but it will at least give me speed feedback on the spindle itself.

You get more info than that from the Hall sensors.

Yes, for commutation and motor position. If the spindles micro-V belt slipped the controller would not see it if it were using the motors Hall's without a synchronous belt or gears for the spindle.

What would be the minimum resolution I could get away with for a spindle encoder for say tapping if it had an index?