8i20/BLDC component, limited spindle speed

(That component has the look of something that I wrote during a lunch break without access to a LinuxCNC machine to test it on). Which is exactly the situation now, too, so again there may be errors.

It needs data types adding to the params
And it needs a closing bracket for the function, and possibly a catch all case

blazini36 wrote: The BLDC lead angle is set in the hal file to 90 (always has been)
if I setp the bldc lead angle while it's running at a speed below the requested (2100 vs 3000) I get no noticeable change in speed.

What numbers did you try? I see the screenshot at 180 degrees (!) and would actually have expected that to stop the motor.

I was suggesting tweaks in the 80 to maybe 110 degree range.

More voltage is worth trying. The 8i20 can handle it and the motor won't care.

andypugh wrote:

blazini36 wrote: The BLDC lead angle is set in the hal file to 90 (always has been)
if I setp the bldc lead angle while it's running at a speed below the requested (2100 vs 3000) I get no noticeable change in speed.

What numbers did you try? I see the screenshot at 180 degrees (!) and would actually have expected that to stop the motor.

I was suggesting tweaks in the 80 to maybe 110 degree range.

More voltage is worth trying. The 8i20 can handle it and the motor won't care.

I tried everything up from say 70 to 180 in 10* increments, didn't really seam to do much.

As for voltage the more I think about it the more I realize it probably does need to me higher. I did wire up the drive for 240v AC (320vDC) last night and changed some nv params on the 8i20. the 1000uf caps I was using were only rated for 250v so I had to make up a bank of 4 470uf 450v caps. I have the line side fused at 5 amps, when the line side SSRs close it blows the fuses from the inrush current on the caps. I'll grab some 10amp fuses today, I didn't really expect the current spike to be that much higher.

I tried double checking the caps but the meter I had on hand doesn't do well with capacitance. I did see the DC voltage drop very slowly as would be expected.

With the PSUs that drive my 8i20s I have a resistor in the circuit that the caps initially charge through, then an SSR bypasses that resistor once the 8i20 reports bus voltage over a threshold.

what type of resistor are you using? Thinking about it it only seems feasible to put a resistor between the rectifier and the caps, but then running a DC SSR parallel to the resistor is going to be tough because I can't say I've come across a DC SSR that can handle 320v.

Not sure how the resistor would work on the line side, If it's proper to put a resistor on one leg obvious what to do with the AC SSR.

I suppose the 8i20 busunderv pin linked to a 7i76e field output to drive the SSR?

blazini36 wrote: what type of resistor are you using?

Looking through old component orders, it seems I chose 620R 50W.
Though that seems a bit larger than strictly necessary to limit inrush. It does not need to be specified for full power handling as it is only used momentarily.

Thinking about it it only seems feasible to put a resistor between the rectifier and the caps, but then running a DC SSR parallel to the resistor is going to be tough because I can't say I've come across a DC SSR that can handle 320v.

No, it is placed on the live AC incoming feed[/quote]

I suppose the 8i20 busunderv pin linked to a 7i76e field output to drive the SSR?

I used a "comp" HAL component monitoring the actual bus voltage, but using busunderv might be simpler.

andypugh wrote: (That component has the look of something that I wrote during a lunch break without access to a LinuxCNC machine to test it on). Which is exactly the situation now, too, so again there may be errors.

It needs data types adding to the params

Code:

param rw unsigned level "state machine state"; param rw float threshold;

And it needs a closing bracket for the function, and possibly a catch all case

Code:

... if (spindle_command - spindle_feedback) > threshold) { spindle_underspeed = 1; } break; default: // not sure how we got here, but fix the situation level = 0; }

Still getting errors during install. Unfortunately I know very little about C or python or any of that but I'd be glad to test this out once I get some pins to play with.

For now I just swapped in 10 amp fuses. Everything in the power supply can handle over 10amps and there's no dry contacts to arc, the line SSRs are rated for 25a so I think I'll be OK since the 10a FNQ fuses seem to handle it so far. It's going to be a pain to stuff anything else in the 10x10 box.

It was definately a voltage limitation because I can now run the spindle up to at least 4k rpms. It sounds ugly but it gets up there. i can keep it pretty reasonable to between 2500 and 3000 by tuning the PID a little but i think beyond that the no load spindle is a little tough for the 7i76e to keep up with. This is with the servo thread at 800000, at 500000 i've had it shut down for following errors.

spindle_monitor.comp:29:9: warning: implicit declaration of function ‘absf’ [-Wimplicit-function-declaration]
The function is "fabs". I get that wrong a lot.

spindle_monitor.comp:38:46: error: expected expression before ‘>’ token Missing bracket, needs to be
Then see if the rest get fixed as a side-effect.

Getting there?

I think if you add up near the top of your code it will pick up the declaration for fabs()
Note I'm not on a computer to check the syntax.
But the errors are telling you to include math.h but we have to use a different include file in .comps so it should work.