Simple stepper tests at 12v, 35v, 1A, 400mA, 1/8 Step and Full step

[Brent Crosby]

ref: http://groups.google.com/group/makerbot/browse_thread/thread/cb326a9a42c84768

Since we got the 50v bench supply, I thought I would get some crude
data on the stepper capability at different voltage, current and step
size.

The setup is MiseryBot (MakerBot Cupcake #1024 http://wiki.makerbot.com/makerbot-1024
) upgraded with Gen 4 electronics ( http://wiki.makerbot.com/smd3 )
and a MK5 extruder, and Thermaltake 120mm fan power supply.

We used only the Y axis for the the tests. We would jog +50mm then
-50mm several times. If the platform came back to the exact spot, we
would count that as no slipping. We would vary the speed to determine
the highest speed that could be jogged without slipping.

Since our bot is now calibrated to 1/8 steps, for the full step test
we used +5mm (*8 = +40mm) and -5mm (*8 = -40mm) and we will multiply
the reported rates by 8 so they can be compared with the 1/8 step
rates.

We have a harness to the stepper motor extruder controllers from the
bench power supply ( http://shop.vendio.com/evan2002/item/770921907/index.html
). The 5v regulators on the stepper motor controllers are used.

Vref = 2.0v (current limited to 1A) (as shipped by MBI)
. . Supply = 12v
. . . . 1/8 step
. . . . . . Maximum step rate = 5000
. . . . full step
. . . . . . Maximum step rate = 5240
. . Supply = 35v
. . . . 1/8 step
. . . . . . Maximum step rate = 11,000
. . . . full step
. . . . . . Maximum step rate = 10,800
Vref = 0.8v (current limited to 400mA) (rated current of steppers)
. . Supply = 12v
. . . . 1/8 step
. . . . . . Maximum step rate = 4200
. . . . full step
. . . . . . Maximum step rate = 5120
. . Supply = 35v
. . . . 1/8 step
. . . . . . Maximum step rate = 11,000
. . . . full step
. . . . . . Maximum step rate = 10,800

Average maximum speed decrease going from full step to 1/8 micro step:
6%
Average maximum speed decrease going from 1A to 400mA: 5%
Average maximum speed increase going from 12V to 35V: 225%

So what did we learn?

Going to microsteps decreases maximum step speed: True but only a bit.
The microstepping driver still delivers full current at the peaks of
the waveform.

Increasing current increases maximum step speed: True, but only a bit.
BUT, increasing current does increase the power dissipated in the
motor by P = I*I*R That is temperature would go up by the SQUARE of
the current.

Increasing voltage increases maximum step speed: TRUE, and in a big
way. Not quite linear, but very significant.

The A3977 controllers and the 7805 regulators will overheat at higher
voltages: FALSE, they are both warm to the touch bit not "hot" by any
means. I can leave my finger on them for any length of time without
discomfort.

In a nutshell, increasing the voltage from 12v to 35v, combined with
decreasing the current from 1A to 0.4 gives 2x maximum step rate, and
decreases the heat in the motor by 6x.

[James McCracken]

AWESOME work, and a great job debunking myths and getting facts re:
microstepping

Just one thought; can you repeat the tests on the x axis, with its
higher mass? Not only will this give you two data points, it'll give
us an idea of how current and voltage scale with mass...

> --
> You received this message because you are subscribed to the Google Groups "MakerBot Operators" group.
> To post to this group, send email to make...@googlegroups.com.
> To unsubscribe from this group, send email to makerbot+u...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/makerbot?hl=en.
>
>

[Brent Crosby]

Thanks for the kind words. This _is_ my "highest mass" axis -- the one
that carries the entire lower mass axis. This is the axis that moves
right to left and carries the front to rear axis. Maybe I have them
labeled wrong!

[James McCracken]

Yeah, that one's typically labelled X, and the one that moves back and
forth is typically labelled Y...

[Erik de Bruijn]

The microstep size does influence how fast you can ramp up motion.
From what I can tell, the smaller the steps, the better the
acceleration. A stepper motor is effectively a spring, so it will get
less of a jerk initially from the inertia.

The voltage has to be higher for high frequency stepping, because the
voltages the motor generates otherwise become higher than the voltages
you put into it. This is why "we" (I won't advertise our name on the
MB forum) don't use 12V but 18V through 20V. 12V really limits the
feedrates you can achieve, even with a free spinning motor without a
mechanical assembly connected to it.

Erik

[Ed Nisley]

the smaller the steps, the better the acceleration

What would really help is blended acceleration limiting: ramp the motors up to speed, rather than starting with a bang, then ramp them down to a stop at the other end of the move. This need not take much time or distance, but it does require coordination between the axes so the extrusion remains on track.

For a given torque, the pull-in step rate is maybe 10-20% of the pull-out rate: a motor that can run at 6000 step/sec can't start at more than 600 step/sec. Running from a higher voltage (thus, with better current control) improves the torque, but that additional power goes to waste after it's up to speed.

If we started that motor gently, we could run it at 1800 rpm and solve a lot of the speed problem!

[Jordan Miller]

yes this method is what's used in reprap and is only possible with so-called 5D firmware and a stepper-driven extruder. this type of extruder can keep up with rapid movement changes that, as you say, necessarily require rapid extrusion rate changes.

jordan

--

[Ed Nisley]

only possible with so-called 5D firmware

Looking through the source code, I really like what I see: limiting the XY acceleration should bring a major improvement to the machine's performance.

After I get a stepper extruder working with Thing-O-Matic 2.7 firmware, would injecting that 5D firmware be more-or-less straightforward?

Thanks!