Z-SUITE 1.8.0 feedback

Noctua is a good fan company.

I think you're right but I've had 2 or 3 Noctua fans fail on my M200.  I've been running the fans at 100% lately and am wondering if that is the reason for the failures.

I think you're right but I've had 2 or 3 Noctua fans fail on my M200.  I've been running the fans at 100% lately and am wondering if that is the reason for the failures.

What part of the fan failed? Bearing? If so, maybe the XY stage motion loads the fluid bearing, doubt the fan was designed to be on a motion stage.

I just wonder: somewhere it was mentioned that Zortrax drives the 12V fans with 24V arguing that the average voltage will be lower due to PWM.

Letting aside that they obviously leave the lowpass filtering to the fan, this reasoning is somewhat contradicted by applying a 100% duty cycle.

At least I would assume that 100% means that the full 24V are permanently applied to a 12V fan ... dunno if this is a good idea.

What part of the fan failed? Bearing? If so, maybe the XY stage motion loads the fluid bearing, doubt the fan was designed to be on a motion stage.

I've no idea - they just stop working and I haven't tried to take them apart to see what happened.  The failed fans do seem to spin about as easily by hand as new ones so I'm guessing that the bearings are OK.  Perhaps the fan motors have burned out.

I just wonder: somewhere it was mentioned that Zortrax drives the 12V fans with 24V arguing that the average voltage will be lower due to PWM.

Letting aside that they obviously leave the lowpass filtering to the fan, this reasoning is somewhat contradicted by applying a 100% duty cycle.

At least I would assume that 100% means that the full 24V are permanently applied to a 12V fan ... dunno if this is a good idea.

I'd hope that Zortrax defines 100% as a 50% duty cycle for the PWM signal, but who knows?  Maybe Marcin can get an anser from the design engineer(s).

I'd hope that Zortrax defines 100% as a 50% duty cycle for the PWM signal, but who knows?  Maybe Marcin can get an anser from the design engineer(s).

That is correct. Full speed (on Auto fan) in the fan speed menu is close to 50% duty cycle using the 24V supply. "100%" in the fan menu is less than 50% duty cycle. I've been running six Noctuas on two M200s for well over a year with no failures or problems.

That is correct. Full speed (on Auto fan) in the fan speed menu is close to 50% duty cycle using the 24V supply. "100%" in the fan menu is less than 50% duty cycle. I've been running six Noctuas on two M200s for well over a year with no failures or problems.

Julia - do you ever print with the fan set for 100% in Z-Suite?  I'd done that for a few weeks before my last 2 Noctua fan failures so I'm wondering if that could be an issue.

Julia - do you ever print with the fan set for 100% in Z-Suite?  I'd done that for a few weeks before my last 2 Noctua fan failures so I'm wondering if that could be an issue.

Good question, and the answer is "no", at least not recently, i.e. in the past year. Occasionally maybe a print on "Auto", which has speeds higher than the "100%" setting, but very very seldom.

Well, soon you will :)

And as a side note: I understand that you measured differences between full auto fan and manual fan at 100%, but the fan speed requested can only be in the range between 0% and 100%, no matter if auto or manual fan mode was used.

Maybe the m200 determines the actual PWM used for the 100% request based on other parameters like temperature or voltage but the m200 can't differentiate between 100% requested by auto mode or manual mode.

Well, soon you will :)

And as a side note: I understand that you measured differences between full auto fan and manual fan at 100%, but the fan speed requested can only be in the range between 0% and 100%, no matter if auto or manual fan mode was used.

Maybe the m200 determines the actual PWM used for the 100% request based on other parameters like temperature or voltage but the m200 can't differentiate between 100% requested by auto mode or manual mode.

I never actually "measured" the difference, but (back in the day, when last I played with it) the difference was audibly obvious. Doing a print at 100% manual made a constant gentle whoosh, while on Auto it would rise to a scream at certain times.

Well you can hear if the PWM is close to 50% or less than 50%? Wish I had your ears ^_^ ...

There might be factors that make the noise annoying other than speed like resonance, interference, beats...

No, it is/was really obvious, no special talent required. Not subtle at all.

I’ve measured 56% duty cycle on 100% fan speed.

Actually I’m using radial fans for both cooling block and printing area. They not survive to 24V PWM, so I’ve used a switching step down for the upper fan and a divider/filter for the lower. I’ve designed the divider/filter in order to safely work up to 60% fan speed. Over 60% the printing process abnormally terminates due to excessive hotend cooling.

So if I need more printing area cooling I can use the hotend coat made by Julia.

But, without the hotend coat (example, hot z-pcabs printing) I cannot use the 1.8.1 because any print that use the lower fan (auto or at least 20%) aborts for excessive cooling.

I strongly hope that Zortrax revise the manual fan management.

Just noticed that 100% fixed fan actually results in a 99.608% request while auto fan goes up to 100%. Most probably a truncation vs. rounding issue.

The difference is small but maybe amplified but firmware behavior.

Side note: it's generally seems like a bad idea to control a brushless fan with a PWM without filtering. Adding a lowpass filter between the PWM output and the fan will average the voltage.

The higher the filter constant (R*C), the smoother the signal will be and the smaller the deviations from the mean value will become. Problem is of course that there will be a voltage drop on the resistors.

Dunno about the original fan, but the NF-A4x10 FLX has a 50mA max. current specification, so e.g. using a 100Ohm resistor would result in a 5V voltage drop at the resistor resulting in 250mW peak power.

And that's at 12V. At 24V I would expect the current to be going somewhere close to the double which is also doubling the peak power.

Of course due to the ~50% max, PWM, these values will be halved in average again. Still, something like 10Ohm would be probably preferable but then you need a high capacitance. Like at least 220µF or so.

Quick'n'dirty LTSpice hack (untested, just meant as a starting point and for visualization):

4070

noctua ltspice.png

generally seems like a bad idea to control a brushless fan with a PWM without filtering

I think it depends tremendously on the PWM frequency. If it's sufficiently high, I'm fine with the fan doing the filtering.

The main thing to worry about is overheating the fan - if the PWM frequency is high compared to the time needed to warm things significantly, there ought to be no problem.

It was about 250-300Hz last I checked in 2014.

Who says the fan needs to filter anything? It's brushless fan - that's not a simple DC motor. Sure, the inductance and resistance of its coils will also provide some filtering of the current through the coils in a way, but the commutation process will be heavily disturbed by the PWM on its supply voltage. Actually the 24V could even destroy the electronics needed for the commutation if there is no proper overvoltage protection. Then again, if there's e.g. clamping diode on the input this could mean the output stage is shortened.

Anyway, the inrush current with a 24V voltage will be obviously much higher than with a 12V voltage for which it was designed.

Besides, the influence of PWM frequency on overheating should be close to zero. What counts is the average power- and this is related to the duty cycle, not the frequency.

I think it depends tremendously on the PWM frequency. If it's sufficiently high, I'm fine with the fan doing the filtering.

The main thing to worry about is overheating the fan - if the PWM frequency is high compared to the time needed to warm things significantly, there ought to be no problem.

The pwm frequency is 280 Hz.

Hello,

I've noticed as well that the printing time has been considerably increased on same part between V1.8 vs V1.7.

For e.g. on the same model, printing time is:

V1.7: 5h15

V1.8: 7h45

Does anyone noticed it?

I believe there was a bug in the earlier version returning incorrect print time