48 volt steppers board selection

deckingman

@luke-slaboratory It's true the a) motor torque will drop as speed increases and b) more torque is required for higher accelerations. But I'd have to disagree that beyond a certain point, increasing acceleration makes much difference to FDM printing.

You misquoted me by saying "Your example of heavy head printing slowly......" because I distinctly said "I accelerate this mass at 1500mm/sec^2 up to 350mm/sec " and I doubt many other people have machines that can actually achieve a melt volume flow rate to be able to print at 350mm/sec.

Admittedly my 1500 mm/sec^2 acceleration rate is modest but even at that, it takes a mere 0.2 seconds to go from zero speed to 350mm/sec and in reality, very few moves start from zero speed due to the jerk setting. So increasing that acceleration rate makes very little difference to the overall print time. And this all assumes that volume flow rate out of the nozzle is capable of matching the rate of change of velocity of the print head, which is seldom the case.

Luke'sLaboratory

Tl;DR

Whoops on the tool speed.
Acceleration is incredibly important to achieve faster prints and the primary path
35mm3/s is pretty achievable nowadays
High voltage is an important feature to accommodate high speed tools and increasingly quick printers
All of this is moot if the intended toolspeed is 200-300mm/s or less.

@deckingman said in 48 volt steppers board selection:

"I accelerate this mass at 1500mm/sec^2 up to 350mm/sec "

My bad - not sure why I saw 150. Maybe I conflated the 1500 with the 150 in my mind. not sure. I'll own it.

Regardless - 350mm/s tool speed (assuming that you're printing with a .4mm nozzle, .48lw, .2mm layer height) is only 30mm3/s, pretty medium-flow nowadays. Slice Magnum, Takoto, HF Dragon, Rapido, HIC, Nova, and the standard volcano is all in that range, with some in this list well above those capabilities. There are even higher-flow hotends in the market, starting with the Rapido, HIC, VolcanoMosq, Mag+, and a few others out there. I'd say a sizable portion own a hotend capable of 30mm3/s, especially if augmented with something like a CHT.

Re: acceleration - the below plate is assuming 400mm/s print speed, 15k acceleration, 8mm/s instant velocity change, except for the first layer. 4 perim, 40% infill. Its predictions are pretty bang-on compared to as-printed, within ~2% of predicted with this profile and this plate.

This is the same plate, except I've dropped the acceleration to 1500mm/s/s

Now - what happens when increasing the acceleration to 4500?

You save an estimated 5hours of print time, and use 13.5N of force instead of 4.5. That adds up after a while.

Lets go to the ultimate "theoretical" limits - does it fall off? you have 3kg gantry, and a conservative estimate of 234N of force, so you can accelerate at a theoretical 78k mm/s/s.
Obviously mechanics won't allow this to happen, but at least lets see what happens to the print times. I've verified that the slicer decreases time alloted beyond 78k accel, so this model should be accurate.

All the way at 78k, you've only saved another 2hrs beyond the 15k estimate. This is where these diminishing returns are really found, assuming your hotend is limited to ~35mm3/s.

But, increasing acceleration is the best real way to increase speeds - increase your jerk too much, you get ovals instead of circles, lose corners, and all sorts of other bad behavior. Acceleration is something that can be increased near infinitely, assuming that you have the power and rigidity to do so, among other plastic-melting limiters.

But, the point stands - at 15k, I can double my productivity over 1500. At 4.5k, I save an appreciable amount of time that adds up. its not inconsequential, and assuming that your machine is built correctly, quite repeatable.

All of this was done without fiddling with the Z-axis speeds - if I switch to a belted or a faster Z, this gets significantly better as alot of this time may be calculated from z-hop.

But, many Common hotends out there can easily go past 35mm3/s - This is where the high speeds and high voltage come into play, and something that is important to address. One can't effectively use high-flow hotends if their tool never gets to the high speed, necessitating high acceleration, demanding high torque.

o_lampe

@luke-slaboratory
to be fair, your example prints faster because of the many travel moves. I'm not sure if "I" ever had a print plate with so many individual parts, nor if it's the average usecase.

...but it's impressive to see what's possible in theory

deckingman

@luke-slaboratory TLDR but ref volume melt rate, I recently demonstrated printing a vase in under two hours with an average volume flow rate in excess of 70 mm^3 / sec. Or to put that into perspective, that was a 1.5mm nozzle, 1.8mm layer width and 1.2mm layer height at 80mm/sec print speed.

At those melt rates, the challenge is not print head speed, but cooling the part sufficiently before the next layer.

I have since tried a 1.8mm nozzle and estimate the volume flow rate as being around 100mm^3/sec but I haven't been able to cool that 2mm layer width/1.5mm layer height @80 mm/ sec head speed sufficiently to actually produce anything other than a collapsed blob

Edit. The 350mm/sec print was done years ago but it was with a 0.5 mm nozzle and 0.3mm layer height, not a 0.4mm nozzle and 0.2 mm layer height.

Luke'sLaboratory

@o_lampe

Shouldn't matter - single large part, many small, and to be clear - my print and travel accel is the same - 15k. No difference there. If its a big part with many infill sections, its still 15k. perimeter? 15k.

I'm not doing this to speedboat either - these are for parts that go out the door.

@deckingman

One of my common profiles is to print at 70mm3/s with a .8mm nozzle, in PLA. To do this, I admittedly drop to 5k accel for now since the plastic lays down neater (it tends to drag, but this is probably from the magnum+ hotend getting too cold on the lower portion of the block, causing the material to be less molten and not wanting to bind)

I'm hoping to get there on my smaller 400x400 machine that currently is limited to 400mm/s by swapping out for lower inductance motors.
Or, I could swap up to 48v, which would get me the tool speed I need to keep my hotend fully flowing. That again, would match the 70mm3/s on just a .4mm nozzle.

I use CPAP blower - WS7040. Its sufficient to cause a jam in the hotend itself - I can turn on and off flow in the hotend by ramping up and down - this is something that's been captured in other setups than mine, and is just a weird thing I have to account for with the mangum+ hotend.

I understand the cautionary tale of not going for omega upgrades when you're not going to be able to use them, but honestly the market for 48v boards is being driven by many more people than just me - its a real need as melt zones continue to grow and mechanics(and software!) are more able to support the higher flows.

I could use the current-providing capabilities of the 6hc's 2/5160's, but it doesn't match what I need it to do otherwise, and for the premium and market position of the 6hc, I feel that its a valid concern that it doesn't support this option.

deckingman

@luke-slaboratory I understand what you are trying to say but I don't think you've made a convincing case for using 48V. My point is that I use 1500mm/sec^2 with 3Kg + of moving mass but that's still well within the torque capability of those 2.8Amp motors (running at 2.3A) even at 24V. If I had a "sensible" hot end which weighed say 300 gms instead of 3Kgs, then I could use 15,000 mm/sec or more for the same torque. That's still less than half the 6 + Amps that the Duet board is capable of. I guess we'll just have to agree to differ.

rexx

@deckingman sorry for the super late reply life been a round latley but i am back in and read the posts and if you are correct i should be ok with only a duet 2 not Duet 3 hc6 but (duet 2) as you said only 2.3 A actual @ 24VDC i have the 178 oz 1.8 deg i forget the MH resistances but i would love to use the duet 2 instead of 3 i think ???..

the duet 2 is rated to 2.8A-firmware set to 2.4A do you think i could use this under your limits above?

i am only driving 3 nema 23x57mm and 2 pankcake nema 17's at 24VDC the duet 2 does goto 32 VDC
feature wise the duet 2 is fine less the close peek amps.

rexx

@deckingman Also i dont need 5 amps way over kill your correct i have 2.8 amps nema 23 (x3) can i try a duet 2 in your opinion or go for the 3 ?

rexx

@deckingman the nema 23's are 23HS22-2804S MH2.6 @1KHz inductance with .9 Ohms resistance @25C 1.8 deg rotor inertia 300 g-cm2
on duet 2 or 3 6hc?

deckingman

@rexx said in 48 volt steppers board selection:

@deckingman Also i dont need 5 amps way over kill your correct i have 2.8 amps nema 23 (x3) can i try a duet 2 in your opinion or go for the 3 ?

IMO, either generation will drive those motors. So the choice comes down to other features you might want and the price you want to pay. Check the specks of each board but ultimately the decision has to be yours. One minor consideration is that later generation products are likely to be supported for longer than earlier generation products but AFAIK, Duet are not planning on dropping support for gen2 any time soon.

jay_s_uk

there will be features that don't make it to gen 2 boards going forwards though, such as multiple gcode streams

A Former User

This post is deleted!

jay_s_uk

if 48v is really required, I can see 6 options

Use a duet 2 with an external breakout board, allowing external drivers to be used at higher voltages
Use a duet 3 6HC/mini+ with 1HCL boards to run at 48v
Use a duet 3 6HC/mini+ with 1XD boards to external drivers to be used at higher voltages
Wait for the duet 3 6XD, allowing external drivers to be used at higher voltages
Use a supported STM32 board to run external drivers to be used at higher voltages
Use an Octopus Pro or Super8 v1.1 to run 5160HV stepsticks at voltages up to 60v

rexx

@deckingman thank you for your reply i am comfortable with set up of the duet 2 and ideally would be great if it could i may just giver her and try again thanks for your guidance here

Cheers

rexx

@jay_s_uk great view on the 48volt side thank you,
the stock stratasys i have was 36-48vdc hybird steppers i had planned on your #6 super fly 8 with (x3)HV5160 mellow board but kinnda wanted to look at keeping the printers on the same platform(duet) thus the duet line up investigation i had no idea the 6XD exsisted thanks to you again
but option 2 seems best as is does not need external drivers but maybe not so...
is the Duet 3 much diffrent then the 2 in set up ease diffrances?

xyz on the nema 23s 3 HV (would this mean i could only send 48v to the 1HCL and the duet 3 can be 24vdc still?)or up to 32vdc
ext1 ext2 nema 17 ht pancake 2 24v regular 5160 board mount

heated chamber with thermal fuse is on ssr mains power but i have like 8x 24vdc fans (11.5 watt-.48A each) 6 of wich on 100% is fine and 2-3 require PWM control +6 optical switchs one is ABL so the duet 2 meets the needs hummmm....

#2 #4 and #6 will be like sheep jumping over the fence tonight mayeb tomorrow too!

Cheers Jay

rexx

@arnold_r_clark
do you know materials with near 0 z layer adhesion loss are around blown film and injection are old and still used today for good reasons but the why nots of FDM printing limits are dropping off every day from flow rate to speeds and accelrations and HT thermos (500C)wich are not getting made in your garage for less than 50+K in machinery and tooling plus the three phaze power transformer power it draws and so on , but FDM will be 95% of your product agreed but the parts gets in where it fits in,
have a look at PEEK PEK PEI PPSU PESU materials and data sheets do you mould any of thease? note the temprature resistance and ASTM Tests injecton small parts in ht materials ok but to make large parts in thease HT tempratures are simply not practical the tooling alone is not fair for less than 100 pcs runs or even 10000 pcs

jay_s_uk

@rexx The super8 1.1 allows you to run 3 drivers at 48v, 5 at 24v or 5 at 48v and 3 at 24v or all 8 at 24v or 48v. The octopus pro allows you to pick and choose which driver runs at 48v. Mellow have just released an external 5160 driver that allows you to use any stepstick type board with 48v. These 2 boards essentially run the same firmware as on the duet boards but don't have the expansion capabilities provided by CAN-FD https://www.aliexpress.com/item/1005004229433552.html

xyz on the nema 23s 3 HV (would this mean i could only send 48v to the 1HCL and the duet 3 can be 24vdc still?)or up to 32vdc

Yes, you could just operate the 1HCL's at 48v and the duet 3 at up to 32v.

is the Duet 3 much diffrent then the 2 in set up ease diffrances?

I would definitely go duet 3 over duet 2 as the duet 2 is almost (but not quite) at its flash limit so it won't be seeing certain new features going forwards (such as concurrent gcode streams). Setup is just the same between them.

rexx

@jay_s_uk

Jay you have been a great help again cheers and thank you

rexx

@jay_s_uk

Good Day!

so i went with the 6HC yes.... feel good about it sofar but this swap is for a heated chamber BL touch is not gonna happen and i want to use a optical 4 wire end stop (OPB- 880T51) and would like to use this for ABL do you know how i would set it up on the 6HC if i can? there are side stops that will deply and retract the interaction with the optical switch from in use to not.

red,blk,grn,white
rd,bk,gn,we

2 anode and two cathode

it is the stock switch for the z foam in a stratasys 768

i will just have to insert the G code for where to travel to the bumps to deploy and retract when done! (one on each side at the rear out side the bed as travel permits)

any thoughts are much appreciated
cheers

dc42

@rexx thanks for choosing the Duet 6HC.

That opto switch you linked to is obsolete, but is similar to many modern slotted opto switches. What you need to do is:

Connect the IR LED side (pins 1 and 2 according to the datasheet I found at https://datasheetz.com/data/Sensors, Transducers/Optical - Photointerrupters - Slot Type - Transistor Output/OPB880T51Z-datasheetz.html) with a series resistor) across either +5V or +3.3V and ground. Choose the value of the series resistor to pass about 10mA through the LED. That means about 180 ohms if you are using 3.3V or about 360 ohms of using 5V. The exact value is unlikely to be critical. The datasheet indicates that pin 2 is the ground side.
Connect the emitter of the output (pin 4) to ground, and the collector (pin 3) to one of the IOx_IN pins.