Software suggestions and requests

[DoulosDS]

Amoose136,
Thanks for your suggestion and drawing. I think the alternating rollers would reduce the error by half, but the accumulator only smooths the dynamic flow. Once the pump stops, if a roller is engaged in releasing a pinch (between 100% and 0% pinched), the pressure would equalize and any error would be transferred to the print tank. However, it is all probably moot, based on my back-of-the-napkin analysis. And, if we can hack the slicing software, we can stop the pump at regular intervals between pinches and eliminate the problem entirely. See my reply to CescoAiel.

[CescoAiel]

Additionally, it is not a recurring error that keeps impacting... It s something that may be recurring as a repeated event, but it is not going to grow progressively larger every time it happens! And as DoulosDS said already: if we can make the pump stop just before it releases the roller-clamped section, we can avoid it altogether... That can be as simple as putting in a sensor that stops it at set locations, or otherwise leaves it running until it reaches that point. (Which would be every 1/nth rotation, where n=number of rollers)

The simplests and cheapest option for that would be a normal motor, and a circuitboard disc with as many interruption lines etched in to it as there are rollers on the roller-body, and a drag-contact on it... As long as the contact touches copper, the motor keeps running, when it gets to an etched section it stops... All you need to do then is modify the contact's position to make it stop in the right place...
If you go the stepper motor route, it is as simple as counting steps!

[DoulosDS]

Peristaltic Pump Amazon.JPG Found this pump on Amazon for $12, free shipping. I was thinking I would need to replace the motor with a stepper, but I don't think that's necessary. I can add an optical sensor to detect the rotor arms as they pass (looks like there are 3). The software can count them and run the motor forward and reverse as needed to adjust the resin level. If the flow rate and RPM specs are accurate, it should transfer 0.2 ml per 1/3 revolution. Conveniently, that is the same volume I calculated for Z-axis error due to tube pinch displacement in my previous post. So, if the print tank is approximately 3" in diameter or 75 mm, the area is 4418 mm2, and 1/3 rotation of the pump results in a Z-axis increment of 0.0456 mm. That's not great resolution, but it means the maximum resin level adjustment rate is only 13.5 mm/min or 0.228 mm/sec. I think that a second pump could run in parallel with this one, maybe with a higher flow rate. Both pumps could pump extra brine to overcome resin surface tension and viscosity issues, ensuring all the previous layer is covered with fresh, liquid resin. Then remove all but the increase for the current slice, within a few seconds.

[DoulosDS]

Peristaltic Pump Ebay.JPG Here is the best inexpensive peristaltic pump I could find with a relatively high flow rate: 500 ml/min for $30 and free shipping. It doesn't provide much info on specs, like rotor RPM, number of rollers (3? 4?), and tubing inner diameter. It does say the in/out hose barbs are for 1/4" tubing, so I expect the pump tube is 1/4" ID also. It may be possible to soup it up with thinner wall tubing, or speeding up the rotor. Regardless, at 500 ml/minute, it could speed up printing.

[CescoAiel]

So, if the print tank is approximately 3" in diameter or 75 mm, the area is 4418 mm2, and 1/3 rotation of the pump results in a Z-axis increment of 0.0456 mm. That's not great resolution, but it means the maximum resin level adjustment rate is only 13.5 mm/min or 0.228 mm/sec.

0.045mm is a good enough resolution for most prints (0.05 is considered typical layer thickness for laser based printing), and can be doubled (ie rise halved) to 0.0225mm by increasing the tank diameter to 8,5 cm, and almost tripled (approx 0.015mm) by going to 10cm.
Of course that also decreased the max rise speed, but then again, how many filament 3D printers complete a 0.2mm (standard thickness for those printers) layer in 1 sec?
Also faster pumps usually also mean less accuracy...

[harpo99999]

cescoaiel, in my limited experience with fff printers a 0.3mm layer height is poor res, a 0.1mm layer height is ok, a 0.03mm is very fine (note all of these are on FFF printers), and the peachy should be able to get into the 0.003mm layer height(depending upon the container size and drip speed) so should be able to produce prints that would please almost anybody that can accept a single colour object with the ability to add dyes to the resin to colour it

[CescoAiel]

cescoaiel, in my limited experience with fff printers a 0.3mm layer height is poor res, a 0.1mm layer height is ok, a 0.03mm is very fine (note all of these are on FFF printers), and the peachy should be able to get into the 0.003mm layer height(depending upon the container size and drip speed) so should be able to produce prints that would please almost anybody that can accept a single colour object with the ability to add dyes to the resin to colour it

First off, I had a typo which I corrected (0.5mm -> 0.05mm), secondly these are pretty small containers he used for the calculations, so you can work with bigger containers to decrease the layer height accordingly (a doubling of the size, reduces the height increase by 75% (ie leaves it at 25% of the original), as we're talking squared (to the power of 2) increase of volume when we're increasing the surface area of the container...
Additionally, the smaller the inside diameter of the tube in the peristaltic pump, the smaller the per-segment volume. The more rollers on the main body, the smaller the per-segment volume... So there's a lot of variables that have NOT yet been considered in DoulosDS's calculations. A bigger body with more rollers may be able to provide both higher precision as well as higher max volumes...
Nobody said this would be easy!

PS: 0.2mm is typical layer height for a filament based 3D printer, although many can actually do 0.1mm, and some even 0.05... The peachy's higher Z-resolution is a great premise to start from, but we'll need to see in practice what Z-resolution will be the most practical...

[DoulosDS]

0.045mm is a good enough resolution for most prints (0.5 is considered typical layer thickness for laser based printing)...

I was searching the whole forum for this info; then I found in on my own thread! I'm glad to hear my $12 PS pump can meter out one slice per 1/n revs in the smallest print tank I expect to use, and could slice it a lot thinner in the largest tank I envision. And, how much faster can the pump move a slice volume of brine than the drip system? This could speed things up a lot.

As for accuracy, the peristaltic pump hose degrades gradually over time, but that change should be negligible during a print. If the pump is run at high speed, the rate of relaxation of the pinched tube may cause an error relative to running at slower speed. If the tube is unable to resume the full, unpinched volume before the next roller pinches it off, the volume between pinches will vary with pump speed. If two pumps are used: one low flow rate for slice volume, and one high flow rate for displacing the resin to cover the previous slice, I think there's no problem. The slicer pump can run slowly, and the displacement pump can run at a uniformly high speed in both forward and reverse. If there is any reduction in pinched volume due the speed of the pump, it should be cancelled out. Of course, the displacement pump would run an equal number of revolutions in both directions and stop between pinches.

!!! Whoa !!! I just enlarged the image of the motor label for the high displacement PS pump on eBay... it says the flow rate at 24 VDC is 4500 ml/min! (editing: WRONG! That's a typo on the label.) That's 75 ml/sec or 7,500 mm3/sec. In the previous example of a 7.5 cm diameter print tank, that is a Z rate of 1.7 mm/sec. That should be plenty fast for displacement, even in a larger tank, or the pump can be run at a slower, but uniform speed. I have sent the seller a question whether the pump listed is the same as pictured, and can it run on 12 VDC at 1/5 the flow rate.

[CescoAiel]

Note there was a typo in that quote... There was a 0 missing:

0.045mm is a good enough resolution for most prints (0.05 is considered typical layer thickness for laser based printing)

[DoulosDS]

I have sent the seller a question whether the pump listed is the same as pictured, and can it run on 12 VDC at 1/5 the flow rate.

No, the pump is NOT as pictured. It is a 12 VDC pump with 500 ml/min max flow rate. It has 3 rollers on the rotor. I have asked the seller if he has any more of the 24 VDC pumps left... I will let you know what he says.