Using a load sensor to accurately load resin into the tray

The cartridge dispensing valve that squeezes a rubber split cup could use an upgrade.
Take look at this

This is an auto-BAR dispensing a precise amount of Gin and Tonic into a glass of ice. The Bar uses a load sensor to accurately weigh and measure the amount of fluids poured into the glass using step-motor controlled mini-pumps.

No split rubber cup that wears out after being squeezed many times.

The method used to determine how much fluid has been dispensed has little if anything to do with the valve that does the dispensing. But you can bet that the valves used in the auto-Bar are significantly more expensive than the bite valves used by FL. FL sells their resin for the same $/L whether you buy it in a bottle or a cartridge. I’m pretty certain that the valve they’re using was the best they could get and still hold their price targets. I for one would not be happy paying a premium for resin in a cartridge vs. bottle in order to have a longer lasting valve.

As for metering, the printer has a sensor to tell it when the tank is to the fill line. The reason the fill sensor can’t be used with 3rd party resins is because the sensor is capacitive/inductive (I don’t remember which but it doesn’t matter), not optical. Different resins will have different capacitive/inductive properties. The reading that corresponds to a correctly filled tank of Tough resin may not correlate to the reading when the tank is filled with MadeSolid Vorex resin. Depending on which “way” the 3rd party resin is compared to what the printer thinks the resin is, you could get over filled or under filled if auto fill was used. A load cell doesn’t help much with this. Resins will have different densities and sensing the weight of the resin in the tank is only useful for determining volume if you know the density.

But also, the printer pushes down on the resin tank during the initial base printing process. Having a load cell that could accurately measure the weight of resin in the tank, and do it repeatably, when the tank is getting pushed on by the build platform between each layer, would IMO be difficult to do without significantly increasing the mechanical complexity of the printer. That would also increase costs to us, the user base. And more complexity means a higher probability of failure.



This is my son’s autobar. He programmed and built it…

The autobar uses the viscosity of the liquid to control the pumps to pour the drinks.
So it knows that Gin , tonic and other liquors have different weights even though they MAY have similar volumes. That difference is calculated within the pumps.
The mini-pumps are cheap but they are pumps that calculate volumes passing thru. so the accuracy is fantastic.

The only failure in the FL2 would be the split-rubber cap failing to close or open.

The split-rubber caps as a metering device is way last century!!

It looks like your son has some serious Maker skills.

The bite valve doesn’t meter anything. It opens and resin flows, it closes and resin stops flowing. If you watch the printer fill the tank, you’ll see it doesn’t just run resin through the valve continually until the tank sensor says the tank is full. It runs a little and then waits to see if the sensor is triggered. If not, it runs a little more. Lather, rinse, repeat until the tank sensor says the tank is filled.

Metering fluid flow requires additional components, which would increase the costs of the printer. As I said before, I don’t want to have to pay extra for features that I will rarely use. The system FL has now works for the FL resins it’s been calibrated to work with. FL can’t possibly calibrate the printer for all the 3rd party resins that are available so their best option is to not calibrate the printer for any 3rd party resin. The alternative is to increase the cost and complexity of the printer to add something that can sense the fluid characteristics of the resin being used.

Your son had to calibrate his load cell in order to compute volume of liquid dispensed. The task he had is a bit simpler since the density of the libations being dispensed will always be pretty much the same irrespective of who’s bathtub gin is being pumped. Water is water, alcohol is alcohol, the ratio is known and any flavoring doesn’t contribute in a meaningful way, so there won’t be a lot of variation. I don’t think the same can be said for SLA resins. I’ve seen a wide range of viscosities across resins, so there is undoubtedly a wide range of densities, too.

Using a Load Cell to meter what’s been put in to the tank would, IMO, be an extremely challenging mechanical design task that would also increase the cost of the printer and probably reduce its reliability.

And it’s important to remember that just because something can be made to work in a prototype environment doesn’t mean it’s suitable for actual field use. Your son’s design looks really cool, I don’t want to minimize his degree of inventiveness, but without extensive testing there’s no way to say if it will accurately dispense liquid and function reliably in a real-world operating environment. I’ve seen plenty of designs in an engineering lab that looked great on the bench and turned in to a train wreck when delivered in to the hands of actual customers.

IMHO what it needs is a Peristaltic pump. There are working designs that can be downloaded from Thingiverse.

Or a Progressive Cavity Pump also has working designs on Thingiverse.

Too bad the firmware for the Form 2 isn’t open source. :’(

There are working designs that can be downloaded from Thingiverse.

This is what I’m talking about.

And it’s important to remember that just because something can be made to work in a prototype environment doesn’t mean it’s suitable for actual field use.

Just because a design for something exists doesn’t mean that the design is suitable for the application. That pump on Thingiverse uses a vinyl hose for the transport tube. It will wear out and fail, too. Possibly in more-catastrophic ways than the bite valve. And it needs a lot more parts, including an electric motor/gear drive, which adds significant cost compared to gravity feed (which has no cost at all). It also adds components that increase the ways in which the printer can fail. If my cartridge bite valve fails to open, I can always swap in another cartridge. If my peristaltic pump fails, I’d need to send the printer back to get it repaired.

I’m still firmly in the camp that the existing design is pretty optimal. It adds virtually no cost and with a minimum of components ought to be very reliable. Yeah, the bite valve wears out and you need to toss the cartridge once the printer refuses to accept it any more. But since the Form2 cartridge costs me the same as a bottle of resin did for my Form1+, I’m perfectly OK with that limitation. It doesn’t cost me anything more, and that seems like the right priority call for FL to have made on this design!


I think silicone would last a very long time - much longer than the rubber valve. And if it needed replacement - it would be very easy and inexpensive to slap in a new tube. Firmware could even insist on replacing the tube after so many liters have been pumped.

I disagree. If the pump doesn’t work, it can’t spill out if it was being pulled above the cartridge fluid level. Unlike when the bite valve fails it is catastrophic. During a recent webinar, the speaker even said that once he had the whole bottle emptied destroying his printer.

It works pretty good - but only if you are using Formlabs resin which I try to avoid. I have also had a number of problems with using even Formlabs resin. I have had it overflow the tank and ask for me to empty it (one of my first prints), and I have had the firmware warn me that there might not be enough resin even though the cartridge was practically brand new and I was printing something quite small.

Anyhow just my ¢2.

I agree with that!

But the pump has multiple failure modes. The motor can die, so the pump can’t pump. The drive wheel can break, so the pump can’t pump. And the tube can split so the pump pumps resin in to the printer instead of in to the tank.

A mechanical pump is fundamentally less reliable than a simple gravity feed. A design off of Thingiverse is going to be fundamentally less reliable than a commercial pump (since otherwise, the commercial pump manufacturers would use the cheap Thingiverse design).

The idea you’re espousing has merit, it would be irrational to deny it. But reliability is something I for one don’t want to sacrifice. A reliable pumping system would add cost. A quick search of the Internet shows that commercial grade peristaltic pumps are >$300. Given that FL has to have some cost margin on the sale of the assembled printer, we’d probably be looking at something north of a $500 adder to the price of the printer for this kind of enhancement.

Not exactly the same thing, but my house, hot water, driveway etc. are all heated from water pumped from my hot spring. The pumps I use last millions of gallons before the bearings in the motor gives out, I run it 24/7 for years before failure. As expensive as resin is, a motor is unlikely to fail before the printer is absolute or other more expensive failures occur.

Also - I don’t think the pump parts would wear out before its useful life.

Then they could make that optional and let the buyer decide if they want to pony up. One failed or defective bite valve could easily cause more than $500 in damage.

Even better yet - they could make the Firmware open source and allow us to add things like this ourselves. I think the number of crowd sourced innovations would make it very beneficial to Formlabs.

I agree, pumps can be very reliable. The circulating pump on my forced-hot-water heating system is 25 years old and still works fine (though I’m losing confidence in it continuing to do so). Those pumps of yours aren’t cheap I bet.

I’m not saying FL couldn’t use a pump, only that a pump that’s reliable enough would add a fair amount of cost. It would only benefit those of us willing to pay more for the ability to work “outside the box”. But the majority of Form2 users aren’t like us. They want an appliance that comes out of the box and works as advertised. They’re not interested or even capable of expanding the operating envelope.

Like anyone who’s bought an Apple product in the last 20 years. :grin:

Making these things optional seems like a decent idea. But in practice it also costs more. The engineering development (NRE) costs are a significant part of the product’s final cost. And when you try to develop a product that’s attractive to all potential market segments, it usually turns out to not be that attractive to any of the target segments.

Open Source has it’s advantages, but not so much for commercial products. MakerBot is a great example of why Open Sourcing a commercial product is a bad idea. Software and Firmware development is generally the most expensive part of a product design process. Having access to the source code and at least one physical example of the product, it becomes trivially easy to clone. Not releasing your code into the public domain protects your Intellectual Property and customer base from clone manufacturers. If I were FL, I would not consider OS to be a good idea at all.

Well actually - Makerbot’s decline started when they decided to make everything proprietary IIRC.

As for clones - I’m not suggesting they open source their substantial hardware engineering. So opening the firmware wouldn’t be practically useful unless someone reversed engineered all the hardware - and at that point I don’t think much would be gained from copying the firmware. In fact, they could open the firmware and even keep PreForm proprietary in addition to the hardware (but we could still add features that way to the firmware - eg. a peristaltic pump).

That way, they would gain crowdsourced improvements while keeping their investment secure.

One problem with using pumps is that they have to be purged every time the resin type is changed. There are some printers around which can use up nearly a whole cartridge just to purge the machine. Even if this was not true of a FormLabs machine having to purge the delivery system would add significantly to the support costs.

It could just purge it back into the vat which has to be manually emptied anyways. Or directly back into the cartridge - same effect.

If you are worried about contamination after the pump has been emptied - in my experience - a few drops of a different type of resin mixed in to a vat with an other, doesn’t cause any discernible side-effects.

If that was still a cause for concern - it could be flushed with alcohol - no need to waste resin to get the pumping clear. :wink:

@3DTOPO: You are absolutely right; for sure, purging could be possible. This is exactly what is done with injection molding machines when they change materials to be processed through the machine: the screw & barrel in the molding machine will have residual material of the molding job that was just completed, and the next material shouldn’t get contaminated by the previous material. You just run some purging compound through the screw & barrel to clean out the previous molding material, then put in your new next material…and off you go.

Same thing for the pump in a Form2: put some alcohol (like you suggested…good idea) or other appropriate medium through the pump to purge it, then off you go with your next print job.

Not to start a flame-war but I don’t think it’s appropriate to compare MBI and FL as apples to apples. They were both small companies in the desktop printer market but that’s about where the similarities end.

Open source was the lifeblood of MBI because every single one of their designs hit the market significantly flawed and the community fixed them for free. MBI needed open source, it wasn’t just a nice value added continuation engineering convenience.

I 110% agree with Randy and FL on their cartridge decision. A vastly improved cartridge system would have alienated many users because of the increased cost of the machine, failure modes, and cost of resin.

Interesting discussion regardless.

I think the reason is FL is excluding a sizable demographic - the tinkerers - by keeping the platform 100% closed and proprietary. If they opened up just the firmware, they would attract more people with that mindset.

Even if tinkerers comprised a small percentage of customers, all it takes is a handful of innovators sharing improvements with the community. I think that would benefit a significant number of the current demographic (and thus FL) because they wouldn’t have to experiment to make improvements. Just download, print and install and follow directions made by others.

Even though you agree with Randy 110%, it was actually him who made the comparison. :wink:

I think by being open source, it attracted a whole different set of users than if it was proprietary.

Also, even though the Form is marketed as a “just works” solution, all you have to do is hang out on these forms and see that learning, experimenting and trouble-shooting is still pretty much a requisite.

I don’t think you are grasping what I have proposed. I think the ideal solution is to make the firmware open source so that anyone that wants to could develop and share improvements such as a pump (which is almost a completely printable part minus some vitamins). This means the machine cost entry point would not change at all.

I think the next best solution is to make it an add on. Users could decide if they want to pay extra or not, and ideally it could be added after the initial purchase as an upgrade.

1 Like

My deepest apologies. Looking like an idiot and hypocrite is a necessary risk when forum skimming during lunch. =]

No apology necessary and I didn’t think you came across as an idiot at all!

  1. I’m definitely not in the least bit flamed by any of this discussion and I sincerely hope no one else feels that way, either! If I’ve said something to offend, I apologize!
  2. I take a different view on MakerBot. MakerBot may have benefited from Open Source, but it was also their downfall. It gave a lot of competitors a big initial boost in the development of their own printers. Not just in terms of how it worked. A big part of product development is also seeing what doesn’t work. It also opened them up to competition from clones. They went “closed source” in their second generation but by then it was too late, the market was already severely diluted by similar competitive products.
  3. FormLabs didn’t design their own SoC and I don’t know of any other 3D printer manufacturer that has. Similar or even identical lasers, galvanometers, stepper motors and controller electronics can be easily sourced on the open market, and you can even find someone to design the box and put it all together for you. A mechanical chassis and a printed circuit board can easily be reverse engineered. The value of the Hardware IP is minimal. It’s the software/firmware IP that costs the most to develop. For a commercial product you’d never want to share that. The approach FL has taken with an API is the right approach. It gives users a higher degree of control over the printer without exposing any proprietary software/firmware.
  4. I agree with everything everyone is saying about a pump. The point I’m trying to make is not that a pump can’t work, but that a pump adds cost and complexity and reduces reliability. It’s sole virtue in the context of this discussion is that it eliminates the need to replace a cartridge because the bite valve wears out. But cartridge resin from FL costs the same as bottled resin from FL. So it only benefits the users who want to run 3rd party resins through a cartridge in Open Mode. I would bet that only a small percentage of FL users fall into this category. The rest would pay more for a feature they don’t need and some would opt out because of the higher price. Adding cost with a pump in order to benefit a small portion of your customer base (who by the way if they use the feature they won’t be buying FL resins, which, let’s be honest, also eats in to profit margin) just doesn’t make good business sense.
  5. It has been my experience that you have to be careful about judging a user base from the users present in a forum. Users who have nothing to ask or complain about tend to be the silent majority. It’s only the people looking to do something “out of the box” or who are having an issue that tend to participate. We can quantify the level of activity on the forum but we don’t know what the “denominator” is. If FL has only sold a couple of 1000 printers maybe there’s enough forum chatter on the Open Mode related subjects to suggest FL has missed a key customer demographic. If they’ve sold 10Ks of printers, the users on this forum looking for more control represent a small percentage and the vast majority are perfectly happy with their printers as-is, suggesting FL has hit their demographic spot-on. We just can’t say one way or the other with what we know.
1 Like