Drip Governor - A most valuble first hack

[rylangrayston]

@cephdon and @mike_biddell I Think that a stepper / encoded motor and a pump is almost certainly a better way to do it, however at peachy we must also focus on a method that is inexpesive enught to have a chance at making it into the basic 100 dollar kit, thus the focus on a simple valve system.

Well hopefully "pictures are more than words" works for this post.
I spent quite a bit of time at the lathe today, here are some pics, ill explain more after i have tested these new parts.

The openscad code comes to fruition( it worked! but needs modification to work under syphen presure)
20140315_175758.jpg 20140316_014036.jpg 20140315_012144.jpg 20140315_012248.jpg

[rylangrayston]

And a new small one done in aluminum:

20140315_202600.jpg 20140315_203902.jpg 20140316_020726.jpg 20140316_020928.jpg


Great Work Everyone!!

[mike_biddell]

It is easy to get carried away and forget this has to sell at $100. If we set the valve to have a relatively high flow rate, retain the drop count system, and use the laser to turn the valve ON and OFF (555 timer circuit), we would have a very effective and cheap system. With the high flow rate, each drop will be a consistent size and hence Z wud have good accuracy. So do not allow the orifice to be altered at all and just switch ON/OFF more frequently for fine Z. This overcomes drop size variation and allows very accurate calibration.

[Slatye]

Do we actually need to control the flow rate? I'm wondering whether a better option would just be a float (can be printed) and a potentiometer (under $1) to give an absolute measurement of water level. Then the Peachy can adjust its own speed to match however fast the water is rising, as long as it's within reasonable limits. For feeding back to the computer, just using the potentiometer to drive a 555 timer should work perfectly.

This would also mean that the container shape really doesn't matter. Slightly angled sides (eg. chinese food containers) would mean that the print slows down towards the top (more water volume required to raise the print height by one layer) but since you've got a reliable measurement of the true height it's easy to compensate.

[cephdon]

So, after looking around a bit, I think it is possible to have something land in the $50 range as an add-on with a motor and pump. That would make it a ~$150 kit instead of $100. Still, I understand about the cost sensitivity.

[mike_biddell]

Attachment 714here's the 555 light toggle switch. Just substitute a FET of your choice and the valve as load on pin 3 in place of the 2n2222. This circuit also works well in the simulator. Unfortunately the allowed file sizes are rather small. so it's only just legible. Compared with the first circuit, it's completely analogue.

Oh, forgot to say, to allow light threshold switch point tweaking on the first transistor, change the fixed 10k (R1) to a pot.

[Anuvin]

Ask 10 engineers for a solution...

I have never spent so much of my time thinking about valves. It's tempting to use stepper motors with the peachy printer, in similar and entirely different methods than the ones presented by cephdon, but electric motors are troublesome devils. Look into motor issues with 3d printing, you will find plenty of scary stuff. One of my favorite things about this project is that it contains no motors. That feature alone differentiates the Peachy printer from all others, and I intend to maintain that perspective for as long as is reasonable.

Is it possible to use more than one drip counter on the mic line? If so, we don't have to bother with flow rate, motors, OR in-line valve mechanisms. We would still need to make a valve, but it can be in the reservoir itself, alleviating some typical leak problems that you get with in line valves. Same peachy printer, just with 4 reservoirs instead of 1, or a single reservoir with 4 outlets. If each line were set for sequentially faster drip rates, you can combine them for a large variety of speeds. For example, if line 1 drips 1 per second, line 2 drips 2 per second, line 3 drips 4 per second, and line 4 drips 8 per second, you get every speed from 1 to 15 drips per second. 15 steps of speed is good enough for me, I should think.

I think a pot and float is a really good idea. My fear is that it would take too long to respond. A float is a bit reactionary, compared to counting the drips. Then again, drips are inconsistent, so that may be less accurate than a float/pot could be. Please share any designs for the float you may have, I know I'd like to see them.

[mike_biddell]

Ask 10 engineers for a solution...

I have never spent so much of my time thinking about valves. It's tempting to use stepper motors with the peachy printer, in similar and entirely different methods than the ones presented by cephdon, but electric motors are troublesome devils. Look into motor issues with 3d printing, you will find plenty of scary stuff. One of my favorite things about this project is that it contains no motors. That feature alone differentiates the Peachy printer from all others, and I intend to maintain that perspective for as long as is reasonable.

Is it possible to use more than one drip counter on the mic line? If so, we don't have to bother with flow rate, motors, OR in-line valve mechanisms. We would still need to make a valve, but it can be in the reservoir itself, alleviating some typical leak problems that you get with in line valves. Same peachy printer, just with 4 reservoirs instead of 1, or a single reservoir with 4 outlets. If each line were set for sequentially faster drip rates, you can combine them for a large variety of speeds. For example, if line 1 drips 1 per second, line 2 drips 2 per second, line 3 drips 4 per second, and line 4 drips 8 per second, you get every speed from 1 to 15 drips per second. 15 steps of speed is good enough for me, I should think.

I think a pot and float is a really good idea. My fear is that it would take too long to respond. A float is a bit reactionary, compared to counting the drips. Then again, drips are inconsistent, so that may be less accurate than a float/pot could be. Please share any designs for the float you may have, I know I'd like to see them.

Just how can we get the float reading back to the software, without changing the design to digital? Also, the resolution of a float and a potentiometer would be poor and very unreliable IMHO. We only have to remember the volume controls on our old audio devices crackling as they lose contact with the potentiometer track. A much more reliable measure of depth and which would have good resolution, is an ultrasonic transducer, bouncing a sound wave off the surface of the liquid. Same problem though, no way of getting the reading back with an analogue Peachy.

[Slatye]

Just how can we get the float reading back to the software, without changing the design to digital?

Easy! You just use the potentiometer as one of the resistors (or potentially both of the resistors) driving an astable 555 timer, with the output going to the microphone input. Measure the frequency of that signal and you've got the height. Or, if you can persuade the microphone to just act as a straight ADC, you can put the potentiometer in voltage divider mode and read off voltages directly.

Pots have a fair bit of friction, which would require a long lever and/or a big float to compensate. The other easy (and cheap) option I can see is a 2-axis analogue accelerometer mounted on the arm; that tells you the precise angle of the arm and gives a fair bit of information about whether there are waves in the liquid, whether the Peachy is being moved, etc. An encoder would work too, but high-resolution encoders are expensive.

Also, the resolution of a float and a potentiometer would be poor and very unreliable IMHO.

Yes, I'm not sure that you'd get the sort of resolution required for this job. You should definitely get ~250 steps that can be measured consistently, but with 0.1mm layers that's only a 25mm high object - pretty tiny. Ideally you'd want more like ten times that, but 2500 repeatable steps from a cheap potentiometer is a big ask.

We only have to remember the volume controls on our old audio devices crackling as they lose contact with the potentiometer track. A much more reliable measure of depth and which would have good resolution, is an ultrasonic transducer, bouncing a sound wave off the surface of the liquid. Same problem though, no way of getting the reading back with an analogue Peachy.

I thought about ultrasonics, but the ones I've used have a resolution on the order of 1cm - definitely nowhere near enough for a 3D printer. They also tend to be fairly expensive.

Ideally you'd use either a pressure sensor or a direct liquid level sensor. Unfortunately even cheap pressure sensors are expensive and super-high-resolution ones that can cope with potentially nasty substances are very expensive. Liquid level sensors like this are nice, but again resolution and cost would be an issue.

Edit: here's the float idea. For the accelerometer version you'd just replace the pot with a bearing and glue the accelerometer anywhere along the arm.

Peachy_container_asm.jpg

[mike_biddell]

Slayte..... now I see where you're coming from !!!! But for my part, I think Rylans valve set to large drips (unalterable) and the 555 light switch will give us all we want (fine Z, coarse Z and combinations of both) all under software control. The components and PCB and valve together, probably under $10. No changes whatsoever to the main Peachy. I would be very happy to receive a Peachy with that config.