So should I just throw away the 3d printed parts from this project?

[Justin]

It seems the project is dead.

[Douglas Pearless]

Hi Justin,

The issue isn’t 3D printing (which I have working to a restricted volume of print about X=0, Y=0 & Z=0), but the correction of the real-time calculations to make the mechanism accurate.

I have the scientific paper which describes how to do that by the inventor of the rotary delta mechanism and I have had a maths and physics whizz review it and provide a solution, but I simply have not had the time to code it into Smoothie.

This affects all Rotary Delta mechanisms, whether implemented in Smoothie or Marlin :-(

If you are up to writing some code, I can send you the paper :-) otherwise this will wait until later this year when I may have time to finally implement it at which time FirePick will work correctly and we can move forward again :-)

Cheers

Douglas

On 16/08/2017, at 1:21 AM, Justin <liquidc...@gmail.com> wrote:

It seems the project is dead.

--
You received this message because you are subscribed to the Google Groups "FirePick" group.
To unsubscribe from this group and stop receiving emails from it, send an email to firepick+u...@googlegroups.com.
To post to this group, send email to fire...@googlegroups.com.
Visit this group at https://groups.google.com/group/firepick.
For more options, visit https://groups.google.com/d/optout.

[Gmail]

Hi Douglas,

I would be interested in the paper. I am curious if it takes effector tilt into account and if so how it determines there is tilt. 

Sent by iPhone

[Dustin Kasel]

Personally I tossed the plastic parts and used the the extrusions, motors, sensors, etc for a cartesian machine. Go on over to OpenPnP for tons of information. This project may not have ever been alive, more like a mirage off in the distance.

[Douglas Pearless]

Mine is sitting on the corner of my desk awaiting updated firmware to be written :-)

[Lamont Cranston]

On Aug 15, 2017, at 1:44 PM, Douglas Pearless <douglas....@gmail.com> wrote:

> I have the scientific paper which describes how to do that

What is the citation (or DOI) for the paper? I'm curious to read it.

[Douglas Pearless]

[Bruce Bock]

Douglas,

How did this go? Did you have time to complete this?

[Douglas Pearless]

Well, it might be time to breathe life back into this project; I am discussing the implementation of the calibration routine with another person, and if this can be done, then the technical issues with the Rotary Delta mechanism will largely be solved and if others are interested, then the project can restart ...

Cheers

Douglas

[Joshua Pritt]

Sweet! What sort of things were left to do?

--

[Douglas Pearless]

Foremost we must implement the calibration routine in Smoothie so that the X-Y-Z does not suffer from the bowl effect and then we have an accurate and fast mechanism.

I have restarted the conversation in the Slack hw_motionctrl  channel for firepick.

Let me know if you have lost the link / access.

Cheers

Douglas

[Douglas Pearless]

The on-going conversations for getting the calibration working are going (to be) posted in our slack channel => https://firepick.slack.com/messages/C0MFAASF6 

On Tuesday, 6 November 2018 18:19:44 UTC+13, Douglas Pearless wrote:

Foremost we must implement the calibration routine in Smoothie so that the X-Y-Z does not suffer from the bowl effect and then we have an accurate and fast mechanism.

I have restarted the conversation in the Slack hw_motionctrl  channel for firepick.

Let me know if you have lost the link / access.

Cheers

Douglas

On 6/11/2018, at 17:48, Joshua Pritt <ramg...@gmail.com> wrote:

Sweet! What sort of things were left to do?

On Mon, Nov 5, 2018, 5:19 PM Douglas Pearless <douglas....@gmail.com wrote:

Well, it might be time to breathe life back into this project; I am discussing the implementation of the calibration routine with another person, and if this can be done, then the technical issues with the Rotary Delta mechanism will largely be solved and if others are interested, then the project can restart ...

Cheers

Douglas

On Wednesday, 16 August 2017 01:21:02 UTC+12, Justin wrote:

It seems the project is dead.

--
You received this message because you are subscribed to the Google Groups "FirePick" group.

To unsubscribe from this group and stop receiving emails from it, send an email to firepick+unsubscribe@googlegroups.com.

To post to this group, send email to fire...@googlegroups.com.
Visit this group at https://groups.google.com/group/firepick.
For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "FirePick" group.

To unsubscribe from this group and stop receiving emails from it, send an email to firepick+unsubscribe@googlegroups.com.

[Charles Steinkuehler]

The slack page doesn't appear to be public. They want me to ask the
workspace administrator for an invitation, but don't provide a link.

Can I get an invite, or is the slack thing a closed group?

>>> email to firepick+u...@googlegroups.com.

>>> To post to this group, send email to fire...@googlegroups.com.
>>> Visit this group at https://groups.google.com/group/firepick.
>>> For more options, visit https://groups.google.com/d/optout.
>>>
>> --
>> You received this message because you are subscribed to the Google Groups
>> "FirePick" group.
>> To unsubscribe from this group and stop receiving emails from it, send an

>> email to firepick+u...@googlegroups.com.

>> To post to this group, send email to fire...@googlegroups.com.
>> Visit this group at https://groups.google.com/group/firepick.
>> For more options, visit https://groups.google.com/d/optout.
>>
>>
>


--

Charles Steinkuehler
cha...@steinkuehler.net

[Douglas Pearless]

Invitation sent to you :-)

[Charles Steinkuehler]

Got it. Thanks!

>> cha...@steinkuehler.net <mailto:cha...@steinkuehler.net>

>>
>> --
>> You received this message because you are subscribed to the Google Groups "FirePick" group.
>> To unsubscribe from this group and stop receiving emails from it, send an email to firepick+u...@googlegroups.com.
>> To post to this group, send email to fire...@googlegroups.com.
>> Visit this group at https://groups.google.com/group/firepick.
>> For more options, visit https://groups.google.com/d/optout.
>

--
Charles Steinkuehler
cha...@steinkuehler.net

[Peter Shabino]

Can I get an invite. 

Thanks,

Peter

My projects:
http://www.wire2wire.org/

---

From: fire...@googlegroups.com <fire...@googlegroups.com> on behalf of Douglas Pearless <douglas....@gmail.com>
Sent: Friday, November 9, 2018 8:50 PM
To: FirePick
Subject: Re: [FirePick] Re: So should I just throw away the 3d printed parts from this project?

 

To unsubscribe from this group and stop receiving emails from it, send an email to firepick+u...@googlegroups.com.

[Douglas Pearless]

Of course,  I will send one later today when I am back home.

Cheers

Douglas

Sent from my iPhone

[Douglas Pearless]

Try this link https://firepick.slack.com/messages/C0MFAASF6 

On Wednesday, 16 August 2017 01:21:02 UTC+12, Justin wrote:

[Ristola]

Just thinking about the FirePick the other day and was wondering if any progress was made.

Can I get an invite also to the Slack Group

thanks

[Teddy C]

Can i get a invite to the slack channel please.

[Peter Shabino]

The firepick slack looks to be dead. Last message on there was from me July 4th 2019 asking if it was dead. Before that the last message was from Nov 2018. 

Have mine assembled and have the smoothie upgrade for it. Not sure what to do with it now. 

Later,
Wire

My projects:
http://www.wire2wire.org/

---

From: fire...@googlegroups.com <fire...@googlegroups.com> on behalf of Teddy C <teddy.c...@gmail.com>
Sent: Tuesday, July 7, 2020 5:21 PM
To: FirePick <fire...@googlegroups.com>
Subject: [FirePick] Re: So should I just throw away the 3d printed parts from this project?

 

--
You received this message because you are subscribed to the Google Groups "FirePick" group.
To unsubscribe from this group and stop receiving emails from it, send an email to firepick+u...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/firepick/10071bd3-3688-40d1-9da1-65d8f5252468o%40googlegroups.com.

[Family First]

Peter, 

Why doing you send up some detailed pictures of the doggone thing.  I have no idea how to place the components.  Pictures of the assembled unit would be wonderful.  Also rotational deltas are now applicable to Duet firmware.  My advice, get a duet maestro (90 bucks) I will assemble mine in the next three weeks and we can Pheonix the firepick!  my email address is lawrenece.r...@google.com  

The firepick was a bit ahead of it's time and it was also not popular since the prismatic deltas hit the scene, but I want to get mine running and embarrass some folks with it so.  Please post those pictures.  

To view this discussion on the web visit https://groups.google.com/d/msgid/firepick/CH2PR07MB6758EEA86971592D89F67B25A4640%40CH2PR07MB6758.namprd07.prod.outlook.com.

[Michael Anton]

Assembly instructions can be found here: https://firepickdelta.dozuki.com/.  More info can be found here: https://www.firepick.org/.  You also might find it useful to buy a copy of Alibre so that you can load up the 3D CAD drawings and assemblies, and make modifications.  This is also a good way to find out where all the parts are located, and how everything fits together.

I have to mention that there are good reasons why this project died (I was a contributor to the project in minor ways, and I have a beta kit that is now in pieces ).  Ultimately, in my opinion, the design was based on a bunch of false assumptions, and that is largely why it doesn't work.  Lots of effort was put into trying to make it work, including building a fully machined unit with no plastic parts.  Alas, that didn't function any better.  Some commercial rotary delta pick and place machines might be good enough for SMT assembly, but they are fully closed loop systems.  It is extremely difficult to make a rotary delta machine as accurate as what can be done easily in a Cartesian based machine.  Even the linear delta style that has become common in the 3D printing community is much more difficult to calibrate, and get to the same level of accuracy as Cartesian machines.  This is likely why no commercial SMT PnPs use a delta configuration.  There was a commercial SCARA machine around though, so that was the one outlier.

Also note that the FPD is based on a branch of OpenPnP, that to my knowledge is no longer supported.  OpenPnP has come a long way since the FPD days, and it really only support Cartesian machines.  I'd suggest you look at the OpenPnP Google group here: https://groups.google.com/forum/#!forum/openpnp, and the homepage here: http://openpnp.org/ to see what kinds of machines other people are building.

Attempting to build and FPD to do SMT PnP will likely result in just a lot frustration with little benefit to you, just like it did for the rest of us.  If you do want to build one though, I may be able to supply you with some of the hard to get parts.  Let me know and we can work something out.

The project is officially dead now, and nobody to my knowledge is working on it at all.  Most of us gave up on it years ago, and moved on to other solutions.  I've purchase two commercial machines since then, and have probably built a couple thousand boards in that time.  You can find some videos of my machines here: https://www.youtube.com/channel/UC7sg74_D2iV2wDiK6THBmBQ if you are interested.  In hindsight, the FPD never would have been able to do what I needed the machine to do anyhow, as it just wasn't large enough, but it was fun to work on at the time, and I did manage to exhibit it in a semi working state at a Makerfaire: http://manton.ca/blog:calgary-mini-maker-faire-2015.

If you want to know more about the adventures of the FPD team from the inside, Jason the lead developer over on the OpenPnP group may be inclined to tell some of that story.  I'm sure he would likely provide you with the same cautionary tale that I am.

On Thursday, July 9, 2020 at 7:45:43 AM UTC-6, Family First wrote:

Peter, 

Why doing you send up some detailed pictures of the doggone thing.  I have no idea how to place the components.  Pictures of the assembled unit would be wonderful.  Also rotational deltas are now applicable to Duet firmware.  My advice, get a duet maestro (90 bucks) I will assemble mine in the next three weeks and we can Pheonix the firepick!  my email address is lawrenece....@google.com  

The firepick was a bit ahead of it's time and it was also not popular since the prismatic deltas hit the scene, but I want to get mine running and embarrass some folks with it so.  Please post those pictures.  

On Thu, Jul 9, 2020 at 9:39 AM Peter Shabino <wi...@hotmail.com> wrote:

The firepick slack looks to be dead. Last message on there was from me July 4th 2019 asking if it was dead. Before that the last message was from Nov 2018. 

Have mine assembled and have the smoothie upgrade for it. Not sure what to do with it now. 

Later,
Wire

My projects:
http://www.wire2wire.org/

---

From: fire...@googlegroups.com <fire...@googlegroups.com> on behalf of Teddy C <teddy.c...@gmail.com>
Sent: Tuesday, July 7, 2020 5:21 PM
To: FirePick <fire...@googlegroups.com>
Subject: [FirePick] Re: So should I just throw away the 3d printed parts from this project?

 

Can i get a invite to the slack channel please.

On Tuesday, July 7, 2020 at 1:22:59 PM UTC+10, Ristola wrote:

Just thinking about the FirePick the other day and was wondering if any progress was made.

Can I get an invite also to the Slack Group

thanks

On Tuesday, August 15, 2017 at 9:21:02 AM UTC-4, Justin wrote:

It seems the project is dead.

--
You received this message because you are subscribed to the Google Groups "FirePick" group.

To unsubscribe from this group and stop receiving emails from it, send an email to fire...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/firepick/10071bd3-3688-40d1-9da1-65d8f5252468o%40googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "FirePick" group.

To unsubscribe from this group and stop receiving emails from it, send an email to fire...@googlegroups.com.

[Lawrence Family First]

I am so intrigued by your story!  What were the false assumptions that made the design fail?  Also what cartesians are you referring to i see h bots or core xy machines in the pnp documentation. 

Sent from my iPhone

To unsubscribe from this group and stop receiving emails from it, send an email to firepick+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/firepick/fc445289-d784-48d9-83db-cc8bc25edf10o%40googlegroups.com.

[Michael Anton]

Well, Core XY, and H Bots are both Cartesian machines for the most part, since they have traditional X, Y and Z axis, only the drive mechanism is a little different.  Many of the OpenPnP builds are neither of these, and have separate X, Y, and Z motors, with Y usually driven by a pair of belts on a common drive axis, or sometimes with two motors.

False assumptions as I see them:

1) That 16X microstepping really gives increased resolution.  The general rule of thumb, is that anything over 8 to 10X microstepping gives no additional resolution, and only yields smoother motion.  See this: https://www.machinedesign.com/archive/article/21812154/microstepping-myths, and this: https://www.geckodrive.com/support/step-motor-basics/accuracy-and-resolution.html, for additional information.  In the FPD, it might be even worse, as the load on the motors changes with position.  On a traditional Cartesian system, only Z has to support any continuous force, whereas the other two axis, only have load on the motors during movement, and after that, little torque is actually required.  Variable torque requirements make getting increased resolution from microstepping much more difficult.  Alas, we actually need the most precision in X, and Y.  Z can be fairly inaccurate, as one should use spring loaded nozzles anyhow (something the FPD didn't bother with, though it is fairly easy to do now) to make up for other errors in Z, like variable component height, and PCB height changes.  You could probably tolerate a 0.5mm error in Z, though, that might make picking parts up more difficult at times, especially if the tape is allowed to move.

2) That any mechanical errors can be compensated for in software.  Because this is a rotational delta, these errors will be difficult to compensate for.  We're talking things like joint slop, hysteresis, backlash, etc.  There are so many sources of error that are possible, and it would be best to fix these mechanically, rather than in software if possible.  The original design has all sorts of places where mechanical errors could creep in.  The kits that were sold were pretty terrible from a precision standpoint.  Only two of the delta arms I received were the same length, but 4 of them were fairly close.  The other two were out by 0.3, and 0.13mm respectively (the linear delta guys shoot for better than 0.05mm error here), which is not too bad compared to what others reported, but the result is that the carriage will tilt as it moves, and not stay parallel to the bed.  A design problem example, is building the arm pivots out of 6 separate parts, and relying on bolt holes to get them aligned properly.  I improved this somewhat by making it 3 parts, that were printed at the correct angle to one another, so that it would be more likely to hit the 120 degree included angle between them, but even then there would be lots of sources of error.  Had these been a single part, alignment might have been pretty good, but printing them would have been difficult.  Truthfully, I think if you wanted to build a delta printer, using the linear style delta would be less error prone, and probably work much better, especially since they seem to work fairly well for 3D printing.

3) That it would be possible to build something really inexpensive and precise at the same time.  This generally doesn't hold, and the smaller the parts are that you want to place, usually means that you need to increase the mass of the machine to keep it stable enough to do so.  I suggested to Neil a number of times, that he should have just thrown money at it initially to get it working, rather than trying to make it inexpensive at the start.  As a result, he kept needing to add things to bring it to a level to get it sort of working, when it would have been better to simplify it later once it works.  He didn't need to shoot for a $300 price tag at the start, as anything even close was probably at least $2000, and that would have helped getting something precise much easier.

4) While a delta design is cool, and offers some advantages, the whole design is easier to do as a Cartesian.  Why try to come up with a completely new way to do this task, when other more proven methods just work?

That's all I can think of just now, this many years later.  For me, the critical one is #2, since when I first heard about the project, that was the one that I suspected was likely to cause the project to fail, and as it turned out, that was the part that they never could get working right.  At the time though, I figured that Neil was pretty dedicated to the project, so he might find a way to do it through sheer will, but eventually the project just got the best of him, and he had to move onto other things to make a living.  Since he was pretty integral to the entire project, it just never went very far after that, even though a number of people continued to work on various aspects of it.

It's too bad the original hackaday.io project has been taken down, as that detailed a lot of the design process that Neil went through to come up with the design.  That would have been nice to keep somewhere from a historical perspective, so that if someone wanted to resurrect the project, they could at least go back to the initial assumptions, and see if they should be modified. 

To view this discussion on the web visit https://groups.google.com/d/msgid/firepick/fc445289-d784-48d9-83db-cc8bc25edf10o%40googlegroups.com.

[Joshua Pritt]

I found the hackaday project page on the wayback machine!

Check out all the juicy details:

http://web.archive.org/web/20170331113633/https://hackaday.io/project/963-firepick-delta-the-open-source-microfactory 

 

To unsubscribe from this group and stop receiving emails from it, send an email to firepick+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/firepick/af9613f7-2a66-477a-a698-241c030ef2bco%40googlegroups.com.

[Michael Anton]

Thanks for finding that.  I'd forgotten how much detail Neil went in to.

[Donald Papp]

I had originally assumed that compensating for errors in software would be done by the vision system, but later learned it doesn't work that way.

I put some work towards a crazy idea of embedding a littrow prism into the head with a camera in an effort to give the machine some awareness of physically where the head was positioned at all times, so that a software loop of "go more or less to right coordinates - OK I see the right pads - bit more to the left - bit more to the left - OK good, now down" would be possible, but I only got far enough to learn of more problems, and by then the project had run out of steam.

--
You received this message because you are subscribed to the Google Groups "FirePick" group.
To unsubscribe from this group and stop receiving emails from it, send an email to firepick+u...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/firepick/d28db441-685a-4db9-97a3-42abfb697c00o%40googlegroups.com.

[Family First]

I know that this project started as a pnp but with the slop you all are talking about I wonder if it would lend itself to non-repeatable painting or drawing.  

To view this discussion on the web visit https://groups.google.com/d/msgid/firepick/CADQDMy_j3R6_TC81icN-eE3JShhgprEwuNUpVLNK0%3DGW5i1pTA%40mail.gmail.com.

[Luiz doleron]

Thank you Michael Manton for the detailed explanation.

I studied for a while the forward and reverse kinematics of deltas and build a firepick myself 1 or 2 years ago in addition to other delta models.

I gave up after getting frustrated with my achieved accuracy. however these last two weeks I decided to get another try. In particular, I am working right now on:

- a new kinematics using 4 motors instead 3

- new software calibration to compensate unbalanced arms 

The idea to use 4 motors is not new nort mine. There are commercial deltas with 4 arms. I believe this can improve repeatability. I am on the way to check it. For while I am developing the Fk and IK equations considering unbalanced arms.

To unsubscribe from this group and stop receiving emails from it, send an email to firepick+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/firepick/af9613f7-2a66-477a-a698-241c030ef2bco%40googlegroups.com.

--

Luiz Carlos d´Oleron

[Family First]

I like the idea of using compensation in some way.  I am not a programmer or math guy but I know that the duet hardware uses a few different leveling techniques that will actually calculate the length of teh arms.  Would someone please send me  a video or series of pix of their machines and the assembly.  I know that the older stuff never really got it for me. 

To view this discussion on the web visit https://groups.google.com/d/msgid/firepick/CAAkrH3hrkHcxck0ZB1R3uJiP10M_cO_pjqbED1uPpR%2BQS_d4qg%40mail.gmail.com.

[Luiz doleron]

In order to get helpful feedback, let me share my current status on performing delta manipulators calibration by software:

# Assumptions

The theoretical E(x, y, z) position of the End-effector is a function of the position (theta1, theta2, theta3) of rotational joints given the robot geometry:

E(theta) = F(theta; geometry)

This function is called Forward-Kinematic which basically maps the entry values (theta1, theta2, theta3) in the array (x, y, z). The robot geometry can be modelled as a set of control points: (arms length, home position, angle between fore-arms, distance for ground, ...)

Considering an actual end-effector position A(x, y, z) we can define the error (A, E) by the Euclidean distance:

Error = sqrt((Ax - Ex)^2 + (Ay - Ey)^2 + (Az - Ez)^2)

Considering the following error sources:

- Assembly issues: such as backlash, misposition, misalign, ...
- Material hysteresis: physical properties changing over time, like linkage length and motor stepping errors
- Dynamics and kinematics phenomena: positional errors due the acceleration and deceleration, inertia, etc
- Numerical float point approximation: rounding, trunk and so on
- Numerical controller errors: program bugs
- Misconfiguration: incorrect assumptions about physical properties such as an wrong arm length or angle
- Measurement errors: experimental errors during the robot position acquisition

# Robot configuration

Configuration is a finite set of parameters which represents the robot geometry: [a_0, a_1, ...]. For example, very often firmwares assume that the rods have equal length being 200-220mm popular values for it.Failing to set those rods values accurately has a hard influence in Deltas.

# Calibration

Given the Central Limit Theorem, the sum of those error sources listed above has a Normal Distribution shape with mean m and deviation d: N(m, d). The Calibration Process aims to reduce either m and d to acceptable values. Considering that the robot program is done and the robot is assembled already, the Robot Configuration is the only source of error we can actually deal with during the calibration process.

Thus, the process of calibration can be taken as a Search Problem in the robot parameters vector space [a_0, a_1, ...]. Said that, the calibration can be defined as:

Find the configuration [â_0, â_1, ...] which minimizes m and d from N(m, d)

This minimization can be achieved by using a set of independent and identically distributed observations [e_0, e_1, ...] and an optimization algorithm for quadratic cost minimization.

Note that these observations must be gathered by a feedback system (sensors) in order to provide the readings back to the calibration routine. It is similar in some way to bed-level calibration in cartesians printers.

Once the calibration program has the observations, it can perform the search using the robot current settings as initial values. Then, the robot is configured with the new estimated set parameter that minimizes the error distribution.

I'm still working on which specific type of algorithm performs better on this space and general purpose hyper parametrization. In my current status this approach shows as promising. Unfortunately due to social isolation and everything related to this pandemy I'm without access to the lab or even to a delta so no way to perform controlled experiments right now. If someone is interested, has deltas and cyber-physics experimental skills feel free to drop me a message or get in touch.

Luiz.

To view this discussion on the web visit https://groups.google.com/d/msgid/firepick/CAKnP0BeNoNdSVwQRo%3De8_wwzCpK%3DiCF%3DbGd0wBK6Pp%3DheeaOxA%40mail.gmail.com.

[Michael Anton]

The problem as I see it, is how do you measure where your nozzle tip is in 3D space to be able to do the measurements?  It is not sufficient to just measure it against a plane (the bed), as components to be placed have thickness, so really all positions up to the thickness of the thickest part you want to place also need to be accurate in X and Y.  Note that accuracy in Z is somewhat less important, if you use spring loaded nozzle tips.

To view this discussion on the web visit https://groups.google.com/d/msgid/firepick/af9613f7-2a66-477a-a698-241c030ef2bco%40googlegroups.com.

--

Luiz Carlos d´Oleron

--
You received this message because you are subscribed to the Google Groups "FirePick" group.

To unsubscribe from this group and stop receiving emails from it, send an email to fire...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/firepick/CAAkrH3hrkHcxck0ZB1R3uJiP10M_cO_pjqbED1uPpR%2BQS_d4qg%40mail.gmail.com.

--
You received this message because you are subscribed to the Google Groups "FirePick" group.

To unsubscribe from this group and stop receiving emails from it, send an email to fire...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/firepick/CAKnP0BeNoNdSVwQRo%3De8_wwzCpK%3DiCF%3DbGd0wBK6Pp%3DheeaOxA%40mail.gmail.com.

[Michael Anton]

@Family First: Are the Dozuki docs that I published a link to earlier at FPD Assembly not good enough to allow you to assemble your machine?  That's what I built mine based on, and they are pretty good.  I suspect that many of us no longer have machines, since the project has been dead for many years.  I took most of mine apart years ago, so I can't help you much regarding pictures or video, other than to try to answer questions when you get stuck.  I think the only part that I have that is still assembled, is the rotary delta mechanism at the top of the machine.

To view this discussion on the web visit https://groups.google.com/d/msgid/firepick/af9613f7-2a66-477a-a698-241c030ef2bco%40googlegroups.com.

--

Luiz Carlos d´Oleron

--
You received this message because you are subscribed to the Google Groups "FirePick" group.

To unsubscribe from this group and stop receiving emails from it, send an email to fire...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/firepick/CAAkrH3hrkHcxck0ZB1R3uJiP10M_cO_pjqbED1uPpR%2BQS_d4qg%40mail.gmail.com.

[Joshua Pritt]

Speaking of the error correction not being good enough by just measuring to the bed in order to manage PCB and component height, would it help if we had a pre-calibrated metal spacer of a given height for it to compare the steps to the bed (auto-leveling) vs steps to the calibration spacer?  Then it would have lots of nice numbers to save into the error correction and leveling functions, right?

To unsubscribe from this group and stop receiving emails from it, send an email to firepick+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/firepick/5dc9ce52-9abb-41b5-8acb-f5cdc98737beo%40googlegroups.com.

[Michael Anton]

Or, maybe use a set of four precision spacers, and raise the entire bed.  Then it could be probed at multiple elevations.  Truthfully, the calibrated working range doesn't probably have to be that high.  Just as high as the highest component that you want to place should be enough, providing that the feeders are at the same height as the bed.

I'd suggest that anyone that really wants to get this working should abandon the arms that Tinwhiskers supplied in the kit, and instead get a set of Haydn Huntley's arms.  He sells these in precisely matched sets, and they use magnet/delrin ends.  The machine would probably be much stiffer using these.  The only issue might be that the balls are probably not in the optimum position for using these arms, but that probably wouldn't be too hard to fix.

To view this discussion on the web visit https://groups.google.com/d/msgid/firepick/5dc9ce52-9abb-41b5-8acb-f5cdc98737beo%40googlegroups.com.

[Reef Morse]

I'm in the process of turning my kit into a sample changer for a scientific instrument (an electron spin resonance spectrometer to be precise).  I tossed the legs and mounted the remaining piece on a gantry, added a stepper to move the whole think back and forth, and have been working on developing software to allow it to fetch samples (100 mm long capillary tubes about 1 mm in diameter), place them in the spectrometer, retrieve them, and go on to the next.  

I've been unable to work on it for a couple of years but I'm starting back up again and would enjoy conversing with anyone still considering doing something with their FPD parts.

Reef Morse

[Family First]

Sounds great.  THe system seems to lend itself to many flexible applications.

--

You received this message because you are subscribed to the Google Groups "FirePick" group.

To unsubscribe from this group and stop receiving emails from it, send an email to firepick+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/firepick/a3073841-aa94-48bb-b46d-b761de2a0d8en%40googlegroups.com.