What is the 1F future? We need to know

Machines (Original/PRO)
1

After a few months absence from this forum dealing with many other issues, I’ve now returned and found myself wondering many things, mostly about where 1F is headed. I fully understand the challenges of a small company and abundant opportunities and problems. As we’ve all got many thousands of dollars in our machines I think we should have some insight into the 1F future.

I started out building my own CNC using the MPCNC project as a starting point. 3D printed parts, running a 3D printer controller. Lacked a lot in rigidity but I was able to make a LOT of parts with it. I went with 1F when I discovered their approach to the mechanicals which promised compact, high performance (accurate) production. I ordered the 16 X 16, then got the X-50 upgrade so have an entirely non-standard 48 X 16 machine which is perfect for my shop. I saw lots of problems with the 1.0.9 controller release and used the 1.1.1 instead which is still on my machine.

I still love the mechanical setup and am getting great results. However…

My questions and concerns:

  1. It is pretty clear that at some point 1F will have to EOL (End Of LIfe) the BB controller. I’m sure they found the fork from BB to be lots more complicated and expensive than estimated, plus their expertise is in the hardware. I’m NOT opposed to this, but REQUIRE an upgrade path. If this happens you’ve got to have a stable release that we can live on.

I am extremely concerned that it’s going backward. I have extensive experience with RPI and have never seen a requirement from any RPI-based system for one type of off-brand wifi dongle. This means you’ve messed with Raspbian as this should be entirely transparent. I posted a lot about getting a dongle running and setting a Samba share since it’s a RPI system, but am very concerned that 1F has gone their own fork on the RPI environment also. This is bad strategy.

  1. Why is Masso the best solution for the future? With the Elite rollout there are scads of problems with the controller. What else did 1F consider in their deliberations? And as above, it seems that 1F will want to support only one controller. Again, I’m not opposed to this but I need an upgrade path that works and is as seamless as possible.

In other words, I don’t want to have great hardware left as a boat anchor with a broken BB/RPI fork and no path to anything else. Sure seems that when funds get even tighter in a dropping economy that this might be a problem for us.

  1. Will 1F target small business even more in the future, leaving us hobbyist/entrepreneur types behind? If I were running the company I’d certainly consider it because that’s higher revenue with upgrades, etc. and with tight money the hobbyist may be SOL.

Why am I worried about a dropping economy? Lead times for new 1F machines are coming down rapidly. I don’t have 1F insight but have to assume that this is because orders are falling. That doesn’t mean we’re screwed, but focusing on one best solution makes the most sense for all of us.

Why do I bring this up? Because this is a perfect machine for me and I need to track better technology as I’m using it more and more all the time. The several signs visible to us concern me greatly. I’m not planning on upgrading to 1.2.1 or 1.3 because of the wifi canary in the coal mine. Any reasonable dongle on the outside of the faraday cage box should work at 100ft wifi distance. I’ve set dozens of these up for embedded and temporary systems.

In brief, I want to understand the future. I designed microprocessors at Intel for 21 years, started my own company and sold it, worked at SpaceX for 6 years as a troubleshooter. I kind of understand complex things and product development. I am concerned and want to know more of future intentions.

4 Likes
2

You raise many valid points and I’ve had similar thoughts, in particular around the BB forked controller. At the end of my warranty period I learned that OF won’t sell a controller without a machine, which at that point means if my controller died the only drop in replacement was the BB controller at a cost of $500 for similar functionality. This caused me to look for alternatives, at the advice of a friend that runs a production machine shop I went LinuxCNC first which worked great but required a lot of experimentation to get working, adding a tool setter took a few weeks of finding the downtime to work on so I wanted something more user friendly out of the box so I went Masso G3. The G3 has been flawless for me but I don’t run the Masso closed loop stepper motors (still using the stock stepper motors with a G540 as a driver). I would suspect the issues reported with the Elite are going to be configuration related or stepper motor driver related on the platform, not the G3 itself - perhaps the USB disconnected issue will have something to do with the way it’s wired up on the touch, I use the on controller USB ports. I don’t use any intermediate connectors between the G3 and the Gecko G540 driver, this could be a source of some of the issues as well.

At this point I see the OF as a killer hardware platform with many off the shelf parts I can easily maintain or replace if they break or wear out (I have over 3000 hours on mine with no hardware replacements required)

I would disagree with this assertion it is something non standard with Raspbian, not that other wifi adapters won’t work, it is more likely support doesn’t want to get bogged down troubleshooting an infinite number buggy wifi adapters from whatever manufacture sells them cheapest on Amazon. It’s easiest to for Support to specify the one (or several) they have fully tested. Others may work fine but Support won’t help troubleshoot them if they don’t. I would expect the same of anyone who’s cloned the git repository, edited the code, compiled, installed and then asks Support for help with an issue with their custom code

4 Likes
3

Randy,

I think you need to add @OnefinityCNC (Onefinity Support) so that they are notified about your post since it appears you are asking them a question.

Jay

1 Like
4

This is incorrect. If you are out of warranty, we will, of course, sell you a replacement controller.

What we will not do is sell extra controllers or controllers to users who do not own a Onefinity machine. This is due to the current ongoing chip shortage, nothing else.

2 Likes
5

When I inquired (via email) I was told “We don’t sell controllers without machines”, if that policy has changed that is good news but that is not what I was told when I asked about buying one after the warranty period expires.

3 Likes
6

As for the wifi dongle, there are many that are commonly used by RPI, such as the TPlink, Linksys,D-Link, Netgear, etc. All of these are transparent to the rest of all the system. The requirement of 10ft from an AP is ridiculous. I get that they may be overspecifying the system to remove all variables, but these requirements are certainly not the optimal.

2 Likes
7

Understood, and I agree they may all work fine but from a perspective a support team there has to be some limit around what they’ll troubleshoot. The 10 foot requirement is a bit overly restrictive but again I can see potential frustration troubleshooting wifi from someone’s home router with a max transmit power of 5dBm that would cause issues or separated by 2 8" thick concrete walls etc., easier to just say “max 10 feet”. In theory if I can connect to it with my laptop from the same location as the controller, the wifi adapter should work too.

1 Like
8

WIFI:. My PC is 60m from my 1F. I run a network cable (100m) from near my pc to the 1F along a stock proof fence holding the farmers sheep from my garden. Its been up and running for 18mths , longer than i expected. Hardwiring may be an answer.

2 Likes
9

Hopefully they don’t put goats in the pen or they’ll eat that wire :wink:

1 Like
10

Hey Derek,

[OT] surprised to find someone in this forum that knows the difference between sheep and goats!

4 Likes
11

Hey Randy,

For the project of creating and offering a new CNC machine, forking buildbotics.com was a good way to go, as it is free and open software and hardware (see onefinity-pcb vs. bbctrl-pcb for the open hardware). But Onefinity simply inherited some problems from upstream when they forked it. For example, just as the ‘buildbotics-firmware’, the ‘onefinity-firmware’ package is the only package on the Raspberry Pi OS system inside the Onefinity Controller that is not packed as ‘.deb’ package and therefore cannot simply be downgraded with ‘apt-get install onefinity-firmware=1.0.9’, which would have been very helpful when Onefinity abandoned and withdraw the 1.1x branch and told people they need to reflash the entire OS on the SD card.

Usually it is never necessary to overwrite (re-flash) a Raspberry Pi OS installation (except if the SD card fails). Today, all Unixoid Operating Systems have smart package management systems. Usually unixoid operating systems have an uptime of monthes, even years, and are updated while they run – including occasionally necessary downgrade of a package. On Debian-derived systems like the Raspberry Pi OS, this is the Advanced Package Tool (APT) system which deals with deb packages.

But it’s not trivial to pack a package as deb file, not only you have to learn debian policy, but you first also would need to integrate the bbserial kernel module into the Dynamic Kernel Module Support (DKMS) system so that it is automatically re-built on any kernel upgrade. Currently it is hardwired to an ancient kernel source and as soon you upgrade the kernel, the bbserial module ceases to work and the Onefinity System cannot access the AVR mainboard with the stepper drivers anymore. This is why you are stuck with a Raspbian operating system of 2017 with all the security issues that came up since then.

The good news is, onefinity-firmware is published under a free license, so if someone is willing to free bbserial from the hardcoding to an ancient kernel source, integrate it into the Debian’s DKMS, and finally pack ‘onefinity-firmware’ as a .deb package, that would be worth it to achieve that it becomes an interesting community project that attracts developers.

The same applies to buildbotics-firmware too by the way, and the question is, wouldn’t the community better contribute to buildbotics upstream, as this original version of the controller firmware (and hardware) would easily fit the Onefinity machine, but evolved a bit further [1], [2]. Unfortunately, Onefinity never backports any upstream improvements to its fork.

Disclaimer: I did not yet have a look to what Onefinity develops at the moment in the alpha and beta branches, due to lack of time.

2 Likes
13

Hi Randy, I’m curious what problems you were having with version 1.0.9? My Journeyman came with 1.0.9. I’ve only had it for 3 or 4 months but I’ve had no issues with it so see no reason to upgrade. I’ve added a laser and rotary axis all on this version as well and everything works as expected.

1 Like
14

I think that’s a great idea. I’d do a hardline also if I had issues with the wifi. I got a very early system and it had a few glitches (connectors…), with the biggest the initial reliance on the RPI’s wifi from within the faraday cage…:slight_smile: Adding a dongle from my scrap bin and changing the config to never use the RPI wifi made it completely usable.

Now you need to teach the sheepdog to run jobs when the sheep are not restless…:slight_smile:

15

The main problem was that I needed a different method of probing. As I recall there was some limitation with the initial 3 way probe that I had a problem with. I ended up rewriting the gcode routine and modifying the webpack to get it right. 1.1.1 gave me the needed functionality.

There were a couple other things that got fixed, but it’s been long enough that I don’t recall details. I think one of them made doing a Samba share more reliable, or at least claimed that it would. This involves setting up a shared drive on the RPI (like on my 3D printers) to transfer jobs directly instead of through the GUI. As I’m just now getting into more volume stuff I will shortly revisit this. 1.0.9 had a static view of file directory and I think they talked about 1.1.1 doing a dynamic query but it’s something I have to check. That itself removes a couple of steps in the workflow.

Mostly I followed the recommendation from 1F. I’m skittish on 1.2.1 because it rolls back some stuff and restricts wifi capability which several have said now breaks on their machine. Discussions here lead me to not want to jump onto a moving train. Now I see a 1.3 and we’ll see what happens down that path.

In the meantime I just keep making stuff like rocking horses, jewelry boxes, serving trays, and engraving random stuff. We’ll see what it looks like when it stabilizes.

2 Likes
16

I own several of each and 3 Border Collies to “manage” them

1 Like
17

I expect that’s correct. I have my 1F in the garage about 20 feet (and a steel door) away from the AP. No issues in a couple of years.

Only providing official support for the limited scenarios they can (or have) tested makes a ton of sense for a small organization.

2 Likes
18

Hey Andy,

this is a solution to my taste! I always keep it this way, for what do I need WiFi, as long as I own this:

20_mm_masonry_hammer_drill_bit_50pct
20_mm_masonry_hammer_drill_bit_50pct1000×109 148 KB

Image: 20 mm masonry hammer drill bit

1 Like
19

Hey Andy @AndyP, hey Derek,

Somehow I’m glad about that. As you may think, I am someone who separates the topics, when participating in various forums. But sometimes there is overlap, and then suddenly you smell not mechanic oil and wood dust, but the fields and field flowers.

My personal experience with the difference between sheep and goats is, when you tell a sheep “This is an electric fence” they believe you and will never approach it. When you tell the same to a goat, they will check the voltage every three hours and break out as soon as the voltage drops :slight_smile:

4 Likes
20

would you mind expanding on this masso/ g540 setup - i am interested in changing over to masso, i don’t want to change stepper just yet either, but it is not real clear to my feeble brain how to make the masso/g540 work

21

You’re in luck, Corvetteguy50 on youtube just made a video on this topic. I went the route of the g540 because I tried out LinuxCNC before I went to Masso G3, I might have gone with stepper motor drivers that support differential mode to start. I run mine on 48vdc supply power which has increased the torque of the stock Onefinity motors.

https://docs.masso.com.au/wiring-and-setup/setup-and-calibration/axis-servo-stepper-examples/Differential-Receiver-Module

2 Likes