When circumstances throw a huge opportunity your way, how do you scale production fast enough to take advantage of it?  In this Phase Dock LIVE, Steve Friedl shows how he helped his brother’s small biomedical manufacturing company pivot to help meet the huge volume needs for rapid and accurate Covid test equipment.

Steve tours us through the medical manufacturing process line, showing how he used Phase Dock products, electronics, 80/20 aluminum, and ingenuity to automate industrial processes and increase production throughput. He describes how he prototyped the functionality, then built the working process automation equipment and converted a prototype into a PCB.  Steve was featured in a Phase Dock LIVE broadcast, April 10, 2021.

In Part One (this post), Steve walks us through:

  • The fill and assembly of Covid test vials
  • Fillerbot #1 evolution from a plotter mechanism into a highly reliable production equipment.
  • How 3D printed parts and a custom PCB streamlined the vial labeling
  • How he prototyped and tested functionality on a Phase Dock WorkBench before finalizing the PCB
  • Why 3D printing is a Super Power
  • Ways to leverage 80/20 aluminum frames for prototyping and production

Part Two topics include:

  • Fillerbot #2 — the Universal Fillerbot in development
  • The pumps, the controller, the communication and the software
  • Beaglebone, Arduino, and software choices
  • Continuous improvement and working with a great team

CHRIS: Hi, friends—I’m Chris Lehenbauer, CTO and founder of Phase Dock.

Welcome to the April edition of Phase Dock LIVE. We’re excited to have you join us.

If you’re new to Phase Dock, we’re a manufacturing startup here in Raleigh, North Carolina.  We focus on systems to mount, organize, protect and transport single-board computers and microcontrollers, so we inhabit the mechanical side of projects more than the electronics side.

Our featured guest today is our good friend Steve Friedl aka UNIXWIZ.

Steve is a Southern California software and network security consultant, a community hike leader, and unabashed fan of Phase Dock products for prototyping. He’s been using our products to develop and prototype automated production equipment in his brother’s medical manufacturing facility in southern California.

We invited him to share what he’s been up to. It’s pretty amazing.

Hello Steve, welcome. Tell us a little about yourself and about your brother’s operation and you got to this point.

STEVE: Good morning Chris and Barbara. I’m so happy to be here. I’m at the production facility of Resolution Biomedical in Tustin, California. Thankfully only two miles from where I happen to live.

We make Covid test collection vials. These have a couple of milliliters of fluid. We take these from empty vials, fill, cap, label, package, inspect and ship. We’ve probably done on the order to 3 million vials in the last year.

Resolution Biomedical was started by my brother a decade or so ago doing cell cytology for a different kind of testing, but pivoted when Covid came around. It’s been a really fun time.

I got enrolled to help ramp up production, so built some hardware, built some software, worked with a lot of the smart people here and we’ve made a really nice product. [2:10]

CHRIS: But your background isn’t really mechanical engineering, is it?

STEVE: No, not much. I have a degree in math and computer science. All of the work I’m doing with the electronics and anything mechanical is strictly amateur.

I’ve been playing with electronics since I was in the sixth grade, but I’m still really an amateur at it. I’m getting better though.

CHRIS: I would say that you are fearless, having seen what you have done. And honestly, that counts for a lot…just diving in and persisting.

STEVE: Right.

I guess the first thing I’m going to do is show the general production process of how we make one of these vials. I’ll do a quick little drive-by.

I’m thankful to have production shift leader Mike here, helping me out behind the camera. He’s got a way better eye for this stuff than I do. [3:00]

When we’re filling these vials, we create a stack of empty vials.

Covid test vials in the Fillerbot

Covid test vials in the Fillerbot

I will mention that everything we are doing today, that you see today, none of this is going into actual product. It is all for demo purposes. Everything we are doing is going to go back into rework, sanitization. Normally, masks and gloves. But we’re only two people here today, so it wasn’t necessary to do any of that. Nothing here is going into real product.

We start by filling this tray. The tray gets thrown in here and it will fill up the vials.

We come down to the next station to do a little finger-tight capping of each little vial.

Then to the capping stations. A little torquing.

And then to this wonderful device that was built for us by Roser Products around the corner in Chino. This takes a feed of the vials, prints the (lot and expiration date on a) label, wraps the label on, and kicks it on out into the tub.

After that, it goes into packaging, inspection and then shipment.

I think I said there were 3 million vials.

That’s the general flow. There are multiple capping stations. People work in the various spots.

But mainly we’re hear to see the Fillerbot.

CHRIS: As you walk, Steve, as the pandemic hit, to get to the quantity of production that you are doing now…you guys let the clutch out overnight, didn’t you? It was a really fast ramp-up, wasn’t it? [4:57]

STEVE: It really was. It was quite a pivot. When we realized that the preservative that we used for the other process worked fine (for Covid testing), we had people who knew how to do this, do all the proper validation.

So, we were basically ramping up production in a different building while this suite was being built out, while I was building the Fillerbot. I guess this is as good a place to start as any.

This is Millard…for Millard FILLMORE.

CHRIS: Of course! (laughing)

Fillerbot with Steve

Fillerbot with Steve

STEVE: [5:30] It’s actually based on this XY plotter mechanism that I bought through eBay. The head moves around to take a physical pen.

I built all of this around it to have some nozzles. I think you can see that we have three filler nozzles here. [5:55] I’m going to see if I can actually start this. I should have checked this advance, sorry.

CHRIS: Oh, there you go! Outstanding!

STEVE: This one is doing three nozzles at a time. You can see that it just moves right along. It’s completely unattended while it runs.

Three filler nozzles

Three filler nozzles

Let’s focus over here on the pumps. We use these wonderful pumps from New Era Pump Systems [6:39]. They are peristaltic pumps. They kind of squeeze the fluid through the tube, so the pump itself doesn’t actually touch the fluid.

Perestaltic pump

Perestaltic pump

CHRIS: Right. Completely sterile.

STEVE: Everything is lot numbered everywhere. We pay a lot of attention to that.

[6:55] It has now filled 72 vials.

CHRIS: That was beautiful.

STEVE: This works really well. The obvious question is so what’s the deal about these three (empty spots in the tray)? The reason is that the XY mechanism simply doesn’t have the full range to fill this whole tray.

I’m working on a new version, the Fillerbot Universal, that will cover the whole range.

We really optimized for doing three nozzles. It seemed to work out really well for us. It also divides into the geometry of the tray. [7:30]

Now I’ll scoot on the other side. I want to show the Phase Dock which is in the back.

This is the guy that runs everything. I’ve got a little Beaglebone here, a little controller.  Some glue logic. A couple of serial ports. The pumps all talk serial. So does the plotter. There’s actually stuff coming up to the front panel. The laptop is because I haven’t got the start button on it yet. That’ll be on the next version.

Phase Dock controller for Fillerbot

Phase Dock controller for Fillerbot

This has been super reliable. It fits perfectly into this spot.

This is actually the final, production version. It is not a prototype. We actually do keep the clear plastic cover on it normally just to keep it protected.

This has been unbelievably reliable. I love how it looks nice. It has been way more reliable than it deserves to have been.

CHRIS: As a side note, even with the 3D-printed Clicks and Slides that we have, they have been more reliable over time than I expected them to be. So, it’s good to see some validation from out in the field.

STEVE: It’s really worked well. And the fact that the thing just runs and runs and runs has been great.

As we move on down the production line, I have more or less to do with the various parts. What I’m going to do is trade places with Mike and have him cap and fill a few. [9:10 Steve takes Mike’s place behind the camera.]

We’ll just do a few of them. Right now he’s using a little piece of plastic that holds the vials apart. That was 3D printed as well. It holds the vials firmly, so you can put the cap on finger-tight really easily.

Again, none of these vials are going into real product. They will all be sanitized and reworked.

Finger-tight capping of vials

Finger-tight capping of vials

Unfortunately, this is a manual operation. We haven’t figured out how to automate it yet. But we have a really good factory crew here. They just go to town…the place just runs and runs and runs!

CHRIS: So, staging the caps on the vials is the part that you’d like to improve, right?

STEVE: Yes, that’s right. So normally he would do them all… but now he’s going to run the capping machine. [10:13]

CHRIS: So you’ve got an automated process there. Well, semi-automated.

STEVE: Not sure I turned the compressor pump on soon enough…it torques it down to the exact torque required for reliable filling. We’re a bit low pressure on the air…but you get the idea. [10:14]

Capping torque tool

Capping torque tool

So, he’s just doing a handful. That’s really good. Alright, now we’ll swap places here.

[10:40] Now come over here to the Roser Labeling machine. We have two of these. These are really wonderful pieces of engineering. We love these guys.

They look really complicated and it is adjustable in places where you didn’t even know they had places. Labels feed from the far end, through an Videojet printer. The paper path comes up and around and feeds the collector over there. All kinds of moving parts and so many adjustments.

The pocket wheel right here…you can see that it holds eight vials going around in a circle.

Roser label machine feed mechanism

Roser label machine feed mechanism

The place where Phase Dock came in was a little bit later.

[11:30] We ordered this without any input mechanism, so I had to design this whole input mechanism here.

CHRIS: Oh, how you feed the vials to the machine.

STEVE: When we got it as ordered, there was just this little plate, so you would manually feed three at a time or something. It was miserable.

We were lucky a good friend of mine is a real mechanical engineer and he helped with this a lot.

This is 3-D printed. Anything white is. Even this piece here. If you have a tray of vials that have been capped and you want to keep it here, this (3D-printed part) keeps the tray from blocking the path.

It’s something that you learn by fooling around with it for a while.

But where Phase Dock comes in… this sensor here is what detects that a vial is shows up…and says “go” and it drives the process.

Sensor on labeling machine

Sensor on labeling machine

[12:17] But, sometimes you need to stop it. Obviously over here you have the Big Red One, the big button, it stops the power and everything.

But that is overkill if you just want to pause the actual flow. So, what I built is this box down here that sits between the sensor and the control. It intercepts that signal that detects the vial.

Red means that we’re in the pause mode. When I hit the green button it lets it go. Let’s let it run.

CHRIS: Wow. That’s fast!

STEVE: So I just hit the pause button. This thing will do about 80 vials a minute when it gets cranking. It’s really quite a thing to see a skilled operator load these input trays up. It’s really amazing. They are way faster than me.

[13:15] This has really helped us (pointing to the box he built). One of the things that comes up…Mike, show that the labels are about to run out over there, right? When those run out, obviously we have to put a new roll of labels on it. The way I used to do it is: when it runs out you just put on a new roll. It takes about ten minutes to fully rethread the machine.

Our production supervisor Marisol figured out that you can actually splice a new roll onto the end of the old one, and not have to go through that whole ten-minute thing.

That’s where this comes in (pointing to the pause button on his box).

Button controller box

Button controller box

If I’m watching (the roll of labels), I can tell it to stop without having to worry about holding this back (the incoming vials) by hand. Now you just hit the red stop button (on Steve’s controller) and it won’t go anywhere. [14:00]

But it’s also got the yellow button. It’s the Uno button. (a single vial pops out and the machine stops again).

If I’m watching the labels, I’ll say “go now”, it’ll do one, but no more. So you can single-step until it’s at the end and we can splice the labels or whatever.

CHRIS: A couple of observations, Steve. Just the continuous improvement that I see there and not only that, but the input from all your team members…I just want to give a shout out to those two concepts because I’m a big believer in both of those things.

And, I think, the third thing I’m seeing there is, not only do you have much more granular control over that machine now, but everything you described is a lot safer than it was by manual intervention, isn’t it? [14:55]

STEVE: It’s safer. And it’s certainly faster because if you kill the power, then it shuts everything down and it’s just no fun. Occasionally the machine does skip a label or something gums up…

Chris, before we went live with the camera, you asked about how do we come up with ideas for improvement…the main answer is, I’ve sat at this machine and run full shifts myself…tens of thousands of vials…so while I’m sitting here running the machine, I’m learning what the sounds are like and I can say “OK, something’s different. I don’t know what it is, but something doesn’t sound right.” So then I’m able to talk to the other experienced operators to find out what that means. And I get the idea for the start and stop, the pause, the single step.

The thing where Phase Dock comes in [15:45] (pointing to the start/stop box)… Obviously this is not Phase Dock…there’s a little printed circuit board that I laid out and designed.

Now I’m going to scoot over here to the Phase Dock. This is that same circuit. I laid it out on this to test it. It simulates all of the inputs. Stop. Go. Uno. And then a “fault” button. [16:10]

Prototype of button controller on Phase Dock WorkBench

Prototype of button controller on Phase Dock WorkBench

I’m actually going to wire an input to the real device. If the printer runs out of ink or something else happens, it’ll be like a separate stop motion.

One idea is to build in a sensor for the input labels so that when they run out, that’ll automatically stop it.

I’m not going to hook it up, but it’s got the same little sensors, it plugs into all the same connections, drives its power, you can see my little Arduino nano I removed.

For me, this is actually the bigger win of Phase Dock than for final production because my desk looks like a tornado. I’m working on 18 different things at once, some of them software, some firmware, some electronics.

[17:00] I could not build this if I had each of these separate on my desk, where I’ve got to move it out of my way and I have to pick this up and hope that this wire doesn’t pop loose. Ugh.

This way I can pick it up and put this down somewhere else out of the way.

I think I have three or four Phase Dock’s at home in various states of project development. It’s nice to be able to put it out of the way and know that when I get back it’s good to go.

CHRIS: We’ve told this story before, but honestly that was the original inspiration for the Phase Dock WorkBench. I tell the story on my friend Andre all the time. It was the same thing. He had two or three projects on his desk and he couldn’t move them. I said “You know what, I’ll bet we can solve that problem.” That really was the driving factor.

But I love the fact that people are now starting to see that you can take it beyond prototyping and actually either put them into production, or use them as the basis for a single board design like you did there. You’re like the whole use case in all directions, that’s awesome. [18:11]

STEVE: I’m definitely a fanatic. I actually wanted to come over here and show a couple of my little Phase Dock parts. I keep a large tub of many of the Clicks and Slides and on many of them you can actually see where the screws go in.

I know that some people have this notion that “Oh, I don’t want to mess them up. I don’t want to cut it. I don’t want to put a hole in it.”

Heck no! I’ve cut these up and I reuse them. I even made myself a little box with the all the bazillion little parts your kits come with. And now I’ve got them all in one spot.

CHRIS: I agree. That’s something I always wanted (people to do), and one of the reasons we are driving to injection molding is that we really want people to think of those components as hackable. Just dive in with the drill and the Dremel and whatever. That’s exactly what we envisioned.

STEVE: It’s great for me. I just make sure I have one of everything in my tub. Sometimes I’m mounting a little microcontroller that you folks provide a Slide for, but sometimes it’s a display or anything, so I’ll just find something that fits right. So that’s been nice.

I didn’t think to bring it…I actually 3D printed a holder for scope probes that clip into a couple of spots on the back of the Phase Dock. [19:41] So my scope probes have a place to sit and from there I run a wire into the circuit wherever I need them.

CHRIS: That’s another thing that I would say. Those are printed ABS which means you can superglue either PLA or ABS to it very easily, so if you do print your own adapter it’s easy to quickly and securely put it onto a Click so you can then go and do whatever you want to do.

STEVE: I actually prefer screws. I know that you folks ship with these longer screws, I actually bought some smaller ones to use if I have a part that doesn’t need the clearance it just digs right in. I reuse some of these (Clicks) three or four times.

CHRIS: Yup. Screws are good because you can take things off. Again, just more versatility. It depends on how you want to hack it.

That was a great walkthrough of the whole process. We saw Millar Fillmore and the Roser labeling machine.

Two things I see you leveraging really hard. One is 3D printing. Tell us about that. [21:00]

STEVE: I love 3D printing. I 3D printed my name tag. Everybody here at RBI wears a name tag.

3D printed name tag

3D printed name tag

I use Sindoh. They’re kind of pro-sumer printers or a little bit higher than that. They are fast and reliable. I’ve been shocked how much I do 3D printing. I print everything!

I’ve done hundreds of designs for little piece parts like this one right here to hold a roll of tape. I use the Command adhesive. It’s handy where you need it.

Everywhere you look around here, you’ll find 3D-printed doodads. [21:50]

CHRIS: I have to say when I first started 3D printing, I was pretty skeptical. But I am astonished at how strong the parts are. You are right. As long as you can quickly design something you can solve a problem very, very quickly.

STEVE: I had never done 3D printing before a year or so ago. I dove in. I learned Fusion 360, which as far as I’m concerned, is the best software every created. It’s fantastic. You use something else? You’re a Solidworks guy?

CHRIS: I am a Solidworks guy, but I have used Fusion as well. But I would just put a shout out there that either package…if people are interested in going down this road of doing some mechanical work…please learn either one of those. Some kind of a 3D solid modeler. Because when you couple that with a good 3D printer, boy it really opens the universe up to you.

I think one of the big things that has helped you succeed is your mastery of Fusion because you are able to model complex things very quickly and then go to 3D printer. So either way, I would put a shout out for solid modeling, yes. [23:08]

STEVE: Fusion has got a great Maker program. I heard that Solidworks is coming out with one, I don’t know the details. Those are the really competent, excellent packages. There are also some free versions.

I could not imagine doing this if I had to outsource design, especially since some things I can actually design and print in 30 minutes for a little one-off. It’s fantastic.

CHRIS: That’s something we’ve talked about and simply haven’t had the bandwidth to do to this point…we’re going to open-source the Slides and put some online parametric design software out there so that people can quickly and easily design their own Slides without having to have software just by typing in some numbers, so we can help people with kind of a gentle onramp into that process.

Because, again, you’re not a professional mechanical engineer however, you’re doing really professional level engineering. I’ve seen that with my friends here as well, Andre’s an EE, but he’s got really strong mechanical abilities. My friend, Matt, is an industrial engineer again. So that ability to design and go to 3D printing has been really spectacular and to see it come to fruition there with what you are doing is amazing.

And the other thing you’re doing is using a lot of that 80/20 extrusion, right? [24:37]

STEVE: Ah, I live and die by this stuff. I actually use the 80/20 brand because it seems to be really well supported. They’re really super slow to ship, but I’m lucky that there’s a Grainger within walking distance of RBI so I’ll actually walk down there and pick it up and walk ¾ of a mile with the four-meter-long bars over my shoulder.

Then I’ll cut them up with a miter saw. As I tell people, a miter saw is a device to turn aluminum into glitter. It’s just incredibly messy, but I do a lot of cutting.

These are all parts for the pump stack which you’ll see in a little bit. I make a cut list.

One of the things that I created recently is…let’s say that I’ve got these three pieces…what I normally do if I need them at a certain length, I’ll rough cut them a little longer. I’ll square up the end and tape them, then cut them all at the right length. It’s really common that the exact length isn’t as important as them all being the same length.

I found I really struggled as I would get four or even six to keep them square.

So I designed some 3D printed pieces where I can slide them in the base. Now I know that they are sitting flush, now I can easily put tape around them and then cut them.

Now I’ve got these little guys for the common sizes that I need.

80/20 cutting fixture

80/20 cutting fixture

It worked out really well. It saved me a ton of time. It’s helpful to have that.

I can build four or five pump stacks and we’ll see that in a minute, whenever you are ready for it.

CHRIS: That’s been helpful. We’ve seen the end product, we’ve seen some of the ways you’ve gotten there. So tell us where you are going next.


END Part 1. In Part 2, Steve shows his Fillerbot Universal, and Chris and Steve talk about scalability, continuous improvement, and Beaglebone.