Archive for the ‘Uncategorized’ Category

PowerBelt3D Zero Printer goes Stainless

May 7, 2021

On the private FaceBook group for the PowerBelt3D Zero, Joe Mathiau has had some success replacing the PVC rollers with metal pipe and replacing the rubber belt with thin spring steel. The belt replacement process he used is complex, including taping the steel to the hot plate, cutting both ends of the belt simultaneously at a 45 degree angle, and taping the pieces together with high temperature metal tape. He said that he was finally getting decent prints. So, I started down his path.

I ordered Thin-Wall 304/304L Stainless Steel Unthreaded Pipe, 1-1/4 Pipe Size, 6 Feet Long, 4347K57 from McMaster Carr and cut it down to size to replace the rollers (500 and 484 mm). I designed and printed replacement roller end caps, both because the new rollers are a different diameter than the original PVC, but also to beef them up. They are larger in every dimension, and support the 5mm axle for 15 mm rather than the original 3, to avoid axle wobble. I also eliminated the 5 mm of slop in the original parts to reduce the cantilever of the axles. The end caps are described here.

The stainless steel rollers

We attempted to make a belt, like Joe did, using 316 Stainless Steel Shim Stock, 6″ x 50″ Roll, 0.005″ Thick, 9548K141 and Stainless Steel Foil Tape, 1″ Wide, 9 Feet Long, 6055A21. Joe cut his in place on the printer. We tried a different approach.

We created a paper template, confirmed that it fit, and then cut the shim stock to size in the shop using scissors. We pounded the rough edges down between two pieces of wood.

Our paper template
Cutting the shim stock with scissors
Pounding the edges flat between two pieces of wood

With the belt in place, we teased a piece of tape beneath one of the edges and pulled the other edge in place. We then taped on top of the joint and trimmed the edges.

Stainless steel held in place temporarily with painter’s tape while we used stainless steel tape to attach the edges

Joe recommended using a diagonal seam, perhaps to strengthen the joint, perhaps to reduce the chance of a deviation in the belt affecting the print all at once. In any event, making the cuts accurately and joining the pieces correctly was challenging. The belt does not creep across the rollers, so it is not dramatically out of square, but one side is looser than the other.

And it is not flat at all, which is quite a feat, given that the stainless steel we started with was dead flat. I sprayed the inside of the belt with hair spray, to ensure a good grip on the rollers. It is hard to believe that hair spray could bow stainless steel, but I do not have any other good explanation at the moment. The belt is about 5 mm high in the center, and nothing I have tried has resulted in a flat belt, or eve near a flat belt.

So. Joe did well, and I did not, and I am not sure where to turn next. Perhaps my CR-30 will arrive and I can stop flailing around with the PowerBelt3D Zero. It is an experimental marvel, but it is not a print at all, so far.

As Much As I Hate Amazon, I May Be Done With eBay

May 4, 2021

I always use the filter on eBay so that I only see deals from the USA. On April 16th at 12:14 PM, I purchased some electronics for $135.34 from eBay user “24_7secretdeals“. Realizing that I had made a mistake, I tried to cancel the order at 12:20 PM, 6 minutes later. At 12:37 PM, the seller congratulated me on the purchase that I had asked to cancel. The seller never mentioned the cancellation request.

On April 20th, four days later, I was informed that the seller had refused to cancel my order. On April 25th, 9 days after I asked to cancel the order, the seller shipped the order. It arrived on April 23rd, two days before the seller claimed that they had shipped the order. I initiated a return with eBay on April 27th.

eBay informed me that, despite the sale having been made with the USA-only filter, and despite the item having shipped from New Jersey, I was required to return the item to the following address:

Sahar Bachar

Atizmoret 20

Rishon lezion 75581

Now that looks like a ZIP code, but “lezion” is not a State in the USA. Googling I discovered that “lezion” is in Israel. The seller was requiring me to return the item to Israel, a process that would cost something like $40. I am sure that this technique is not an accident. The seller has a total of 3 different addresses registered with eBay. This was a blatant attempt to try to get me to not return the item.

I called eBay Customer Support on April 29th to complain about the situation. I explained that it seemed unreasonable for me to have to return an item to another country when it was purchased in the USA. The person agreed, and said that I would not have to do such a return, but she said that there was nothing she could do until 3 days had passed from the moment of the return request, to give the seller time to respond. I was asked to call back on Saturday May 1st.

I called back on Saturday May 1st and spoke with Karen. After over 30 minutes on the phone I was told that the seller had 3 business days to respond to me, and that I had to call back on Monday May 3rd. So, the person I spoke with on the 29th was wrong. Karen agreed that I was right, and that the seller was not in compliance with eBay rules. She wrote everything up, so that I did not have to repeat everything again. I doubt that had the effect that we hoped for.

I called back on Monday May 3rd and spoke with Valentina. After another 30 minutes on the phone I was told that, despite what I had been told on the 29th and the 1st, I did have to pay $40 to return the item to Israel. When I explained that I was supremely disappointed in eBay’s logic, I was told that there was nothing that she could do.

I asked to speak with a Supervisor and was told that, due to Covid, all she could do was forward my information and that a Supervisor would call sometime during the next 48 hours.

At this point, I am done with eBay. I cannot work with a company which refuses to provide reasonable protections against this kind of abuse. I specifically asked to deal with a USA seller, and the item was shipped from a USA address. To require me to return the item to Israel is ridiculous on its face.

And the erratic and waffling advice that I was given by the three Customer Service representatives makes me think that I probably should call back a few more times, to see if I can get a different opinion. Very frustrating.

We will see what happens next. But I fear that Amazon may have a new customer.

Added: The supervisor never called back. I called again and was disconnected. I called again and asked to speak with a supervisor and was hung up on again.

You simply cannot trust eBay. If you purchase an item that claims it is shipping from the US, be aware that you may be required to pay for return shipping all the way on the other side of the planet. And eBay does not seem to care.

PowerBelt3D Zero Printer, take Four

May 1, 2021

The designer and fabricator of the PowerBelt3D Zero kits, Adam, runs a Private FaceBook Group where printer owners can trade tips. On that group, the consensus is that the printer does not work very well right out of the box. The belt is wrinkled and the PVC belt rollers quickly deform under the pressure of the belt, resulting in the belt tension varying as the rollers rotate. Rather than a working kit, the feeling is that it is a bunch of parts from which you can start your own belt printer explorations. People have switched out the PVC belt rollers for steel, made their own belt end caps on a lathe, and replaced the belt with spring steel or stainless steel. That is where I am heading now.

So, if you are considering the PowerBelt3D Zero, please be sure that you know what you are getting into. Adam will let you into that Private FaceBook Group, but only after you pay for the kit. Not sure why he wants to hide the threads on that Group.

And the erratic printer freezes continue, at random times. So, after replacing the electronics, even if the new belt and rollers work better, I have another mystery to solve.

I hope my Creality CR-30 comes. It may have its own problems, but at least they are unknown to me at the moment.

PowerBelt3D Zero, Take Three

April 29, 2021

Well, my attempt to validate the hold down rails failed spectacularly. About 24% of the way through the 17 hour print (estimated by measured print length), the printer simply stopped. The extruder and bed remained heated, but the printer stopped moving; the print did not terminate in any reasonable way.

When I removed the print, it was bowed, as you can see, at the end where the print started. This lack of linearity makes the parts useless to me. I don’t see this as resulting from bed wrinkles but of some other origin. I welcome suggestions.

So far, only the smallest test prints have run to completion. I have no idea where the source of the poor reliability might lie. I also welcome suggestions here.

At this point, it seems that using the printer is just a waste of filament.

PowerBelt3D Zero Printer, Take Two

April 28, 2021

Adam never got back to me about my previous blog post.

Turns out that the controller board was not dead at all. Adam said that defects in the end stop cables would cause the LCD display to put out garbage characters. Turns out he was wrong. Incorrectly wired cables cause the LCD to be entirely blank. So, I chased my tail for 10 days, thinking the controller and LCD were dead, until I figured out that his description was wrong.

In addition, the extruder thermistor was shorted out at the extruder because the mounting screw had been over tightened, cutting through the insulation. I backed the screw off.

I think the cables to the LCD had to be swapped, too.

After that was sorted, I was able to get the printer to function.

The calibration prints went without a hitch, although the dimensions were slightly off, so I tried my first real print a 30 inch 17 hour print. It was fun watching it print, but the print lifted up off the belt/bed at times, because the belt was not firmly attached to the hot plate. You can see that here:

When the print was complete, it became clear that there were at least 3 defects. First off, the print failed to complete: there was a read error off of the SD card about 5 minutes before the end of the 17 hour print. Frustrating. Never happened on any of my other printers. There was a section of the print that narrowed from 18 mm down to 15.5 mm. And there was a crack in the print where adhesion was inadequate. Looking at the pictures and videos, it seems that these defects are correlated with uneven areas of the belt.

This shows the dimensional defect:

And these show the problems with the surface of the belt:

I created a hold-down system, to keep the belt on the heated plate. We will see how that works. All in all, a pretty frustrating experience.

The aluminum struts are held in place by magnets on the bases on each end of the belt drive. I will do another 17 hour print tomorrow, and see if the quality improves.

Building the PowerBelt3D Zero 3D Printer

April 20, 2021

I learned about the Creality CR-30 3D printer back in the fall of 2020, but did not have an application for that technology. 3D printers with belt drives for the build surface offer two advantages: you can print large batches of objects unattended, or you can print very long objects (like a 4 foot beam). I sometimes printed batches of objects, but had no use for printing long objects. I had purchased a Raise3D N2+, which can print 2 foot tall objects, but had never taken advantage of that feature, and eventually sold that printer.

But, as luck would have it, I now do need to print long objects. In fact, I need to print 4′ x 1″ x 2″ objects. I could have signed up for the Creality CR-30 KickStarter back in the fall of 2020, and expected delivery in May of 2021, but I failed to understand my need at the time. And, of course, KickStarter schedules are notoriously optimistic: I might not even be able to order a CR-30 in May.

Two other alternatives exist in April of 2021: the BlackBelt 3D printer (here) and the PowerBelt3D printer (here). The former runs around $12,000; the latter is a kit for $750. I chose the PowerBelt3D kit, because I’ve built kits for a long time and the price was right. There is a fairly negative review of that printer here, but a conversation with Adam Fasnacht, the designer of the PowerBelt3D printer, gave me confidence that most of the issues listed in the review had been addressed. I did the research on the printers on April 6th; I received the kit on April 8th. Very responsive!

My first reaction to the box of parts was positive. Everything was packed carefully and neatly, and Adam had created a series of construction videos on YouTube, so I started in.

I do much better with written documentation than with video documentation. It is easy to flip through a manual (or scroll or search through a PDF), but remembering which video you saw something in, and then scrolling through that video, searching visually, is time consuming. In addition, the platform I used to view the videos, a large Samsung Galaxy tablet, had a really annoying quirk: when you pause a video, the image becomes dim, too dim to see easily. So, I found the video approach to be inherently … wrong.

Adam created the printer design back in 2018, and received enough feedback that he is now on Version 2 of the design. I admire his desire and energy to continuously improve the kit, but this created documentation problems. His initial video is 1 hour long, a marathon, and out of date; his V2 videos are split into about 15 parts. But neither of them are complete or current. Finding the required instructions was not only frustrating, but often impossible. I did a lot of things twice.

Some of the problems had to do with the way Adam made the videos. Sometimes he would pick up a [black] part, flip it around to show you the orientation, and then work with the part. Some sections were sped up to reduce the length of the videos. Often I had to single frame through the videos in order to see what he was doing. And I had to repeat sections of the videos, sometimes as often as 8 times, before I could figure out what he had done.

Adam rarely said out loud what was most on my mind. For example, there are 4 ways to lay out the side rails of the printer. He never said out loud how to identify the orientation of the side rails. Sure, he laid them out on the video, but the key points for orientation are the locations of the black holes on the black parts. I eventually figured it out, but it took a lot of time. This pattern of not answering the obvious questions when it came to part orientation continued throughout the videos. He also spent a lot of time showing you the obvious: after positioning the parts on an assembly, he made you watch as he tightened each and every screw; he could have stopped after the first one. Much of the video is not necessary, and much of the necessary parts are missing.

This is less of a kit than a puzzle, or an adventure, or a game. If you go into it thinking of it as a game or a challenge, you will do much better than if you think of it as a kit. Let’s say it is a kit without adequate instructions.

One example is the diagram showing how to wire up the printer. There is no wire layout diagram, so wire routing is left up to the imagination; I wired everything 2 times, and in places 3 times. But more importantly, the diagram showing the location of the connectors on the controller board lacks some essential information. One section is labeled for the 12V DC power input to the board. But there is no information about which of the two connectors is positive and which is negative. I figured it out eventually (it is, after all, a game, not a kit), but that was needlessly frustrating, especially for a V2 kit.

At one point he explains that the computer board is labeled, so you can figure out where everything goes. That is true. The board is labeled. On the bottom. But his instructions are to attach the board to the case and then wire it up. At the time that you need access to the labels, they are no longer visible. On the video he says that he will put a copy of the board layout on his web site. It might be there, but I was unable to find it.

Some parts did not fit. He said to insert the leftover PTFE tubing over an extruding threaded rod, but nothing I did allowed me to get that tubing in place; other buyers on the FaceBook group had similar problems. That left the wire bundle from the extruder assembly dangling a bit. And that meant that the wires were in the way of the moving Y parts. That resulted in my creating another custom part to hold the wiring harness. Again, not a big deal, but annoying.

And the controller case is attached to the rails on one side, and hangs unsupported on the other side. While pressing the connectors onto the control board, I heard a cracking sound: the case had cracked. Putting feet under the unsupported side of the board would be wise, so I did so.

At the end of the construction videos, after everything is wired up, Adam says “happy printing” or something similar. When I turned the printer on, the fans rotated, but the LCD did not light up. After a week of troubleshooting with Adam, including attempting to re-flash the firmware, it is clear that the board is dead. I asked if I could buy a replacement from him, but he never answered the question. The board must have been working when he flashed the firmware on it originally, so it may have been damaged at some point later on.

And then there are the obvious design flaws.

For example, in order to load the firmware, you have to copy the firmware.bin file onto the SD card on the controller board. And in order to do that, you have to pop the SD card out of the board. And in order to do that, you need to be able to access the SD card, physically. The case that was printed to hold the computer board has a solid wall where the SD card needs to be accessed. I had to un-wire everything and cut a hole in the case. I do not understand how a V2 kit could have a case which makes loading the firmware impossible. But, remember, as a puzzle this is kind of cool.

And then there were the mis-labeled or missing parts. Two of the motor mounts were shaped differently than in the V2 videos, which confused me for a while (Adam explained that to me). And the part cooling fan mount was not included at all (he supplied an STL for that part, and I had it printed in an hour, so not a huge deal).

The 12V power supply would only provide 10.4 volts, so I grabbed a spare supply that I had lying around; not a big deal for me, just an annoyance. Adam never offered to replace the power supply.

The power supply has some exposed 110V wiring, so Adam designed a 3D printed plastic cover to make things safe. He supplied a power switch that was supposed to snap into the cover. It went into the cover easily, but fell right out, which was not safe at all. I could have drilled some holes and screwed that cover in place, or glued it in place, but it became clear that there was not enough room inside the cover for all of the wires. Cramming 110V wires into an enclosure is not a great idea. I ended up scrapping the switch and cover and doing things differently. Perhaps the switch was supposed to come with spring clips that held it in place. Even so, the crush of wires would have concerned me. It looks so easy on his video!

The bed heater is mounted on four custom L brackets, but the holes in the brackets did not line up with the holes in the heater. I ended up drilling new holes in two of the L brackets in order to mount the heater. Not a big deal if this is a puzzle, but annoying if this is a V2 kit.

BlackBelt has its own version of Cura for slicing models, and Creality has a similar version of Cura. I found the BlackBelt Cura to be very buggy (it crashed without warning about half the time I used it, and created poor slices). The Creality Cura crashed less often, and produced better slices. But I found Cura, in general, to be much less friendly than either Repetier Host or PrusaSlicer. Fortunately, Adam switched over to IdeaMaker by Raise3D, an slicer that is a bit quirky, but I like the UI much better than Cura. Adam provides simple instructions for configuring IdeaMaker on YouTube (although the downloadable files have some file naming problems).

It is likely that I misunderstood some things along the way; maybe some of the above is my confusion, and not something for which Adam should be held responsible. But the sheer quantity of problems is way beyond what I was expecting. The printer is described as “easy to assemble” on the web site. I beg to differ with this description.

Adam also said “We’re on a mission to empower people with technology. We can’t do that without standing by our products.” and “If anything arrives non-functional, or breaks within 90 days of receiving your products, we will replace those components free of charge.” Adam did not live up to his promises. Even after I offered to pay for replacement parts, Adam ignored my request. He has replaced none of the components, let alone free of charge.

As of this moment (April 19th), the printer still does not work. I have ordered a replacement controller and LCD display (not sure which one is dead). I now have to flash the firmware, which could be quick and easy, but nothing has been quick and easy so far.

This is not a kit for the average Maker. If you have assembled a Prusa MK3S kit, for example (I have done this twice), be prepared for an entirely different experience. If you have never assembled a kit before, this is not the place to start.

Adam started out very helpfully, but after I peppered him with questions (some of which I eventually answered for myself) he fell silent, and I had to finish the build by myself. I can understand his frustration with my dozens of questions, but I hope he can understand my frustration with his poor instructions, bad parts, and design flaws.

I showed this review to Adam, asking for comments about errors of fact. After three days, I gave up on getting a response from him. If he responds at some point, I will consider editing this review.

Johnson & Johnson Covid Vaccine

April 18, 2021

According to the CDC, as of mid April, there have been 31,382,000 cases of Covid, with 563,000 deaths. Lets quadruple the death number as an estimate of people who get “severe” Covid, to 2,252,000. This means that about 1 out of 14 people who get Covid have a severe case, or 7%.

Approximately 4,000,000 doses of the J&J vaccine have been administered, which should have prevented about 280,000 deaths (7%).

And yet we have stopped using the J&J vaccine because it may have caused about 8 hospitalizations due to a special form of blood clotting. That is one clotting incident per 500,000 people.

This seems like an enormous over-reaction. Perhaps someone can explain to me where my math went wrong. My take is that, at the moment, we are putting hundreds of thousands of people in jeopardy for no good reason at all. Why is there no push back? Have we elevated medical professionals to some god-like status where we do not question their decisions? Are we over-reacting to a small number of tragic illnesses? We pretend that we are dealing with this health crisis in a methodical manner, but this does not make any sense to me.

Jericho Ice

March 11, 2021

We took a walk along a stream in Jericho Center, Vermont. You can watch a video here:

I started out just taking a snap shot or two, but ended up shooting a lot of video as well as stills. It was one amazing tableau after another. As usual, the pictures and video do not do justice to the original. And in a week or two, it will all be gone for another year or two.

I guess some wild turkeys walked this way a few days ago

Smuggs Ice

March 9, 2021

Route 108 in northern Vermont (between Jeffersonville and Stowe) is closed during the winter, because the road is so steep, narrow, and twisty that it is impossible to plow. But you can park and walk up and over the notch if you wish. Katrina climbed some of this ice on Sunday, and we walked back to look at it more closely. The pictures do not do it justice.

People were skiing and snow boarding down the snowy bit in the middle.

Chris Wray Is Wrong

March 7, 2021

7 March 2021

Chris Wray testified before Congress this week, as can be heard in this podcast:

https://www.lawfareblog.com/lawfare-podcast-chris-wray-vs-committee-no-bull

During this session, at about 43:00 into the podcast, the following took place:

Congress person: “Is the FBI … are you currently pressuring any of these platforms, these social media platforms and tech platforms to include backdoors in their software that would help defeat end-to-end encryption?”

Wray: “Are we pressuring?”

Congress: “Are you encouraging, are you pushing for, are you … is there a desire to get…?”

Wray: “We are not trying to get backdoors. That is, I think, a criticism that gets leveled our way by people who don’t understand our position often, so I appreciate the opportunity to address it here. We are concerned about end-to-end encryption, especially default end-to-end encryption, in connection with a lot of these platforms. And we are concerned that if these companies continue to move in the trajectory that they are moving in we are going to find ourselves in a situation where no matter how bullet proof or iron clad the legal authority, no matter how compelling the facts and circumstances, no matter how horrific the crime or heart breaking the victim, we will not be able to get access to the content that we need to keep people safe. What we have been suggesting and the cryptologers [sic] and cryptogrifists [sic] that I talk to say this is doable, is for the companies themselves to build in a way to have legal access when confronted with the proper legal authority, so that they can get access to information, and can provide it in response to a warrant or a court order. We are not going to have a key, we are not asking for a backdoor, that is a myth, an urban legend, that has been directed our way, but this is a subject that I think the American people need to understand, because decisions that affect the life and blood of Americans all over this country, which normally are made by our elected representatives, are in effect being made in corporate offices in big technology companies. And different people can come down in different places on that balancing, but I would submit that this is a balancing that should be made up here, and not by one company based on its business model.”

Wray appears to be saying that it is OK if companies have the technical ability to decrypt messages that their users send, so long as the messages are only decrypted when the company is supplied with a court order. Wray does not like the implication that the FBI itself wishes to be able to decrypt the messages; he labels this a “backdoor”. But it is OK if others can do this, and the FBI uses those others.

Wray misses the point, and it is depressing that such a powerful and ostensibly educated man does not understand the situation better.

Just to be clear, a “backdoor” is a way to decrypt a message when one does not have the key normally required for decryption. Think of it as getting into your phone without having your password. The backdoor “breaks” the decryption. It does not matter who controls this process, whether it is a corporation or the FBI. The process is called a “backdoor” in either case.

So, when Wray says “We are not going to have a key, we are not asking for a backdoor”, he is simply wrong. That is exactly what he is asking for. Wray is attempting to make a distinction that does not exist. The FBI wishes to use a backdoor, and is trying to pretend that it is not a backdoor if some other entity (the tech companies) break the encryption scheme.

The danger of a backdoor is that if one entity can break the encryption, then if that secret (the backdoor) ever leaks, all of the messages encrypted using that method can be broken. The mere existence of a backdoor puts all of the messages in jeopardy. It is not a question of whether the entity that controls the backdoor is “good” or “bad”, “corporate” or “government”. Given how many sites we see getting broken into every week (including governmental sites, the NSA, etc), it is virtually inevitable that any backdoor will eventually be revealed and lost.

And, because of this, many of us believe that backdoors are inherently problematic. We understand that when the “bad guys” use encryption, that is also problematic, but both approaches have a down side.

Wray needs to understand that the FBI in fact does want access to a backdoor. It is not that people do not understand the FBI’s position: we do understand their position. It is neither a “myth” nor an “urban legend”. And having backdoors, at all, represents a security threat that many of us find to be too dangerous to endorse.

Wray seems prepared to trust that these companies will always act with honesty and propriety, and that they will safeguard this ability successfully. I suppose there always is a first time. I have no idea why he is so sanguine.

His testimony is frustrating. He is trying very hard to “educate” the American people with information that simply is not true.