One failure leads to four more

Things have been frustrating in the shop lately. It feels that whenever I try to make a little progress on a project, something breaks or another issue rears it's ugly head.

After months of successful operation, the Voron Switchwire suddenly stopped printing. Tearing down the printer and toolhead suggested a thermistor failure. As the printer uses a Revo hotend, the thermistor and heater core are a single unit. I contacted E3D, and they took pity on me and sent out a replacement.

"Not to worry," I told myself, "I have two printers." I went to spin up the Voron Zero and quickly discovered that it too had a problem. When I had built the printer two years ago, the gantry had a notable skew. Yet, I didn't notice anything wrong with the prints, so I ignored it. When I rebuilt the hotend on that printer earlier this year -- another Revo -- the skew issue became even worse.

I spent a weekend trying to use Klipper's Skew Compensation to try and null out the issue. While this did improve things, I discovered that the entire Voron Zero's frame was skewed and non-square. I didn't know this two years ago, simply because I didn't have the ability to accurately measure the printer's dimensions. While troubleshooting that, one of the blower fans failed, again. 

The replacement heater core arrived, I put it in the Switchwire and things came up just fine.

I was hoping that was the end of it. Then the dishwasher failed. Replacing that set my savings back a year.

And then, last weekend, the Switchwire died again with the same pattern. This time, however, it didn't appear to be the heater core. I was able to put 17VDC through the core from my bench supply and it heated easily. The thermistor responded too. After a bit of poking, I discovered it was the umbilical instead.

This was particularly frustrating as I was chasing both print quality and dimensional accuracy to fix the Voron Zero. I ran out of time and energy that weekend, so again, I put it down. Yesterday I was able to trace it down to the +24VDC line which goes into the molex connector for the umbilical. 

This, turns out, was a design failure on my part. I used a lower gauge wire for the primary power lines and they were arguably too big for the connector housing. This suggests that this could have been the failure the entire time. There's no way to know for sure now since I disposed of the previous heater core. I pulled the 24V line, cut off the tip, and then repinned it. That appears to have fixed the problem:

While the print did succeed, the experiment in doing a multi-material print did not. The ABS cooled too much, and the filament just wasn't hot enough to bond to the previous layer well. Thankfully, this was only an experiment.

The most astonishing thing about this entire saga is the single thing I did to improve print quality:

I slowed the printer down. 

That's it.

The Switchwire is a bedslinger, and was originally an Ender 3 Pro. The Ender was never a speed demon while printing, and at best could get up to 100mm/s before noticeably degrading. Unlike other Voron printers, the Switchwire doesn't really have an existing slicer profile. There are several community ones of course. Last year, I must have grabbed one of those and put it to work without considering the original machine's limitations. The Switchwire slicer profile was set to 120mm/s. I dropped that down and down until I got to around 90mm/s.

The result was some truly excellent prints out of the machine. I further tuned the esteps and temperatures until I started getting prints with a good trade off of accuracy and print strength.

The problem with the Voron Zero is that it's like a watch. It's hard to fix something wrong with the frame without disassembling the entire printer. The machine is so small, and the tolerances so tight that trying to fix it in situ may not be an option. It helps little I'm getting increasingly frustrated with the mini stealthburner's delicacy, or how many MOSFETs I burned out on the mainboard.

So, I'm rebuilding it. I really, really, really want to stop short of building from scratch here. My rough plan is build an alternative frame first, and then consider transferring as much of the existing printer hardware as is salvageable. Maybe this time I can incorporate a few mods I've been meaning to implement...