Arcade cabinets, #coffee, #espresso, #3DPrinting, video games, #Linux, and all sorts of other nonsense!
886 stories
·
1 follower

Integrating my 3D Printer into my Home Automation

1 Share

Back when I was involved with our local makerspace, I really wanted to help show off the 3D Printers that the makerspace had purchased. At that time, access to the printers and Pat’s 3D-Printing expertise were among the best reasons to become members of the makerspace. As the guy with the keys to the makerspace’s social media accounts, I really wanted people to share pictures and videos of everyone’s 3D prints so that we could showcase what our members had been up to.

In fact, I was so enamored with the printers’ outputs—I had expressed my intentions of finding a way of capturing pictures of what people printed and automate the sharing of those prints. Unfortunately, the enclosed-style printers that they had at the makerspace (and that I bought to use at home) simply prevented getting a decent picture of the items printed on the printers.

Fast-forward a year or two and I’ve decided to upgrade my old Qidi Technology I — Dual Extruder 3D Printer with something newer. I opted to upgrade to a Prusa I3 MK3 3D Printer. Among its features was an open design that was infinitely more friendly to capturing the time-lapse photography that I had been wanting to do previously.

Enter OctoPrint, Octolapse and OctoPrint-IFTTT

OctoPrint

For those of you unfamiliar, OctoPrint is a fantastic print controller for your 3D printers. OctoPrint runs on a Raspberry Pi and manages your 3D printer; it’s got a fantastic feature set and I can’t recommend it enough to anyone who owns a 3D printer. The moment I’d completed my purchase of the Prusa I3 MK3, I started shopping for a Raspberry Pi, and wound up buying a CanaKit Raspberry Pi 3 B+ Kit whose job it’d be to manage my new 3D printer.

My first 3D print on the Prusa I3 MK3 was Pat’s Mounting Brackets with Swivel for Logitech C270 and C920 Web Camera which works so nicely with the inexpensive IKEA Tertial work lamp.

After a little bit of work, I had my Logitech C922 Webcam attached to the Tertial work lamp in place of the lamp, and mounted to my 3D printer’s stand and ready to record time-lapses of all my prints.

OctoPrint has some time-lapse photography features built right into it, but it’s pretty basic. The pictures are taken throughout the print job and the print head was frequently in different positions, oftentimes obscuring the printed object. I’d dug around the Internet a bit and learned that people have improved the quality of their own prints’ time-lapse photography by working with their slicing programs to generate the appropriate GCode on every single layer change. What they do is send the code to move the bed and print head to the same position on each layer and get that to coincide with the time-lapse photography.

Octolapse

Thankfully, there are options for people who don’t want to have to monkey with settings to insert custom GCode on every layer, like me. A very handy OctoPrint plug-in exists named Octolapse. As I understand it, Octolapse interprets and analyzes the GCode uploaded and handles inserting the correct GCode to improve your time-lapse videos. Just working through the few set-up steps improved my videos dramatically!

But after viewing a few of my first couple time-lapse videos, I still had a few complaints about what I was seeing:

  1. Too Darn Bright: My beer-stein lamp is right next to the 3D Printer, and while it’s a fantastic source of light to keep my office lit to what my specifications are, it was washing out everything in the videos of my prints.
  2. Lights turning on and off!: Over the duration of my prints, my home automation had been turning the lights in my office on and off. In the time-lapse, every now and then you’d see a chunk of frames with the lights in either position and it annoyed me that the videos’ lighting wasn’t consistent the whole way through.
  3. Not Automated Enough: I was thrilled to be getting the time-lapse of my 3D prints captured, finally. But I was still disappointed that I was having to manually find the files, download them to my PC, and share them on my own.

The first two problems were solved pretty easily by just manually turning the lights off and keeping them off during my prints. But as I was turning lights off and manually uploading time lapse videos to my social media accounts, I wondered what it’d take to do this in a more automated fashion.

OctoPrint-IFTTT

And then, out of nowhere, OctoPrint let me know via a notification that a brand-new plug-in had been published, OctoPrint-IFTTT. I’ve been using IFTTT for a smorgasboard of automated activities for years now and I eagerly installed the new plug-in and got started tinkering with it.

Now about those lights!

For whatever reason, the combination of my Logitech C922, the colors of my various rolls of IC3D ABS filament, and my nearby beer-stein lamp resulted in all of my videos looking really washed out. White filament almost appeared to be glowing and was so bright, it was devoid of features. Bright green filament looked practically pastel, and my red filament wound up looking pretty pink. So the first thing that I did was configure OctoPrint-IFTTT to call IFTTT’s Webhook with OctoPrint’s PrintStarted event and then tied it to the trigger I had set up to automate turning off my beloved beer-stein lamp.

Making it a bit darker in the room when my 3D printer is active improved the filament from looking to be so washed out and I was mostly happy with the results. I think ultimately I probably need to do a bit of research and experimentation to find the best lighting for these time-lapse videos, but I’m pretty excited that I can automate that solution to be triggered by my 3D printer’s activity.

Publish the Time Lapse Videos Automatically

After studying the supported OctoPrint events, I knew what I wanted to take a look at next: the MovieDone event. On the surface, it seemed like it’d be a simple task to trigger IFTTT to post my time-lapse videos to Twitter using IFTTT, but in the initial releases of OctoPrint-IFTTT, that wasn’t possible. The file’s path and name was just being passed as a string on to the IFTTT actions and not the actual file content.

I reached out to the developer, tjjfvi on GitHub, and submitted a feature request. The developer was gracious enough to offer me a few tips and in the process we found that (for what I’d been looking) IFTTT was expecting to be passed a URL where the file was accessible. We came up with a solution that’d allow IFTTT to pull it directly from my OctoPrint server, but that’d involve exposing my OctoPrint server to the Internet and that seemed like not the greatest of ideas.

However, by the end of the day, the developer had put out a beta version using the file.io file-sharing. If asked, the OctoPrint-IFTTT would upload the file to file.io, which returned a URL that could then be sent on to IFTTT for performing whatever action you wanted done. Unfortunately, I couldn’t find any actions in IFTTT that’d upload video clips to either my Twitter account or blog’s Facebook Page. Ultimately, I wound up adding IFTTT’s competitor, Zapier, into the mix as well. In the end, what I built seemed convoluted, but it worked!

Brian’s Time-Lapse Sharing Automation

  1. OctoPrint-IFTTT creates a webhook to IFTTT at the completion of creating a time-lapse video and triggers activity that uploads the video to a particular folder in my Google Drive account.
  2. Zapier monitors Google Drive and when a new file is uploaded to the specified folder, it is uploaded to my YouTube channel.
  3. Using IFTTT, I created two new actions to share the YouTube video’s URL to both Twitter and Facebook.

What’d I think?

I was—and still am—pretty excited to have achieved a goal that I’d had in my head for quite a while. But in this particular case, most of the value wound up being found in the travel—not in the actual destination. I knew that automating the sharing of these videos would be formulaic, but I wound up being turned off by what showed up in my social media feed. I also didn’t quite appreciate how convoluted the automation wound up being. I was reliant on far too many different services working independently of each other in the hopes of accomplishing my task.

In the end, a somewhat convoluted process to generate a formulaic result seemed like a bad combination to me. However, I did decide to go ahead and keep the initial step that results in the time-lapse being uploaded to Google Drive. Having the time-lapse videos up on Google Drive would make sharing them to social media quite a bit simpler.

What’s Next?

To be honest, I don’t know! Regardless of the fact that I wasn’t a huge fan of how it turned out, I’m still pretty in shock that one of the more meaningful goals was achieved. I probably would appreciate adding a couple new IFTTT applets to send me a Pushover notificationon my mobile devices when a print finishes or failed. I think maybe it’d be neat just to keep a Google Spreadsheet to log all the different prints I’ve done and how long they took. What sort of functionality would you like to see triggered in IFTTT by your 3D printer?

Read the whole story
patshead
1 day ago
reply
Plano, TX
Share this story
Delete

Hover-1 XLS Folding Electric Scooter Brake Upgrade

1 Share

I’ve had my Hover-1 XLS Scooter for somewhere around seven months now, and I’ve put over 200 miles on it. I mostly take it on short trips to the park—most of the time I take a round trip shorter than 1.5 miles, but I do tend to make those trips almost every day.

In the time I’ve owned the scooter, my only major complaint has been the brakes. Out of the box, they didn’t stop the bike very well. I was able to adjust them, and that improved the situation a lot, but they still weren’t great.

I believe I over-adjusted the front brake, leading it start leaking hydraulic fluid. I noticed this after I was already out of warranty. I could have attempted to get the hydraulic caliper replaced anyway, but from what my research tells me, these cheap hydraulic calipers are horrible.

If you found this post, you’re probably looking to replace a broken brake caliper on your Hover-1 XLS. I’m not going to make you read the entire post to find out what I used.

Here’s the TL;DR. I bought a pair of Shimano M375 brake calipers. They aren’t hydraulic, so they’re less prone to having weird issues. They bolt right on, and they have so much more stopping power than the stock calipers, and they’re inexpensive. I’m able to lock up the wheels and skid to a stop. I wouldn’t recommend doing so, but the original brakes certainly couldn’t do that!

The long story

Do you know the story of the boiling frog? I was that frog. My brakes were performing worse and worse until just about stopped working. By the time any replacement calipers arrived, my brakes weren’t really brakes anymore. They could reduce my speed, but they sure wouldn’t stop me!

I procrastinated quite a bit. I’m not a completely useless mechanic. On cars, I’ve replaced brakes, swapped exhaust manifolds, upgraded fuels pumps, and done all sorts of tuning. Replacing calipers on a bike should be easy, right?

Shimano M374 brake calipers on my Hover-1 XLS

My biggest problem was not knowing what to order. There seem to be all sorts of sizes and mounting options for brakes. What kind of mount does the Hover-1 XLS have? Am I going to have to replace the rotor, too? If I order the simpler non-hydraulic calipers, will I need to replace brake levers and cables, too? Ugh!

I still know very little about bicycle brakes. I just looked for a set that were meant to fit the same size brake rotors—the Shimano M375, and I eyeballed the mounting hardware in the photos. The calipers I chose were less than $20 each with free 2-day Prime shipping. Even if I didn’t or couldn’t return them, I’d only be out $20, right? It seemed worth the risk!

I only ordered one caliper. I replaced the dead, leaking front caliper, and adjusted it using the business card trick. The I went for a ride. I didn’t even wait until I got home. I ordered a second caliper to replace the rear brake while I was still at the park!

My Hover-1 XLS Folding Electric Scooter

These calipers are an amazing upgrade and simple. They’re a little more sensitive, especially compared to my dying front caliper, so they did take some getting used to.

Two months and another hundred miles later

It has been two months, 100 more miles, and a fresh set of brake calipers since I last wrote about my Hover-1 XLS scooter. Was it the right choice? Should I have bought something else?

I know this isn’t a review. If you found this post, you probably already own one of these scooters. I’m still going to give you my opinion anyway!

It is disappointing that I had to replace the brakes, and that the less modern, less complex parts work better than the original hardware. Hover-1 wanted to be able to list hydraulic brakes on their spec sheet, and they didn’t even use proper hydraulic brakes—the good ones have hydraulics right up to the brake levers.

Shimano M375 brake calipers on my Hover-1 XLS

I always tell everyone that I probably should have bought an electric kick scooter, like the Xaiomi M365. They’re nearly as fast, have almost as much range, but they weigh half as much. Loading the 50-pound XLS scooter into the back of the SUV is an awkward maneuver. I di

Another good option would be converting a regular pedal bike into an e-bike. Electric conversion kits are available at Banggood for around $400, and they look pretty easy to set up. With some time and effort, you could definitely DIY a proper e-bike that won’t cost much more than the Hover-1 XLS. You’ll have a much more capable bike that way, but you’ll need to invest an unknown amount of time. Ultimately, I imagine this is the best choice.

My dirty Hover-1 XLS Scooter

All that said, I’m not unhappy with my purchase. So far, I’ve invested something short of $600 and a couple of hours into my e-bike purchase. That includes the scooter, a helmet, and the brake calipers. Weather permitting, I ride the bike almost day. I’ve probably been on around 100 journeys this year, and I expect to continue to use the e-bike just as often over the next year.

What do you think? Do you have a Hover-1 XLS? Do you enjoy it? Are you happy with your purchase? Do you have some other form of electric transportation, like the Xaiomi M365? I’d like to hear about it in the comments, or you can stop by our Discord server to chat about it!

Read the whole story
patshead
38 days ago
reply
Plano, TX
Share this story
Delete

FuriBee GT 215mm Fire Dancer

1 Share

Those of you who follow my blog may have noticed that my blogging has been a bit sporadic over the past few months. The primary reason for this has been the birth of my son back in 2016 and the fact that I accepted a new job a few months back. But a big contributor to my absence has been my introduction to the world of drones. Almost exactly a year ago, I built a 450mm quadcopter, and ever since then I’ve been pretty quadcopter crazy. In fact, I now have a military-grade 55-liter backpack full of goggles, batteries, tools, spare parts, transmitters, action cameras, and most importantly, anywhere between four to six quadcopters which can be found inside or strapped to the outside.

My friends at Gearbest discovered early on that I’ve been quite obsessed with quadcopters and they’ve sent me a couple to review, starting with the KingKong 90GT micro-quadcopter that I’ve enjoyed so much that I still carry it around with me every time I go flying. When Gearbest offered to send me the FuriBee GT 215mm Fire Dancer FPV Racing quadcopter, I leapt at the chance to review it.

At the moment, I have two other different 5-inch FPV racing quadcopters: the Holybro Shuriken X1 and the insanely economical BFight 210. I really like both of these quadcopters. How would the FuriBee GT 215mm Fire Dancer measure up with my favorite two racing quadcopters? Which of the two quadcopters do you think the Fire Dancer would be the most similar to?

Specifications and my First Impression

On paper, the FuriBee GT 215mm Fire Dancer compares well with both my Holybro Shuriken X1 and BFight 210 quadcopters. While reading over the product’s specifications, my gut seemed to be telling me that the GT 215mm was more likely to remind me of the BFight 210:

Component Description
Frame 215mm X-configuration made of 4mm thick carbon fiber
Flight Controller Omnibus F4
Motors 2306 2400KV brushless motors
Propellers 3-blade 5048
ESC 4-in-1 BLHeli_S 35 amp
VTX 5.8GHz VTX, 48-channel w/ broadcast power 25mW/200mW/600mW
Camera 960 CCD camera with a 2.1mm lens
Weight 325g

About the only thing I’d squabble with from the product’s specifications is their claim about 1080P HD video. Perhaps the lens and camera sensor are indeed capable of 1080P, but there’s nothing onboard to record at 1080P and certainly nothing to broadcast in high definition either. If you’re wanting some HD video recorded on this quadcopter, you better strap a GoPro to it.

Unboxing the FuriBee GT 215mm Fire Dancer was pretty exciting. I was impressed with its light, yet sturdy build. The composition of the BFight 210 and the Fire Dancer are pretty comparable, but the frame seemed a bit more rigid than the BFight 210’s frame. More importantly, the Fire Dancer has bigger motors, bigger ESCs to feed those motors, and a better flight controller than you’ll find on the BFight 210.

However, I was not quite impressed with tiny little antennas on the cheap little Frsky receiver that was provided with the GT 215mm Fire Dancer, so I immediately swapped it out with a Frsky R-XSR that I had here in my collection of spare quadcopter parts. I also found that the quadcopter’s VTX was wired into the SBUS on the flight controller. I assume this was done to set up the possibility of controlling the VTX from my transmitter. However, in practice what I found was that my band and channel for the VTX were changing on what seemed to be each and every flight. Because the VTX lacked any documentation to explain why it was wired up to SBUS, I ultimately decided to just clip that wire. I was more than willing to set the VTX’s band, channel, and broadcast power by using the buttons built into the VTX.

Initial Flights

A quick note: as you will soon find out, I am still quite the novice at flying FPV. There’s a lot that I can’t do well, like soft landings or traversing our drone gates (aka kids’ soccer goals sans nets). While I feel like I’ve made a ton of progress since my first few FPV flights with the KingKong 90GT, I definitely lag behind many of my contemporaries. I’m not brave enough to follow in Pat’s shoes and do power-loops over trees, at least not quite yet!

I was super impressed flying the FuriBee GT 215mm Fire Dancer. My first flight was uneventful; I zipped around one of our favorite wide-open spaces and didn’t experience any kind of funkiness that can accompany a flaky quadcopter with suspect parts or a bad tune. It stayed up in the air for the entire duration of my 4S 1500mAh battery. If you’ve seen me fly a quadcopter then you’d recognize the magnitude of this accomplishment. My quadcopters seem to find their way to the ground long before their batteries have been drained!

My concern with both my BFight 210 and the FuriBee GT 215mm Fire Dancer was definitely durability. Not necessarily because the drones were flimsy, but because I’m hard on my quadcopters, as my BFight 210 can testify to. Moreover, the frames on both quadcopters are sleek and a bit on the slight side. Plus, I’d heard Pat’s horror stories about breaking his beloved BFight 210 in a particularly hairy crash one day.

Given my track record, it wasn’t entirely unsurprising to me that on my first day I wound up wrecking the FuriBee GT 215mm Fire Dancer. I was attempting at taking the GT 215 through one of our drone gates and winged the side pretty good—hard enough that it tore the propeller right off the motor. I was pretty pleased when I was able to put the propeller right back on and get flying again. It’ll be quite some time before I’d declare that the FuriBee GT 215mm Fire Dancer is durable enough to withstand my treatment, but the fact that it bounced back from this first-day crash so well seemed like a pretty good indicator.

So what does Pat think?

In Pat’s most recent blog about the demise of his own X1 and how awesome he thinks the BFight 210 is, he mentioned the arrival of my FuriBee GT 215mm Fire Dancer. Specifically, he said that he was looking forward to being able to kick its tires and take it for a flight. Naturally, I was excited to hear what he thought, so I handed him my transmitter on the condition that he write a couple paragraphs with his thoughts.

Brian let me try out his new Furibee racing quad. He had the camera angle set lower than I’m used to, but I had no trouble flying it. I did a power loop over a tree, and I hit a few aps in and around “The Lady Tree.” It was fun!

The Fire Dancer reminds me of the BFight 210. They’re both light drones with skinny arms, and they feel quite similar in the air. The Furibee has bigger motors, bigger ESCs, and an upgraded flight controller compared to the BFight 210—I didn’t do a good job putting those to the test in my short flight!

I like the Furibee GT 215, and it doesn’t cost much more than my BFight 210. It seems like a “no-brianer” to pay a few dollars for all those little upgrades!

Conclusion

There were two things about reviewing the FuriBee GT 215mm Fire Dancer that really stuck out. For starters, I really thought that GT 215 was excellent both in its features as well as its value. Like Pat said, it is only a few dollars more than the BFight 210, but it has bigger motors, can use bigger batteries, and has a better flight controller.

My biggest complaint was with the FrSky receiver that it came with. It was cheap enough that I simply decided to replace it before even trying it. I think my advice to others would be to buy the receiverless “plug and play” version of the quadcopter and spend the few dollars you save to buy a better receiver like the FrSky R-XSR.

But altogether, I think it’s a pretty fantastic quadcopter. To be able to find such a powerful and entertaining racing quadcopter for as little as $160 seems pretty incredible to me—especially when you consider the cost of quadcopters like my beloved Holybro Shuriken X1.

Naturally, Gearbest is selling the FuriBee GT 215mm Fire Dancer with a variety of receiver types:

But more importantly, with the help of the FuriBee GT 215mm Fire Dancer I was able to put together what I think is easily my “best” FPV flights so far. Perhaps I’m beginning to get enough stick-time in that my piloting skill is improving, or the GT 215mm is just that nicely put together and tuned than a clumsy novice like me would be able to fly it. Whatever the reason, it resulted in what I think is my best FPV video to date.

Read the whole story
patshead
394 days ago
reply
Plano, TX
Share this story
Delete

DIY NAS: EconoNAS 2017

1 Share

Ever since building my first NAS over five years ago, I’ve been keeping current by building a pair of new NAS builds each year: a powerful, budget-oblivious NAS build and an economical, price-efficient NAS build in an effort to demonstrate to prospective do-it-yourself NAS builders thata NAS can be built and assembled to suit their needs and budget. The DIY NAS: 2017 Edition was off the charts in storage, processing power, throughput, and also price! It was fun to build and blog about and even though its price tag was incredible, I thought it still had a pretty good price-to-performance ratio.

For 2017’s more budget-friendly NAS build, I decided to double down on my yearly goal to exceed the prior year’s EconoNAS performance while keeping the total price as close to $500 as I can. Every year I set this difficult goal and every year I wind up missing that goal. Will I wind up hitting that goal in 2017?

CPU & Motherboard

In shopping for the motherboard, I found out right off the bat that my new dedication to my budget goal was going to likely make this year’s EconoNAS remind everyone a bit of last year’s EconoNAS. I wound up picking an ASrock B150M-DVS R.20 (specs), which ultimately wasn’t all that different in specifications from the 2016 EconoNAS build, the Asus B150M-K D3. Chief among its feature set for me was the support for Intel’s LGA 1151 CPUs, its support for up to six SATA 3.0 devices, and its affordable price point. I’ve built a few DIY NAS machines (including my own) using ASRock motherboards and I’ve been pretty pleased with the results so far.

I am always stingy when it comes to paying for a CPU in the EconoNAS machines. I usually wind up seeking out the least expensive CPU that I think can handle taking care of FreeNAS and that fits in the motherboard. For this year’s EconoNAS, I wound up picking the Intel® Pentium Processor G4400 (specs). The Pentium G4400 wound up being a tiny bit more powerful than the G3920 which was used in the prior year’s EconoNAS.

RAM

As I did in 2016, I decided to go with 16GB of RAM. In the FreeNAS hardware requirements the minimum requirement is 8GB of RAM, but the recommended amount of RAM is 16GB. With my $500 budget in focus, I was tempted to dial the memory back down to 8GB in an effort to save some dollars, but then I decided against that. The two Kingston Technology ValueRAM 8GB 2133MHz DDR4 DIMMs (specs) were also an additional tiny upgrade over the 2016 EconoNAS, coming in at a few hundred megahertz faster at 2133MHz.

Case, Power Supply, and Cables

In looking for a case, I had one thought in mind: value! I pinched as many pennies as I could and sought out the least expensive case and power supply that I could find which would accommodate a decent number of 3.5” hard drives. My search turned up a manufacturer that I hadn’t been aware of before, Logisys. The Logisys CS305BK (specs) is a mid-tower ATX case that features a total of 10 bays (six 3.5” bays and four 5.25” bays). In my goal of being thrifty, I couldn’t avoid the temptation of this very inexpensive 10-bay case. At 480 watts, the included power supply should be more than enough to drive the rest of the components I picked out for the EconoNAS.

As it often happens, being thrifty came at an expense. In order to hook everything that I bought up, I wound up needing a set of extra SATA cables, a 3.5” to 5.25” drive bay adapter, a molex to SATA power adapter, and a 4-port SATA splitter. Regardless, I would’ve had to buy more SATA cables because motherboard manufacturers are notoriously skimpy in sending SATA cables with their motherboards. However, I did wind up spending another ten dollars buying adapters that I may not have had to buy if I’d chosen a different case entirely. Is it possible that this very inexpensive case was deceptively more expensive in the long run?

Storage

I go about buying storage for each NAS system with two goals in mind: first, to exceed the prior build’s storage capacity, and second, to spend less money on storage than the prior year. As you can imagine, these two goals conflict with each other. Bigger hard drives are a better value in terms of dollars per terabyte, but bigger hard drives carry a bigger sticker price too. Considering how similar the 2017 EconoNAS was winding up to the prior version, I wound up spending most of my effort shopping for hard drives.

FreeNAS Flash Drive

As is the case nearly with every NAS I’ve built since my first one, the SanDisk Ultra Fit 16GB was my choice for using to store FreeNAS on. I was tempted to add a second drive so the EconoNAS could have a mirrored pair of OS drives, but I opted to save the few bucks a second USB drive would wind up costing. I originally wound up picking the SanDisk Fit because of its compact size and reliability and because their performance over the years has earned my loyalty.

NAS Hard Disk Drives

To say I agonized over shopping for hard drives would be putting it mildly. The 2016 EconoNAS featured 6x2 TB harddrives for 12 TB of RAW storage. Adding a seventh 2TB hard drive in 2017 seemed like too much of a cop-out, and unsurprisingly, the prices of 2TB HDDs had not really changed much since building the 2016 EconoNAS.

Instead, I wound up going with generic “White-label” 3TB HDDs for the same price as the Hitcahi 2TB HDDs I’d purchased the prior year. In seeing them, I asked nobody in particular, what exactly is a “white-label” hard drive? Wikipedia defines a white-label product as “A product or service produced by one company that other companies rebrand to appear as if they had made it.” So effectively, the hard drives were manufactured by an unspecified, but well-known manufacturer likely intended for the use inside other products or by anyone who buys hard drives in volume.

This seemed like a riskier option, but at only $4 more per hard drive than I paid for the storage drives in the 2016 EconoNAS, it seemed like a risk worth taking. Having decided to buy the White Label 3TB hard drives, I then had to pick how many to put in the system. I wound up opting to buy five hard drives, for a raw total of 15TB of storage. Because I advocate at least using two drives of redundant data, the net storage will wind up being 9TB.

In comparison to last year’s build, it’s an increase of 3TB raw storage (15 vs 12), and 1TB of net storage. But I actually wound up spending less money on hard drives this year (~$300) than I did last year (~$340) because I bought one fewer drive total and their relatively close price points. Final Parts List

Final Parts List

Component Part Name Count Cost
Motherboard ASRock B150M-DVS R2.0 specs 1 $64.19
CPU Intel® Pentium® Processor G4400 specs 1 $53.60
Memory Kingston Technology ValueRAM 8GB 2133MHz DDR4 Memory specs 2 $89.95
Case and Power Supply Logisys Corp CS305BK w/ 480W PSU specs 1 $36.11
SATA Cables SATA Cable 26AWG SATA III 6.0 Gbps with Locking Latch for HDD 10 inch SATA Cables (10 pack) N/A 1 $8.50
Power Adapter StarTech.com 6in 4 Pin Molex to SATA Power Cable Adapter N/A 1 $2.74
Power Splitter StarTech.com 4x SATA Power Splitter Adapter Cable N/A 1 $6.27
OS Drive SanDisk Ultra Fit 16GB specs 1 $9.99
Storage HDD White Label 3TB 7200 RPM Hard Drive N/A 5 $59.99
TOTAL: $652.24

Hardware Assembly, Configuration, and Burn-In

I’ve been pretty proud that I’ve yet to hit a serious issue when assembling my different NAS builds. They typically go right together, boot up, and burn-in without any problems. It would seem that in 2017 my lucky streak would hit a bump in the road. How big of an issue do you think that I ran into?

Assembly

My very first thought about the assembly came in working with the case. The Logisys CS305BK is inexpensive for a reason: it’s cheap! The first case I received got abused in transit and the fascia of the case had completely shorn off its posts. The power supply wound up coming loose and had banged around hard enough to bend the sheet metal that held the power supply in place and bowed the hard drive cage a tiny bit. The metal used in the case is so incredibly thin, that I jokingly referred to it as foil on a number of times—it almost seems like it should be protecting a pumpkin pie rather than a host of computer components. That being said, I expected this case to have warts. No matter what way you look at it, it’s an inexpensive case with a ton of room inside for a plethora of hard drives.

Everything went together quite easily and I didn’t have any issues assembling the computer, but I did wind up having an issue on its first boot up. It wouldn’t post at all! After quadruple-checking all of my work and removing/disconnecting everything but the CPU, a single stick of RAM, and the power, I had no luck whatsoever. I swapped in a known good power supply that I had lying around, hoping that I’d just received a defective power supply. But that different power supply didn’t solve my problem.

I wound up being in a bit of a pickle, because I didn’t really have any way to verify whether the CPU, RAM, or motherboard was to blame for my problem. In doing some research into the ASrock B150M-DVS R.20, I learned that a BIOS update was required in order to support the newest Kaby Lake CPUs and I just so happened to have originally picked the Intel Celeron G3920, a Kaby Lake CPU. Without a spare Skylake CPU laying around, I wasn’t going to be able to flash the BIOS in order to support the CPU that I’d originally bought. Instead, I opted to buy an additional CPU, the Intel® Pentium Processor G4400 . The price was a few dollars more, but it had the dual benefits of being supported by the motherboard in its current form and it benchmarked a little higher than the Celeron G3920 that I picked.

Overall, not a huge issue. But I was a bit bummed that it snapped my lucky streak in building NAS machines for my blog! Onward to the next lucky streak, I hope.

White-Label Hard Drive Roulette

Among the things I was most interested about was exactly what kind of hard drives I was going to wind up receiving. One of the things I do in my the DIY NAS series of builds is buying hard drives from multiple manufacturers to try and avoid a bunch of drives that could’ve come from the same bad batch at the same manufacturer. When I placed the order for my five white-label drives, I briefly imagined receiving hard drives from three or four different manufacturers and getting the best of both worlds.

In looking at the five different drives, it was quite obvious to me there were two distinct models of drives. I had four of one model of drives, and one of a different model. This wasn’t quite ideal, but it is an upgrade from EconoNAS builds in the past where I’ve frequently bought what seems like the most amount of storage for the fewest dollars.

A little bit of detective work in the BIOS and taking a peek at the drives’ S.M.A.R.T data showed me that I wound up with:

  • 4 of MaxDigital MD3000GBDS
  • 1 of Hitachi HUS724030ALE64

The MaxDigital drives are a bit of a surprise to me. To be completely honest, I hadn’t heard of MaxDigital prior to my white-label hard drive roulette experiment in building the 2017 EconoNAS. When I ordered the hard drives, I bought the absolute best deal I could find on the 3TB hard drives.

My bottom line on hard drives is that I’ve been burned and massively disappointed in every single hard drive manufacturer at one point or another. I’ve had failed hard drive experiences and painful RMA process war stories to carry a grudge against all of the hard drive manufacturers at one point or another. My preference would be to use the hard drives from a manufacturer I’m at least familiar with, so I’m not entirely sure how I feel about these MaxDigital hard drives.

I am happy with the price I got; they were by far the best deal among 3TB hard drives when I did my shopping. Ultimately, the EconoNAS is all about deciding how much risk you’re willing take in order to save a few dollars. I’ll be curious to see how these drives perform and how reliable they wind up being.

Hardware Configuration

When I built my first NAS, getting the BIOS configured to boot up only via USB was tricky enough that I dedicated a paragraph or two to the hardware configuration. But in the years since, this has gotten better and simpler. Pretty much the only change I wound up making in this year’s EconoNAS was setting it up to boot off the FreeNAS USB drive. Beyond that, I didn’t make any other changes.

If I wanted, I could’ve theoretically updated the BIOS and switched back to the original CPU that I purchased for this year’s EconoNAS. But that CPU’s only benefit was being a CPU-generation newer and saving between $5 and $10. Personally, I would’ve been happier to spend the extra money and not have to deal with swapping the CPU. Ten bucks to save myself some time and frustration seems like a good deal. I decided to leave the older, but more powerful CPU in the machine and decided to forgo updating the BIOS.

Burn-In

For burn-in, I really like to focus in on the RAM and the CPU, primarily because they’re kind of a hassle to replace and because there’s no redundancy for this hardware. If I were truly anal-retentive about the burn-in, I’d also try and run some sort of burn in test on the network card and also the storage array once it is put together. But, I think that burning in the NIC and the hard drives might be a bit over the edge of reasonable. This is just my opinion, I wouldn’t blame any of you for wanting to also burn in those other components too.

My first boot is almost always into Memtest86+ to check the system’s RAM. This is for any computer I buy, but in particular for my NAS builds. I always have a bit of lingering doubt in my mind about the RAM until I’ve seen Memtest86+ complete at least three passes without any errors. Because I usually wind up doing something else, like flying drones, I usually wind up doing quite a few more passes than just three. Keep this in mind when you see the number of passes of Memtest86+ I wind up going through—I think anything over 3 is pretty much overkill.

For the CPU, I use my UltimateBootCD—but on USB—and I run the Mersenne Prime Test. My goal in running the test is to peg the CPU at 100% for a long time and look for any kind of instability or lock-ups. I usually do my testing in “waves” a shorter test of only 5 minutes or so and then subsequent tests of 10, 20, and finally 30 minutes. I’ll usually hop over into another console and monitor CPU usage and temperatures to keep an eye on anything getting out of hand.

As I expected, the 2017 EconoNAS survived my round of burn-in tests and was ready for the installation and configuration of FreeNAS.

FreeNAS Configuration

Installing FreeNAS is a snap. I always wind up downloading the latest FreeNAS ISO and then writing it to a USB drive, since I always seem to have about half a dozen USB devices floating around. I then boot up off that USB device, and then installing to the actual intended USB drive for running the FreeNAS OS. The trickiest part of this is remembering which USB device holds the FreeNAS installation and which USB device is going to hold the FreeNAS OS. Once the installation is finished, I remove the USB drive with the FreeNAS ISO on it and let the machine boot up for the first time.

Once it’s booted up for the first time, I do almost everything from the FreeNAS web interface. The default server name is FreeNAS, so you should be able to pull it up entering in the URL of http://freenas. On occasion, I’ve been industrious and figured out the IP address by logging in to my router and reviewing what IP addresses have been handed out to which machines via DHCP. Once the FreeNAS web interface has been pulled up, I usually go through these steps:

  1. On the first screen of the Initial Wizard, I hit exit because Initial Wizards aren’t complicated enough to make me feel like I’m accomplishing something.
  2. I set the hostname to: my machine name.my local workgroup name (example: econonas2017.lan)
  3. Create a new volume under storage. I always add all of the hard drives to the volume and I pick RaidZ2 for my Volume Layout. I opt for RaidZ2 because the data gets written to the different drives in such a way that you have two hard drives’ worth of redundancy.
  4. Add a group to contain all the users who’ll be able to access my share.
  5. Create a user for myself. Because I want to keep things simple for authentication, I make sure my username and password matches that of my computer(s) I’ll be accessing the NAS from and I also add myself to the group that I created in the prior step.
  6. Create a dataset belonging to the volume created earlier. I fill out the Dataset name and Comments, set the Share type to Windows, and made sure the Compression level is set to Inherit.
  7. Update the permissions on the dataset I just created, setting the Owner (group) to the group I created earlier. I set the Permission Type to Windows and I check the box for Set permission recursively
  8. Underneath the Sharing > Windows (SMB) Shares, I add a new share pointing at the path of the Dataset created earlier. I updated the Name and make sure Apply Default Permissions is checked.
  9. Following the share creation, I choose to enable the Windows SMB Service
  10. Under Services > SMB I configure the service by making sure the NetBIOS name matches the server name and the workgroup matches what I put in before (in step 2.)
  11. Because I’ve had luck in improving performance in the past, I choose to Enable Autotune under System > Advanced.
  12. I’m lazy, so at this point I just go ahead and reboot the NAS after making all of these configuration changes. This probably isn’t necessary and everything that I accomplish in a reboot can probably be achieved by releasing/renewing the IP address from your DHCP server.
  13. Following the Reboot, I go browse to the NAS from Windows Explorer and I validate that I can see the share and make changes to it.

Initial Wizard Updating Hostname Creating a Volume Adding a Group Adding a New User Create a Dataset on the Volume Setting Dataset Permissions Creating a SMB Share Enabling SMB Service Configuring SMB Enabling the wizardy of Autotune Reboot! Exploring the Share

Way back when I built my very first FreeNAS box, I was really surprised at how simple it was for a dingbat like me to install FreeNAS, get it configured, and be able to access the shares from my Windows machines. Over the past few years, I’ve gotten a bit better at it and hopefully it shows in the few steps above. But I think it’d be negligent if I didn’t also point out that this is only beginning to scratch the surface at what you can do with FreeNAS.

Benchmarks

When it comes to my NAS builds, only two numbers matter to me: power consumption and throughput. Power consumption is interesting to me because I intend to run my NAS perpetually. The only time my NAS is off is when the power is off at my house and the battery in my UPS has drained. Because of that, I like to keep an eye on how much electricity it uses. And beyond that, I like to see throughput numbers. I especially like to see the NAS builds fully utilizing the NIC on the NAS during a file transfer.

Power Consumption

Electricity is a sneaky, hidden cost of owning a NAS. Typically, I’m so eager to try and hit my $500 goal that I completely ignore the cost of electricity. I wouldn’t be surprised that spending a few dollars more for a low-power CPU and motherboard wouldn’t ultimately wind up being a cheaper option once you calculate in the cost of electricity over the lifetime of the device. To help everyone make a better decision on their own with regards to power consumption, I did a little bit of power-consumption benchmarking.

I’ve been using my Sonoff POW for these power benchmarks for quite some time. I really like being able to check and see how much power a particular outlet has used just by pulling it up from my mobile phone. I monitored the power consumption remotely and grabbed the highest number that I saw (while monitoring) during each of the following scenarios:

  1. Boot Up
  2. Idle
  3. Memtest86+ burn-in
  4. CPU burn-in
  5. NAS write throughput testing
Bootup Idle Memtest86+ Mersenne Prime NAS Write Test
116 watts
75 watts
98 watts
102 watts
80 watts

Altogether, I was satisfied with the power consumption. Using 80 watts during the sequential file write (more on that below) seemed like a pretty decent number for making the disks work. Ultimately, I felt that 2017 EconoNAS was up to the task of fulfilling it’s primary function without causing exorbitant utility bills.

Throughput

When benchmarking a NAS, it is important to remember that your primary bottleneck is going to be your network interface. For example, the 3TB HGST Ultrastar 7K4000 HDD in this year’s EconoNAS has a sustained throughput of 171MB/s which converts to about 1.368 Gbps. All by itself, that single hard drive is capable of saturating a gigabit connection all by itself.

I like to do a throughput test to see how the NAS performs on a sequential write and a sequential read. I use IOMeter to perform the test. I used my primary desktop computer to perform the test, so I set up 2 workers for each of my 8 CPU cores (16 total). And I ran two different tests: a sequential write and a sequential read, both of which used a 512Kb transfer request size.

  • Sequential Write: 92.21 MB/sec with transfer speeds as high as 810Mbps
  • Sequential Read: 91.24 MB/sec with transfer speeds as high as 876Mbps
Sequential Write Throughput: 92.21 MB/sec Sequential Write Performance: 810Mbps Sequential Read Throughput: 91.24 MB/sec Sequential Read Performance: 876Mbps

Overall, I though that the throughput testing was a bit of a mixed bag. I was pretty bummed out that I wasn’t fully utilizing the 2017 EconoNAS’s Gigabit link. It’s certainly something that I’ve accomplished in prior years’ EconoNAS builds, including last year on the 2016 EconoNAS. But I was pretty pleased to see the write tests perform in the same neighborhood as the read tests.

Conclusion

Quite a few things disappointed me this year about putting together the 2017 EconoNAS. For starters, I was going to miss my $500 price point … again. I thought the case was pretty chintzy and cheap enough that one was even damaged while it was shipped to me. I was more than a little bummed when it didn’t completely saturate my gigabit link. And most of all, I was extremely let down when it seemed like there wasn’t a whole lot of difference between this year’s EconoNAS and last year’s EconoNAS.

I think that the 2017 EconoNAS is a fine machine. For right around $650, you get 15 TB of raw storage, an Intel Pentium G4400, 16 GB of RAM and quite a bit of room to grow with 3-4 empty drive bays inside the case. If you’re looking to get the most storage possible for as few dollars as you can, I think this build is a very good direction to go and that my disappointment shouldn’t wind up also being your disappointment. Compared to the current pricing of the 2016 EconoNAS, this new build is better in nearly every way and a better value to boot!

The root of my disappointment is that I’m simply building these NAS machines too fast. Newer hardware isn’t pushing the prices of older hardware down fast enough to justify a new EconoNAS build every 12-18 months. I even purposefully postponed the EconoNAS build several times to see if the parts got any better or the prices got any lower—they didn’t! This also probably applies to my “regular” DIY NAS builds too. In 2018, I imagine that one of my New Year’s resolutions will be to come up with a new approach on the frequency of DIY NAS builds.

How do you guys think I did? Where would I be able to trim some money, but also exceed what was built 18 months ago for the DIY NAS? Or should I have spent even more money to truly set it apart? Do you think I’ll ever build a $500 EconoNAS?!

Giveaway

BrianMoses.net’s 2017 EconoNAS #FreeNASGiveaway

Read the whole story
patshead
405 days ago
reply
Plano, TX
Share this story
Delete

Quick and Easy Solar Hot Air Balloon

2 Shares

[Becky Stern] likes to harness the power of the Sun. Most of us will immediately think of solar cells and other exotic solar energy techniques. But [Becky] shows how to make a hot air balloon using nothing but tape and garbage bags.

The idea is quite simple. You form a large envelope from black trash bags and fill it with air. Becky does that by just running with it, tying it off, and topping off with a little manual blowing. Once the sun heats the black bag, it floats.

We were hoping she’d show us if there was any margin for payload, but she didn’t. However, electronics are getting smaller and lighter all the time. Sure, batteries weigh a lot, but maybe there would be an opportunity to mount some solar-powered device on the balloon to further harness the sun’s energy.

If you want to have a go at that, [Becky] has a free solar class that covers using the sun to engrave wood and using solar panels to charge a battery or power an Arduino.

We doubt this solar zeppelin will perform like a propane-powered balloon but it is still fun and educational. It doesn’t seem likely that you are going to reach space with it either.


Filed under: solar hacks



Read the whole story
patshead
442 days ago
reply
Plano, TX
Share this story
Delete

Maker Update: Self-Centering Drill Bits

1 Share

In Donald Bell’s latest Maker Update video, he looks at acoustic levitation, an Arduino made by Sony, a new kit by Anouk Wipprecht, self-centering drill bits, and a turning old monitors into a video wall. See show notes here.

From the video transcript:

For this week’s Cool Tools review we’re going to take a look at this self-centering drill bit set from Bosch. Great for makers and DIY home repairs. I wish I’d bought these years ago. This set cost me around $23, and if you want the same one, clicking the Amazon link in the show notes helps to support my videos and the Cool Tools blog.

If you’ve ever tried to attach some pre-drilled piece of hardware to something — a hinge, a latch, a coat hook — you’re probably familiar with the challenge of placing the hardware, marking where the holes need to be, and then carefully drilling. But if your pilot holes are just a little off — even in just one hole, the whole placement of the hardware will shift. It bugs the crap out of me.

A self-centering drill bit makes this process foolproof. The bit has a spring-loaded collar that sits into the hardware you’re attaching and keeps the drill bit dead-center. So long as your hardware doesn’t shift, the holes will be perfectly center.

This set from Bosch comes with three common bit sizes — a #6, #8 and #10. They’re also a quick change design that can just drop into an impact driver, making it easy to drill and screw with the same tool.

I will say that if you’re only using this occasionally, you could spend less and just get the smallest size. That will give you a centered pilot hole that you can expand with any regular drill bit. That said, for bigger projects, having the right size bit keeps you from having to drill the same hole twice.

-- Mark Frauenfelder

Self-Centering Drill Bit Set ($23)

Available from Amazon

Read the whole story
patshead
517 days ago
reply
Plano, TX
Share this story
Delete
Next Page of Stories