DIY Active Sub-GHz Differential Scope Probe

Fancy measurement gear is often expensive to buy, but some bits of kit are entirely DIY’able if you’re willing to put a little work into the project. [Christer Weinigel] needed to get some measurements of a differential clock signal that was ticking away around 500 MHz. El-cheapo probes aren’t going to cut it here. They won’t have the bandwidth and most off-the-rack probes are single-ended, that is they’re referenced to ground. [Christer] needed the difference between two balanced signals, neither of which is grounded. In short, [Christer] needed a high-frequency active differential oscilloscope probe, and they’re not cheap. So he built one himself.

simulation-schematicThe circuit in the probe is really just an instrumentation amplifier design with a modified input stage and a 50 ohm output impedance. (See this article on in-amps if you need to brush up.) With higher frequencies like this, it’s going to be demanding on the op-amp, so [Christer] spent some time simulating the circuit to make sure it would work with his chosen part. Then he made up a bunch of PCB designs and had them made. Actual results matched fairly well with the simulation.

With some minor tweaking on the input damping resistors, he got a tool that’s dead flat up to 300 MHz, and totally usable up to 850 MHz. If you tried to buy one of these, it’d set you back the cost of a few hundred lattes, but this one can be made for the price of one or two if you get the PCBs done cheaply. Of course, the design files are available for your own use. Kudos [Christer].

Edit: By total coincidence, Bil Herd just posted a video intro to differential signals. Go check it out.

And thanks to [nebk] for the tip!

Mid Century Modern Speaker from 90s Road Trash

[BarryAbrams] found some 90s speakers on the side of the road.  At first he thought he might have made a real score, but his coworker who knows about this sort of thing (we all have one) let him know they were merely average. Undaunted, he removed the speakers from their MDF housing, fixed a small dent in one of the tweeters, and got to work.

He cut a new frame for the speakers out of plywood. He adorned the plywood box with maple and walnut from a local supplier. The box then got a coat of urethane. His skill at sign making showed in the final finish, and the wood looks very good. Our only complaint is the straight legs instead of the slightly angled and tapered ones common to mid-century modern furniture style.

The electronics are a Chinese amp and a Sonos knock-off. [Barry] only needed to control the volume and power for the speaker set. He came up with a clever 3D-printed knob and switch configuration. When the volume is turned all the way down the speaker set turns off.

The end result sounds and looks better than anything he could get for the $125 US Dollars he spent on the project. We certainly wouldn’t complain if this were a fixture in our living space.

Continue reading “Mid Century Modern Speaker from 90s Road Trash”

A Home CNC Built By Someone Who Knows Their Stuff

[thisoldtony] has a nice shop in need of a CNC. We’re not certain what he does exactly, but we think he might be a machinist or an engineer. Regardless, he sure does build a nice CNC. Many home-built CNCs are neat, but lacking. Even popular kits ignore fundamental machine design principles. This is alright for the kind of work they will typically be used for, but it’s nice to see one done right.

Most home-built machines are hard or impossible to square. That is, to make each axis move exactly perpendicular to the others. They also neglect to design for the loads the machine will see, or adjusting for deviation across the whole movement. There’s also bearing pre-loads, backlash, and more to worry about. [thisoldtony] has taken all these into consideration.

The series is a long one, but it is fun to watch and we picked up a few tricks along the way. The resulting CNC is very attractive, and performs well after some tuning. In the final video he builds a stunning rubber band gun for his son. You can also download a STEP file of the machine if you’d like. Videos after the break.

Continue reading “A Home CNC Built By Someone Who Knows Their Stuff”

Build Yourself an Awesome Modular Power Supply

You may think you’ve built a power supply for your bench. Heck, we all do. But until you check out [Denis]’s bench power supply build, you may not even know what you’re missing.

[Denis]’s design is nearly entirely modular and targeted to the intermediate builder. It’s built on easily available parts and through-hole components. It’s got an Arduino running as the brains, so you’re going to be able to hack on the code when you feel like tweaking it. But easy doesn’t mean light on features. Let’s walk through the build together.

6203871454520455862_thumbnail

It starts off with a pre-regulator: a switching MOSFET that gets the voltage down to just a couple volts above the target value. Then it’s off to the post-regulator that includes all of the fine adjustments, the DAC and ADC interfacing to the microcontroller, and some fancy features like a “down-programmer” that turns the output off extra quickly.

On the user end of things, [Denis] made a very sleek board that incorporates a TFT touchscreen for the controls, Arduino connections, and the obligatory banana plug outputs. There’s opto-isolation on the SPI bus, a real-time clock, and a bunch more goodies on board. He’s in his third revision of this module, and that level of refinement shows. It’s even SCPI compliant, meaning you can control remotely using an industry-standard protocol.

So what would you do with a ridiculously fancy power supply under microcontroller control? Test out battery charging algorithms? Program test routines to see how your devices will work as their batteries drain out? We have no idea, but we know we want one!

3D Printer Tool: Set Your Extruder Steps With Ease

My printer has other issues that i'm still tuning out, but the warping in PLA and excessive surface roughness has all the signs of over extrusion.
My printer has other issues that I’m still tuning out, but the warping in PLA and excessive surface roughness has all the signs of over extrusion.

I have an old Prusa i2 that, like an old car, has been getting some major part replacements lately after many many hours of service. Recently both the extruder and the extruder motor died. The extruder died of brass fill filament sintering to the inside of the nozzle (always flush your extruder of exotic filaments). The motor died at the wires of constant flexing. Regardless, I replaced the motors and found myself with an issue; the new motor and hotend (junk motor from the junk bin, and an E3D v6, which is fantastic) worked way better and was pushing out too much filament.

The hotend, driver gear, extruder mechanics, back pressure, motor, and plastic type all work together to set how much plastic you can push through the nozzle at once. Even the speed at which the plastic is going through the nozzle can change how much friction that plastic experiences. Most of these effects are somewhat negligible. The printer does, however, have a sort of baseline steps per mm of plastic you can set.

The goal is to have a steps per mm that is exactly matched to how much plastic the printer pushes out. If you say 10mm, 10mm of filament should be eaten by the extruder. This setting is the “steps per mm” in the firmware configuration. This number should be close to perfect. Once it is, you can tune it by setting the “extrusion multiplier” setting in most slicers when you switch materials, or have environmental differences to compensate for.

This little guy lets you tune the steps per mm exactly.
This little guy lets you tune the steps per mm exactly.

The problem comes in measuring the filament that is extruded. Filament comes off a spool and is pulled through an imprecisely held nozzle in an imprecisely made extruder assembly. On top of all that, the filament twists and curves. This makes it difficult to hold against a ruler or caliper and get a trustworthy measurement.

I have come up with a little measuring device you can make with some brass tubing, sandpaper, a saw (or pipe cutter), a pencil torch, solder, and some calipers. To start with, find two pieces of tubing. The first’s ID must fit closely with the filament size you use. The second tube must allow the inside tubing to slide inside of it closely. A close fit is essential.

Continue reading “3D Printer Tool: Set Your Extruder Steps With Ease”

Crappy Robots And Even Crappier Electronics Kits

Robots and DIY electronics kits have a long history together. There probably isn’t anyone under the age of forty that hasn’t had some experience with kit-based robots like wall-hugging mouse robots, a weird walking robot on stilts, or something else from the 1987 American Science and Surplus catalog. DIY robot kits are still big business, and walking through the sales booths of any big Maker Faire will show the same ideas reinvented again and again.

[demux] got his hands on what is possibly the worst DIY electronics kit in existence. It’s so incredibly bad that it ends up being extremely educational; pick up one of these ‘introduction to electronics’ kits, and you’ll end up learning advanced concepts like PCB rework, reverse engineering, and Mandarin.

Continue reading “Crappy Robots And Even Crappier Electronics Kits”

Make PCBs with DLP, OMG!

There’s so many ways to skin the home-fabrication-of-PCBs cat! Here’s yet another. [Nuri Erginer] had a DLP projector on hand, and with the addition of some reducing optics, managed to turn it into a one-shot PCB exposer.

If you’ve ever used photo-resist PCB material before, you know the drill: print out your circuit onto transparency film, layer the transparency with the sensitized PCB, expose with a UV light for a while, dissolve away the unexposed resist, and then etch. Here, [Nuri] combines the first three steps in one by exposing the board directly from a DLP projector.

The catch is that the projector’s resolution limits the size of the board that you can make. To fab a board that’s 10cm x 10cm, at XGA resolution (1024×768), you’ll end up with a feature size of around 0.004″ in the good direction and 0.005″ in the other.

For DIP parts, that’s marginal, but for fine-pitch or small SMT parts, that won’t do. On the other hand, for a smaller board, optimally one in the same 4:3 ratio, it could work. And because it exposes in one shot, you can’t beat the speed. Cool hack, [Nuri]!

When you need more precision, strapping a UV laser to an accurate 2D robot is a good way to go, but it’s gonna take a while longer.