Design engineer · Milwaukee, WI

Ideas to working prototypes, quickly.

I'm Danny Blau. CAD, fabrication, electronics, and test, all from one pair of hands. I match the fidelity of each prototype to the question it has to answer.

ConceptSketch and CAD
RoughSize and shape in hand
BuiltLooks-like, Works-like, and/or fully integrated
RunningTested and documented

Projects

Personal builds.
Work from Milwaukee Tool and AndyMark client projects is confidential, so the public record of that work is in the patents section below.

Battle bots

Ongoing

Two combat robot projects with different jobs. ExtraVERT is my ideal kit bot build, made to get newcomers into the sport. HammerHead is my personal experiment with a new layout and form.

1-lb and 3-lb classes/custom PCBs/RP2040/custom firmware

01

ExtraVERT

Kit project

My ideal kit bot build, and the latest version is the most functional yet on my path to a combat-ready starter kit. It asks for no soldering, no CAD, and no 3D printing from the builder, and it is priced so it could be a holiday present.

RP2040 main board/3D-printed parts/routed UHMW chassis rails/solderless connectors/budget-first

Hero photoimages/extravert/hero.jpg

The latest version, with 3D-printed parts and routed UHMW chassis rails.

Desired outcome

A builder who is opinionated and motivated about where to go next. The kit is an attempt to bring more people into the sport with a lower barrier than the existing kits.

I want to break down the idea that combat robotics starts with the build. A journey can start with a competition and inspire a traditional build from there.

Problem
Combat robotics has a steep skills barrier. Newcomers have to solder, learn CAD, and print parts before they ever drive. I wanted to remove those steps so people get hooked first and learn the skills because they want to.
Constraint
Cheap enough to give as a gift, tough enough that it isn't destroyed in its first fight. It doesn't have to win, only to survive and stay fun. Target price: $150.
What I did
  • Designed the main board in KiCad around an RP2040: battery in, drive motor drivers, 5V regulation, solderless headers for the radio and an external weapon ESC, and a through-hole on/off switch of my own design.
  • Priorities, in order: safety, reliability, cost, performance.
  • Early versions ran a Nano-style board in a printed chassis. Later versions moved to the custom RP2040 board, and some of those had custom injection-molded wheels and motor mounts.
  • The latest version, with orange wheels and a mustard body, is the most functional so far: it uses 3D-printed parts and chassis rails routed from UHMW.
  • Revision 4 adds the weapon motor ESC to the board and removes the redundant wheel motor drivers, because a two-motor drivetrain doesn't need an onboard backup. Together, the changes added no size to the board.
Why 3D printed
The printed parts mean simple production with no tooling, easy spares, and a straightforward way for new builders to start modifying their kits once they are ready.
Result
  • Team GRAFFITI used the kits as their offseason skill-building program. The students split into three teams, and each built a copy of the kit and decorated it before our inaugural Fight Night.
  • After seeing the robots in action, they had many ideas for improving them. They modified the CAD models in Onshape, 3D printed and routed new parts, and fought the robots again in the arena.
  • The students never had to solder or worry about the electronics, so they could put their motivation into the hardware upgrades.
  • There is also room on the RP2040 for anyone who wants to modify the software for added functionality or tuning.

Process

Roughimages/extravert/rough.jpg
Rough · early generationFirst custom PCB next to the printed chassis half it was designed to fit. The assembled chassis on the right runs a Nano-style board, motors, and a receiver.
Builtimages/extravert/built.jpg
Built · later generationCustom RP2040 board with XT-style power connectors and terminal blocks, plus custom injection-molded wheels and motor mounts.
Internalsimages/extravert/internals.jpg
Internals · running on the custom boardThe custom main board in its tray, with a microcontroller module, terminal blocks, and XT power connectors, wired to a small ESC and the weapon motor.

Board design

KiCad layoutimages/extravert/kicad-layout.jpg
3D renderimages/extravert/kicad-3d.jpg

The main board in KiCad: the layout (left, labeled "rev4 ESC extension") and a 3D render (right). Revision 4 adds the weapon motor ESC to the board and removes the redundant wheel drives, and together the changes added no size to the board.

In the arena

Arena stillimages/extravert/arena-1.jpg
Arena stillimages/extravert/arena-2.jpg

The original kit builds, driven by high school students, fighting this summer.

After the students' upgrades

Modified robotsimages/extravert/modified.jpg

The students' modified robots, after the upgrade round: new parts designed in Onshape, then 3D printed and routed.

02

The leap to twin weapons

ExtraVERT is where I refine the kit. HammerHead is where I try something different.

Weapon layoutimages/battle-bots/weapon-layout.jpg

One weapon in the center, then one at each corner. ExtraVERT (right) carries a single weapon in the middle of the robot. An early HammerHead mockup (left) moves to a weapon at each of two opposite corners.

With a weapon at each corner, more of the ways an opponent can come at the robot end with a weapon on target, so it has to be lined up less precisely to land a hit.

03

HammerHead

Personal project

My personal experiment with a new layout and form: a 1-lb antweight built around twin corner weapons, with its next generation already underway, running the same main board as ExtraVERT.

1-lb antweight/twin corner weapons/ExtraVERT main board/RP2040/custom firmware

Hero photoimages/battle-bots/hero.jpg
Problem
In the 1-lb class every gram is spoken for. The challenge with HammerHead is fitting a second weapon motor and blade into that already tight weight limit, while the robot still has to survive being hit repeatedly and stay easy to drive and repair.
Constraint
One pound, total. Chassis, electronics, battery, and weapon all share that budget.
What I did
  • Designed and built the bodies, control boards, and motor control, then wrote the firmware.
  • HammerHead runs the same main board as ExtraVERT, so the two robots share one electronics design.
  • I rewrote the firmware to be user-friendly, with a labeled tuning zone, failsafe handling, spin-in-place correction, a non-blocking weapon ramp, and proportional weapon control from the radio.
  • I reshaped the body to look more like a shark without adding weight, studying it in cardboard and clay. With a tournament approaching, I then simplified the design to simpler shapes in CAD.
Result
What held up

The robot held up well. It took a beating and kept on fighting.

What I learned

The bite on the weapons wasn't large enough, and the weapon motors weren't ideal for this use. I also learned I don't have enough battery to last a full match with both weapons running at full blast.

Next generation

It's underway, with a single weapon motor spinning both warheads. That cuts battery drain, simplifies the robot, and gives back weight for a larger battery or more armor.

Process

Roughimages/battle-bots/rough.jpg
Rough · drivetrain optionsThe first decision. Three layouts tried on the same CAD body: two wheels, mecanum, and a roller, sketched over the model. I picked the two-wheel version for simplicity and robustness. Click to enlarge.
Conceptimages/battle-bots/concept.jpg
Concept · cardboard and clay studyA clay form over a cardboard core, testing how far I could push the body toward a shark.
Builtimages/battle-bots/cad.jpg
Built · simplified CADWith a tournament fast approaching, I simplified the design to simpler shapes. Blue body shells over white chassis plates, a wheel at each end, and a gold weapon hub at each of two opposite corners. Modeled in Onshape.
Runningimages/battle-bots/running.jpg
Running · after the fightsHammerHead after competition, with the scuffs and scrapes to show for it.
Watch the FUN interview on YouTube

Video

At the tournament I was interviewed for the FIRST Updates Now (FUN) YouTube channel. I talked about my robot, the new things I was doing with it, and my background from FIRST Robotics to combat robotics.

FIRST Robotics

Six weeks from game reveal to a working robot

Team 8701 · GRAFFITI

The FIRST Robotics Competition build season is six weeks long. It is where I learned to prototype fast, decide early, and work as one team.

6 weeksfrom game reveal to a working robot
2026Creativity Award, Team 8701
2009when I started mentoring
Team
FRC team 8701, GRAFFITI, based in the Greendale, WI area. 2026 Creativity Award winners.
My role
Co-Lead and Lead Technical Mentor for GRAFFITI. I have mentored FIRST teams since 2009, and I have held a lead, co-lead, or lead technical mentor role for 13 years, since the 2013 season. I also lead the team's yearly goal setting and tracking, so the whole organization moves as one cohesive group with a shared purpose.
How we compete

We are a small, young team competing against older, larger, more sustainable teams, and we don't shy away from that.

We find our success because of our size, not in spite of it.

The six weeks

  1. WK 1Read the game, commit the chassis

    Read the game, discuss strategy, and start prototyping new functions. The chassis is committed to manufacturing.

  2. WK 2Refine prototypes, start the chassis

    Prototypes get refined. The chassis comes back from manufacturing and assembly begins.

  3. WK 3Fabrication gets underway

    Game-specific parts are well into fabrication, and they are assembled onto the chassis as they are ready.

  4. WK 4Fab and build

    Fabrication and assembly continue side by side.

  5. WK 5Electrical work begins

    Wiring and electronics go onto the robot.

  6. WK 6Final assembly and testing

    Final assembly and all-up testing, with the first scrimmage at the end of the week.

Software, in parallel

The code team runs alongside the hardware build, starting with high-level dummy code and moving to the real thing. This year we plan to have a practice chassis the software team can use for initial testing before the main robot is ready.

2026 season photos

Prototypeimages/first/prototype.jpg
PrototypeAn early cardboard-and-foam mock-up with two hooked arms reaching up out of a box of yellow balls. It was built to see how many of the game balls we could hold alongside the other space claims on the robot. Click to enlarge.
Buildimages/first/build.jpg
Build · chassis wiringAn aluminum frame and a self-reinforced polypropylene belly pan, partway through wiring. The four yellow caps are the swerve modules, which drive the robot around the field.
Buildimages/first/build2.jpg
Build · week 5The frame cage is up around a clear center tube, and the swerve modules are wired. The climbing system is installed in the robot. The ball thrower and hopper are being built separately, to be installed later.
Version 2images/first/v2.jpg
Version 2 · rebuilt for StateMuch of the robot stayed the same, but the ball launcher was made wider for more throughput. The climber was removed to make room. It worked well, but climbing wasn't worth as many points as better, faster launching.

The team

Team photoimages/first/team.jpg

Team GRAFFITI receiving the Creativity Award in 2026 for their use of a spooled climber and integrated, nested packaging of multiple sub-systems.

Photo credit: FIRST Wisconsin

Beyond the robot

  • Alliance selection tool

    A browser tool with live-sortable stats, a draft simulator, a double-elimination bracket with projected winners, and an AI pick advisor. Built with real qualification data from MROC 2026 and a prior Wisconsin event. Team 8701 was Alliance 6 captain in the MROC playoff field.

  • Match video scouting In progress

    A review system for the 2026 game, REBUILT. It scores each robot's points by match phase and sorts robot roles into offense, defense, and defended. The goal is to rank teams for alliance selection and prep the drive team with opponent clips.

Rethinking how we run build season

Problem
We lost a large number of seniors and recruited mostly freshmen. In the past we designed and built the entire robot ourselves, with help from a sponsoring machine shop and publicly shared resources from the larger FIRST community. That approach no longer fit the roster.
What I did
  • In the debrief of the 2026 season and the run-up to the 2027 kickoff, I led conversations first with the mentors and then with key students about changing how we move through build season.
  • I steered the team toward more off-the-shelf designs and parts, so we can get running sooner and leave more time for programming, drive practice, and testing.
  • The change was in mindset and theory of success as well as in the plan, so I didn't rush it. We worked through it over several weeks, including backup plans and what-if scenarios.
  • Key students from both the hardware and software teams weighed in and voiced their concerns. I worked to show them that the upside outweighed the downsides, and that there would be room for them to do what they are most passionate about.
Where it stands
Planning is aligned and underway. We will see how it plays out when the 2027 season kicks off in January.

More projects

DMX-to-servo module

Toward commercialization

An 8-channel converter that lets a standard theatrical lighting console drive robotic stage elements.

8 channels/Seeed XIAO RP2040/DMX-512 at 250 kbaud/configurable start channel

Hero photoimages/dmx-servo/open.jpg
Problem
Robotic elements on a stage need a control path the crew already knows. Theaters already run DMX from the lighting console, so the robotics should speak it.
Constraint
Under $100, and about the size of a deck of cards or smaller. It also has to be reliable in a live production and simple for a crew to address.
What I did
  • This is the first functional mockup: a working 8-channel DMX-512 to servo converter, hand soldered and much larger than it needs to be. A custom PCB would shrink it.
  • When an off-the-shelf DMX library didn't hold up, I wrote a custom raw-UART receiver on the RP2040's hardware UART (250 kbaud, 8N2) with gap-timing break detection.
  • I replaced a faulty RS-485 shield with a bare breakout and added a configurable start channel.
  • I'm now looking at channel-select options (display, DIP switches, shift registers) for a productized version.
Result
This unit is running on the stage I built for my daughter. It drives a winch that opens and closes the curtains, through a touch screen control panel that also controls the lights.

The full system

Control boximages/dmx-servo/control-box.jpg
Control panelA wall-mounted touchscreen with color-sequence buttons for the back and arch lights. It also controls the curtain winch.
Builtimages/dmx-servo/closed.jpg
DMX-to-servo moduleThe module in its enclosure, mounted on the wall under its power supply.
Curtain winchimages/dmx-servo/winch.jpg
Curtain winchA gearmotor driving a cord spool that opens and closes the stage curtain.

Costume and prop lighting

Ongoing

Addressable LED builds for costumes and props, sized and shaped for the person wearing them.

NeoPixels/QT Py SAMD21/PlatformIO/wearable power packs

Hero animationimages/wearable-lighting/hero.gif

The Ember lit from inside, pulsing on and off in hand.

Problem
Lighting for costumes has to be controllable, comfortable to wear, safe for kids, and invisible until it matters. A prop has to look like a prop first and electronics second.
Constraint
Size and shape were driven by the story. The Ember is roughly 4 × 2 × 1 to 1.5 inches, and I needed to stuff the electronics in without access after the prop was made.
What I did
  • Project Ember is a prop for this year's family Halloween costume: Hades, as he appears in Disney's Descendants franchise. This version of Hades is full-tilt punk rocker, with a leather jacket and spiked studs, and his Ember is the source of his magic.
  • I sculpted the Ember at a similar size to the one in the movie, with more close-up detail and a more pocket-friendly shape.
  • I built a silicone mold from the sculpt for casting, and packed a string of independently sparkling NeoPixels, a small LiPo battery, and a QT Py SAMD21 controller inside.
  • A second project is in the idea phase: lighting effects for a costume, controlled from the tech booth.

Process

Conceptimages/wearable-lighting/concept.jpg
Concept · design boardReference crystals and flames, a sketched crystal, the flame-carved body, glowing light spots, and a hand sketch for scale. Click to enlarge.
Roughimages/wearable-lighting/rough.jpg
Rough · molding the sculptThe original sculpt inside a mold box, with two-part silicone poured around it to make a mold for casting the Ember.
Electronicsimages/wearable-lighting/electronics.jpg
Electronics · inside the EmberIndependently addressable NeoPixel boards, wired to a small LiPo battery, all of it small enough to live inside the Ember.
Builtimages/wearable-lighting/built.jpg
Built · the moldThe silicone mold of the sculpt, strapped shut, with a pour sprue on top.

The finished Ember

Closeupimages/wearable-lighting/ember-closeup-1.jpg
Closeupimages/wearable-lighting/ember-closeup-2.jpg

Two closeups of the cast Ember, held in hand for scale. It was cast in glitter-filled epoxy and painted to highlight the hotrod flames in the sculpt.

Patents

The public record of my employer work. Each link opens the filing on Google Patents, drawings included.

NumberTitleStatusAssignee
US12697707B2 Trigger mechanism for power toolA trigger that senses squeeze with a force sensor instead of moving, so the tool powers on without physical trigger travel. Granted 2026 Milwaukee Electric Tool
US12519372B2 Pass-through brushless motorA motor whose stator frame lets the output drive pass through to the rotor, removing the direction-changing gears in a reciprocating saw. Granted 2026 Milwaukee Electric Tool
US12402723B2 FIRST LEGO League tableDesigned at AndyMark after discovery research with the FIRST LEGO League community. Granted AndyMark
US20250079610A1 Single cell battery pack platformA single-cell battery pack that delivers at least 170 watts to a handheld tool through a stem-and-base design. Pending Milwaukee Electric Tool
US20230390837A1 Coping systemA power tool for coping cuts on trim, driven by an outer-rotor brushless motor. Pending Milwaukee Electric Tool
US20260102894A1 Power tool including an internal auxiliary power supplyAn energy storage circuit inside the tool housing that a controller can connect with the battery pack in series or parallel. Pending Milwaukee Electric Tool
US20260163444A1 Power tool having busbarsRigid metal busbars replace wires between the battery terminals and the circuit board, and can be insert-molded into the housing. Pending Milwaukee Electric Tool

I am a named inventor on each filing, alongside co-inventors. Pending applications are published but not yet granted, and status can change.

About

I'm a design engineer in Milwaukee. At Milwaukee Tool's front-end innovation group I built prototypes, often as the only engineer on a team, turning open research questions into physical things people can pick up.

Before that I spent seven years at AndyMark running a product line of 700+ parts and a yearly game design program, and before that I designed multi-layer insulation for satellites. I have mentored FIRST Robotics teams since 2009.

This page is where I keep the work I can show. If you have a fast-moving project and need someone who can take it from sketch to something running, I'd like to hear about it.

Speakingimages/about/speaking.jpg
Speaking at the Dr. Woodie Flowers Mentor Conference in Detroit, MI, 2026. The talk was on goal setting when the wins aren't obvious, and finding success on any team.

Contact