Close

Hack Chat Transcript, Part 1

A event log for Life at JPL Hack Chat

It ain't rocket surgery. No wait, it is...

lutetiumLutetium 08/21/2019 at 20:020 Comments

Hi everyone, welcome to the Hack Chat. It's time for us all to learn about what for a lot of us would be a dream job - working at the Jet Propulsion Lab. Arko joins us today to talk about what it's like to be a robotics engineer at JPL.

Welcome Arko! Can you start us off with a little about your background?

arko12:00 PM
Howdy everyone!

Aaron Waruszewski joined  the room.12:00 PM

Aaron Waruszewski12:01 PM
Hello

Nicolas Tremblay joined  the room.12:01 PM

gmarmolburgos12:01 PM
Hello

arko12:01 PM
Sure! I'm an electrical engineer with a BS from Cal Poly Pomona. I have 12+ years in circuit design, robotics, and random hackery things.

arko12:02 PM
Worked on many projects ranging from Mars rovers to DeLorean Time Machines.

Aaron Waruszewski12:02 PM
Is there any way for someone with years of professional software experience to get a job working at JPL even if they don't have a degree?

arko12:02 PM
For more info, I have a short portfolio online: http://arkorobotics.com/portfolio.html

Aaron Waruszewski12:03 PM
Nice

Christopher Bero12:03 PM
Aw yeah, you have FRC on your site!

arko12:05 PM
Hmmm good question... I personally don't know of any SW engineers without a degree and it maybe that there's a policy on having a degree. I would suggest reaching out to our HR department for a better answer: https://jpl.jobs/

arko12:05 PM
Oddly enough, some of the best SW engineers I know don't have degrees...

Tanya F. Hakala joined  the room.12:05 PM

Christopher Bero12:06 PM
What's been your favorite project at JPL?

arko12:06 PM
@Christopher Bero Yes! I was part of Team 589 for all 4 years of high school and went back to mentor :)

Aaron Waruszewski12:06 PM
Ok cool.

Christopher Bero12:06 PM
Nice!

arko12:06 PM
hmmm, I'd have to say the Mars 2020 Rover's Lander Vision System:

Christopher Bero12:07 PM
That looks like a pretty intense setup.

arko12:07 PM
I had the opportunity to fly in a helicopter at 17000ft to test hardware destined for mars.. so that has to be my favorite (so far!): https://twitter.com/arkorobotics/status/1132341128848498688

arko12:08 PM
Yeah, we packed *a lot* of hardware into that helicopter

Aaron Waruszewski12:08 PM
What would you say is the best project that has already been launched?

Nicolas Tremblay12:09 PM
DeLorean Time machine???

arko12:10 PM
For context, we designed and developed a very powerful flight computer/vision system for the M2020 Rover (https://mars.jpl.nasa.gov/mars2020/mission/technology/entry-descent-landing/) so that it can figure out where it is during landing and to divert away from hazards

Christopher Bero12:10 PM
Sweet

arko12:11 PM
We packed an engineering model of the flight hardware and a bunch of "ground support equipment" hacks into the helicopter as well as a gimbal with a camera and IMU on it

arko12:11 PM
and used the gimbal to simulate the parachute motions during descent.

Aaron Waruszewski12:11 PM
Nice

Were you testing more of the algorithms that will be used for landing or the actual hardware that will fly on the mission too?

arko12:12 PM
JPL has this philosophy of "Test as you fly. Fly as you test", meaning you should test your system like how you plan to fly it

Randall D joined  the room.12:12 PM

arko12:12 PM
and remember to fly as you tested it :)

arko12:13 PM
@Dan Maloney Great question. It was a full systems test.

arko12:13 PM
Up until then, we had tested everything in pieces. So this field test used final flight software and hardware in the most 'flight like' conditions possible

arko12:14 PM
The actual flight hardware itself is actually already in the rover () and was tested separately

gmarmolburgos12:14 PM
What has been your hardest proyect to achieve?

arko12:14 PM
more so a hardware design verifcation test (TVAC/EMI EMC/Vibe/Shock)

arko12:15 PM
what we flew in the heli was a flight replica

So 2020 is like the size of a Mini Cooper? Maybe a bit bigger?

arko12:16 PM
@gmarmolburgos hmmm, I'd have to say M2020 was the most difficult overall due to complexity

arko12:16 PM
Not to say my other tasks haven't been difficult

arko12:16 PM
:P

gmarmolburgos12:17 PM
Thanks 😊

arko12:17 PM
@Dan Maloney M2020 is the same size (ish) but with more instruments - Humans for scale

Christopher Bero12:17 PM
What's the project lifecycle like at JPL? What kind of teams are there? Do you have multiple tasks/projects going at once?

arko12:17 PM
Love this question

Adam12:17 PM
How does the component selection process change when designing for environments as harsh as space? What are some surprising things you need to take into account?

Christopher Bero12:17 PM
That's a big rover.

gmarmolburgos12:18 PM
What would you recommend a high school student to read about if he wants to study aeroespacial engerneering? (Any books or authors you could recommend) (podcasts even)

arko12:18 PM
So project lifecycles vary from task to task, as do teams

Aaron Waruszewski12:18 PM
What programming language has been the most proven to use when making the rovers/probes?

arko12:18 PM
So JPL, in general, has two types of projects: Flight and Non-Flight

arko12:19 PM
most folks work Flight only

arko12:19 PM
some work Non-Flight only... and there are a few who work both

arko12:20 PM
I happen to be "both". Which can be really challenging from a time management perspective, but it allows me to live in both worlds and to exchange experiences between them.

arko12:20 PM
As for teams, again it varies, but my rule of thumb is to pick working with good teams.

arko12:21 PM
cool tasks come second because... lets face it.. this place is a candy store of tasks. It's hard to find a lame task.

Christopher Bero12:21 PM
Awesome haha

arko12:22 PM
As for component selection, we have an internal catalog of not only known good parts, but designs as well

Nicolas Tremblay12:22 PM
*Bows down in reverence*

Nicolas Tremblay12:22 PM
JPL sounds a bit like nerdvana

arko12:23 PM
Need 5V in your design? Boom, known working design with traceable testing and stocked parts

arko12:23 PM
workign footprints and schematic, etc

Adam12:23 PM
I'm not sure you're allowed to answer this, but who do you guys go to to get your PCB prototypes/finals done

arko12:23 PM
Lots of different vendors

Christopher Bero12:24 PM
Huh

arko12:24 PM
Flight tasks require ITAR compliance, so that sorta limits us.

arko12:24 PM
but for prototype projects, I like to use US vendors

Frank Buss12:24 PM
does JPL use some formal verification when programming, like Spark/Ada?

arko12:24 PM
hah, Ada

ap-tech12:25 PM
Chinese or American PCB vendors?

arko12:25 PM
American. Purchasing out of the country is a MAJOR pain and requires justification

Frank Buss12:25 PM
was not Ada's fault that one rocket crashed :-)

arko12:25 PM
there are formal verification processes

arko12:25 PM
lots of code reviews

arko12:25 PM
for the most part, we use C/C++ for SW and Verilog for HW

Captayne joined  the room.12:26 PM

arko12:26 PM
@Frank Buss

arko12:26 PM
^ My favorite talk on Flight SW at JPL

Aaron Waruszewski12:27 PM
Awesome

Frank Buss12:27 PM
cool, will watch it. But formal verification is very difficult for C, but I guess you use some standard like MISRA C?

arko12:27 PM
We def do

arko12:28 PM
We use a bunch of tools like coverity for static analysis

arko12:28 PM
and dynamic

arko12:28 PM
internal tools too

Adam12:29 PM
@Frank Buss From wikipedia it looks like: "The NASA Jet Propulsion Laboratory C Coding Standards[6] are based on MISRA-C:2004."

Frank Buss12:29 PM
you can still do a lot of silly things in C, why not Ada? It is supposed to be much safer

arko12:29 PM
we have a lot of 'lessons learned' from past missions which we incorporate into flight code rules

arko12:29 PM

https://en.wikipedia.org/wiki/The_Power_of_10:_Rules_for_Developing_Safety-Critical_Code

WIKIPEDIA

The Power of 10: Rules for Developing Safety-Critical Code

From Wikipedia, the free encyclopedia Jump to navigation Jump to search The Power of 10 Rules were created in 2006 by Gerard J. Holzmann of the NASA/JPL Laboratory for Reliable Software. The rules are intended to eliminate certain C coding practices which make code difficult to review or statically analyze.

Read this on Wikipedia

arko12:30 PM
everyone one of these rules traces back to a mission failure

arko12:30 PM
Ada's cool, but I'm a Rust person myself :P

arko12:30 PM
*waits for fight to ensue*

Adam12:30 PM
I know that recently a lot of software dev has been moving over to the cloud, is there anything being done using cloud computing at JPL?

arko12:30 PM
LOSTS

Christopher Bero12:30 PM
Is there anything JPL/NASA publishes that you feel doesn't get seen enough? Cool images or interesting datasets?

arko12:30 PM
LOTS*

arko12:31 PM
we do more than just code for flight hardware, in fact those teams are much smaller than the other apps we make inhouse

baldwinngo081312:31 PM
hey arko, im not sure how relevant this is to you, but what are your thoughts on the future of RF engineers in the space industry? Would look to work at JPL or NASA one day on their satellites/comms. Currently at an internship in Pasadena lol, literally down the street from JPL.

baldwinngo081312:31 PM
love*

arko12:31 PM
I cant really speak to it since im not a cloud person :P

arko12:32 PM
@Christopher Bero need to think about that. The JPLRaw youtube channel is pretty cool

arko12:32 PM

https://www.youtube.com/channel/UCYt2JN2z3ZXQpgPHt_eKcTw

YOUTUBE

JPLraw

JPL Raw is a channel used by NASA's Jet Propulsion Laboratory to deliver on-demand video to news media and others. The majority of video on this channel cons...

Read this on YouTube

morgan12:32 PM
I work at Chabot Space and Science Center in Oakland, we take advantage of the JPL assets being available in a lot of our work. From images, to 3d model to code and data

Christopher Bero12:32 PM
Thanks!

Christopher Bero12:32 PM
Oh, nice!

morgan12:33 PM
And being able to reach out to The Studio has been a great resource in exhibit development

arko12:33 PM
@baldwinngo0813 Nice! I know we are always looking for great RF engineers, I'd recommend applying if you are interested. For background, I applied to JPL 12+ times

Christopher Bero12:33 PM
Woah, to different areas inside?

arko12:33 PM
sometimes it's just that I didnt have enough experience or that they had already filled the position

Christopher Bero12:34 PM
Ah.

Adam12:34 PM
What was different about the last time (other than getting the job)

arko12:34 PM
@morgan nice!

morgan12:34 PM
@arko do you ever get to do work with The Studio?

arko12:35 PM
@Adam Fresh out of college. Timing. Made connections at JPL so I had a foot in the door.

arko12:35 PM
@morgan Not familiar, you mean the media department?

Discussions