- Aug 25, 2020
- Reaction score
We are talking about the mainstream companies....so I didn't list the smaller manufacturers.
Yes, it may be "software", but it still needs to be tested with lab equipment at least. Not every engineer has the required equipment at home because the company may not have enough to send home or the inclination to disperse it's equipment. But that said, it can still be written at home to save some time.There's no excuse for not doing software updates. That sorta thing can be done in your underwear at home.
Yes, it may be "software", but it still needs to be tested with lab equipment at least. Not every engineer has the required equipment at home because the company may not have enough to send home or the inclination to disperse it's equipment. But that said, it can still be written at home to save some time.
I'm curious, how often your software developers need to do the lab tests? Debugging DSP things probably require lab equipment to be available every 30 minutes, but I'm pretty sure everything else can be tested without such a gear.Most people have no idea how much money lab equipment costs to do work in the 30-40ghz range. If you are starting from scratch and doing it on the cheap, I would say it costs about $300k to fully outfit a rudimentary high frequency lab, not including software. A single VNA or Specan can be $100k+, depending on your needs. Connectors capable of passing 40ghz cost $100-200 each, and six inch cables are $300. Assortment of X, K, and KA band horns (few grand easy). And you need an assortment of connector types and cable lengths (90 degree, straight through, edge-mount, surface mount, 2.92, 2.4, etc). Signal generators are thousands. Calibrated hybrid dowconverters, same thing. Tem cells, filters, amplifiers, test fixtures, power supplies, scopes, load testers, microscopes, soldering stations, hot air stations, SDRs, reflow ovens, LPKF machines. And you don't need just one of these things, otherwise you will have engineers constantly bumping into each other.
Then, if you want to talk about "nice to haves" which help for making things like text fixtures, backing plates etc:
Dual extrusion 3d printer
CNC machine for precision backing plates in copper and brass
Full assortment of power and hand tools
Complete hardware assortment (nuts/bolts/washer)
Stock of different thickness copper and brass sheets
Not to mention $150k+ in annual software licenses:
CST Microwave Office
Plus a bunch I am forgetting
And, you need multiple seats of each (otherwise only one engineer can use it at a time) and you need multiple operating system licenses (linux and windows).
Low frequency stuff is cheap...but once you start getting above the 18-20ghz range things exponentially increase in price.
"Working from home" can only get engineers so far when you need access to this equipment.
I'm curious, how often your software developers need to do the lab tests? Debugging DSP things probably require lab equipment to be available every 30 minutes, but I'm pretty sure everything else can be tested without such a gear.
Yes firmware is low-level software while things like UI are high-level. High-level software is less hardware-dependent. That IC thing was new to me, I thought most detectors have separate DSP chip like Blackfin?Totally depends on what they are working on. Adding a feature to UI? Can be done easily at home. But "firmware" also includes things like setting gain levels sometimes, and that you may need equipment for.
There is also a difference between detector architectures. With something like the Pro M, the same IC that handles the UI is also handling signal processing. This is why stupid errors could creep into firmware sometimes - engineers at Genevo would go change the laser tone, but screw up false filtering or something.
With Theia, the UI is completely abstracted from the signal processing hardware, so there is zero risk of that happening. It's just linux, and the DSP is contained in the FPGA.
Honestly, with Theia we will stop using the term "firmware" as a catch all. There will be OS updates, and then there will be DSP updates. Theia is the first detector where what the forum calls "firmware" updates will actually be software updates. There are an incredible amount of benefits from this from many standpoints, including reliability and scalability.
The OS we developed for Theia will be the same one in use on all future in-house Radenso detectors, and it will be refined over years and years as opposed to starting from scratch with each new detector like everyone does currently.