Shooting the breeze with some engineers recently, asking hypothetical questions like “what would you do if you won a billion dollars in a lottery?” This led to discussing what would be a perfect one-man lab for prototyping and experimenting. Unencumbered by the thought process, I put forth the following assertion: with the various modern equipment available these days, and with a not-so-crazy expenditure, one could equip a home lab that is capable of building things almost as complex as a smartphone. Do you agree, or was the soju (a Korean traditional alcohol drink) going to my head?
What equipment would you put in such a dream lab?
The inspiration for my statement, besides the soju, was a couple of business trips I had in 1999 to a small Danish company. Their electronics lab was operated by one technician, but was extremely well equipped with amazingly compact SMD and machining equipment. This one guy made a wide variety of industrial and military grade circuit board modules, all in-house and quicker than the ordinary outsourcing times of the day. While I think he did the PCBs out of house, his did everything else in-house, pick and place, reflow, testing, and computerized milling machines to make enclosures. Like my hypothetical super-lab, this guy’s lab was an ordinary sized company lab, not a whole building or wing.
Note: clearly, with a billion dollars you could build factories to have at your disposal. Instead, I’m talking about a modestly sized home lab, say a large room or basement, at what I believe could assembled at less-than-a-millionaire budgets.
I’ve become a huge fan of Keysight’s Benchvue for the Agilent/Keysight stuff and would love to have the latest and greatest version of the Rigol, Siglent, and Keithley instruments I use in a Keysight brand. Having to switch between the different software for a single test run is a pain, especially with VISA and IVI conflicts
I tried LabView, but I found that BenchVue is just easier and faster for a novice to get results.
More specifically, it’d be a dream to replace my current gear with…
34470A 7.5 Digit DMM
B2983A Femto/Picoammeter (because why not?)
33622A Waveform Generator
N6705C DC Power Supply/Analyzer (check out all the options!)
N3300A DC Electronic Load Mainframe with all the modules as options
DAQ970A Data Acquisition System (DAQM907A, DAQM904A, DAQM902A)
Obviously everything would be fully optioned out and under calibration and full warranty
I know there’s ultra-high-end stuff, I don’t need much really. Having no limit to what I can work on would be awesome though
Fun thought exercise
On the production side, something like a Makerbot Z18 along with Markforged Metal X and X7 would be insane. Imagine printing titanium!
I probably wouldn’t do PCB manufacturing, Seeed and others in China do a great job cheap enough and you don’t have to deal with the hazardous side of things. I’m on the fence with assembly, it’d be interesting to see how well Seeed does compared to doing it myself.
Also, soju bombs have created some great stories for me too
== EDIT ==
Although, now that I put more thought into it, with massive cash resources you could build a company (or contract out) to create a product similar to Makerbot for producing PCBs and maybe even include pick-and-place in assembly line fashion. Does one already exist?
Good data. I think I’ll start a BOM for this dream lab concept.
In my dream lab I would not have much equipment. Only the stuff I need to use; and when I need it. Reality is a different matter – Having useless things like MakerBot Replicator 2 and reflow oven made out of Chinesium lying around and wasting space. Seriously, in a lab little bit of “less is more” goes a long way. Very hard to buy free space. Also known as “you can not put simplicity in, but have to take the complexity out!”
I may be a bit biased, but I’m surprised that you think a FDM printer is useless. I find it very helpful for prototyping purposes – not just for the product, but for making fixtures and tooling to help achieve better results.
Right now, I’m debating whether I want to spend ~$3,000-$4,000 on a low end pick-and-place machine. I do a lot of onesies by hand right now most o my clients will typically ask for 2 to 5 pieces.
There are also semi-manual prototyping precision placement machines – in some ways, I think those would be better for the onesies – but they are pricey! (https://www.manncorp.com/component-placement-and-handling/manual-pick-and-place/manual-pick-and-place-systems/smt-place-2000-manual-pick-and-place.html for an example.)
On the contrary. A FDM printer is super. Just not the specific model. Oh, and need not to get stuck in the “death by rapid prototyping” loop with FDM…
Eurocircuits manual pick and place is pretty nice lower budget option. Real full pnp goes for 40…100k and is too difficult to program for 2-5 protos anyway. Not even talking about the cost of feeders and never having the correct ones. Also good stencil printing is a must if you want to do finer reflow. 98% of SMD errors are due to bad paste printing, IIRC