Author Topic: Total cost for a lab for marketable work (freelancing)  (Read 5742 times)

0 Members and 1 Guest are viewing this topic.

Offline embeddedguy85

  • Contributor
  • Posts: 38
  • Country: us
Total cost for a lab for marketable work (freelancing)
« on: October 23, 2017, 06:36:55 pm »
I was under the impression that an electronics lab setup would cost around $1,000 according to the eevblog video.

But recently ive talked to someone who says he is spending $6k+ on a lab.

Question is, how much would I have to spend to setup a lab capable of doing the most common freelance jobs?

I'm just not seeing WHY someone would need to spend so much on equipment. Maybe a more sophisticated oscilloscope? But which projects require such a piece of equipment?
 

Offline ataradov

  • Super Contributor
  • ***
  • Posts: 6015
  • Country: us
    • Personal site
Re: Total cost for a lab for marketable work (freelancing)
« Reply #1 on: October 23, 2017, 07:47:23 pm »
Depends on the job. If your job is design and board bring up, then $1000 seems reasonable.

If your job is medium run assembly, then $6K may not cover it.
Alex
 

Offline embeddedguy85

  • Contributor
  • Posts: 38
  • Country: us
Re: Total cost for a lab for marketable work (freelancing)
« Reply #2 on: October 23, 2017, 08:02:05 pm »
I was thinking not of production runs but the design and debugging, etc.
 

Offline ataradov

  • Super Contributor
  • ***
  • Posts: 6015
  • Country: us
    • Personal site
Re: Total cost for a lab for marketable work (freelancing)
« Reply #3 on: October 23, 2017, 08:07:41 pm »
For general purpose electronics, I see no reason to spend more than $1000 on electronics stuff. Microscope and other mechanical tools may not fit into this number. But $2000 will be more than enough. You can buy for the future, of course, and get more expensive equipment right away.

If you are doing RF design, then it will be way more expensive.

Special equipment, like thermal chambers, will blow $6K budged immediately.
Alex
 

Offline Pinkus

  • Frequent Contributor
  • **
  • Posts: 608
Re: Total cost for a lab for marketable work (freelancing)
« Reply #4 on: October 23, 2017, 08:09:04 pm »
No way you will go away with $1K!
For hobby usage: yes, 1K might be OK

But if you want to work as a freelancer (and you say 'marketable work') you need better and more equipment. If somebody shall be able to produce and sell what you developed you need to do much more task than a hobbyist.

What you will need is a
· reasonable 4 channel scope with misc. options
· logic analyzer
· reliable 2 channel power supply
· reliable bench meter
· good soldering station
· fume extractor

when it comes to smd:
· stereo microscope or Mantis
· soldering oven
· lots of parts and storage for them

$6K sounds pretty fair to me - even if you just want to start with the minimum, you better think about $2K. But you should know, after a few years your purchases will sum to several $10K!!

btw: don't forget the need of some mechanical stuff like a mill for prototyping housings. Also not a must but sometimes helpful (but not needed at the beginning): spectrum analyzer, electronic load, etc.

I would guess my lab did cost at least $20K (for my peace of mind I better not go and sum all purchases ... and for a lot of stuff I had great deals... if I would add SRP I easily would be at $30K). Of course I did not purchase it at once. Step by step I sold stuff and purchased better stuff when it was needed. However I better should have left out the cheapies at the beginning - but this you learn by yourself with time: good work need good tools. But investing a lot of money in tools only makes sense, if you are sure that you can assert on the market.... there are a lot of freelancers out there.

Good luck!
« Last Edit: October 23, 2017, 08:11:14 pm by Pinkus »
 
The following users thanked this post: embeddedguy85

Offline dmills

  • Super Contributor
  • ***
  • Posts: 1686
Re: Total cost for a lab for marketable work (freelancing)
« Reply #5 on: October 23, 2017, 08:10:17 pm »
It depends quite a lot on what you are doing...

RF stuff is expensive, spectrum and network analysers (And the cal kits!) can be thousands, even the inter series connector adaptors that you WILL need can add up to a few hundred bucks by the time you have a reasonable selection, then pads, bridges, filters...

Audio, again expensive if you want the gear to really measure what you have built, even an old AP system 1 is a couple of grand, something similar if you are doing very low noise or high precision analogue.

Digital of the microprocessor and hang a few peripherals off it sort is cheap and you could probably tool up for this for a grand, however as soon as you add DRAM and high speed busses you are back into the RF space again, but there is money in the simple stuff as long as you do not mind the software side.

Do not forget the software aspect, Matlab, PCB CAD of choice, Solidworks? Maybe ADS or such if that is how you roll? You can spend a scary amount on software licenses very, very, quickly.

And then we come to the reference books, that copy of "Handbook of filter synthesis" or "Knuth" can pay for itself fast, but they are still a couple of hundred bucks of outlay.

Finally, it is worth owning some mechanical tools, a lathe and smallish mill are **Useful** in this game.

You can do all of this for much less of course, but what you will find is that the guy with the right tools can work faster and with less false starts, and that translates into income.
 

Offline Pinkus

  • Frequent Contributor
  • **
  • Posts: 608
Re: Total cost for a lab for marketable work (freelancing)
« Reply #6 on: October 23, 2017, 08:12:45 pm »
The above is very true - I forgot the software  :palm:
 

Offline rstofer

  • Super Contributor
  • ***
  • Posts: 6649
  • Country: us
Re: Total cost for a lab for marketable work (freelancing)
« Reply #7 on: October 23, 2017, 08:14:46 pm »
I would think you could get a nice hobby level lab for around $1000.  I seriously doubt that hobby grade stuff is going to go very far in a professional environment. 

Any professional measurement you take will need to be backed up by a calibrated instrument and the instrument will have to be accurate enough to justify the measurement.  I don't know if cheap chinese meters can be calibrated to a traceable standard.  You can't expect the cal lab to characterize the equipment so the published specs will have to be acceptable in the industry.  We see from some of the DMMs that CAT ratings are bogus, why are the accuracy specs any better?

All of this will clearly depend on the project and the customer but I would expect to spend several times $1000 to get equipment anywhere near good enough for professional use.  I guess I'm thinking that Keysight might be my new best friend.  They would surely be taking a lot of my money.

Perception is everything!  As a customer, which would you rather see on an equipment list, Rigol xxxxx or Keysight yyyyy?  Hobby versus professional.

You can struggle around making certain measurements with less than the required equipment but if a measurement requires a 4 channel scope, a 2 channel won't produce the image required by the customer.

It used to be that if you wanted to see something you bought Tektronix, if you wanted to measure it, you bought Fluke.  The industry has probably changed a little but not a whole lot.  We can certainly include Keysight for either function.
 

Online blueskull

  • Supporter
  • ****
  • Posts: 12479
  • Country: cn
  • Power Electronics Guy
Re: Total cost for a lab for marketable work (freelancing)
« Reply #8 on: October 23, 2017, 08:17:35 pm »
I think you've mentioned before you want to do embedded? Then you don't need a fancy scope.
Just an Analog Discovery, a programmable power supply, a soldering station, a smoke absorber, one or two good ($100+) DMMs and a hot air station should cover your equipment usage.
Then you also need solder wire, solder flux, solder paste, solder wick, 2 pairs of good ($20+) tweezers, a good workbench with good lighting and a good chair.
You will also need a fast computer to compile Linux from scratch if you do the embedded Linux thing, prepare at least 32GB RAM if you run Linux in a VM, and 24GB if you run Linux as a physical machine.
You will find yourself in needing of a printer, at least two monitors and a good set of keyboard/mouse.
 

Offline ataradov

  • Super Contributor
  • ***
  • Posts: 6015
  • Country: us
    • Personal site
Re: Total cost for a lab for marketable work (freelancing)
« Reply #9 on: October 23, 2017, 08:17:51 pm »
Any professional measurement you take will need to be backed up by a calibrated instrument and the instrument will have to be accurate enough to justify the measurement.
That is some very idealized scenario. I've seen so many contractors that have no idea what they are doing, and they have no problems winning the bids.

Equipment matters way less than your brains.

Perception is everything!  As a customer, which would you rather see on an equipment list, Rigol xxxxx or Keysight yyyyy?  Hobby versus professional.
With attitude like this, you will always overpay for stuff. Good on KS marketing though.
Alex
 
The following users thanked this post: Mr. Scram

Offline TimNJ

  • Frequent Contributor
  • **
  • Posts: 797
  • Country: us
Re: Total cost for a lab for marketable work (freelancing)
« Reply #10 on: October 23, 2017, 08:21:20 pm »
What do you consider "common freelance jobs"? There's no one-size-fits-all electronics lab. Most labs are highly tailored to a certain type of project. As you start to work on more advanced projects, you will quickly realize that in order to validate your design, you may need some fairly pricey equipment. If you are designing small microcontroller based widgets, then chances are a $1k lab setup will do the trick (at least for a while).

For "high-performance" designs (that is, high speed, low noise, high precision, etc.), you need your test equipment to be "better" than the circuit you are trying to design. For instance, my next big project is a programmable voltage reference which can output any voltage between 0 and 20V in 0.00001V increments. I can only confirm that the linearity and accuracy of my device with a high precision meter with high level of measurement certainty. Such a meter might cost $3000, by itself.

In industry, working on high speed communications (4G/5G, for example) requires some pretty insanely spec'd test equipment.

But I would start off small, being working on some projects, and eventually when you realize "Hmm, I'd sure like to make 'x' but I can't test it", then look for something that can make that measurement.
 

Offline ataradov

  • Super Contributor
  • ***
  • Posts: 6015
  • Country: us
    • Personal site
Re: Total cost for a lab for marketable work (freelancing)
« Reply #11 on: October 23, 2017, 08:23:57 pm »
But the growth is kind of embedded in the question. There is no way you will get a job doing something way more complicated that what you did before.

Even if you just spend $100 000 and buy the best equipment ever, you will not get jobs to use that equipment without some track record.
Alex
 

Online blueskull

  • Supporter
  • ****
  • Posts: 12479
  • Country: cn
  • Power Electronics Guy
Re: Total cost for a lab for marketable work (freelancing)
« Reply #12 on: October 23, 2017, 08:24:58 pm »
Digital of the microprocessor and hang a few peripherals off it sort is cheap and you could probably tool up for this for a grand, however as soon as you add DRAM and high speed busses you are back into the RF space again, but there is money in the simple stuff as long as you do not mind the software side.

A good scope and 2 sets of active probes are all what is needed to do this.

Do not forget the software aspect, Matlab, PCB CAD of choice, Solidworks? Maybe ADS or such if that is how you roll? You can spend a scary amount on software licenses very, very, quickly.

I don't use Matlab, nor SW. ADS is only for serious RF guys.
I use Octave to do Matlab's job, and when I need a specific tool box that's not there, I write one in C from scratch.
I use TinkerCAD to do rough modeling, and OpenSCAD to do fine modeling. I then use FreeCAD to export the files. I also use SpaceClaim (included in Ansys AIM) if I want to do Ansys AIM simulation ($14k, 1/10 the cost of a full Ansys Workbench and thermal, mechanical and electromagnetic solvers).
 

Offline katzohki

  • Frequent Contributor
  • **
  • Posts: 354
  • Country: us
    • My Blog
Re: Total cost for a lab for marketable work (freelancing)
« Reply #13 on: October 23, 2017, 08:36:25 pm »
It absolutely depends on what you're going to do, what capabilities you need to have and so on. What can you do for a client right now? Add to that what you think you need.
 

Offline Mr. Scram

  • Super Contributor
  • ***
  • Posts: 8169
  • Country: 00
  • Display aficionado
Re: Total cost for a lab for marketable work (freelancing)
« Reply #14 on: October 23, 2017, 08:45:15 pm »
Not all customers are prepared to deal with any issues that can and sometimes do arise when you use an alternative CAD package. There is a reason companies pay stupid amounts of money for Solidworks licences and that is because it's the industry standard. Not being able to deliver 100% compatible files will hurt you a lot in some areas.

I'm less familiar with other software packages, but I imagine it's often the same there.
 

Offline Mjolinor

  • Frequent Contributor
  • **
  • Posts: 289
  • Country: gb
Re: Total cost for a lab for marketable work (freelancing)
« Reply #15 on: October 23, 2017, 08:50:10 pm »

I set my lab up between 1991 and 1998. I spent a total of £148,000 doing it.

The total worth of all that equipment today is probably in the order of £5000 I would guess.

:)

It depends what you are doing.
 

Offline cdev

  • Super Contributor
  • ***
  • Posts: 5082
  • Country: 00
Re: Total cost for a lab for marketable work (freelancing)
« Reply #16 on: October 23, 2017, 08:56:15 pm »
Electronics is like art. Which is to say its also a bit like being stricken with a disease. Or rather an addiction.

If you want to do it, you'll figure out a way to do it.

But once you are asking for money from people then your time is money.

So you will likely want to have the most optimized tool set for what you do. Optimized means not spending more than you need to in any area, so you will have the money to spend where you need it. Every person's needs are different.

Its always been my observation that startup companies that spend a lot of money on impressing customers with fancy [whatever] often fail.

Personally, I think there is a sweet spot for value with everything. Including every kind of tool.
Some tools are exceptional values and they make all sorts of things easier but you learn less.

Other tools help you learn and you understand more faster.
« Last Edit: October 23, 2017, 09:09:30 pm by cdev »
"What the large print giveth, the small print taketh away."
 

Offline cdev

  • Super Contributor
  • ***
  • Posts: 5082
  • Country: 00
Re: Total cost for a lab for marketable work (freelancing)
« Reply #17 on: October 23, 2017, 09:00:27 pm »
I just got a nifty little stereo microscope for $35 new on ebay. Its like a kids microscope, but the optics are quite decent.
For general purpose electronics, I see no reason to spend more than $1000 on electronics stuff. Microscope and other mechanical tools may not fit into this number.
"What the large print giveth, the small print taketh away."
 

Online AndyC_772

  • Super Contributor
  • ***
  • Posts: 3497
  • Country: gb
  • Professional design engineer
    • Cawte Engineering | Reliable Electronics
Re: Total cost for a lab for marketable work (freelancing)
« Reply #18 on: October 23, 2017, 09:00:38 pm »
Don't focus too much on the cost of the lab equipment. It'll seem like a lot if you add up the costs of everything you think you might need all at once, but that's not the right way to equip a lab.

The key thing is to buy what you need, when you need it. Always be ready to buy a new tool, but don't actually do so until you have a compelling reason to spend the money.

I started out with a simple 'hobby' lab, consisting of:

- a good, solid bench with ESD mat
- a 4 channel analogue 100 MHz scope
- a reasonable soldering station
- a very cheap '830' multimeter
- a good selection of hand tools
- a laptop PC running free schematic and PCB software

That's about it. It was enough to tinker and play with at home, but really wasn't up to the job of working professionally.

No PSU, no function generator, no spectrum analyser, no logic analyser.

First was an update to the IT equipment. Out went the free software, and in came a copy of OrCAD PCB Designer, which has been an indispensible staple of my business ever since. Also a NAS unit with a UPS, and a robust backup solution. You'll need somewhere to store customer jobs, and you can't afford to ever lose data to a drive failure.

Next, the scope had to go, replaced by a Tek TDS754D. I needed the 500 MHz bandwidth, for a specific series of customer jobs.

I bought a used, and quite old, Fluke multimeter too, for confidence and the sake of appearance in front of customers as much as anything else. It confirmed that my '830' was in fact perfectly accurate and usable.

It's very, very rare indeed that anyone actually questions the accuracy of a measurement I've taken. There's no rule that says you "must" have a calibrated meter just because you're a "professional".

That said, several years later, all my equipment has been upgraded apart from the bench.

I sold the Tek scope and now have two MSO-X3000A scopes instead; one 1GHz model because I needed the extra bandwidth for a specific job, and one 500 MHz model which I take to customer sites.

I have a few multimeters now, of course. Most often I use a 34465A, mainly because it has a permanent place on my desk so it's convenient. I like the data logging features - but more often than not, I could use a much more 'ordinary' handheld meter

I also have a Fluke 289 which I bought for a job that required logging a vehicle's battery voltage while it was in motion It's now a 'travel' meter which I always tend to have with me when I visit customers, on the basis that although it's not actually that nice a meter to use (rather slow, poor display contrast), it can do pretty much anything.

The same goes for the signal generator and impedance analyser. I don't use them much, but for the jobs that require them - and justified their purchase - they're invaluable.

For a time I did keep a few saved searches on Ebay, and picked up good quality, general purpose equipment (power supplies, microscope) when it came up for sale at a good price - but although I use my power supplies and microscope regularly, the fact that they weren't bought for a specific job makes them unusual.

Offline hammy

  • Supporter
  • ****
  • Posts: 435
  • Country: 00
Re: Total cost for a lab for marketable work (freelancing)
« Reply #19 on: October 23, 2017, 09:10:44 pm »
2000-3000€ PC and CAD graphics card + two Monitors
500€ Bench DMM with data logging
400-800€ DSO with or without LA
500€ Salease Logic Pro 8
320€ Amscope SE400-LED + some more WF eyepices
220€ Brymen BM869e Handheld DMM
90€ Desoldering Station
240€ First Power Supply
240€ Second Power Supply
160€ Soldering Station
120€ Voltage Standard
200€ Peak Atlas DCA75 + LCR45
500-4000€ ECAD Software

5270-10170€ -> 6185-11937USD

not included: soldering tips, solder, cables, coax cables, wire, cutter, pliers, esd mat, good light, second DSO, second LA, some more DMM,  development boards, some more software, NAS for data storage, components, breadboards, 3D printer, 50Ohm termination, coax T-pieces, function generator, spectrum analyzer, some good books ...


Cheers
hammy
« Last Edit: October 23, 2017, 09:18:56 pm by hammy »
 

Offline cdev

  • Super Contributor
  • ***
  • Posts: 5082
  • Country: 00
Re: Total cost for a lab for marketable work (freelancing)
« Reply #20 on: October 23, 2017, 09:36:34 pm »
As an embedded guy, you could in theory make almost any kind of lab or testing device with a computer, and a few additional parts. There is no use reinventing the wheel when you already own some wheels that just need an axle to work.

A recent thread on cheap ebay gadgets had a discussion of some really good value cheap gadgets like the $10 sigrok compatible logic analyzers and the $7 AVR component tester that identifies almost everything with two or three leads. You cant go wrong with them. Also at least one good multimeter, but really two or more are quite useful, especially if you can log their output. I like to use an analog multimeter to show a device's  current draw.

A bench power supply with both constant current and voltage is essential to have. Its good if it has a switchable meter or even better if it has two, so you can see voltage and current draw simultaneously. An ESD mat and a wrist strap to ground yourself.  Some small cup like containers for parts in your current project (I use the lids from peanut butter jars.) Solderless breadboards and solid wire.. Lots of different kinds of wire and cable.. Headers.. a number of different kinds of headers.. male, female, for the stuff I do, I use some RF connectors.. always need to use USB-UART converters.. I keep a bunch of plastic boxes which hold common various small parts, resistors, capacitors, LEDs, some common transistors, ferrites and toroids..

Good, bright lighting-is essential for soldering.. I have a florescent light on a boom with a magnifier, its handy for inspecting a board in detail.. I use a really bright clip on light for soldering.. I also have a small strong fan to pull the fumes away and more generally a double fan with a heat exchanger to the outdoors to keep the air fresh.

A scope is super useful, you can start out with a used analog scope.  A good soldering iron with a grounded tip and adjustable temperature. Ideally it should have changeable tips and its good to have an assortment of tips. Some kind of holder for boards. A vise is good to hold a board you are working on while you are working on it. its nice to have an all-metal one that can take some heat (for use with a preheater) I use a "stick vise" for that. It was $6.

A hot air gun with adjustable temperature and a hot plate with controllable temperature are nice too. Cables and interconnects, probes.. "Dupont cables" alligator clips.  Small tools, side cutting pliers, wire stripper, tweezers, small screwdrivers, solder in several different sizes, solder flux (pen), solder wick, solder paste. other supplies, qtips, IPA.

Really, electronics as a hobby is not all that expensive compared to many other hobbies. All told I really doubt if I have spent even $1000. Probably more like $750. And its been spread over several years.  Most of the stuff I order on ebay is under $5. You can do a lot with inexpensive tools.

But - I am just starting out, really. I'm just fooling around. Were i doing anything professionally, it would be a wholly different story.

You will have to spend more money, probably much more, eventually, but do it sensibly.

Use your inexpensive equipment to its fullest and you'll know when you have outgrown it.

« Last Edit: October 23, 2017, 10:19:52 pm by cdev »
"What the large print giveth, the small print taketh away."
 

Offline ataradov

  • Super Contributor
  • ***
  • Posts: 6015
  • Country: us
    • Personal site
Re: Total cost for a lab for marketable work (freelancing)
« Reply #21 on: October 23, 2017, 10:56:23 pm »
2000-3000€ PC and CAD graphics card + two Monitors
Sure, you'll also need to lease a personal jet. How would you visit customers otherwise?

You can rack up the bill, of course. But I don't think this was the question.
Alex
 

Offline embeddedguy85

  • Contributor
  • Posts: 38
  • Country: us
Re: Total cost for a lab for marketable work (freelancing)
« Reply #22 on: October 23, 2017, 11:05:38 pm »
Well, thanks for all the replies. I have to say it is discouraging because my plan was to learn and build projects on my own and then hopefully get into freelancing... but the issue I see now is money! Because I am cash poor but time rich as it is right now. How could I afford such professional equipment to provide the services companies need?

So, maybe there is a middle step I could take? Perhaps only writing firmware on a freelance basis? This wouldnt require that equipment, right? Just basic electronics knowledge.
 

Offline ataradov

  • Super Contributor
  • ***
  • Posts: 6015
  • Country: us
    • Personal site
Re: Total cost for a lab for marketable work (freelancing)
« Reply #23 on: October 23, 2017, 11:08:03 pm »
Well, thanks for all the replies. I have to say it is discouraging because my plan was to learn and build projects on my own and then hopefully get into freelancing... but the issue I see now is money! Because I am cash poor but time rich as it is right now. How could I afford such professional equipment to provide the services companies need?
You are contradicting yourself. If you don't have the knowledge to use all that equipment, how you will use it? How do you plan to get customers?

At this point it is a hobby, and you can start electronics hobby with $10.

It is like saying "I want to be an auto mechanic", but I can't afford to buy a 4 car garage right now.
Alex
 

Offline rstofer

  • Super Contributor
  • ***
  • Posts: 6649
  • Country: us
Re: Total cost for a lab for marketable work (freelancing)
« Reply #24 on: October 23, 2017, 11:09:15 pm »
If you stay in the firmware end of the business, you will still need a scope and logic analyzer but perhaps not as high end as if you were designing hardware.

After all, if you define a PWM output, it would be nice to see if it really changes pulse width.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf