Turn an ATX Power Supply into a Bench Power Supply
jaygoel.com
jaygoel.com
To use HN parlance, one might say that I have launched! (Any VCs out there want to discuss other ideas I'm working on?)
This kit look me about 6 months to put together, from prototyping the PCB to iterating over an enclosure design. Waiting for components to arrive in the mail (PCBs, laser-cut designs, resistors) was what took the most amount of time.
When I started, I was fully convinced that I would have a working product within just one or two iterations. Putting together a polished hardware kit (and I'd say that this has some degree of "polish" to it, though far from perfect) is much more difficult and time-consuming than I would have ever imagined. Quite different from the agile! lean! iterate! that we talk about in software-land.
Oh. So I have 4 or 5 PCBs from the first few iterations. If you end up ordering one (and I encourage you to do so!) then send me an email and I can throw one in the box. You know, as a collector's item.
If you have any questions or anything, do let me know. Be on the lookout for a post describing some numbers behind this traffic spike.
1. The measurements do not take the laser's kerf into account. That is to say, it does not account for the thickness of material burned away by the laser, so the resulting faces are a bit loose.
2. If you look at the design, the tabs (where the faces fit into each other) go straight across. It would be better if the tabs were smaller, and then have several of them fit together across the length of the box, rather than one long tab.
Most people who have a use for a bench supply should also be able to figure out how to make this box for $2 instead of $45.
Then I spent a whopping $200 on a real lab power supply with adjustable current limiting and variable 0-30 volts and I've been using it for the last 15 years. Sure is nice to be able to set a comfortable 50mA when bringing up new hardware knowing that you can't blow anything up! That alone is worth the price.
It's a nice project, but $45 for something you could build from $10 in parts in literally less than 10 minutes? Really?
Only $80. I was shocked. (By the price, not by lack of proper grounding :-) It is even fully isolated, something I figured would have been cut for cost savings. Mastech HY1803D if you are curious. Working great so far, my only complaint is 0.1V thermal drift over a few hours.
Edit: I mean a variable voltage variable current bench/lab power supply.
Instead of this box you could just clip the appropriate wires for the same effect.
What I really want is a variable voltage supply easily made from any old ATX PSU. As an amateur that would be an invaluable tool.
http://hackaday.com/2010/12/09/atx-psu-turned-into-an-adjust...
It might be a good starting point for a kit which doesn't require a modification of the power supply itself.
Posted my thoughts regarding this supply elsewhere in the thread. Good reviews on Amazon too.
I can't recommend any specifically simply because I have never seen a bad product come out of hp's Test & Instrumentation division. They're all good.
That seems totally impossible to me.
I bought the fully assembled product because I felt the kit version was significantly overpriced compared to the bare PCB and the $5 savings was not worth having to assemble it myself. I hope you're making a good margin on it! :) Regarding comments from others about the possibility of poorly regulated PSUs: you are absolutely correct! I verify the performance of the (higher quality) PSUs that I use before I hook them up to my projects. I've already created an integrated version of this that has worked wonderfully and I am excited for this more versatile solution. [0] Still, when I need a verified good, current controlled bench supply I use a 40+ lb (18 kg) HP bench supply.
[0]: https://picasaweb.google.com/100509155963341533738/BenchSupp...
The margins on this are not bad, but I wouldn't say they are stellar either (for some definition of "stellar") You can look at all of the components/PCB/acrylic and back-of-the-napkin estimate its unit cost, not to mention shipping+assembly (hand crafted!)
You raise a good point in terms of the non-assembled version only being $5 cheaper, which is further confused by not wanting to undercut distributors' (okay, one distributor thus far) prices.
It's interesting, because I've never really had to worry about these kinds of issues (pricing, distributors, expensive unit costs) in the past; it is kinda fun to look at a project from these other angles, orthogonal to the pure engineering component!
I love the trend of people making well-documented kits that I have a good chance of fixing if they break. I'll have to scrounge up an old ATX supply-- I think I just threw my last old one away a few months ago.
The documentation took a long time. I went from no documentation, to a 30 megabyte PDF I created with OpenOffice (not recommended), to borrowing a friend's light box and buying InDesign to create a pretty neat assembly guide.
Incidentally, the Adobe Suite might be the first proprietary software I have ever purchased (well, besides an occasional video game.) I have not regretted the purchase and it runs well enough for my purposes on a VM.
<salespitchmode>You never know when an old ATX supply will come your way, or when you'll get the urge to work on a neat electronics project. Better to have the tools on hand, order today!</salespitchmode>
I'm not sure about InDesign. Wouldn't it have been pretty easy to make a web page that had roughly the same appearance as the PDF?
But it is a good point; I may change this in the future?