Power supply teardown / Upgrade (suggestions needed)

[Gridstop]

Hi all,

After watching the power supply threads here and elsewhere not really turn into anything yet, I'm going to go through the process of taking an older HP power supply (usually various kinds are available on ebay for less than a hundred bucks), doing a tear down and repair (the one I got is 230v, needs to be changed to 110, has some other minor issues). I'm hoping to document the process as well.

After getting it up and running, I'm also going to convert it over to full digital control. The end result will be some modular PCB's that could be used with almost any older HP/Agilent power supply (they all use virtually identical feedback/control circuits), and then daisy chained together to a master controller that will interface either via Bluetooth directly to an android device, or via usb to a PC (or r-pi) and then to a Linux or android front end via tcp/ip.

So I'm just interested in feedback as far as:

1) Should I just do a blog with pictures/text, or is it worth doing some videos? Combination of them? There will probably be long gaps in the process as I'm busy with school and/or getting PCB's made and such.
2) Is it worth trying to document software/firmware design process and/or PCB layout? I'm planning on switching from eagle to diptrace so I'll be kind of a newbie with the software, so that might be pretty ugly to watch.
3) I'd like to experiment with the android front end (I've done tons of java, I just need to learn the new UI layout stuff), is it worth making it work on a phone or just stick with tablet designs (probably a Nexus 7 as reference device).
4) Is it worth exploring Bluetooth SPP? It seems generally sketchy making it work, but then if it works on the standard nexus devices that's probably good enough. This avoids all the usb nonsense as well. The alternative is connecting to a PC or r-pi (probably using MCP2200 or FT232 for vid/pid's sake) and then using tcp/ip after that, which is fine but just adds more layers.
5) I'm most comfortable with atmel/avr, though I'd like to get into PIC too, but I feel that might be just one too many things to learn & deal with. Just wondering if people had strong opinions on the matter.

My plan is to make all the designs & software open/free, of course.

[Sleepwalker3]

With HP stuff there's a fair chance you may be able to save yourself a lot of time and effort by just hooking in to the HPIB bus (essentially GPIB) if it's one of the models that has this. If you search the blog (actually blog search not working properly, use google) and forum you'll find a bit of discussion about GPIB last year and a project from one of the guys (Sasigrande I think it was), which is OS and he's selling them. That in itself would most likely remove the need for any Micro work at all, just hook to the HPIB and let your PC, Rasbperry Pi, Chocolate Sundae <insert other silly food related micro name here>, or whatever, do the simple control and display stuff - unless of course you really want to do the micro thing. Chances are the HP designed internal setup would work much better than a lot of the stuff you could do yourself, so it's likely more practical to just interface to that and tell it what you want to do with a simple app, but it depends on what you are actually trying to achieve by doing this project.

As for linking to tablet/phone/electric toaster ;) etc. I don't really see a lot of use for it myself, but others may. I guess it might be good for logging or something, but if I were you I'd forget that for V1.0 and just get the basics done, you can always add the bells and whistles later.

Blog is probably good, but the odd video thrown in never hurts, as long as it's also covered in the blog. If you're comfortable with a particular line of processors, then run with that if it will do the job. You can always make V2.0 later if you decide to go with the PIC.
Again I'd stick with Eagle if you know that, tons of people use it (I don't, don't like it, but I have it just to view stuff others have done).
You'll have to make it all open and documented or you'll either get flamed by every FOSS nut or pestered no end by thousands wanting you to email them. You should also prepare youself for being bombarded by hundreds wanting 'this and that' support for their own project that's more important than anything else (at least to them it is!) and expect you to do all the hard yards, so just beware. I'd suggest getting done before starting to trickle the info rather than a running blog, but up to you.
Anyway, that's just my 2 cents worth.

[Gridstop]

Mainly I'm talking about the HP analog supplies (no digital circuitry/controls at all). Most of them have a pretty standard set of terminals along the back that configure the voltage & current feedback circuitry. With a couple DACs and ADCs they can be converted to digital control & readout fairly simply. What I found is most of the older even simple non-programmable digital supplies tend to be very low wattage or very expensive.

At this point I will probably just target some kind of serial protocol that can either go over USB via any rs232<->usb adapter, or via bluetooth bridge if so desired. And leave that stuff until later.

For micros I am indeed thinking I'll stick with atmel for now and if there's enough interest someone wants to redo it with pics, so be it.

For PCB my main reason for switching to diptrace is pcb size limits of eagle are causing me problems, not in this project but in a few others. Diptrace also has just a much better feel than eagle so I'm liking it so far. I would go to KiCAD but it's just a nightmare still.

Thanks for the input. I'll take your advice on waiting until I have more finished (maybe at least have PCB's in hand) to start putting real info out.

[Sleepwalker3]

You may want to try Design Spark (via RS), I've heard good things - Free, but not FOSS. PIC have some nice ADC/DAC in the micros, so worth considering, but that doesn't mean go that line. Best to stay with what you know for V1.0.
For protocols (or anything else for that matter), yes best stick with known/simple for the first lot (e.g. SPP) and go fancy later if necessary. Will you be programming micro in C? Just in case you're more a BASIC person, you may want to look at Great Cow Basic (Sourceforge) - I haven't tried it, just found it and it had very good response from comments. You could likely also do all this with a PicAxe, but you sound like you're comfortable with the AVR, in which case that's the best choice - Always better to actually get something happening and out there and modify it later, rather than trying for perfect on V1.0 - That's my take on it anyway, others may disagree - Don't shave the Yak!

Yes, I see a lot of project people getting swamped by requests and it's amazing how some people think that you somehow have to do everything for them, just because you showed a bit of what you did, so better to have the information on hand and say "Refer to that".

[hak8or]

Is there any reason you chose to use diptrace instead of kicad?

Also, for videos, I found that combinng both videos and written content is extremely difficult and hard to orginize. An alternative would be to make gif's of very short clips to demonstrate something which is hard to do in text, but too short to make an entire video for. I have never personally tried it though, and gif's might be problamatic for things such as phones, not to mention they are somewhat bandwidth intensive. If gif's are done wrong, the filesize will be huge, but if done right, they can be managable.

I originally tried to do a timelapse of the design of a simple board design, all the way from selecting parts, through designing and making a schematic, and to the end of the PCB design. I had a program which took a screenshot of my entire desktop every few seconds and saved it to a folder. Unfrutnetly the program fumbled itself, resulting in corrupted files half way through, and it did not tell you that something was wrong untill you are done and trying to render it! It wasn't fraps, but something else instead, sadely I forgot what it was. But, what I did have was very cool. It really does show what the entire process is, including getting side tracked like going on reddit, listning to pandora, and a game here and there. It was fun to watch, but with only half the content it was not worth it to upload. Keep in mind that you will rack up data on your hdd somewhat quickly, so make sure to have plenty of space beforehand!

For communication, I had a similar dilemma a while ago. How much do you value your time, and how quickly do you want to finsh the project? Are you willing to allow the communication aspect become a project in of its own? Will you use the new communication medium in future projects, and will it be warranted considering the increased cost for the communication module, or if will be usefull to you in the future. What will you use for communication before you get the bluetooth working, and will you have enough energy left to work on communication after getting the basic power supply stuff done. And lastely, do you value the expirence and latter know how of the usage regarding the new communication medium enough to warrant the time and project delays? In the end, I decided against using wireless because I am cheap and don't want to pay four bucks for every blue tooth module. But, it does make yoru projects look infitly more professional and brings a much more "cooooool" factor.

I have dabbled with both PIC's and ARM's, and both have their perks. You should also consider MSP430's! The questions are, do you do most of your projects based on battery power? If so, try out the MSP430's.

How much money are you willing to shell out for the initial investment for the new platform, mostly in terms of programming tools. ARM's can be programmed by buying a bus pirate, using their onboard boot loader via UART, or buying a TI launchpad and using the programmer on the board, so all in all it is extremely cheap to get started with the ARM platform. Also, it is getting extremely popular, the M4's are a beast in terms of preformance, and IDE's are starting to improove for arms. Not to mention that the cheapst arm is about three bucks on digikey in single quantaties, and it has a good bit of RAM and flash. The libraries are starting to mature well also. Arm's have a full blown free very mature compiler, so that is good.

PIC's are an intersting creature. It was my second MCU, if you can consider the basic stamp an MCU. You need either a PICkit2 or PICKIT3 to program PIC's. I reccomend a pickit3 so you can program PIC32's. The programmer is a somewhat expensive, but it includes some cool stuff like a low speed logic analyizer and bus interface. PIC's also come in DIP packages and are very cheap for the low pin count packages, such as DIP8. They use some sort of "opensource" compiler thing which I am fuzzy about. It is some sort of GCC thing with Microchips stuff integrated in it, and if you don't pay you dont have access to optimizations. I do not thing there is a fully free compiler for PIC32's, but I might be wrong. Microchip is also very popular, and has some cool libraries.

MSP430's have the launchpad, which has its own programmer on the board, and costs only twelve bucks or so. It is VERY beginner friendly. They have a fantastic community on 430h, and they have some dude that made a full tool chain setup thing. All you do is download eclipse, and via maybe four clicks install a plugin, and bam, you are ready to start coding and flashing. It is also continously updated, but it is lacking some features, such as seeing variable values when debugging under certain conditions, but otherwise you can easily start using it. It is the least popular of the two though, and has less full blown libraries, but I have never had a problem because of it. C is C, the libraries do not change that much between microcontrollers, you just need to muck around with the lower level components a bit to get it working on your platform.

For ultimate power stuff, you should check out the gecko line, they have some super awesome power consumption tools while debugging.

Damn, my professor got back, so I have to end it here. Good luck! :)
Edit: Holy batman, this is long, sorry! Though, this might be a front page worth small look on microcontrollers. :D