Community Bench Meter

[Mastro Gippo]

http://www.polycase.com/item/vm-36boot.html
this looks nice and cheap. I would like a handheld multimeter, not a bench one...

[GeekGirl]

http://www.polycase.com/item/vm-36boot.html
this looks nice and cheap. I would like a handheld multimeter, not a bench one...

Nope, Can not use them.... They don't come in PURPLE

I think if we can make it modular then we can have a simple front end display for hand held and a fully optioned desktop model ?
The prices of the hand held case is VERY GOOD, for low qty.

[GeekGirl]

When it comes to the enclosure I am going to suggest that we use Farnell / Newark, this is so that we can all have everything FIT, Front Panel overlays the right size etc.

When doing a project like this I like to start with the form factor and user interface first and how that's all going to work, and then work backwards to the detailed schematic. Because a schematic does not a project make.

http://www.hammondmfg.com/
http://www.serpac.com/
http://www.polycase.com/
are some of my favorites.

Also, I'd consider a budget too, too high a price and it's going to be a show-stopper, so cost should always be factored in.

I'd probably go with PIC myself for familiarity and simplicity. 32bit ARM power is not needed for such a thing, and the development tools are harder to work with for the average punter.
It's hard to beat the $40 PICkit, MPLAB, and free (limited) C compilers for the PIC combo.
A lot of people say ARM and GNU C is cheap and free, but when it comes down to it the tools are not consistent, and harder for a beginner to use.
But of course that may not really matter if most people just want the finished pre-programmed product.
Heck, I'd even seriously consider jumping on the Arduino bandwagon here. Separating the custom front end board from the processor board entirely. Gives options and lowers the development risk. And it will excite the hacker and makers (don't underestimate that market appeal)

Dave.

I want to try and use a chip and "common" compiler, I am not a huge fan of GCC (although I use it the most) as it is not like a commercial product that you install and it just works. I know GCC is getting better, but things are not there yet.

I agree we do not need an ARM processor. Just an 8 or 16bit uC with a good amount of Flash and RAM.

Even though I am an Atmel Gal, I see no reason to not to consider a PIC or MSP serries, (horses for courses

I am starting to picture this project as a "modular" system, basic is display and Multimeter front end, more advanced would be to add the 4wire resistance, ESR, LCR, Logic Analyser (to give us the ability to spec # of channels (less channels = more samples/longer period)) Frequency counter, Function generator, maybe even a terminal adapter ie send and receive RS232 for exercising projects.

I think that we need to cull the list and not make this project a tool for every situation.

I am not sure, but I am sure that in the next few days I will make a list and description of every idea so we can work out what we REALLY want in this product.

[armandas]

http://www.polycase.com/item/vm-36boot.html
this looks nice and cheap. I would like a handheld multimeter, not a bench one...

Nope, Can not use them.... They don't come in PURPLE

I think if we can make it modular then we can have a simple front end display for hand held and a fully optioned desktop model ?
The prices of the hand held case is VERY GOOD, for low qty.

I think the bench-top would be most appropriate and easy to build. But hey, it's going to be open-source, so anyone can build their own version, if they like!

[Andrew]

When looking for the mentioned "original instrument enclosure" I think I went through everything mentioned here and more (Hammond, Serpac, Polycase, Bopla, Pactec, Proma (now gitec), Prio, Combiplast, Teko (brrr), ...).

And because of that I am of course fully educated and informed where to get PURPLE enclosures GeekGirl so desperately needs: http://www.elpac.de/categories.php?cat=200&lang=en

[charliex]

ARM GCC is a lot better now, since ARM paid for it to be fixed/upgraded since it was absymal.

How about a PSOC ? While i agree you don't need much of a uC to run it as is, it'll allow more room for growth. If its C then the differences aren't as difficult to overcome.

[Thermal Runaway]

I think that we need to cull the list and not make this project a tool for every situation.

I am not sure, but I am sure that in the next few days I will make a list and description of every idea so we can work out what we REALLY want in this product.

I agree.  Most of the things mentioned would be really cool to have on a meter actually, but it's probably better to get a decent system working first then worry about cool extras at a later time.  That's what new versions are for.

Someone else mentioned handheld.  I have to agree - I'd prefer a handheld project meter as well.  Mainly for selfish reasons, in that I already own a really cool and accurate Agilent bench meter which in all fairness would really take some beating.  I guess there are some extra features that could be added to it, but the features it *does* have are pretty much perfect.

Handheld... well I have quite a nice Fluke meter but... I dunno I think it could be improved quite easily.  I don't think any one manufacturer has totally nailed the handheld market yet.

With regard to compilers, if it is decided that the PIC route is the way to go, then I would recommend the BOOST C compiler.  It's not the *best* compiler in my opinion (that award goes to Hi-tech PICC), but it's certainly the best bang for buck.  Even the professional version is well within most people's budgets.  I used to use Hi-Tech but found that it was priced outside of my range for my hobbyist stuff.  So I've recently moved to BoostC.

Brian

[jahonen]

BTW, This looks attractive for a DMM, a 24-bit ADC with ARM7 µC (however no personal experience on this chip):

http://www.analog.com/en/analog-microcontrollers/analog-microcontrollers/aduc7060/products/product.html

Like NXP ARM7's, this one too has an integrated boot loader, so programming can be done just with level-translated RS232, even without any software pre-programmed.

Nice thing about sigma-delta ADCs is that by setting the modulator rate correctly, you can get a very good rejection of 50 and 60 Hz mains-borne noise. Good for DC- and rectified AC-like (like measuring output of that true RMS converter) measurements.

I agree that feature list needs to be prioritized to "must have", "nice to have" and "not needed" features. Otherwise this will not lead to anything Feature creep is very much poison to the product development, especially on the HW side, where a seemingly small change can cause a major re-design.

Regards,
Janne

[EEVblog]

http://www.polycase.com/item/vm-36boot.html
this looks nice and cheap. I would like a handheld multimeter, not a bench one...

For a bench meter, you could do worse than this one:
http://www.polycase.com/category/zn-series.html
A stackable bench instrument case.

Dave.

[EEVblog]

With regard to compilers, if it is decided that the PIC route is the way to go, then I would recommend the BOOST C compiler.  It's not the *best* compiler in my opinion (that award goes to Hi-tech PICC), but it's certainly the best bang for buck.  Even the professional version is well within most people's budgets.  I used to use Hi-Tech but found that it was priced outside of my range for my hobbyist stuff.  So I've recently moved to BoostC.

If you went PIC then you'd need a good reason the stray from the free Microchip C compiler.
Sure it's one of the most inefficient on the market, but hey, it's free (optimizations crippled only) and available to everyone.
You almost certainly wouldn't need a top quality compiler for this project (with the huge Flash sizes available these days), just something that works.

But everyone can fight over processors choice until the cows come home, that's why I'd be looking at Arduino actually. It's fast enough, all tools are free, and adds a more "community" feel to the project I think.
But it's not the best choice if you want ultra-low optimized power consumption. But that might actually be a reasonable trade-off.

Dave.

[Ferroto]

heh most of these posts are well above my head.

I'd recommend physically intergrading the uCurrent adapter in the project.

[badSCR]

If someone does not have a PICkit to program the microcontrollers,  They could just order the PIC  with the code already programmed into it.  I have a pickit2 so its not a problem for me.

"Looking for a low-cost programming solution? Let Microchip do the programming in a fast, cost-effective, secure and proven method."

[Ferroto]

If someone does not have a PICkit to program the microcontrollers,  They could just order the PIC  with the code already programmed into it.  I have a pickit2 so its not a problem for me.

"Looking for a low-cost programming solution? Let Microchip do the programming in a fast, cost-effective, secure and proven method."

I think digikey offers that option as well.

[dmlandrum]

I concur with Dave on this one. Using an Arduino for this project (and for the record, I'd like a bench meter) would get the attention of Make Magazine and you'd have a large DIY contingent descend upon this place like Martin Sargent at a Megan Fox convention. That can be both a good and bad thing. Arduinos also have the advantage of being relatively cheap and fairly easy to program. A shield could be designed with the extra components needed for true RMS and so on.

But I'm not the expert here, so I'll leave it to you guys.

[EEVblog]

My train of thought on this one...

Stick with bench mounted instead of handheld, because handheld has too many variables that will muddy the requirements.

So, a bench meter. Well, it's gotta be more than a bench meter of course, and with the suggestions so far it's actually becoming more of a general purpose "bench instrument/lab system".

So what would the average punter want from such a system?
Well, all the usual multimeter stuff and data logging that's been talked about, simple very low end oscilloscope functionality, maybe a simple logic analyser / serial protocol system, some form of simple functional generator, and a lab power supply. Bingo, you have a complete mini lab in one unit.

But I wouldn't try and make it all the one unit, I'd have them as separate modules. One as the bench meter/data logger/oscilloscope, one as the bench power supply (that could also power the other modules), one as a PC/Ethernet/WiFi comms module etc.

Such a system would need some thought into any required comms between systems, or they could just be made separate and stackable.

I'm actually working on a very low cost bench power supply at the moment, and was actually considering those stackable bench enclosures (http://www.polycase.com/category/zn-series.html). This is a spin off my almost already developed uSupply project which is another very low cost but quite capable power supply (the uSupply has made a cameo blog appearance, but everyone almost certainly missed it)

Food for thought.

Dave.

[Curtisbeef]

I definitely agree that this should be a Bench Meter and not handheld. There are too many hurtles to overcome with the handheld system. Plus with the increased size there will be more options for expandability and also case options. I think at first we should make the internals only, and let people decide what kind of enclosure they want. Then as we get going maybe someone will find a standard case that we can all use.

I think that we should try to do some type of modular design. Make a Main board with some type of a standardized connector that will carry data/power/etc to the modular boards.

I'd really like to do the wireless display it could be done very simply with a nRF24L01 transceiver and would have a range of 50 foot easy indoors. I think the display should sit in a cradle or something like that on the bench meter. And we could design module for a non wireless display as well, for people who do not need that functionality.

I dont know if we wanna go the Arduino route, those things are so expensive for what you get. Since we will most definitely be making PCBs for modular boards and things like that, making a Main board with a ~$5 PIC Chip seems like a better option to me then having to make the Main board and the have a 30 dollar Arduino attached to it anyway. Unless you are saying that we should just use a Arduino compatible chip on our board and use the Arduino bootloader and IDE to write our code.

My Vote is for a PIC24 but that is one thing that we should definitely decide on soon. I would like to make a development Board setup for our micro controller of choice as soon as possible.

[Andrew]

I am repeating myself here with regards to several aspects:

But I wouldn't try and make it all the one unit, I'd have them as separate modules. One as the bench meter/data logger/oscilloscope, one as the bench power supply (that could also power the other modules), one as a PC/Ethernet/WiFi comms module etc.

One of my first posting in this thread pointed to such a DIY system. You can get some inspiration by looking at the pictures in

http://www.heise.de/ct/projekte/machmit/ctlab/wiki/LayoutSeite
http://www.heise.de/ct/projekte/machmit/ctlab/wiki
http://www.heise.de/ct/projekte/machmit/ctlab/wiki/AlleModule

I think at first we should make the internals only, and let people decide what kind of enclosure they want. Then as we get going maybe someone will find a standard case that we can all use.

This usually doesn't work out. Too many parameters depend on the enclosure. In my experience, if you go with an off-the-shelf enclosure, and if you don't want to go through many design iterations, you have to take care of the enclosure early in the project, not late. Otherwise you end up with the typical DIY enclosure: A huge, half empty box, taking up much more bench space than necessary, with PCB(s) somehow mounted on glued standoffs or other makeshift mechanical support.

[EEVblog]

This usually doesn't work out. Too many parameters depend on the enclosure. In my experience, if you go with an off-the-shelf enclosure, and if you don't want to go through many design iterations, you have to take care of the enclosure early in the project, not late. Otherwise you end up with the typical DIY enclosure: A huge, half empty box, taking up much more bench space than necessary, with PCB(s) somehow mounted on glued standoffs or other makeshift mechanical support.

I agree, as I mentioned before.
Get the enclosure, looks, and user interface/experience wrong and it may not matter how good your electronics is.
I'd actually rank the electronics at least 3rd or 4th down the list of things to do, which is getting pretty close to the bottom end :->

Dave.

[EEVblog]

One of my first posting in this thread pointed to such a DIY system. You can get some inspiration by looking at the pictures in
http://www.heise.de/ct/projekte/machmit/ctlab/wiki/LayoutSeite
http://www.heise.de/ct/projekte/machmit/ctlab/wiki
http://www.heise.de/ct/projekte/machmit/ctlab/wiki/AlleModule

That looks very nice and professional, but probably mega expensive.
I think cost should be a huge driver behind any such DIY project.
After all, we probably aren't talking about anything blazingly new, so maybe help differentiate with bang-per-buck low cost?

And also, who is the target audience?
Just GeekGirl to fulfill her fantasies? (;->), professionals?, hobbyists?, education?, Hackers/Makers?
etc.

Sure, projects can be done "just for the heck of it", but it's nicer if you can target an intended niche.

Dave.

[GeekGirl]

And also, who is the target audience?
Just GeekGirl to fulfill her fantasies?

Lets no go there

(;->), professionals?, hobbyists?, education?, Hackers/Makers?
etc.

Sure, projects can be done "just for the heck of it", but it's nicer if you can target an intended niche.

Dave.

My original idea was to build a "Bench Meter" with good specs (higher than my Fluke 29).

Now people are indicating that they would like an entire DIY lab.... I have no problem with this. I personally think that if we design an open protocol for comms, we can hang anything off it

I like the cases that Dave pointed at (as long as they make a BIG one for a decent PSU (2 variable and 4 fixed outputs (ir 2x 3-30V, +/- 5 and 12V))

But tomorrow I am going to try and get this idea in some shape

[Mastro Gippo]

The ADUC7060 looks Great! ..if only I knew how to program ARMs...  
Dave, can you take some pictures of the analog frontend boards of your most expensive multimeters? Only board layout can be copyrighted, schematics can't, so if we can get a ready to go schematics we can design a slightly different board and have a product as good as the cool guys, for the price of bare components.
I was curious about the implementation of the auto-range too, tonight I thought about it a lot and I had some ideas so I'd like to see if I got it rigt in my mind.

I hope the end product will be so cool and modular and cost effective that the target audience will be ANYONE. I suggest a vibration function in the continuity mode, so we can really fulfill GeekGirl's fantasies!   (Sorry, couldn't resist.)

[Andrew]

I suggest a vibration function in the continuity mode, so we can really fulfill GeekGirl's fantasies!   (Sorry, couldn't resist.)

That would explain the request for a purple case. Well, partly 

[badSCR]

I definitely agree that this should be a Bench Meter and not handheld....

I think that we should try to do some type of modular design....

I dont know if we wanna go the Arduino route, those things are so expensive for what you get. Since we will most definitely be making PCBs for modular boards and things like that, making a Main board with a ~$5 PIC Chip seems like a better option to me then having to make the Main board and the have a 30 dollar Arduino attached to it anyway. Unless you are saying that we should just use a Arduino compatible chip on our board and use the Arduino bootloader and IDE to write our code.  .....

Yap,  I agree.
I don't like the Arduino, BasicATOM, BasicX, BASIC Stamp... They cost to much for what you get.

I like the wireless display module, idea. As an option.

The "Service Manual" sometimes has the schematics.  They cost less then $20, some are free.

[Thermal Runaway]

If you went PIC then you'd need a good reason the stray from the free Microchip C compiler.

I must confess I've never actually tried Microchip's own compiler.  The simple reason being, they don't support (or at least they didn't) the 16F range of devices.  While I was learning C, these were the devices I was playing with, so I had to find a compiler that worked with them.  That's how I got on to PICC.

Once you're used to a compiler, you don't really want to have to swap to a different one (at least, not just for the sake of it anyway).  Hence I never got around to trying Microchip's own compiler.

Brian

[jahonen]

If you went PIC then you'd need a good reason the stray from the free Microchip C compiler.

I must confess I've never actually tried Microchip's own compiler.  The simple reason being, they don't support (or at least they didn't) the 16F range of devices.  While I was learning C, these were the devices I was playing with, so I had to find a compiler that worked with them.  That's how I got on to PICC.

Once you're used to a compiler, you don't really want to have to swap to a different one (at least, not just for the sake of it anyway).  Hence I never got around to trying Microchip's own compiler.

Brian

That is probably due to that anything less than 18F is not C-friendly, as there is no RAM or stack in the usual sense. Stack is very essential thing for a typical C compiler as the arguments for a function are usually pushed on the stack and function local variables also reside there. That makes a 16F a very hostile environment for a C-compiler. Instead there is a big bunch of registers.

Regards,
Janne