Open Source Multimeter

[firewalker]

I've said it before. I'm convinced the optimum number of people to work on a project like this and make the decisions rounds down to 1, almost always.

Dave.

And usually the outcome is better than a same project with many peoples taking decisions.

And as we say in Hellas "When to many rosters are singing, it delays the dawn" (it really looses its value in English).

"Opou laloun polloi kokoroi, argei na xhmerwsei"

Alexander.

[XynxNet]

Some sort of a generic sensor interface would be nice.
At the top of the meter you could plug in little addon sensor modules, which consists of a small mikrocontroller for callibration data and conversion tables and of course the sensor (light, sound, vibration, pressure, etc.). The meter suplies the power for the sensor module and displays and logs the sensor data.

[ruku]

I've said it before. I'm convinced the optimum number of people to work on a project like this and make the decisions rounds down to 1, almost always.

Dave.

Dave, that's the way to make decisions quickly! Not necessarily the best decision, perhaps... but certainly better than design by committee. However, I still wonder if Open Source Hardware can gain a little momentum from how Open Source Software works. Perhaps once these folks get it rolling, they can bring it in for community assistance.

[Frant]

Some sort of a generic sensor interface would be nice.
At the top of the meter you could plug in little addon sensor modules, which consists of a small mikrocontroller for callibration data and conversion tables and of course the sensor (light, sound, vibration, pressure, etc.). The meter suplies the power for the sensor module and displays and logs the sensor data.

The family of standards IEEE 1451.x suggests several solutions. To be more specific, 1451.1 and 1451.2, or 1451.4 could be considered, the others are too complex and don't make sense in this case. However, adding more features has to stop at some point.

[Mechatrommer]

I've said it before. I'm convinced the optimum number of people to work on a project like this and make the decisions rounds down to 1, almost always.
Dave.

Dave, that's the way to make decisions quickly! Not necessarily the best decision, perhaps... but certainly better than design by committee. However, I still wonder if Open Source Hardware can gain a little momentum from how Open Source Software works. Perhaps once these folks get it rolling, they can bring it in for community assistance.

during most meeting i've attended, the more efficient way is someone (leader?) to prepare a complete program tentative for everybody to agree/discuss, rather than meet together and talk from nothing, which risk in the end of meeting you still have nothing. so, imho, in order for open source to work, someone has to prepare a working circuit/product to present and improvement or open source effort will start from there. and someone who started it has to be competent or "soundly" knowledgable so other dont have to improve trivial things/aspects.

[EEVblog]

Dave, that's the way to make decisions quickly! Not necessarily the best decision, perhaps... but certainly better than design by committee.

Not if you get one really smart person who knows that they are doing.
Such a person is worth a hundreds grunts working a project.

However, I still wonder if Open Source Hardware can gain a little momentum from how Open Source Software works.

Likely not, they are two different beasts.
Ok, in theory it can, but in practice it never works that way.
Code is just code, you don't have to worry about manufacturing, physical usability, part availability, power consumption issues etc, and everything that goes along with that. It's a whole different ball-game.
For example, open source software like this BBS forum works because you can build a nice code base, and then everyone can pitch in and write plug-in modules to help expand it, building block style. The project gets bigger and better with time, but still remains the same beast underneath.
You can't really do the same thing with hardware. By it's very nature, hardware lends itself more toward complete physical change with each design iteration. It's not easy to reuse and build upon existing physical blocks in the same manner.
Sure you can reuse bit of schematics as building blocks, and some firmware, but that's about as far as it goes practically.

Dave.

[Harvs]

Thoroughly agree with Dave's comments about open-source hardware.  Lets be realistic, open hardware has been around since valves and electronic magazines, this is no new concept and it hasn't revolutionized hardware development yet.  The only people who really seem to be pushing this concept are the PCB "studios" selling the boards made from other peoples designs.

Maybe I'm just a cynic, however when I see phrases such as "Agile development for hardware", I tend to think they don't have a clue about how decent hardware is designed.

[zunklimt]

wouldn't a modular approach be best the solution?

a solid base to begin with and then add new plug ins with new capacities

if that works for audio synthesis and arduinos, it should also work for professional measurement equipment.

using the smallest smd parts and good design, i'm sure it would be possible to do very small modules, maybe as small as lego pieces or with few squared centimetres

everyone could then arrange its metering system as desired: hand-held or bench, big or small, vertical or horizontal, simple or complex, wired or wireless, with as many channels as desired, with or without that or that capacity...

a well thought unified system for interconnection and physical attachment between the small modules would be the key to get a solid piece of gear at the end...

i know this wouldn't be the easiest to design (both ergonomically and electronically) but the result would then have the best for everyone...


but then, maybe simply doing a good arduino shield with the adequate circuitry to use it as a serious multimeter would also be a way...

if arduino itself is not good enough, maybe using the maple from leaflabs.com ...

[EEVblog]

wouldn't a modular approach be best the solution?

a solid base to begin with and then add new plug ins with new capacities

if that works for audio synthesis and arduinos, it should also work for professional measurement equipment.

It almost certainly won't.
The design will just end up big, expensive, and clunky, and likely not satisfying anyone requirement.
It been tried before many times over the decades, and it just never ends up working, there are just too many design intricacies involved with most projects.

You could possibly have some module thing with a multimeter like this, in fact I was thinking separate modular boards for the inputs. But it's not like you can them go do the same thing with say displays etc.

but then, maybe simply doing a good arduino shield with the adequate circuitry to use it as a serious multimeter would also be a way...
if arduino itself is not good enough, maybe using the maple from leaflabs.com ...

Arduino or any system like that is not the right solution for something like this. It has to be fully custom.

Dave.

[EEVblog]

Thoroughly agree with Dave's comments about open-source hardware.  Lets be realistic, open hardware has been around since valves and electronic magazines, this is no new concept and it hasn't revolutionized hardware development yet.  The only people who really seem to be pushing this concept are the PCB "studios" selling the boards made from other peoples designs.

Yes, nothing new at all. Every magazine project ever published for the last 50+ years is open hardware, it just didn't have that name.
Some projects even got "built on" by other people.
My DSOA project for example, Jim Rowe at EA magazine took that design and produced the Mk2 unit. Used the same PC interface and core circuitry for software compatibility, but got a whole new PCB, new case, and new controls, turning it into a better unit (albeit, much more expensive).
Ok, so doesn't that prove that open hardware works like software?
Well, yeah, for that one iteration only.
If you wanted to go beyond that (which I did with the Mk3), it quickly becomes clear you have to scrap that design entirely and start again from stratch. There is only so far you can build upon a hardware project before it become non-optimal and a better new approach is required. That usually means an almost complete re-design.

Dave.

[jahonen]

Btw, in case nobody remembers, we seem to had this kind of discussion in the past, see Community Bench Meter. It didn't seem to lead to anything, as nobody did not really want to spend some time on it. Fancy things are usually 99% hard boring work, 1% of the fun stuff.

Regards,
Janne

[Chasm]

99% of the fun stuff is usually doing the design, once you move beyond it the work starts and the fun is over. That the the moment where most remember that they don't need the device anyway and that "Because I can!" alone is not sufficient motivation.

[Uncle Vernon]

I'm just a little curious on peoples perceptions of open source an on the comparisons between OSH and OSS.

Open source doesn't dictate design by committee it's more about freedom and transferal of effort and expertise an alternative to "get what your given" and "don't reverse engineer or we'll sue" corporate models.

The comparisons between a proposed multimeter project and open source Linux don't make a lot of sense. If you think for a section about what a Linux distribution is, you can see it is a collection of tools. And just like our multimeter should be, many of those software tools are solo or small team efforts.

To build the ultimate multimeter is like trying to build the best car or the ideal home, there is no such thing and there never will be.  You can only achieve your ultimate, a committee driven thing will always be a collection of compromises. Look at the software comparison the Linux world is full of many similar tools with differing feature sets, open source gives the freedom to select your own best fit collection of tools.

If we were to achieve the hardware equivalent of Linux, we'd be bringing together lots of different open source hardware items and sharing the knowledge. Cant find a power supply that perfectly fits your needs? Then borrow the best bits of existing designs and build your own and share the knowledge you gain along the way. Same for meters or any other piece of equipment.

Open source has found favour from those who felt penned in by commercial offerings. Microsoft anyone? The need exists for any operating system that doesn't demand it's users to the company line. I'm not so sure the same need exist with a multimeter, Fluke aren't trying to tell me how to do my work, Microsoft is, and that's why I utilise Microsoft alternatives wherever I can.

Where open source hardware would really thrive is in areas like domotics, where anyone could select the collection of products and designs that suits their particular want, need and cave size to achieve their unique personal build.

I'm enthusiastic about open source and see it as a worthwhile pursuit, and while I see this open source meter as a bit like pushing string uphill, I'd still love to be proven wrong. Seeing what doesn't work and understanding why can still be some of the most useful engineering there is.

Feel free to discuss, dismiss, agree or deride to your heart's desire.

[Zero999]

I agree with Dave that you can't compare open source software with hardware.

One fundamental difference is with hardware, there's the physical cost of components and manufacture which doesn't exist with software. A software producer doesn't have to worry about cost per unit whether their open source or not. All a commercial software developer needs to do is ensure the revenue they make from sales covers the cost of development, plus a bit of profit.

The physical cost of a single unit of an additional software licence is always near zero. A hardware developer needs to get their product made and sold. The physical cost per unit will always be the price of the components, plus manufacturing and shipping. This is why software piracy is not really stealing. A software vendor doesn't directly loose the value of 1000 licences because someone uploads their software to a file sharing site. A hardware developer would loose the value of 1000 DVMs if someone hijacked a lorry carrying their meters.

A software developer is always foolish, if they're unwilling to negotiate on price. An extreme is example: Altium would be better off selling me a copy of their latest design package for £100 than me either not buying it or pirating it. If they don't sell the software to me the get nothing, if they do sell it to me for £100, they get £100 and their only cost is shipping a DVD to me, £10 at the most in administrative costs,, shipping etc.

In my opinion, the best open source project would be something like an oscilloscope developed by a commercial seller who made the source code for all of the firmware public and encouraged people to improve it, fix bugs and write translations.

[Bored@Work]

This is why software piracy is not really stealing.

Bullshit. You are just doing the usual word games. Word games software thief do to cover up that what they do is wrong. It doesn't matter how you call it, and how many arguments you pull out of thin air. It is wrong.

In my opinion, the best open source project would be something like an oscilloscope developed by a commercial seller who made the source code for all of the firmware public and encouraged people to improve it, fix bugs and write translations.

Oh, you mean the Wittig/Wellec oscilloscope? Three years or so of community "enhancements" of the software and still doesn't beat a cheaper Hantek or Owon.

[Zero999]

This is why software piracy is not really stealing.

Bullshit. You are just doing the usual word games. Word games software thief do to cover up that what they do is wrong. It doesn't matter how you call it, and how many arguments you pull out of thin air.

Wow, what an informative post!

Thanks for wiring a detailed and well reasoned argument as to why it is stealing.

It is wrong.

I didn't say that software piracy is not wrong. I just said it isn't stealing. You can't call it stealing because the unit value of a piece of software is effectively zero. Copyright infringement, voiding of a contract yes but physically stealing isn't the right word for it.

Of course if everyone pirated software no one would make any money from it but they do and software companies have actually encouraged piracy in the past and still do today, especially in emerging markets to increase the user base.

Anyway, I'll stop now. I don't want to turn this thread into flame war on piracy. I just wanted to point out it isn't as simple as many say.

Oh, you mean the Wittig/Wellec oscilloscope? Three years or so of community "enhancements" of the software and still doesn't beat a cheaper Hantek or Owon.

I admit I have no experiences of such oscilloscopes so can't comment. Have you?

[Mechatrommer]

why dont we just limit the OSH definition to schematic level, not up to pcb/product manufacturing. with only schematics to improve, we can later make our own way on how to manufacture it. and with OSH, you should not expect profit as much as the CSH (closed source) can. something like 50% will buy, 50% will diy or pirate, the figure could be wrong, i just want to demo the idea.

[joegtp]

I still don't think group OSH will work.  Simple scenario, I'm a project creator/lead and I want to build a single channel, no graphic LCD, really precise, handheld DMM. If you don't share that same exact goal, hardware design fall apart. If you as a contributor, add channels it will increase not only your cost but mine. If you add a graphics LCD again it would completely change the design possibly requiring a new uC more PCB space, etc.. You get the idea. 

In OSS if you add a wizbang GUI it doesn't really effect me. As the project leader the cost and change to what I use and want it for does not change. Although sometimes it does and that's why you see 20 flavors of linux distros.

Software, at least good software, is very modular and parts can be swapped in and out without a whole lot of problem. That's why I think Arduino is so popular, shields can be added and removed, IO pins are exposed to be expanded upon. Which for a hobby, just playing around scenarios, that works great. But once you try and build a product or do something fancy it starts to fall apart.

Look at Arducopter (http://code.google.com/p/arducopter/),They started off using a standard arduino mega board but the requirements required a custom board which is where they are today. I also believe it started as one guy designing the whole hardware. But now that the hardware is mature all sorts of people are contributing software. Sure the board is OSH but in the end one guy design and built it to the point that it fully worked before others started stepping in.

[Zero999]

why dont we just limit the OSH definition to schematic level, not up to pcb/product manufacturing. with only schematics to improve, we can later make our own way on how to manufacture it. and with OSH, you should not expect profit as much as the CSH (closed source) can.

Schematics aren't very helpful for modern oscilloscopes, at least half of the design is in the firmware.

something like 50% will buy, 50% will diy or pirate, the figure could be wrong, i just want to demo the idea.

What do you mean? Being open source means you give permission to redistribute so piracy would mean someone violating the GPL such as making modifications and selling the design without releasing the source code.

If you're talking about piracy of proprietary software it probably varies widely. Software with a larger user-base will obviously be pirated more and more expensive software is probably likely to have a higher piracy vs legitimate user ratio. Other factors such as the type of demographic of the user-base and DRM also make a difference. The answer is no one  really knows.

[HLA-27b]

Hello there, first post on the forum...

I've been pondering with the idea of data acquisition for over a year. You know, cobbling together some sensors, do the signal munching and data crunching and pass the data to a PC to be displayed and recorded.So I was thinking whether or not it would be a good idea if I brought the subject to Dave, whom I don't know except watching his videos on YouTube. Since I don't know how to design electronics it occurred to me that it may not be such a good idea to push half baked ideas to people I don't know in order to get them to design it for me.

Anyway PXI was mentioned in the thread and since I was drooling over it for long time I decided to go ahead and draw things as I imagined them. So I put together most of the ideas I liked from the forum to see what it would look like.

Design criteria as follows:

- Should use PXI form factor.

I really like PXI. It has everything necessary for measurement and control work. Basically it is a PCI bus with a few additions like precision timing, star triggering etc.

The physical layer of PXI is taken from IEEE 1101.10 which is the standard dealing with Eurocard physical dimensions which are ment to go in 19" racks. Since this is worldwide standard finding cases and hardware should be easy and cheap.

http://ph-dep-ese.web.cern.ch/ph-dep-ese/crates/standards/1101_1.pdf

The backplane of PXI is largely based (but not exactly the same) on CompactPCI which is also industry standart.

Other than that everything revelant is published on
http://www.pxisa.org/Specifications.html

All in all a community designed multimeter has best chances of survival if it is based largely on industry standard  hardware. Thus it can live inside many standard cases side by side with commercial stuff and other community gadgets like LCR's, Load cell amplifiers, Thermocouple modules, Frequency generatos, counters, relay boards, SDR's etc. etc.


- Standart Commercially available case

As I said, I am all for PXI and if this multimeter is ever developed I'd stick it in a 3U Rack with all the other community developed things to go with it. However the conversation is about a hand held device so i drew it as such. My point is that you can have both PXI and "Hand Held" at the same time. This case can house two PXI units. If it ever becomes available, users can switch one of the channels for a LCR for example.

   The box on the pictures is BOPLA Aluplan 51020 it measures 245mm Lenght x 143mm Width x 52mm Height. It is made from aluminium extursions. Top and bottom covers are 3mm thick everything else is thicker. Corner pieces are zinc castings. I don't know how much it costs but sure beats 30.000$ tooling cost for custom case.
    http://www.bopla.de/en/product-catalog/enclosure-technology/product-category/aluplan-13/product-detail/ap-51020-2.html

It is ment to be held sideways so that when sitting inclined on a bench the probe cables are sticking out from the right side - away from where your hands would be when pushing the buttons.

- Enough space for batteries

   Currently it has 4 AA cells inside. Maybe it can accomodate 4 D cells if very slightly bigger box was chosen. I think a power input should be made available as well. So that the gadget can work for months measuring away some mundane thing an give alarm and start logging data only if some alarm condition is met.

Something like "Give alarm and log data as long as Channel #1 voltage is below 6.25V AND Channel #3 Temp is above 99 Degrees Celsius" Which would mean that your chemical reaction is underway even when you are not applying power. Which may as well mean that you hit the Jackpot and you are going to be rich.

- Good User Interface

Dave said he likes 7 segment dispalys but 4 dispalys with 5 digits each? I'd prefer a good ole monochrome LCD please. The one on the picture is 100mm x 70 mm (say 4 by 3 inches). Button arrangement is inspired from aircraft interfaces with a screen and soft buttons around it. A slight twist is the two row buttons on the right side which may be used to alter preferences without entering menus.
   





Front View



Side View




Side View - Connection Jacks



Cross Section - To show what goes where


The small PCB behind the batteries is the "PXI Back-plane" of course it is not standard.it connects only 3 modules (the main board with the display is also a module)






This is standard Eurocard size 3U. PXI, CompactPCI, IEEE 1101.10 and probably many others all use this form factor. Sockets at the back side change from standard to standard.



Well, I'd like to hear what you think

Edit: Dave mentioned these sketches on The Amp Hour.
        Starts around 0:20:05

[Frant]

Well, I'd like to hear what you think

Impressive work!

[Joshua]

Awesome Job! What CAD software is that?

[HLA-27b]

Awesome Job! What CAD software is that?

Google SketchUp
it is free and by far the easiest one to use.
http://sketchup.google.com/

Edit: I attach the 3D files used to make the stuff above.
Please note that the non free version of SketcUp was used to create them.

[EEVblog]

Hello there, first post on the forum...

You win the first post of the year award!
Awesome.

Would it fit 2 x D cells?
That would still kill the capacity of 4 x AA's

Dave.

[HLA-27b]

It could fit 4 D cells if the case was some 10 - 15mm longer. So definitely possible with another case. In fact i pick this one a bit arbitrarily - it was the first one that could take 2 PXI boards so I used it straight away. 

I probably should post the 3D files so everybody can have a go at it. (Done look above)

On the other hand I would appreciate if you Dave (and all of you guys of couse) can elaborate on some questions I have in mind.

How much stuff could fit on 100mm x 160mm Eurocard?
2 separate channels with amps and everything or maybe just one volts channel and one amps channel?

What do you think of the PXI thing?
Do you think that once a community designed multimeter and a backplane is made available to the amateurs it would spur development of many other devices to go with it? LCR's Freq Generators etc?

Would you buy one if it was sold as a kit of pieces ?

For example a kit of
1 - PXI multimeter board (assembled and tested)
1 - Main board with processing LCD screen and buttons (assembled and tested)
1 - 1 backplane (to be assembled by the buyer)
1 - Box to be drilled and assembled by the buyer (takes two pxi units)

Any prognosis on the prices?