Community Bench Meter

[jahonen]

BTW, just checked local Farnell price for STM32F103C6T6 (72 MHz ARM Cortex-M3, 32 kB flash, 10 kB RAM), 4,27€/each. Compare that to 18F-series PICs, which tend to be more expensive, although just 8 bits. I have programmed PICs using assembly, ARM7s and MSP430s with C. I think that the initial difficulty of the ARM is usually just getting things up and running (like configuring the PLL etc). After you figure that out, ARM tends to be much easier. But much of that can be avoided if you bother to spend more than 15 minutes to read the MCU user manual (and that errata Dave mentioned)

I guess that one would like to use floating point arithmetic for the calibration calculations and result processing to utilize the ADC resolution optimally. 32 bit ARM is much more suitable to that task. I'm not saying at all that is not possible with 8-bits and with fixed point arithmetic.

I have seen so many times at work that it is better to initially choose somewhat overkill chip to do the stuff, since price of the hardware tends to come down with the time, but the complexity of the software tends to go up Customers not realizing that will often find themselves doing re-design in panic at the very near the deadline, when it turns out that software has grown much beyond initial expectations

Regards,
Janne

[charliex]

I wish ethernet was in use a lot more on a lot of equipment, its so much better. There are so many stacks and solutions available as well now.

[EEVblog]

Usb and ethernet cables on a multimeter are just a disaster waiting to happen. It's not a good idea to connect a multimeter to the common ground, because if you don't isolate everything you will end up shorting things to ground if you measure a signal with a floating reference. I think wireless (bluetooth, zigbee, custom 2.4ghz) is a better option.

I don't see any benefit in using an ARM, I think all the tasks we need can be easily handled by an easier micro.

I suggested the microSD because they're, well, micro, but a regular slot will be ok too because you can find adapters everywhere.

Sorry Dave, but I think that a rechargeable battery is a must if we want an OLED diaplay, wireless, SD slot.... But I wouldn't go the SLA route, li-ion are cheap and easy to work with, SLA weigh a lot and has Peukert and is big...

We should try to focus on building a meter, and keeping it simple and useful (and CHEAP!!); I understand the enthusiasm, but adding tons of features that don't belong to a meter will only be frustrating in the end imho. 

Frustrating like rechargeable batteries!
You can easily have an a low power dot matrix LCD display with a nice backlight, best of both worlds. And even with the backlight on you could still shoot for 1000 hours operation off D cells.
The Newhaven LCD display on my new uWatch design for example takes under 200uA, and backlight works well at under 10mA.
OLED would likely not allow long term data logging with continuous display.

A agree with SD over micro, much less fiddly and more readily available.

No built in direct connect PC comms (Ethernet or USB), it needs to be external and isolated. So some form of clip-on box with a serial IR interface.

Dave.

[badSCR]

Anything hard-wire to computer scars me, about the Isolation problems $$$.


*Serial IR ; slower speed, but the SD-Card could be used as a buffer.
*An SD-Card or usb flash drive  is a must for data logging. (for high speed data logging maybe 1~2Gb RAM)
*Auto Ranging?
*3~5 channel Volt meter/logic    And     Dual channel Amp meter.
*Would like the ability for 10~20 DC amps continuous.
*Ability to measure Resistance, Diodes, Inductors, and Capacitors.
*maybe a Transistor tester.
*micro volts/amps
*maximum 600 Volts to 1Kv ?
*phase angle (for three phase power) ?
*able to use Fluke Clamp-on Current Probes (for three phase power) ?
*low frequency Oscope ability (less then 200Khz) ?


*A Wall-adapter for a PSU.
*Able to use a Lithium-Ion battery pack from a cordless tool (non-charging).  If your some where and the battery dies then you could use the one from a power tool.
*D cells or the Big 6volt battery, both are easy to find.


*In Resistance mode it could tell you the color code and if it is within tolerance.
*Maybe a Composite video output.

[badSCR]

Could do the Ethernet and use a Wifi Bridge to connect to a PC.  And use a buffer.

[xani]

well, ethernet use isolating transformer, just get a good one ;]

[Curtisbeef]

For the USB data logging could it just be Opto-isolated serial lines going to a USB powered and Isolated from the rest of the board FTDI Serial Chip. Is that enough?

[charliex]

just put RS232 out on it, after that its trivial to add on Wifi or Bluetooth.

[EEVblog]

*Serial IR ; slower speed, but the SD-Card could be used as a buffer.
*An SD-Card or usb flash drive  is a must for data logging. (for high speed data logging maybe 1~2Gb RAM)
*Auto Ranging?
*3~5 channel Volt meter/logic    And     Dual channel Amp meter.
*Would like the ability for 10~20 DC amps continuous.
*Ability to measure Resistance, Diodes, Inductors, and Capacitors.
*maybe a Transistor tester.
*micro volts/amps  or  pico volts/amps
*maximum 600 Volts to 1Kv ?
*phase angle (for three phase power) ?
*able to use Fluke Clamp-on Current Probes (for three phase power) ?
*low frequency Oscope ability (less then 200Khz) ?

*A Wall-adapter for a PSU.
*Able to use a Lithium-Ion battery pack from a cordless tool (non-charging).  If your some where and the battery dies then you could use the one from a power tool.
*D cells or the Big 6volt battery, both are easy to find.

*In Resistance mode it could tell you the color code and if it is within tolerance.
*Maybe a Composite video output.

Add an adjustable constant current output for testing LEDs.

Dave.

[Thermal Runaway]

Of the features already mentioned, I have experience with the following:

• LCD Interfacing via serial and parallel methods, character and graphic
• SD Card interfacing for data storage and retrieval (currently working on a project that uses this)
• Infra-Red communications
• General Microcontroller firmware design
• General electronic circuit design

However, I am unlikely to be available to contribute to fun stuff like this until the academic year has ended (around about May/June).  I'll follow this thread with interest until then, and see how it's progressing in a few months time.  If there are still tasks I can help with at that time, I'll volunteer for some jobs.

Brian

[SockThief]

It feels a bit like a kitchen in a cooking school at the moment, with everyone throwing in their own wishes, stirring it about, its going to be interesting to see what comes out. This is the part I always like about community based open-source projects, it can be anything at all! It's like giving a list to Santa and getting it all! (and yes there are probably a few people reading this at the moment, that feel a little underwhelmed by the iPad unveiling last night. The actuality had no hope of living up to the expectactions, we have no such limitations!) But this brings to me a small idea that has been floating around for a while, I always have my iPhone with me. And currently there are iPhone bases Oscilloscopes, meters etc. But as far as I can recall (I actually haven't looked at them for a while) they use the internal H/W to perform A/D conversion etc. So there are strong limitations there.

But! What if there was an external module that communicated by Wi-Fi to the iPhone (or any other smart phone or Wi-Fi equipment for that matter) but let's assume an iPhone, passed all the measurements to the iPhone, then the iPhone could postprocess the results for display, emailing, sms etc. There is no reason why a PC could not stand in for this function, or another custom build module. In any case, by using an iPhone you have
- data which is as high-a-quality as possible (after all its coming from a purpose designed piece of equipment)
- a display is already available
- a ready made communication chanel between the data-collector and the display
- a ready made data analyser
- 3G on the phone allows you to phone-home data to HQ (perfect for in the field work)

just another idea to throw into the pot. I guess this is something Im more than willing to look into and play around with. And can largely be done in isolation from the rest of the project, merely relying on a communications protocol.

Til next time!

[xani]

Well, if u have ethernet u can connect it to pretty much everything everywhere
Another idea:
Maybe split it into 2 parts, analog input + simple uC to interact with A/D Conv. + "head" with screen and all other measurement functions connected by infrared or optocoupler. So when something fails on analog part, digital part with SD/USB/Ethernet and other fancy stuff is safe. Would be easier than isolating USB

[jahonen]

Well, if u have ethernet u can connect it to pretty much everything everywhere
Another idea:
Maybe split it into 2 parts, analog input + simple uC to interact with A/D Conv. + "head" with screen and all other measurement functions connected by infrared or optocoupler. So when something fails on analog part, digital part with SD/USB/Ethernet and other fancy stuff is safe. Would be easier than isolating USB

Agilent power supplies controllable via GPIB or Ethernet are done just this way. Probably their bench DMMs too. One should be very careful about the isolation. Handheld DMM is especially critical since it is "hand held". Bench thing is somewhat easier in that sense.

Ethernet line transformers are tested 1 kV RMS isolation, but that is not enough for safety. Something like 5000 V RMS would be more appropriate there. But anyway without extensive testing, there is no way to guarantee safety (or any CAT-rating), as Mr. Maxwell tends to be more imaginative than we poor design engineers

Add capacitor ESR measurement to the functionality list. This should work downto 1 milliohm (4-wire measurement required?), since most ordinary electrolytics are rated something like 10-20 milliohms.

Regards,
Janne

[GeekGirl]

Well, if u have ethernet u can connect it to pretty much everything everywhere
Another idea:
Maybe split it into 2 parts, analog input + simple uC to interact with A/D Conv. + "head" with screen and all other measurement functions connected by infrared or optocoupler. So when something fails on analog part, digital part with SD/USB/Ethernet and other fancy stuff is safe. Would be easier than isolating USB

Agilent power supplies controllable via GPIB or Ethernet are done just this way. Probably their bench DMMs too. One should be very careful about the isolation. Handheld DMM is especially critical since it is "hand held". Bench thing is somewhat easier in that sense.

Ethernet line transformers are tested 1 kV RMS isolation, but that is not enough for safety. Something like 5000 V RMS would be more appropriate there. But anyway without extensive testing, there is no way to guarantee safety (or any CAT-rating), as Mr. Maxwell tends to be more imaginative than we poor design engineers

Add capacitor ESR measurement to the functionality list. This should work downto 1 milliohm (4-wire measurement required?), since most ordinary electrolytics are rated something like 10-20 milliohms.

Regards,
Janne

Hi Janne, I would like to add ESR, should not be too hard

I have isolation for my Tek TDS220, I use an isolation box I built as a project for my networks course I did back in '98. It is about 4m of optic fibre cable and TX / RX pairs, it isolates RS232 (in each direction Max232 - Tx Led - Fibre Cable - Rx photo diode - Max232. Both ends were powered by 9v batteries, an would run for about 16 hours)

That could be an option for this project. I am yet to find any decent voltage rated isolation devices (ie >10Kv)

[Mastro Gippo]

Yes, ESR and Dave's uCurrent are a must! 
I'm a bit scared of ARMs, that's why I was supporting PICs.. But if someone manages to get an easy bootloader working, I can try to join the fun! I've been thinking about learning ARMs for a while... Maybe we can start on the mBed platform, http://mbed.org/! All the hard part is already done.

[Mastro Gippo]

By the way, I just got myself an iphone (yes, I know, shame on me, but i got it for just 100€ broken and I repaired it so at this price it's worth it), and I think that an OS-based user interface like SockThief said can do great things.
But not everyone has an expensive iphone... well, last year I got this from dealextreme: http://www.dealextreme.com/details.dx/sku.19609
it has a GREAT screen, visible outside in bright sunlight (I'm not kidding; I wrote an app for it to measure and display speed and direction to skippers for a sailing boat), and it's cheap (there are even cheaper models). The nice thing is that you can program it in c# REALLY easily, debug code line by line via USB, use wide available libraries, make graphics easily, and the interface is based on win CE so it's very flexible. I even found an I2C port inside it, by looking at the radio transmitter module datasheet, and tapped into it. The port is mapped in the OS, but I was not yet able to do anything with it yet.
For 5$ more you can get bluetooth too.
We can do the "detachable display" trick by having a black box that deals with all the analog stuff and communicates with the display via bluetooth.
It's a hack, I know, but it's the biggest bang for the buck!  

[jahonen]

Yes, ESR and Dave's uCurrent are a must! 
I'm a bit scared of ARMs, that's why I was supporting PICs.. But if someone manages to get an easy bootloader working, I can try to join the fun! I've been thinking about learning ARMs for a while... Maybe we can start on the mBed platform, http://mbed.org/! All the hard part is already done.

Most ARMs have a bootloader built in already. For example NXP ARM7s can be programmed by just connecting RxD and TxD via voltage level shifter (e.g. MAX232) to PC (USB serial ports work just fine since all communication is RS232-friendly). ROM Bootloader is started if certain pin is pulled down when reset is released. Programming software is also free, e.g. http://www.flashmagictool.com/.

Regards,
Janne

[Mastro Gippo]

Cool!! Thank you!  Now I just need a project to justify spending money for a dev board - and arm ones are pretty expensive.... But I've seen some nice options from olimex.

[EEVblog]

It might be worthwhile considering the LXI standard interface for LAN based instruments:
http://www.lxistandard.org/

Dave.

[charliex]

I'm happy to offer to help with the code for ARM, or most processors.  I've also got developer accounts for iphone, and most other phones OS's,, embedded systems ( even the nintendo wii and ds! ) for any kind of remote stuff, we use the GBA a lot like this.

http://www.mikroe.com/en/tools/ have an excellent selection of dev boards.

[Thermal Runaway]

@charliex

Every single time my Elektor issue comes through the post, I see the adverts for those development boards and think that it'd be pretty useful to have one around - they do seem quite comprehensive.  But I've never got myself one.  Each time I do a new project, I build a quick and dirty breadboard based dev board to suit the specific purpose.

I should get one of those though, I think.

@thread

I've no previous experience with ARM at all, although in fairness once you get into the realms of C... I guess changing processors is not much of a problem.  I've never tried it, because PIC has always suited my particular needs.

I'll watch this thread with interest.  After all the ideas have been thrashed around, someone (GeekGirl) is going to have to take control of the situation and decide on a specification for the design.  Then organise people to contribute to it.  Without the spec or the organising, nothing will get off the ground so I think it's pretty important. 

Brian

[GeekGirl]

I'll watch this thread with interest.  After all the ideas have been thrashed around, someone (GeekGirl) is going to have to take control of the situation and decide on a specification for the design.  Then organise people to contribute to it.  Without the spec or the organising, nothing will get off the ground so I think it's pretty important. 

Brian

Hi all,

I am watching this thread closely, I was going to do a block diagram, but I am hanging off for a little while more, as there is great discussion going on I am going to ask people to nominate areas they would like to help with If anyone wants to make up parts of the PCB layout or Schematic, as long as they provide a PDF I can incorporate it into the "master" schematic and PCB (I use ADS09 (Altium Designer Summer 09) and I do not expect other to have this package) then I will publish these back to this thread as both native ADS09 and Pdf

I will hope by Monday to have a list of features that have been suggested so far in one POST so we can all follow (Sunday for me is going to be hectic lol)

If the group would like I can also set up a project page on Sourceforge so we can use CVS for tracking hardware, firmware and software revisions

Regards,

Kat.

[Andrew]

I should get one of those though, I think.

I find them, how should I say, unbalanced. 70 or 100 buttons on a board looks great, but what on earth are they good for on a dev board? Sure, if I think hard enough I can find (artificial?) applications where I would need 70 buttons on a dev board to prototype. But if I keep it real, a prototype application needing 70 buttons warrants a separate keyboard (incl. communication protocol), and even 20 buttons on a board aren't really what one needs on a dev board.

IMHO these boards are made to impress, not to deliver the most value for the developer.

I'll watch this thread with interest.  After all the ideas have been thrashed around, someone (GeekGirl) is going to have to take control of the situation and decide on a specification for the design.  Then organise people to contribute to it.  Without the spec or the organising, nothing will get off the ground so I think it's pretty important. 

I want to add another aspect  to the discussion, more basic and down to earth than just a feature list.

Since this is going to be a bench multimeter it needs an enclosure. I assume that a custom enclosure is out of the questions, so an off-the-shelf enclosure is to be used.

From my experience, when you work with an off-the-shelf enclosure it is unavoidable to settle early on in the project on a particular enclosure, because the enclosure will dictate to a certain amount how things can be build. It will to a large extend dictate the mechanical dimensions of the PCB(s), their mounting, and even partly the separation of functions into different PCBs (e.g. a separate front-panel PCB, a power-supply PCB, etc.)

I actually like the classic bench-top multimeter form factor like in

http://www.home.agilent.com/agilent/product.jspx?cc=DE&lc=ger&nid=-34037.899260&imageindex=3
http://www.bkprecision.com/products/photos/large/2831D_front_lrg.jpg
http://www.fluke.co.uk/comx/show_product.aspx?pid=37725

in particular when they come with the rubber protection (BTW, does someone know who invented that form factor for instruments?)

These enclosures are approx. 300 mm deep (including the rubber), 260 mm wide (including the handle), 110 mm high (including the rubber).

Some time ago I tried to find an affordable off-the-shelf enclosure of the above kind, but I wasn't very successful. Just finding something with the right dimensions was a problem (typically they had the wrong depth / width ratio), most didn't have a handle bar, or they had handle bars which were completely different compared to the ones in the examples (making stacking impossible). And I didn't find a single one with the rubber.

So, if someone knows a reliable source for affordable enclosures of this kind, I'd like to know it. And I would suggest that a particular enclosure is adapted very early for the bench meter.

[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.

I think even DigiKey maybe even mouser will have similar sizes and maybe even models.

Regards,

Kat.

[EEVblog]

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.