Community Bench Meter

[GeekGirl]

Hi all, in the EEVBLOG #56 comments I said I would start this topic so here it is.

Lets design a Bench Meter, that can do what we want, completely open source firmware (for micro) software for *nix, Mac Windows, Open source schematic and PCB.

Something that someone with reasonable soldering skills can build (SMD is ok but we will keep it to 0805 and larger) The PCB would probably need to be commercial as we will need PTH and Double sided.

OK some specs to kick the ball off, These are open to debate

Battery / Mains powered.
Back light LCD Graphics
5000 count (we could aim for 50000, but I do not know with a DIY project if we can get it foolproof enough for the average hobbyist to get it to work)
20A DC range, uA range (maybe dave would give permission for the uCurrent to be included inside)
Ohms, LOW OHMS
Lightning fast continuity range
RMS AC
Graphing function
Data Logging

I would like a retro push button range switch

What are your Ideas ?

I will keep tabs on this topic, and I am prepared to run the hardware side, keep the master schematic and PCB uptodate, I will issue the Schematic and PCB in both PDF, ADS09, Gerber format (sorry I do not use Eagle)

Ok do you want to do this ?

[joelby]

Jaycar actually sell a DMM kit for $24.95. The specs are undoubtedly rubbish, but it comes in a standard-looking multimeter case!

L & C meter? There are plenty of schematics floating around for making these.

[GeekGirl]

Yeah, I have seen the Jaycar Kit, But the specifications are not shown (maybe they are that bad

I am thinking of a design that is more accurate, that we can design as both a learning experience and a way to get a "decent bit of kit".

There are lots of projects for just about everything but a "DIY" Bench meter.

I think the whole design would be fun and we can all learn things along the way (I have never done high precision ADC designs).

[SockThief]

Hej!

Long time reader, first time caller. Actually when I saw this post, I couldn't help buy post a response. Personally I think it's a great idea, as you surf the web, reading blogs, there is, inevitably, a comment to a post "why isn't there an OpenSource X?". And, there is no reason why there can't be! It seems Dave's popularity is growing and a legion of experts find themselves following (remember the experts arguing over the CRO probe interference?), so there is the talent out ther to achieve this.

While I appreciate the Jaycar kits, etc (I remember Dick Smith Funway had a Multimeter kit somewhere in the pages of time) I do feel GeekGirl is on the money when we say we should aim our sites a little higher. A good piece of test equipment is essential, and there is no more piece of kit more fundamental than a good meter (why is it Dave reviews 'em so often, and seems to get so riled up when they aint up to spec?).

As a beginner, not only out of this project will you get a good meter, but you should also, out of following the design, get an appreciation out of good design practices, electronic theory etc.

As a professional, you will be able to give back to the community that we all love, and pass on that knowledge that our mentors and experieces has given us.

Me, personally, have been out of the electronics game for a LONG time! Seemed there was a shortage of software guys when I needed a job... so I am longing to get back into electronics scene (which is why i hang around these dark corners of web), and am always looking for a project to help and learn from. I am open to helping design any part, though, at this stage, i will confess my ability to design acurate meaurement circuits is not the best, so I will be interested to see what others come up with.

As a comunity project, my personal feeling is that we should be as modular as possible, not only will this make design easier, as each part can be designed and tested independently, but it will also allow switching of various modules. For example, not confident enough to get a 50,000 count circuit running and don't need the accuracy? No problems, drop in a lower speced module. Can't get part XYZ? Easy, design one around what you can get. etc, etc, etc. So I guess the first stage we need to be looking at is some block diagrams....

Looking forward to this, should be fun!

[Andrew]

German computer magazin c't did a whole series about building a modular DIY measurement and control lab http://www.heise.de/ct/projekte/machmit/ctlab/wiki

It is (multiple) MCU controlled, with LabView integration, USB, Ethernet RS232 interface. Among the published modules are general purpose I/O modules, lab power supply modules, DDS module, audio amplifier module, electronic load and whatnot.

And a voltmeter module. The voltmeter module schematics are at http://www.heise.de/ct/projekte/machmit/ctlab/browser/Schematics/schem_DIV3.pdf (24 bit ADC of which approximately 18 to 20 bits are usable, approx 5 1/2 digits). The TrueRMS extension schematic is at http://www.heise.de/ct/projekte/machmit/ctlab/browser/Schematics/schem_TRMSC-DIV.pdf (links don't load the PDF's directly, but point to wiki pages holding links to the PDFs)

The hardware design is copyrighted by the publisher. But there shouldn't be anything wrong looking at them and getting some inspiration.

BTW, ICL7xxx based DIY multimeter schematics were popular in the 80th. However the same ICs and schematics are used today in $7 rubbish 3 1/2  multimeters. So there isn't much to gain going the ICL7xxx route. You can't beat $7.

Oh, and I would go for a bench-top open-source multimeter to avoid the mechanical problems of squeezing everything into a handheld case. Going with the classic bench-top meter form-factor or the 19" form-factor (like c't) gives a lot of room.

[Valhallasmith]

I've been thinking about building a similar tool.  Except I am aiming more at something small and fits in a round Altoids tin.  First project in this line is probably going to be a 24V-~0 regulated charge pump run off a wall wart.  The one after that is a multimeter/lcr with a couple DAC/DDS outputs.

[Curtisbeef]

I'm so happy that a little comment Inspired this thread

I'm pretty Amateur but I would love to help out in anyway I can. I'm pretty familiar with LCD technology/programming

For a project like this i think the Best thing we could use would be a KS0107/KS0108 Compatible LCD They can be found everywhere in just about every color and even in OLED and VFD, the most standard size is 128x64. And they are extremely cheap for the non-fancy color ones. Also there is code out there for driving them on just about every platform you could imagine.

The prices on this site are a little high but it has the largest selection, just to get a idea http://tinyurl.com/yfqwg6y

Also for a bonus here is some video of a KS0108 Compatable VFD
and OLED


Edit: Found a great Pic of a PLED KS0108 display

[Mastro Gippo]

Count me in! I have some experience with very low voltage precision measuring (http://www.mastrogippo.it/eeg/) and I'm pretty good at firmware if we use PICs. But I think a ti MSP430 will fit better. Good way to learn!

[GeekGirl]

Count me in! I have some experience with very low voltage precision measuring (http://www.mastrogippo.it/eeg/) and I'm pretty good at firmware if we use PICs. But I think a ti MSP430 will fit better. Good way to learn!

DAMN you Now I want to go and work on my ECG system I have another project to put up on the board now Years ago I was developing an ECG monitor (similar to a halter monitor) which you could wear and it would link back to a PC / mobile phone wirelessly that if you electrocuted your self, it would shut power off to the area and place a call to the ambulance service giving a pre recorded message (stating where you are, what has likely happened etc.)

One day I may actually get it finished and on the market lol.

[dmlandrum]

I'm a beginner electronics designer, but I am a physics major, so I hope that counts for something. One of the classes I'm taking this semester is Physics of Electronics, focusing on instrumentation, so this might be a great thing for me to get involved in. I have some software experience, but not with embedded, but I've been wanting to get into that for some time as well.

[xani]

im thinking, maybe "dual display" ? Some nice LED 7-segment for "main" display and 4x20 (or graphic) LCD for other things like menus and other measures.
So for example u can have Vrms on nice big LEDs + other measures like Vp-p, Min-Max, or freq on LCD. Or even dual-measure, for example I on LED and V on LCD + power and resistance calculated from it

[GeekGirl]

im thinking, maybe "dual display" ? Some nice LED 7-segment for "main" display and 4x20 (or graphic) LCD for other things like menus and other measures.
So for example u can have Vrms on nice big LEDs + other measures like Vp-p, Min-Max, or freq on LCD. Or even dual-measure, for example I on LED and V on LCD + power and resistance calculated from it

I am liking this idea, but I do not want to use a character LCD, I want to use a graphic LCD, with one like the one posted by Curtis, you could have both in 2 lines on the LCD

[EEVblog]

Cool idea.

I think it's gotta be different somehow, not just a bench multimeter.
I'm not sure what or how, just offer something compelling, either in form factor and/or combination of features.
Be careful not to slide down the oscilloscope functionality slippery slope...

Forget mains use, any meter worth building can be battery powered.

This is more of a hack idea, but what about a project to design a new engine/user interface for the Fluke 233?
It's got a re-flashable micro in the head (don't know about the base), and you could probably replace the LCD with a graphic one. Not exactly low cost though.

Dave.

[Mastro Gippo]

Hehe Kat, I had the same idea sometime ago, but I wasn't going to wait for the heart to stop to call for help, I was thinking about a bunch of electrodes that trigger the power shutdown when they detect a voltage difference between them, caused by an electrocution.

I think we can easily make a wireless display like the one Dave reviewed, but it would be hard to design the case. The wireless interface could be used to interface to a PC too. I think we can just start to design the PCB around a cheap and widely available multimeter case, maybe from Dealextreme.

As Kat says, graphic LCD/OLED is a must. A micro SD slot is tempting too.

I think hacking a Fluke is a bit blaspheme  , but maybe if we can find a good multimeter with a nice analog front-end to harvest for cheap, it would save a ton of time.

I always wanted to build one of these:
http://mondo-technology.com/super.html
we can get a lot inspiration from that project! 

[GeekGirl]

Cool idea.

I think it's gotta be different somehow, not just a bench multimeter.
I'm not sure what or how, just offer something compelling, either in form factor and/or combination of features.
Be careful not to slide down the oscilloscope functionality slippery slope...


Dave.

Hi Dave,

I definitely do not want to make this a scope lol (Rigol and others have nice cheap units )

I am thinking one idea to set it apart is to make the current input separate, so we can compute Power Also so we can display current and voltage at the sametime, compute power factor and VA for AC, Power for DC.

[jahonen]

What I would like to see, is some modern computer interface, USB or even better for long distance logging, ethernet. Of course, the isolation/safety issues will arise then.

Actually, what I like about mains powered meters is that I have never had to run around finding fresh batteries before starting measuring with them Unlike my Fluke 89IV, which seems to eat AA's like bread. That is why I like the Agilent 34401A (or any similar) so much.

Regards,
Janne

[GeekGirl]

A micro SD slot is tempting too.

I would just use a normal SD card, it is not that hard

I think hacking a Fluke is a bit blaspheme  , but maybe if we can find a good multimeter with a nice analog front-end to harvest for cheap, it would save a ton of time.

Unless we are just going to be "Hackers" I do not find much appeal in modding a meter that is so cool already

I always wanted to build one of these:
http://mondo-technology.com/super.html
we can get a lot inspiration from that project! 

Nice project, some great options if we used one of the newer Pics (24 Serries IIRC, I am not a Pic Chick, I am an AVR Chick, but if the community wants to design with a Pic that is fine )

Their case does not have enough room

[GeekGirl]

What I would like to see, is some modern computer interface, USB or even better for long distance logging, ethernet. Of course, the isolation/safety issues will arise then.

Actually, what I like about mains powered meters is that I have never had to run around finding fresh batteries before starting measuring with them Unlike my Fluke 89IV, which seems to eat AA's like bread. That is why I like the Agilent 34401A (or any similar) so much.

Regards,
Janne

Two very good points, we could do both Ethernet and USB in the one design, just drop a WizNet Module in (I think they have USB and Ethernet in the smae footprint). I would like to slip in a small PSU, if only to charge the battery. I have built in 1.2Ah 12V gel cells in equipment (they only way a couple of hundred grams .

Keep the great thoughts up

I hope tonight or tomorrow (I am GMT -8) to be able to produce a block diagram (nothing to complex just get some ideas on paper

[EEVblog]

I definitely do not want to make this a scope lol (Rigol and others have nice cheap units )

I am thinking one idea to set it apart is to make the current input separate, so we can compute Power Also so we can display current and voltage at the sametime, compute power factor and VA for AC, Power for DC.

Yes, dual inputs would be a nice differentiator.

It's got to do data logging though, and hence low frequency scope stuff, but that would be it. Dual channel data logger would be a handy tool.
Idea: Would be nice if it had a serial protocol analyser of some form too.
And taking that a bit further, a small and simple 8 channel logic analyser input maybe?
Or maybe the two primary inputs plus a simpler 8 input channels for logging.

Powering it from a single D cell would be neat.

Dave.

[EEVblog]

I would like to slip in a small PSU, if only to charge the battery. I have built in 1.2Ah 12V gel cells in equipment (they only way a couple of hundred grams .

Nooooooooooo! If you go down the charging dark path, forever will it dominate your destiny!

I think it's possible to power a multi channel meter with a small graphic display for many hundreds of hours.
If it was say direct D cell powered, a target of 1000 hours of operation would give you a current budget of 18mA. Obviously more once PC connectivity kicks in.

Dave.

[xani]

Maybe use some small ARM instead ? More horsepower and interfaces, like for example  AT91SAM7X256 , builtin Ethernet and USB, ok price and plenty of space in flash and RAM.
Then we could add things like USB<=>SPI/I2C/USART debugger or serial bus monitor on LCD. Or telnet into your meter from work

[Curtisbeef]

I like the MicroSD card idea. Its simple enough to put a 8GB storage card in there for many hours of data logging and they are super simple to interface with almost 0 extra parts(besides the connector). Id definitely say go with the MicroSD over the standard SD the size makes it way better and there are no downsides.

USB PC Connection is a MUST.
I love Dave's idea about about having a simple serial terminal. That would be a Awesome feature. If we are gonna go that route might as well add I2C and SPI Analyzers Etc...

My opinion about the Ethernet is that it kinda adds a huge amount of complexity / expensive parts for a feature that I don't know how useful it would be to most people. But like was said before we could always go for a modular type of a system where that could be a addon feature or something.

For Micro I would probably suggest some type of MSP430 because they are know for using like insane low amounts of power. And there are a TON of options. But if we are going for Horsepower ARM would definitely be a good option.

I love the idea of making it handheldish..../portable but if you guys wanna make it a bench tool thats cool too.

[xani]

SD have one advantage over MicroSD, you can use card from your old phone/camera in your meter ;]. Or just make connection on PCB to both SD and MicroSD socket, problem solved ;]

About ethernet, i think that using something like "Ethernet shield" for  8 bit uC is a overkill (more expensive than main CPU) but if we use some stronger CPU we could just use one with buildin Ethernet/USB (like that AT91SAM7X256 which is about $10 here in Poland). And Power over Ethernet/USB option would be nice also, one less cable on bench ;]

About power consumption,  we could add option to power off some parts of meter, like USB and Ethernet when on battery power and/or downclock CPU when on battery and not doing anything computation heavy work.

[Mastro Gippo]

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. 

[jahonen]

USB PC Connection is a MUST.
I love Dave's idea about about having a simple serial terminal. That would be a Awesome feature. If we are gonna go that route might as well add I2C and SPI Analyzers Etc...

My opinion about the Ethernet is that it kinda adds a huge amount of complexity / expensive parts for a feature that I don't know how useful it would be to most people. But like was said before we could always go for a modular type of a system where that could be a addon feature or something.

The advantages of ethernet over USB are:

1. No need for writing device drivers at computer end. All necessary things are already in every operating system on the world. Yes, you can use those RS-232-USB converters, but they still require a driver.
2. Easy galvanic isolation (in fact ethernet standard requires transformers at both ends).
3. Is not operating system dependent.
4. Long distances (>5m) easily achieved, communicates easily to other side of the globe if required

That serial terminal is also easily implemented over TCP/IP using telnet. Ethernet is replacing the venerable GPIB slowly, most modern serious measurement equipment is equipped with LXI (LAN eXtensions for Instrumentation). Some kind of compatibility with HP34401 etc command set would be also nice. So you can actually say "MEAS:VOLT:DC?" to the meter (using that telnet) and it will perform a DC voltage measurement and return the result using plain ASCII. That makes it easy to program your automated testing system

Of course, that becomes with the price tag of more complex embedded software, but 32-bit ARMs (7TDMI, Cortex-M3) etc. are very cheap nowadays. Actually so cheap that one should think very carefully before using 8-bit ones, from the price standpoint.

Regards,
Janne