EEVblog #184 – Open Hardware Multimeter ConceptPosted on July 6th, 2011 44 comments
Dave’s off-the-cuff open source hardware multimeter concept.
Inspired by the thread on the EEVblog forum: http://www.eevblog.com/forum/index.php?topic=3978.30
Don’t be too quick to dismiss USB (or other wired connection) completely. Ability to perform measurements controlled by computer comes pretty handy sometimes, and I’d rather have wired link than wireless for that purpose.
(have you actually tried to use stock bluetooth on windows? I have no idea how they have managed to make the pairing process so f***ing difficult, compared to for example bluesolail)
that’s a really interesting concept. I also concur with no wired interface existing on such a concept device as in real world using a multimeter even on the bench (not to mention the field) is already cumbersome without the PC. So please forget the USB but consider some wireless communication method (Bluetooth, XBee, etc.).
As for the range switch, I think this would be a weak point for an Open Source Hardware design with limited manufacturing resources, when even big and long established brands fail miserably sometimes. How about using a simple digital rotary encoder instead?
Sleep over the idea…
Id rather see 802.11 style gpib than Bluetooth.
Certainly, low-power wireless would be great to have!
Dave, have a look at MC1322x chips – my guess is that
it would be able to handle all of your needs here.
Well, may be coupling it with a CPLD if you need more
You can use MC1322x with Contiki OS and that will give multitasking and wireless networking.
In my opinion the best data interface for instrumentation would be Ethernet. All data signals (but ground) are already isolated, it’s darn fast and nearly realtime, it’s easily available on most not too small microcontrollers, it allows point to point and point to multipoint communications and, most important, it’s a billion times easier to program compared to USB.
In stead of building one huge, expensive do-it-all box with a display and 6 channels, build one main multimeter with just the simple Voltage/Ohms input, then add the extra channels as small, display-free boxes with one channel each that automagicly link to the main box via low-energy Bluetooth.
That way you are no longer limited to 6 channels, so if you really want 17 voltage channels, get an extra 16 voltage modules.
You can distribute your measurement modules around the work site so the probe wires can be shorter.
If a short happens and an amp channel blows up, you only have to replace one, cheap box, not the entire 6-channel-jesus-box.
If you are working on high-voltage stuff, then you are much safer as you don’t need to fiddle with the probe boxes, only with the main unit.
I’d want to use low energy Bluetooth for the wireless part, so the probe devices could be paired up with a PC or a smartphone for fancy-pants logging or simply to avoid having to invest in the big, expensive display box.
Having simple, standalone units also means that development is much cheaper as you can work on perfecting one probe box separately from all the other kinds.
The cost for separate modules isn’t necessarily any greater than an integrated solution, as a 1000V isolated DC/DC converter and optocouplers needed for an integrated channel could end up costing the same or more than the separate battery holder, box and BT module.
YES! I saw this before going to bed, and as I was sleeping on it, I came to the same conclusion. In fact, this is the kind of thing the open source community often does better than commercial enterprises. In the end, the important thing that’s standardized and open is the communications/logging protocol — maybe a thin layer on top of xbee. Once that’s done, anyone can make a small (open) device that senses another channel, for temperature, frequency, current (either through a shunt or a current clamp) etc. And anyone can make nice logging/display software. And it’s easy to put different channels on physically separated devices (in different rooms, or in different compartments of a vehicle) where running wires to a central box would be cumbersome.
It’s morphed beyond a simple multimeter into a general multi-channel wireless data logging system, but that’s OK, the world needs general purpose datalogging more than it needs another cheap multimeter.
Very Interesting, cool project.
Check this display if you do not know it alredy:
I use for my botmat projekt this displays:
super low power, super thin, nice to watch in the sun. SPI single supply 5v or 3.3v great stuff.
14:00 into the video is very relevant.
Multimeter Dongle plugged into a super cheap Chinese Android pad. Support plugging the dongle into the PC too.
We have lots of USB oscilloscopes. Why not USB multimeters too?
Why bury perfectly good (but already outdated) displays and buttons into a meter? Back then when first analog multimeters were invented, the idea was to create a box which will multiplex and reuse the costliest part of measuring device – moving coil indicator – for a number of measuring tasks. This concept survived well into DMM era but we are not obliged to stick to it any more. Don’t bundle everything together if it isn’t much cheaper to do it that way and still properly (safe and without compromising accuracy).
IMHO, “business” (measuring) end should be decoupled from “presentation and user input” end. Right now small pocket-size commercial gadgets like smart phones are all the rage and they have wireless connectivity, good GUI and API, as well as future migration path. They could be used as controlling device for single one or multiple instances of members of a family (from simple single channel to combined, like power measurement, gain/loss measurement or power source internal resistance measurement, battery capacity, etc.) of bare “meter” sensing boxes. That should take care of the problem of displays, buttons and SD card interface. Just make sure there are aids to troubleshoot connectivity between “boxes” and “gadget”. A sole LED lamp indication per box would probably suffice.
And for the users who can’t afford modern gadgets (like myself perhaps), there could be a separate nice minimal MultiDMM controller Open Hardware project.
Voice and/or Morse code output would be a killer feature for certain applications.
We once had a discussion with a laptop repairman in LJ and this was the best solution we could come up with. The difficulty was that one needs to position the probes looking into a microscope and at the same time look at the multimeter display. We went over the whole spectrum of different solutions: color flashes, ultra-small displays and mirror systems to bring readings into field of view… etc, etc. Turns out, the voice/morse announcer is the simplest thing one can do.
Id rather see USB for pen drives instead of SD.
Pen drives are largely available and are pretty much standard and everybody has one.
So you can transfer data between the meter and PCs, MACs, Workstations, etc..
Stupid thing I would like to see a meter do.
Go back to where you were adding an LED to your microphone and ask yourself what would be the fastest way to measure the I and R of that circuit?
You know the input voltage to the LED’s dropping resistor. And you can probe the output voltage that was dropped by either the resistor or the pair consisting of the resistor and the LED.
So if you can measure the voltage drop, and remember the source voltage at the same time; what could your meter calculate from that?
And if you measured the total circuit current? Close? God knows if we could only measure the heat that a small voltage drop caused. Though that may be usable in hi power applications.
Sorry I am rambling.
I think that with 4 input plug spaced like this can be very hard to stick standard probes (that have an L shape) in it
I think that the graphical display can be put on the external “visualizer modules”
Why not a 2 input Voltage Current ground multimeter whit the option to add external input /and or visualization instruments via zigbee or similar wireless protocol???
Multimeter (2 input) – Data Visualizer (via wireless) – GPIB wireless interface for pc remote control/synch with other instruments
Or with not a bench top multimeter that have a remote wireless probe input?? I have my bench multimeter and a remote “client” where I put my probes on and transmit wireless the data to be displayed
Nicola that is no problem. The high of the bottom of the multimeter case (ground plugs) can be made lower ie. 10 mm Then the the top wires ca go right over the button row wires !
Please consider an ON/OFF switch instead of the OFF position of the range switch (in case you add one).
I really hate to click/click/click/click back and forth for a quick measurement on my multimeter, I feel like I’m abusing the sucker
Excellent idea Dave. What I like is that this meter could be quite thin with no range switch. A lot of bench multimeters use relays to change ranges. One of my favorite meters was a fluke 12 which had 4 push buttons instead of a range switch.
The wireless comments inspired an idea. What if instead of multiple displays, you had multiple sending units. You could have several voltage senders several current, several temperature etc. You could even have output units like a signal generator. Then the base unit could be essentially just a big display and communications hub. That way, you could easily replace the senders if they “blew up”. You could monitor several locations at a time. Maybe have one multi function sending unit that snaps into the base unit to make it more like a normal meter.
This is a good Idea, build a telemetry net with blue tooth like the CAN bus. Then you could have modules for current, voltage, ohms, L, C, R. And the power capacity would not be limited by an expensive board. The module housings can be standardized to reduce tooling cost. http://www.blackbox.com/Store/Detail.aspx/LanTest-Pro-GB-Cable-Tester-Kit/TS4000A
Something with this form factor, could the units be solar powered, and charge when not being used? If the modules are cheap enough they could be built into say a battery box or a piece of equipment that would allow a technician to walk around with their meter display and check reading without connecting wires, or having to remove protective shields to take readings . Have the modules turn on when blasted with IR to allow them to sleep until needed.
A simple mod I am thinking of doing to my cheaper fluke.
The bargraphs on digital meters are sometimes just not quick enough when I am looking at fast outputs. I am thinking of adding a LED for low voltage and a Neon for high. These would have to be switchable in case you did not want to load the circuit.
I think you guys are getting away from the essence of a handheld DMM for use in electronics. Most of the time I use my multimeter is to quickly check a resistor value or a voltage or two and a device that requires booting and going through menus just to take a simple reading ruins this. Such a device would not replace a multimeter.
It’s absolutely true that most of these ideas are straying from a simple basic multimeter. I think with good reason, as Dave explained in his video. You can already pick up a good basic multimeter for $50.00, or a nicer one for more money. There’s no way an open hardware device will be better at the same functions for the same price. A limited resource, low volume, open project can’t compete regarding testing against CAT III safety standards, nor can it easily compete in areas of designing custom injection molded plastic case pieces. For the most common everyday needs, the common everyday commercial tool will do well, and its high volume gives it great economies of scale.
But a new open hardware project CAN potentially add some useful and perhaps extensible features not currently available on mass market meters. And that’s what much of this brainstorming is about.
1. Bluetooth (and an iPhone Client, i am willing to contribute)
2. Magnets inside the case (for data logging inside metal cases)
I work in manufacturing and I’ve lost count of the times that we’ve needed to test a small production run of connectors/cables for short circuits and stuff. We have test machines that can handle that, but they’re massive things and it means that the product has to be moved off the line to do the test – bit of a pain.
Often you just need a very simple <16 channel connection test or something like that, so I've always wanted to design a multimeter and include a 16-way IDC connection on it that I could plug cable looms into. I could design in a self-learning feature so that all you have to do is take a known good cable or board, connect it to the meter, press learn, and then you can test your short run of production stuff against the good reference quickly and easily.
I like your design, but the thing definitely needs PC connectivity (Either low power bluetooth (class 3), or perferably USB). From a software point of view it should provide a simple terminal like interface to take readings and access all commands avaiable to the meter.
One reason for this is I spent several months working on a test application using a National Instruments test rack which would test some high end electronics we designed (20K dollars per unit). Each slot in the test rack consisted of a windows PC, another slot was a relay matrix, another a programmable DMM and another was a programable power supply. The tests would run for up to 20 minutes and log all of the results to a database. Manufacturing reports would be generated and sent back to us weekly basis by our assembly house.
Having a programmable power supply(easily gotten) and a programmable DMM (as above) and providing your own PC would mitigate some of the high cost of the test rack equipment.
I’ve noticed that with your new camera your videos are interlaced. You should switch to one of the progressive modes so that you don’t have to deal with deinterlacing.
I’d like to see rechargeable batteries so I never have to open the case (well, rarely) and an isolated charger input that has a DC-DC converter built in so that I can have it on charge (or while operating it as a data logger over a long period) and not have to worry about coupling of the input power to the inputs of the meter. It would also allow me to power the thing from AC, DC, various voltages etc (maybe from min 5V so I can use USB), so I don’t have to find the right charger
Another feature that would be nice is a semiconductor test function. This could identify the type of 3 terminal device attached. Perhaps it could even determine basic characteristics such as Vgs(th) hfe, etc.
And if you’re going to have a capacitance function, adding ESR would be great.
Oh, and make it simple and intuitive to operate
I like your concept of a mltimeter with several
isolated input channels, but i agree with Flemming Frandsen: The measuring frontends
should be separated from the actual “main unit”.
I think one reason for the success of popular open source projects is a modular expandable structure.
If you want to add too many functions into a single device then it will eventually become too big and complicated to be useful. On the other
hand you will never have a multimeter that will satisfy all your needs. With a modular concept,
everyone can design customized measuring fronends, even for “exotic” applications
(such as pH probes, ion sensitive probes i.e. for measuring the nitrate content of water, probes for the measurement of ionizing radiation and many more).
You just build or buy what you really need.
I totally agrree, plus i think it would be nice if there was not only one type of “main unit”, but one full-featured handheld device and one minimalistic, “black box” which would just connect the front ends to measurement software via bluetooth, wi-fi or wired ethernet.
Regarding the ethernet option, I don’t think it’s that
much appropriate, unless it’s going to be a bench unit.
I had thought about it and it would be perfectly reasonable to use TI’s WiLink (which comes with the
OpenLink stack) especially cause it features WiFi,
BT and WPAN – http://ti.com/wilink7
So the user could really switch between whichever wireless protocol is most suitable in a situation.
I am a WPAN/WSN/6loWPAN fan, though I understand
that we are still a long way from there and no
proposals had been seen whether mobile phones and
PCs will spot WPAN anytime soon :((
Ignore complaints about negativity, we need good criticism. Hardly anyone gets it right first time, that’s how we improve things. Now, about the meter idea: A talking meter… It’s a great idea! In this age of equal opportunity workplaces, it means sight impaired technicians could measure high voltages! Seriously, I know you are brain storming the idea, but I can smell feature creep. The SD card is spot on; for most field work, you can’t have a PC nearby or leave it monitoring overnight. The data logging setup must be simple and intuitive. These days, every designer acts like we only ever use his megabox, and never anything else. The designer loves his baby, knows it inside out, but the end user has a job to do and uses other megaboxes. You get swamped with menus and buttons and the boss doesn’t give you time to sit around reading the manual. The old linux concept of having only one function per module, but do that function really well, still holds true. Don’t use wireless connections. Many techs work around large RF fields that swamp data receivers. I like the multichannel and DC power measurement, but trying to fit everything in makes it a megabox. Questions: how many digits, mains immunity, accuracy, true RMS, true AC power, power factor, closed case calibration? Cheers, Colin
So I took a little from what everyone said and came up with this:
A small compact Wifi enabled DAQ that has no screen but a host of connectors allowing for probes, thermocouples and BNC leads.
Multiple units can attach to each other and communicate over SPI/I2C. Extra units such as bigger battery packs or LCD displays can convert the DAQ into a stand alone multimeter.
As a single unit, the device will run off a li-po battery and communicate to a PC, phone or anything else with wifi, which will act as a host controller and display.
Wilink would be a great option.
Have a look anyway…
Why muck around with pushbuttons or rotary encoders at all ?
TFT touchscreen is the way to go. They cost very little these days and open a ton of opportunities.
you can have a whole GUI and exptend it on the fly. something that cannot be done with ‘buttons’.
Now, on to the serious elements of the design
The reference for the a/d convertor.
I suggest looking at an LM399 type reference , although for battery powered designs this may be problmatic due to the large consumption of the heater in that reference system.
Communication. USB is fine , but make it GALVANICALLY isolated ! analog devices have chips that can isolate a USB port.
ANd now the real key part : the A/D convertor.
please don;t make the mistake of using an off the shelf A/d Convertor chip. They all stink for real measurement applications.
You need to construct the convertor based on a dual slope of multiplope architecture.
This is the base principle used in ALL high performance meters sucha as the Keithely 2000 2001 2002 , Agilent 34401 34405 34410 34411 3478 and 3456. Fluke desktop multimeters use that principle as well.
A good low noise , lof drift opamp , a precision comparator and a few good analog switches is all you need.
Combine that with a stable reference and a good quality integration capacitor and you can easily build a 6 1/2 digit machine.
The convertor becomes tunable since you have full control over the digital portion.
I would build the guts of the system around an FPGA. This allows for on the fly reconfiguration. You can embedd a cpu core like a nios(altera) or microblaze (xilinx) or even one of the opencores processors.
Hardware accelerators can easily be built as well as filters and other digital elements.
the analog section can be confined to a shielded sections.
A programmable currentsource ( ohmmeter , diode test ) and the multislope convertor + and ad673 for true rms is all you need.
Study the scematics of the keithley and agilent machines to have an idea of what is involved in the front end.
Good quality analog switches are things like Siliconix DG444.
A mulitslope system can be made to be fully self calibrating.
just my 2 cents.
For batteries I would go with 18650 LiFePO batteries – the stuff notebook batteries are made of.
I suggest using a rechargeable lithium ion battery.
A battery life of say 7 days on a charge would be a good target.
Don’t know if you have seen this ‘USB analog electronics multi-tool’ project on kickstarter? Lot of great functionality in this, not exactly portable, but very cool for purpose, 10 days left.
I don’t think anyone has mentioned this feature. I’d like to see a Hz mode. Real time, low/high and average readings.
I’d like to see masurements and calculations in some sorte of dB.
I prefer a meter with two isolated inputs and n wireless connected like Flemming Frandsen described.
And don’t forget the iOS & Android App
cu Michael K.
1 Trackbacks / Pingbacks
[...] EEVblog #184 – Open Hardware Multimeter Concept | EEVblog … [...]
Leave a reply