Author Topic: Open Source Multimeter  (Read 237774 times)

0 Members and 1 Guest are viewing this topic.

Offline ejeffrey

  • Super Contributor
  • ***
  • Posts: 2485
  • Country: us
Re: Open Source Multimeter
« Reply #125 on: July 11, 2011, 02:48:11 pm »
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?

I am not sure.  On the one hand, PXI is indeed really nice, and could be useful for more advanced test gear.  However, it isn't exactly very hacker friendly.  Even many advanced hobbyists are not likely to have the equipment to develop and test PCI devices.  It also seems like a low speed device like a DMM doesn't gain a lot here.  I would prefer a low speed serial bus (SPI or CAN) for communication between the modules, and then USB and ethernet on the controller /display module for interface to the outside world.  You loose the ability to directly plug into a PXI chassis and integrate with other PXI test gear, but that stuff is usually pretty expensive and out of reach of most hobbyists.
 

HLA-27b

  • Guest
Re: Open Source Multimeter
« Reply #126 on: July 11, 2011, 04:17:16 pm »
...However, it isn't exactly very hacker friendly... 

On the contrary, PXI lets you leave out features that you don't use. As a bare minimum you have to implement J1 socket.
J1  is for power, 33MHz clock and 32 bit data, 10MHz precision clock and triggering.

You can omit any of these unless you need 'em.
If you don't need precision clock and triggering you simply don't connect to the pins.
Likewise if you don't need 32 bit data you can limit yourself to 8,16,24 bits.
Other boards connected to the same back-plane are not affected.
(you probably knew all these but I didn't until yesterday) 

This only leaves out the backplane, which happens to be passive,
 and the processing board which I imagine as something akin to Beagle board

So while not exactly piece of cake it is not hacker unfriendly either.



 

Offline Mechatrommer

  • Super Contributor
  • ***
  • Posts: 10307
  • Country: my
  • reassessing directives...
Re: Open Source Multimeter
« Reply #127 on: July 11, 2011, 06:24:33 pm »
that thing looks like oscilloscope.
It's extremely difficult to start life.. one features of nature.. physical laws are mathematical theory of great beauty... You may wonder Why? our knowledge shows that nature is so constructed. We simply have to accept it. One could describe the situation by saying that... (Paul Dirac)
 

Offline ejeffrey

  • Super Contributor
  • ***
  • Posts: 2485
  • Country: us
Re: Open Source Multimeter
« Reply #128 on: July 11, 2011, 06:40:50 pm »
...However, it isn't exactly very hacker friendly... 

On the contrary, PXI lets you leave out features that you don't use. As a bare minimum you have to implement J1 socket.
J1  is for power, 33MHz clock and 32 bit data, 10MHz precision clock and triggering.

You can omit any of these unless you need 'em.

Interesting.  I haven't looked much at PXI from the hardware side, and it has been a few years since I had to deal with PCI.  The precision timebase and trigger bus aren't what worry me, it is the PCI.  I am under the impression that you still have to implement a complete PCI interface -- arbitration, enumeration, and so forth.  That means an FPGA, a dedicated bridge chip, or a processor with PCI built in.  None of those is out of the realm of a hobbyist to implement, but it can be tricky to debug without a logic analyzer.
 

Offline Bloch

  • Supporter
  • ****
  • Posts: 444
  • Country: dk
Re: Open Source Multimeter
« Reply #129 on: July 11, 2011, 06:51:29 pm »
Hello there, first post on the forum...

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

So true.

But the PIX thing is not for me. I think there are a lot of work in that. But if I can buy one finished then no problem.

And it seems for me to be an "old" standard. Data in parallel is not the "new" way to do it. Think on IDE / SATA or Parallel port / USB.

At work i uses a lot of time on PLC's. And there is the "new" standard EtherCat. Buzzwords free and open / fast / True Time stamps / and just works on  Ethernet.

But if i get a vote  :) then i think on a RS232/USB >  KISS = keep it simple and straightforward
 

Offline JimHorn

  • Contributor
  • Posts: 6
  • Country: us
Re: Open Source Multimeter
« Reply #130 on: July 13, 2011, 06:15:17 am »
One DMM feature I've always wanted has been standard on oscilloscopes for decades: 1, 2, 5 range switching.  With modern digital calibration and LSI, it should be simple to provide.  And it would mean that a small increase in a measured value wouldn't suddenly cost you 90% of the instruments resolution.

Yes, it means many measurements would have only even or just 0 or 5 as the last digit but the microprocessor running the show could easily show the error band or resolution if asked (another button function?).  Most manuals include the instrument accuracy and resolution specs but who remembers those in use?  The instrument should...
So many signals, so little bandwidth!
 

Offline ejeffrey

  • Super Contributor
  • ***
  • Posts: 2485
  • Country: us
Re: Open Source Multimeter
« Reply #131 on: July 13, 2011, 07:27:34 am »
I would always rather just have more counts.  Unnecessary range switches are too annoying, and the resolution and dynamic range of the modern sigma-delta converters is such that the number of displayed counts is set by the display size and the component accuracy, not the ADC.
 

Offline hisense999

  • Contributor
  • Posts: 39
Re: Open Source Multimeter
« Reply #132 on: July 14, 2011, 04:06:29 am »
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.

It's a wrong point of view, Altium is not a fresh computer game which selling in milions of copies monthly but dedicated software - trust me even if this cost £10 ppl still will preffer to use pirate copies. Dedicated software must be expensive and base on selling it into companies, Altium don't care about you and your £100 Altium care about big companies selling them hundreds of seats Altium also don't care if you pirate it even this is good cuz then is possibility to make you stick with Altium software and one day maybe you open a company and must to buy original one, or in further work you push your boss to get Altium for PCB design.

Anyway as I also selling software I know is better to sell 1000 copies than 10000 is same profit but more small amount of customers to handle.

Just last time my company maked an experiment and we selling software really cheap, belive me we never make this mistake again amount of customers to handle was groving up day by day but amount of income money was a nothing compared to previous products, was more headache than profit on this. And at final you can discover you reach a peak point where sales only will go down cuz you cover with your dedicated software ~40% of possible end customers and then profit is decrasing month by month and wishes of existing customers are incrase whole team start to be tired of tips going from sales and huge amount of work needed to make first existing customers happy and then improvements to catch new customers.

And remember dedicated software is not an game you cannot leave it with bugs and without updates or design a "Your Software 2" cuz dedicated group of customers have already this which make them happy and for prepare improvements and something which can be push as new product for which you can charge a new money is not only to redesign levels in game and a little improve graphics but built dozens of complicated things from beginning.

P.S.
Sorry for my poor english :)

B.R.
 

Offline hisense999

  • Contributor
  • Posts: 39
Re: Open Source Multimeter
« Reply #133 on: July 15, 2011, 03:53:41 pm »
About whole project, is it not better with Open Source Multimeter to just make a real fun hack - just to take a most cheap multimeters which are flooding a market and sometimes cost 2$ and redesign HW inside (maybe even to add own MCU) giving for it super accurate results and new features... I know only 4 digits and no special connectors but if this must be a real open hardware multimeter then this must be something which other ppl are able to make in home without ordering special housing, 2pcs of LCD, custom PCB (the best soldered already cuz then for sure will match specs), so why is not better to pick up most cheap multimeters on the market and check which can be most easy to hack. I speak about perfect hack - additional small board with own MCU inside and only modify existing PCB for replace some crappy components with quaility ones adding some wires and then is a big chance for make open source multimeter for ~20USD which anyone can rebuild in home.

B.R.
 

Offline FreeThinker

  • Frequent Contributor
  • **
  • Posts: 791
  • Country: england
  • Truth through Thought
Re: Open Source Multimeter
« Reply #134 on: July 15, 2011, 06:34:54 pm »
About whole project, is it not better with Open Source Multimeter to just make a real fun hack - just to take a most cheap multimeters which are flooding a market and sometimes cost 2$ and redesign HW inside (maybe even to add own MCU) giving for it super accurate results and new features... I know only 4 digits and no special connectors but if this must be a real open hardware multimeter then this must be something which other ppl are able to make in home without ordering special housing, 2pcs of LCD, custom PCB (the best soldered already cuz then for sure will match specs), so why is not better to pick up most cheap multimeters on the market and check which can be most easy to hack. I speak about perfect hack - additional small board with own MCU inside and only modify existing PCB for replace some crappy components with quaility ones adding some wires and then is a big chance for make open source multimeter for ~20USD which anyone can rebuild in home.

B.R.
Problem is that there is very little inside to hack & the build quality is crap! Non starter sorry :-\
Machines were mice and Men were lions once upon a time, but now that it's the opposite it's twice upon a time.
MOONDOG
 

HLA-27b

  • Guest
Re: Open Source Multimeter
« Reply #135 on: July 15, 2011, 07:10:07 pm »
I think we should start to spec out this thing already.

So here I ask the question:

What should the analog section look like? What should be the specs?
 

Offline rfdave#gmail.com

  • Contributor
  • Posts: 22
  • Country: us
Re: Open Source Multimeter
« Reply #136 on: July 16, 2011, 02:49:30 am »
How about starting out with a simple USB/1 channel dmm box. Once that's done and working, it's a fairly simple module to plug into other systems.

Dave
 

Offline Zero999

  • Super Contributor
  • ***
  • Posts: 15767
  • Country: gb
  • 0999
Re: Open Source Multimeter
« Reply #137 on: July 16, 2011, 01:35:53 pm »
It's a wrong point of view, Altium is not a fresh computer game which selling in milions of copies monthly but dedicated software - trust me even if this cost £10 ppl still will preffer to use pirate copies. Dedicated software must be expensive and base on selling it into companies,
Yes, may be I took it to the extreme but you know what I mean.

Quote
Altium don't care about you and your £100 Altium care about big companies selling them hundreds of seats Altium also don't care if you pirate it even this is good cuz then is possibility to make you stick with Altium software and one day maybe you open a company and must to buy original one, or in further work you push your boss to get Altium for PCB design.
You're certainly right there - I've said that before.

Quote
Anyway as I also selling software I know is better to sell 1000 copies than 10000 is same profit but more small amount of customers to handle.

Just last time my company maked an experiment and we selling software really cheap, belive me we never make this mistake again amount of customers to handle was groving up day by day but amount of income money was a nothing compared to previous products, was more headache than profit on this. And at final you can discover you reach a peak point where sales only will go down cuz you cover with your dedicated software ~40% of possible end customers and then profit is decrasing month by month and wishes of existing customers are incrase whole team start to be tired of tips going from sales and huge amount of work needed to make first existing customers happy and then improvements to catch new customers.
Interesting, so you don't want too many customers because it increases support costs.

So why not just reduce the price of the licence and charge more for support?

Surely that sounds like the best business model? Sell relatively inexpensive licences to hobbyists with terms such as not for commercial use and no support and make the real money from big  companies.

Offering small companies special deals would also be a good idea.

Quote
And remember dedicated software is not an game you cannot leave it with bugs and without updates or design a "Your Software 2" cuz dedicated group of customers have already this which make them happy and for prepare improvements and something which can be push as new product for which you can charge a new money is not only to redesign levels in game and a little improve graphics but built dozens of complicated things from beginning.
I don't know if I agree with that, I use AutoCAD electrical at work and it's as buggy as Windows 95 was at its initial release and crashes just as often.
« Last Edit: July 16, 2011, 05:23:23 pm by Hero999 »
 

Offline Joshua

  • Regular Contributor
  • *
  • Posts: 194
  • Country: us
Open Source Multimeter
« Reply #138 on: July 16, 2011, 02:08:21 pm »


So why not just reduce the price of the licence and charge more for support?

Surely that sounds like the best business model? Sell relatively inexpensive licences to hobbyists with terms such as not for commercial use and no support and make the real money from big  companies.


I don't know if there's anything I despise more than when a company won't answer a simple phone call question unless you pay for that support. I.e. Microsoft, apple, HP...
 

Offline hisense999

  • Contributor
  • Posts: 39
Re: Open Source Multimeter
« Reply #139 on: July 16, 2011, 02:47:19 pm »
So why not just reduce the price of the licence and charge more for support?

Surely that sounds like the best business model? Sell relatively inexpensive licences to hobbyists with terms such as not for commercial use and no support and make the real money from big  companies.

Offering small companies special deals would also be a good idea.


Is more easy to decrase than incrase prices especially when all is based on distributors network which bought stock already and software is not license based but fisical security dongle based, everything always have more deep story inside even when looks very simple.

And with Altium and non-commercial use this is very good idea especially when this also include limitation to two layers boards then this can be perfect deal for hobbysts and good protection for Altium.

B.R.

P.S.
About OpenSource multimeter the most important is to choose enclosure\housing also exist another OpenSource possibility - 3D project which everyone with MakerBot, RepRap or UP! 3D plastic printer can be able to print, just now many ppl have 3D printers and price of it is decrased every year, so community with 3D printers can be supplier for it :)
« Last Edit: July 16, 2011, 02:54:11 pm by hisense999 »
 

Offline bobski

  • Contributor
  • Posts: 36
  • Country: us
Re: Open Source Multimeter
« Reply #140 on: July 16, 2011, 08:47:50 pm »
Here's my own take on the subject. Feel free to pick it to pieces.

As this thread demonstrates, everybody's going to have their own requirements for an ideal multimeter. As such, the thing should be made modular so the end user can swap out parts to fit their needs, rather than trying to create a one size fits all meter as everybody and their brother has done.
In that case, the base meter would be a rack-like case with some basic infrastructure that all designs would need: Physical structure for the meter, battery holder(s), a processor, slots for modules (consisting of mechanical supports and a standard set of electrical connections) and such. The modules would be in two or three from factors. A small, simple input module that could handle basic functions like voltage, current, resistance, etc. These modules would ideally interlock horizontally so 3 or 4 of them could be lined up rack style (no dead space between modules) at the bottom of the meter, resulting in the multi-input design dave is after. Measurement circuitry requiring more space could be built in double or triple-wide modules.
Display modules would cover the full width of the meter and as I said above and possibly be designed around a bus so they could be stacked as necessary.
Controls could be integrated into these display modules, exist as a dedicated display module, or perhaps a third control panel type module is in order.
Point being, people can easily assemble a meter to meet their own needs and leave out the extra stuff.
This would turn the open-meter into a set of standards… The mechanical specs of the modules, communications between each module and the processor, current demands and such. Once you've got that, anyone can create their own modules to expand the meter's abilities. To facilitate that, blank modules (just the plastic housing and maybe standard connectors) could be sold, as well as populated V/I/R/C, LCD and control modules.
Since the modules and services would be standardized, alternative form factors for the meter "rack" could be created as well. Rather than the standard meter housing with lots of module slots, a compact probe-style meter could be made with one or two input slots and a display slot. In that case, the input could be either banana jacks, a point probe with common jack, or tweezers like those obnoxiously expensive LCR meters. A bench-top model would also be perfectly reasonable… Power supply module maybe? That would allow disposable batteries, rechargeable or plug-in sources.


The battery is going to be user replaceable.
Is that particular battery a standard battery that will be in production x number of years from now?
Use something that is ubiquitous like the 18650 Li-ion cells. They're cylindrical, have AA-like proportions (though with larger dimensions) and typically provide around 2300 mAh @ 3.6V. In that case, AA-style battery holder designs (snap in, cradle with cover, caped tube, whatever) would be easy to adapt and would eliminate the need for device-specific battery packs. While they don't quite make the run-out-to-a-drugstore cut, they're used as the base cell in a lot of consumer devices. My older Acer netbook uses them. Judging from the shape, dimensions and specs of the battery pack, my dust buster uses them, and my cordless drill probably does as well. Point being that even if they become hard to find new, it should still be fairly easy to scrounge them out of old battery packs. Using individual removable cells means one can carry around a few extra if spending a lot of time in the field.
« Last Edit: July 16, 2011, 08:56:16 pm by bobski »
 

Offline rfdave#gmail.com

  • Contributor
  • Posts: 22
  • Country: us
Re: Open Source Multimeter
« Reply #141 on: July 16, 2011, 10:15:20 pm »
A proposed Starting spec

  • usb interface
  • powered from USB
  • 4 1/2 digit resolution
  • volts/amps/ohms scales
  • autoscaling
  • no buttons/controls on case, just banana plugs as input
  • 3 plugs-Common, Volts and Amps
  • Continuity test, fast response time
  • bench focused, so input voltage levels limited to 50v

This could serve as a module in some of the fancier setups, but without this basic functionality, you've got nothing.
 

HLA-27b

  • Guest
Re: Open Source Multimeter
« Reply #142 on: July 16, 2011, 11:41:36 pm »
Here's my own take on the subject. Feel free to pick it to pieces.
....
In that case, the base meter would be a rack-like case with some basic infrastructure that all designs would need:

I agree and I don't agree. Racks are my preferred form factor as well. I wrote about PXI previously and maybe it can be made open source. Not the PXI itsel of course but a board designed to PXI may as well be open source.

What I don't agree is that maybe we are trying to tie this thing to a particular shape too early. It is not that voltmeters need racks (but racks need voltmeters...and other things).

This would turn the open-meter into a set of standards…


 IEEE - 1101.1 is one standard you would like to stick to. It governs the mechanical aspects of rack equipment. On the other hand what governs backplanes is not standards but industry specifications like PCI PCIe CompactPCI PXI etc. PXI is not a standard yet. At the moment it is a specification maintained by a consortium of companies. It may become standard in the future. Standards and specifications are quite an entangled mess. Maybe we should keep them beside the point for the time being. For all I know The Open Source Multimeter consists of a few guys trying to decide what ADC to use.
 

HLA-27b

  • Guest
Re: Open Source Multimeter
« Reply #143 on: July 17, 2011, 12:38:42 am »
A proposed Starting spec

  • usb interface       Check
  • powered from USB        Likely an option
  • 4 1/2 digit resolution        Make that 6          https://www.eevblog.com/forum/index.php?topic=4087.msg55214;topicseen#msg55214
  • volts/amps/ohms scales   Check plus maybe Sourcing and LCR
  • autoscaling     Check
  • no buttons/controls on case, just banana plugs as input
  • 3 plugs-Common, Volts and Amps  4 plugs would enable Kelvin 4 wire measurements
  • Continuity test, fast response time   Check
  • bench focused, so input voltage levels limited to 50v     That seems dangerous
[/li]
[/list]

This could serve as a module in some of the fancier setups, but without this basic functionality, you've got nothing. Check
 

Offline bobski

  • Contributor
  • Posts: 36
  • Country: us
Re: Open Source Multimeter
« Reply #144 on: July 17, 2011, 01:47:30 am »
IEEE - 1101.1 is one standard you would like to stick to. It governs the mechanical aspects of rack equipment.
I see your confusion. I'm saying the meter would be like a rack in the sense that a rack provides multiple slots that can be occupied by a variety of equipment. I suppose what I'm thinking of has more in common with PLCs. For those not familiar with them, PLCs are computers used for industrial automation. The hardware consists of a rack with built in power supply and a series of slots for cards. The slots are all interconnected by a backplane board. The first card is a CPU card with processor, memory and basic run/stop controls. The remaining slots can be used for whatever is needed - input cards (analog or digital), output cards (relay-based digital, or analog), or hybrids of the two. If an assembly line gets re-tooled and different I/O connections are needed, you pop out one card, plug something else in it's place, update the software to reflect the change and away you go.
The form factor I'm thinking of is handheld multimeter-sized, though there's no reason you couldn't make a bigger or smaller "rack" that makes use of the same plug-in modules.
 

alm

  • Guest
Re: Open Source Multimeter
« Reply #145 on: July 17, 2011, 02:03:11 am »
USB power would require at least one isolated DC-DC converter, more if you want multiple channels. Probably not worth the complexity IMO.
 

Offline Zero999

  • Super Contributor
  • ***
  • Posts: 15767
  • Country: gb
  • 0999
Re: Open Source Multimeter
« Reply #146 on: July 17, 2011, 07:28:40 pm »
USB power would require at least one isolated DC-DC converter, more if you want multiple channels. Probably not worth the complexity IMO.
Good point.

If you design the power supply it would need UL approval and buying one with a with a high enough isolation voltage for mains won't be cheap.

I was going to suggest the one in RS linked below but it's only rated for 3kV which won't be good enough for mains, although it'll be fine for SELV i.e. automotive or telecoms use.

http://uk.rs-online.com/web/p/dc-dc-converters/4943838/
http://docs-europe.electrocomponents.com/webdocs/0518/0900766b80518157.pdf
 

HLA-27b

  • Guest
Re: Open Source Multimeter
« Reply #147 on: July 18, 2011, 02:15:50 am »
USB power would require at least one isolated DC-DC converter, more if you want multiple channels. Probably not worth the complexity IMO.

Here is some reading material about isolation and why it is required.

http://zone.ni.com/devzone/cda/tut/p/id/3410

For people like me who didn't know.
 

Offline Fox

  • Regular Contributor
  • *
  • Posts: 54
  • Country: de
  • Electronics Engineer
Re: Open Source Multimeter
« Reply #148 on: July 18, 2011, 10:33:48 pm »
A proposed Starting spec


This could serve as a module in some of the fancier setups, but without this basic functionality, you've got nothing. Check

Going from 4 1/2 Digits to 5 or even 6 Digits  will increase the complexity dramatic.
The ADC alone won't guarantee  the 6Digit resolution, one will need a very good analog signal conditioning circurtry.
e.g the Agilent U1242 has an 24bit ADC (ADS1242) like the LTC2415 but gives one "only" 10000 counts with its relatively "simple" input section, but the Fluke 28II with an 20Bit ADC (LTC2435) gives one 20000 counts with a better input section.
So keeping 4 1/2Digits resolution would increase the possibility of success of an open source Multimeter
A closed Switch should have zero Ohms or less!
 

HLA-27b

  • Guest
Re: Open Source Multimeter
« Reply #149 on: July 18, 2011, 10:59:47 pm »
Going from 4 1/2 Digits to 5 or even 6 Digits  will increase the complexity dramatic.
........
So keeping 4 1/2Digits resolution would increase the possibility of success of an open source Multimeter

I agree. Trying to increase resolution places high demands on every aspect of the design. Voltage reference, clock jitter, PSRR, Analog noise etc etc...

However I think we have some advantages here compared to the commercial guys. Open Source multimeter will likely be an evolutionary project. There will be many iterations and refinements. By and by it will get better. Another advantage that we have is that we don't have to worry about profit margins and product placement. Therefore we can use (hopefully justified) pricier components.

What worries me is that no designs popped out yet. It is like all of the more experienced guys prefer to just watch and keep quiet. As if they are afraid of criticism or something.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf