Author Topic: Open Source "Lab instrument system" Idea  (Read 34250 times)

0 Members and 1 Guest are viewing this topic.

Offline FenderBender

  • Super Contributor
  • ***
  • Posts: 1115
  • Country: us
    • The Solid State Workshop
Re: Open Source "Lab instrument system" Idea
« Reply #25 on: September 13, 2012, 10:59:26 pm »
Not sure if I'm too fond of the stacking enclosures..

Think National Instruments. That might even be easier than getting actual enclosures for each. Building the module onto a "card" instead of giving it it's own box might actually be easier. How you do the backplane and interfacing is up to you. I don't know much about interfacing but I'm sure you could figure out how to do it correctly. If you see in this National Instruments...instrument, you have one big module which pretty much just looks like a PC motherboard I/O and then the rest are interfaced with some sort of busses to that board...presumbably.




Again, what is the intended audience of this product? What price would it start at? If you are thinking $1k+, then maybe I should go away, but what kind of person would be expected to buy this?
« Last Edit: September 13, 2012, 11:14:44 pm by FenderBender »
 

Offline David_AVD

  • Super Contributor
  • ***
  • Posts: 2797
  • Country: au
Re: Open Source "Lab instrument system" Idea
« Reply #26 on: September 14, 2012, 01:24:52 am »
Here's a portion of my eurocard frame and modules:
 

Offline Rerouter

  • Super Contributor
  • ***
  • Posts: 4694
  • Country: au
  • Question Everything... Except This Statement
Re: Open Source "Lab instrument system" Idea
« Reply #27 on: September 14, 2012, 02:34:30 am »
David would those be held in my the screws at the top or only friction? and overcome by tugging on the tab down the bottom? if so a similar design could probably be mimicked by using blank expansion card slots and folding back the guide tab at the bottom and removing the top securing part, to give a compact 1x wide slot design, though i am struggling to find a source of them with mounting holes provided :?

still would love to hear some feedback on what specifics should we start off this design aimin for :/
 

Offline David_AVD

  • Super Contributor
  • ***
  • Posts: 2797
  • Country: au
Re: Open Source "Lab instrument system" Idea
« Reply #28 on: September 14, 2012, 06:24:48 am »
The units pictured are standard 3U eurorack card fronts and modules.  There are a few different manufacturers.  The frame is the biggest expense.  The card fronts are relatively cheap.

Each card front / module is held in with 2 or 4 screws (depends on the width).
 

Offline BlochTopic starter

  • Supporter
  • ****
  • Posts: 453
  • Country: dk
Re: Open Source "Lab instrument system" Idea
« Reply #29 on: September 14, 2012, 03:56:21 pm »
The units pictured are standard 3U eurorack card fronts and modules.  There are a few different manufacturers.  The frame is the biggest expense.  The card fronts are relatively cheap.

Each card front / module is held in with 2 or 4 screws (depends on the width).


It it this ?

http://uk.farnell.com/schroff/24563-132/subrack-3u-235mm-84hp/dp/1370392
http://uk.farnell.com/rittal/3653040/module-3u-160mm-14hp/dp/1198884


Will a so expensive case be worth it ?
 

Offline senso

  • Frequent Contributor
  • **
  • Posts: 951
  • Country: pt
    • My AVR tutorials
Re: Open Source "Lab instrument system" Idea
« Reply #30 on: September 14, 2012, 03:56:51 pm »
Those Eurorack card fronts really seem nice, but I cant seem anyone selling them, only DJ/audio related things that go on Euroracks.  :-[
 

Offline BlochTopic starter

  • Supporter
  • ****
  • Posts: 453
  • Country: dk
Re: Open Source "Lab instrument system" Idea
« Reply #31 on: September 14, 2012, 04:01:51 pm »
Then all you need is a "caddy" to plug them into.. 


 :o http://sine.ni.com/nips/cds/view/p/lang/da/nid/202664


that is about 6000 dollars !! just for the case.....
« Last Edit: September 14, 2012, 04:11:55 pm by Bloch »
 

Offline shebu18

  • Frequent Contributor
  • **
  • Posts: 309
  • Country: ro
Re: Open Source "Lab instrument system" Idea
« Reply #32 on: September 14, 2012, 05:06:07 pm »
I like the idea with the black boxes stacked. The bottom one is the PSU and communiction and the other one are the extensions with extension pins in th bottom and the female counter part on top covered by a rubber thing for dust protection. If you want to add a module you have just to take the rubber thing out and plug the module in.


The real problem is, like someone sad, the communication between the modules. What offers the best speed vs pin count vs noise vs reliability?
RS-485 looks interesting. 32receivers and 32 senders, 10Mbit/s speed. In this instrument system you have one receiver and more senders. 12V would be maybe good for canceling noise out.


LE: Comparison of 8 serial data protocols http://www.maximintegrated.com/app-notes/index.mvp/id/3438
Enclosure good for some projects http://www.ebay.de/itm/Desktop-Instrumentation-Project-Enclosure-Box-Case-Plastic-/150870440462?pt=UK_BOI_Electrical_Components_Supplies_ET&var=&hash=item2320943a0e
board interconnectivity: http://ro.farnell.com/multicomp/2001s-20g-420-01/socket-0-5mm-5-5-5mm-20way/dp/1661180
http://ro.farnell.com/multicomp/2000p-20g-270-01/header-0-5mm-3-5-5-5mm-20way/dp/1661169


fixing could be done by magnets(i don't know if they interfere with the electronics)
« Last Edit: September 14, 2012, 08:25:36 pm by shebu18 »
 

Offline FenderBender

  • Super Contributor
  • ***
  • Posts: 1115
  • Country: us
    • The Solid State Workshop
Re: Open Source "Lab instrument system" Idea
« Reply #33 on: September 14, 2012, 08:49:04 pm »
Well perhaps you could find cheaper ones. Maybe they make something similar in ABS or something?
 

Offline David_AVD

  • Super Contributor
  • ***
  • Posts: 2797
  • Country: au
Re: Open Source "Lab instrument system" Idea
« Reply #34 on: September 14, 2012, 09:09:21 pm »
With Eurocard racks, the frame is the most expensive part, but you only need one.  The enclosed modules are also fairly expensive, but the plain card fronts are much cheaper.

The card fronts simply bolt on to the front of the circuit board.  I have designed a few products that used Eurocard frames.  Some used 160mm depth cards, some used 220mm depth cards.

You can use a backplane and DIN41612 connectors on the back of each card so they just plug in.  I have done that for most of the Eurocard framed units I've built.

The left picture below is the one that goes on the card and the right one goes on the backplane.  They are about $3 - $4 each retail, much less in quantity.

 

I don't know if people are interested in going this way, but I thought I'd explain what the system is like anyway.
 

Offline crispus

  • Regular Contributor
  • *
  • Posts: 131
  • Country: ro
Re: Open Source "Lab instrument system" Idea
« Reply #35 on: September 17, 2012, 03:03:21 pm »
I think the major obstacle in a DMM project is the analogue part.
For me, the rest are details.

First thing I thing is schematics. For schematic (especial the analogue part) you don't need case, display, communication, etc...

Dare someone to start (at least analogue part)?
I know I'm numskull, but I look around me and I feel better.
 

Offline robrenz

  • Super Contributor
  • ***
  • Posts: 3035
  • Country: us
  • Real Machinist, Wannabe EE
Re: Open Source "Lab instrument system" Idea
« Reply #36 on: September 17, 2012, 03:52:14 pm »
As in life, conception is the exciting fun part. Then the unending hard work begins. :)

Offline gxti

  • Frequent Contributor
  • **
  • Posts: 507
  • Country: us
Re: Open Source "Lab instrument system" Idea
« Reply #37 on: September 17, 2012, 05:06:32 pm »
The problem with eurocard is the PCBs all need to be pretty large, which rules out inexpensive professional PCB fabrication. But there's little point arguing about chassis or PCBs or stacking or video or anything if nobody has any real vision about what's being built, otherwise this thread is just a place for people to trot out their little ideas that, if all implemented, would result in a hulking, bloated monstrosity. Even something as simple as a lab power supply can easily spiral out of control unless you set out at the beginning with an idea and stick to it. If you want to build it, build it! Otherwise nothing will happen...
 

Offline ptricks

  • Frequent Contributor
  • **
  • Posts: 671
  • Country: us
Re: Open Source "Lab instrument system" Idea
« Reply #38 on: September 25, 2012, 09:08:41 pm »
I am doing a project now that will comprise a logic analyzer with protocol analaysis, DSO oscilloscope, multimeter, transistor hfe , capacitor, inductance, frequency meter, and function generator, that will connect directly to a monitor and has a keypad and rotary encoders for an interface.

The board I am using is by embeddedarm.com. It has an ARM9 processor , MAXII CPLD, and a CycloneII FPGA, ethernet interfaces, and a vga port driven off of part of the FPGA, the vga has its own dedicated memory. The cost was right and I am not looking to do anything high end, more hobbyist type specs. The board is designed to very high specs for temperature and stress and I have really been impressed with the support the company offers. They provide full open linux source to everything from bootloading from an SD card to the VGA code for the display.  It also has a pc104 interface connector as well as pinouts for lcd, spi, and digital IO and it runs off 5VDC @ 1A.  The company even includes examples on how to interface 1 wire devices, SPI and I2c from within linux.

I got the board used for about $75

http://www.embeddedarm.com/products/board-detail.php?product=TS-7300



My goal is to do most of the work inside the cpld and fpga and have as little support parts as possible.  The cool thing about this board is they have developed a driver for the embedded linux run by the ARM cpu that allows me to push a new design to the fpga on the fly without external tools. So I don't have to have everything fit in the FPGA , I can load a specific module as it is needed, allowing the use of the entire FPGA for each function if necessary.

If I had the funds I would go with one of their newer boards , complete with SATA ports and a ton of ram for an embedded board with open design.
http://www.embeddedarm.com/products/board-detail.php?product=TS-7800
« Last Edit: September 25, 2012, 09:20:34 pm by ptricks »
 

Offline Praxis

  • Contributor
  • Posts: 22
Re: Open Source "Lab instrument system" Idea
« Reply #39 on: January 11, 2013, 12:17:59 am »
I bought tone of the pis. After looking over the docs I can't really say it's all that.

The GPIO is only a partial bit of what's available form the chip, and is not well documented. They don't even tell you how fast the lines can change state with any certainty. The network is just a usb dongle; the "camera" interface (that DOES allow high speed input) is proprietary, you're not going to be able to develop for it without ndas and binary blobs.

Repeat after me ... the Pi is not a microcontroller, it's a linux box.  That said, it's a very cheap linux box that does HDMI.  The place of a Pi in something like this is as a user interface, simply handling the display (and perhaps user input) of data, in addition to doing data analysis and logging.  It just sucks info from the MCU and processes it and displays it.  Also, it doesn't have to be shit-hot, it just has to be good enough for the job.

Quote
The pi should not be the basis of ANY "open hardware" system - it can't, because it's not open hardware. You'd be way better off picking an intel motherboard as the basis for an "open" system. HD3000/4000 video is fine for any application like this, and the drivers are fairly robust. You'll get way more computing power, low price, and a well documented and free software development platform.

Sometimes I wonder if the Open Hardware movement is shooting itself in the foot by not being able to compromise -- you propose a solution that's going to wind up as well more than twice the price just as a premium for ideological purity, when it's frankly the least important part of the system.

Sure, parts of the RasPi aren't open.  On the other hand, no one's talking about using the NDA'd parts of the RasPi hardware or software for this.  Anything written for the RasPi that relies as minimally as possible on hardware-specific bits should be readily portable to any other similar mini linux computer.  (Sure, the low-level RasPi graphics stuff is NDA'd.  Who cares?  Unless you're a driver designer or demoscene freak you never come that close to the video hardware for it to matter.)  Perhaps considering the display (as partially closed hardware) as being seperate from the acquisition hardware would satisfy purists?  Then it's really an open hardware (the acquisition bits) and open software (the display and interpretation bits) project. 
 

Online Marco

  • Super Contributor
  • ***
  • Posts: 6693
  • Country: nl
Re: Open Source "Lab instrument system" Idea
« Reply #40 on: January 12, 2013, 08:26:28 pm »
Why not simply use the arduino shield method of interconnecting boards? Pin headers are only rated for 100 insertion cycles, but meh. I was actually looking if there were any cheap LCR shields and was surprised there is very little like this.

If I had the perseverance to tackle a system like this I'd start with the Freescale KL25 board (the only one of the ~10$ uC boards with a 16 bit ADC AFAIK) add a bluetooth module and make an android app for UI. Design a standard sized enclosure for 3D printing with gaps for the pin headers which can be customized for whatever connectors the shield needs and then start with a single shield for an auto-ranging V/A/LCR meter (is going to be pretty cheap,since the ADC/DAC are on the Freedom board the shield only needs attenuation, buffering and protection).

Assuming the circuit board for the shield is sold without too much markup,  cheap access to a 3D printer and ownership of an android phone/tablet with bluetooth this could get someone a useful and expandable measurement platform for 30$ in parts with assembly required.

That's the only way you are going to get some mindshare ... low cost.
 

Offline sub

  • Regular Contributor
  • *
  • Posts: 107
  • Country: au
Re: Open Source "Lab instrument system" Idea
« Reply #41 on: January 29, 2013, 10:31:45 am »
In my mind, low cost and ease of development can be achieved if one is willing to choose sufficiently simple connectors and spread their circuitry over as many instruments as possible.  The Arduino would make a good starting point, as it provides the USB to serial interface as well as some power.  The presence of ADC pins would mean that there is a digitiser available, even if it is nothing special.

I have started working on a SCPI parser, so that we can make existing software that can talk to the instruments easily.  The standard has the facility for multiple virtual instruments, so it would make a good start for this type of system.
 

Offline sub

  • Regular Contributor
  • *
  • Posts: 107
  • Country: au
Re: Open Source "Lab instrument system" Idea
« Reply #42 on: February 07, 2013, 12:09:08 pm »
I have built a prototype of the SCPI parser, which I have placed online at https://github.com/LachlanGunn/oic (MIT licence).

It is a bit of a mess, using great amounts of dynamically allocated memory and leaking it like a sieve.  However, if nothing else it gives me something to work with as I build something for an embedded platform (Arduino to start with).  Once this is done I will provide a multimeter-like interface to the ADC (just as a demonstration), and an interface to control an AD9835.

The analogue bandwidth of the micro's ADC isn't high enough to make this into a dodgy VNA, so if I decide to go down that path I will probably start by building a digitiser.  This might be a good step towards making something useful.

After making things work satisfactorily on the Arduino, I would like to get it working on the PIC32 via Microchip's USB-to-serial library.  This would bring us a bit closer to being able to plonk down a chip or a module on the board or in a chassis that would handle all of the PC interface requirements.  Eventually I would like to use USBTMC instead of resorting to USB-to-serial, but that will take some time to develop.

For those interested, I have placed my plans for development into README.md in the repository listed above.
 

Offline MysteryBunny

  • Contributor
  • Posts: 11
  • Country: ca
Re: Open Source "Lab instrument system" Idea
« Reply #43 on: March 01, 2013, 01:24:50 pm »
If you want to make an OSHW oscilloscope you should consider these boards... They are more expensive than a $20 one because of the PCB itself... ( I'm working on a 386 processor in Verilog/VHDL, and then a cross-assembler FPGA to convert x86 to ARM... So I wouldn't even start this project yet, so this is an interesting idea I thought of for the future)...

This is MUCH MUCH faster than the raspberry pi which the clock speed of the pi is very deceiving... This is also COMPLETELY OPEN SOURCE: http://beagleboard.org/bone

The BeagleBone uses a TI AM3358 ARM Cortex-A8-based microprocessor. Announced on Oct 31, 2011, the main processor is available for as little as $5, uses a 0.8mm ball-grid array and standard DDR2 memory, making this board easier to clone than other BeagleBoard designs.

I also have some M/F jumper wire going to my DE-0 so I have programmable logic as well. 3.3v compatible and cheap...

http://www.altera.com/education/univ/materials/boards/de0-nano/unv-de0-nano-board.html

This is like the Xilinx Spartan-6 series but cheaper... This particular model is 22k LUT-4 I believe and the FPGA itself is around $10-20 ( maybe less now ) ... That's quite a lot of gates. I never started this project in 2004 because back then you only had maybe 6-10k LUT ( For reference a lookup table is sram cells I believe + 4-6 MUX or multiplexers... You can emulate basically anything, the hard-IP block DSP is also fairly good )...

I think if you were to make an OSHW oscilloscope you could get it into the $50-100 range in total with a custom PCB. The real work is just making the ADC front end since you might need to blow around $20-50 of prototype money selecting one with the right sample rate while keeping the cost low... That sounds like a joke but you could easily do it to those who are critical... If you have the I/O with serdes/transceivers in PLL arrays I'm sure a 400-700mhz A8 is enough ( not to mention it blows the doors off of the pi in processing power and I/O )... Once again this is only prototype cost to DESIGN an oscilloscope kit ( with another kit )... You're easily going to have to make a separate board since you don't know what voltage/current scenario you're going to get out of your front end even buying these two things... I think it's an interesting idea though... Then you can just feed it back to either the beaglebone or FPGA board... Analog devices also makes a $150 board that has a spartan 6 + blackshark on it @ 600mhz+ ...Which those are both budget lines that can be scaled up or down depending on needs. All you'd need as the hobbyist to solder BGA would be a cheap IR... OR even just the ATTEN hot air, reballing stencil, gel, etc... That sounds extremely fussy and risky, but I think if people had the technique down ( I've seen a lot of attempts to teach it on this site so I think it's the best place to talk about it) it wouldn't be that risky to try at least since the investment wouldn't be that high ( once the community has a premade PCB and the parts are standard for the kit )... The only real investment is getting together a kit based on these parts. But we'd have to prototype the system first, and THEN build an OSHW board specifically for test gear...

All the processing power is easy to get and an LCD display isn't that surprising to add... Even an 800x600 display is laughable for a cyclone IV FPGA since it can handle an overlay too... I'm not sure about the cost of that though... It would be fine even with monochrome I think. But once again it could be MODULAR and hardware-agnostic with risers... MUCH MUCH better than the weak amtel in the arduino, and it's a huge improvement over the pi... But the raspberry pi is also a $35 kit and it's ultra low cost. Beaglebone is around $90 but the processor itself is only $5... These particular FPGA and microcontroller are perfect I think for a test piece. I saw examples on the texas instrument site showing how to design your own washing machine ( commercially ) and everything with block diagrams... Which basically is a cry for help saying it's so cheap and fast it can be, and should be tried in anything almost... Xilinx has low-cost platforms, but they start out at $150 usually through digilent design, or they have no features like easy-access GPIO... I have also looked at analog devices blackshark boards ( which surprisingly is only $50 or so for one, but they only work with the BEmicro board, which also doesn't have the 40+ GPIO pins like the DE-0 by terasic )...

Anyways, I hope that offers some insight...

This is a rather tricky thing to find on google: http://www.analog.com/en/evaluation/sdp-bemicro/eb.html

The SDP platform starts at $50 and goes up to around $175 I think? But this isn't easy to find on the analog devices website... Look at the H1 series...
 http://www.analog.com/en/system-demonstration-platform/evaluation/index.html

The only problem with the $50 one is getting the connector to mate with your board. The cost of the nicer ones is for I/O, bigger PCB, and then the final one you're paying for the FPGA too... http://www.analog.com/en/system-demonstration-platform/controller-boards/evaluation/EVAL-SDP-H1/eb.html ...They are almost a waste though compared to the beaglebone I think...
 

Offline Hobgoblin612

  • Contributor
  • Posts: 47
  • Country: au
    • Youtube Channel
Re: Open Source "Lab instrument system" Idea
« Reply #44 on: March 06, 2013, 08:32:03 am »
I think a project like this needs to start with a standard bus and case design and work from there. One of the hardest things is going to be the user interface. An old 17" LCD monitor would work great as the screen. But the input is a lot harder. For some things like logic analysers a keyboard and mouse work fine but for a scope or function gen you need buttons and knobs. It is not an easy project because different test gear has different requirements for case size, power, interface, bandwidth to the main controller and so on. Any case system chosen or designed would need to have multiple compatible sizes and probably a metal or shielded version. Also keep in mind that if you incorporate a DMM or logic analyser it won't be isolated (unless you isolate all the data and power). In my experience open source projects generally need to be tightly managed or the user interface gets out of control.

Just my 2c worth.

Timothy.
ooooohh... what does this button do???
 

Offline sub

  • Regular Contributor
  • *
  • Posts: 107
  • Country: au
Re: Open Source "Lab instrument system" Idea
« Reply #45 on: March 06, 2013, 10:57:57 am »
My inclination was to use one of those sub-$100 tablets for the display, since you get buttons there too.  It is potentially a bit irritating (I use occasionally an Agilent 9000-series with a touchscreen, and hitting the right button can be difficult), but probably the most space-efficient way to get a useful interface.  This would mean writing an Android USB driver if one were to use USBTMC, but you can't have everything.
 

Offline Hobgoblin612

  • Contributor
  • Posts: 47
  • Country: au
    • Youtube Channel
Re: Open Source "Lab instrument system" Idea
« Reply #46 on: March 06, 2013, 01:30:02 pm »
Im not so sure about the tablet idea.
  • The bandwidth into the tablet is a problem since most of them just have a USB interface
  • I'm not sure how much use the touch-screen is in this application because mostly you need precise pointing like a mouse or actual physical knobs
  • the rest of the instrument is not going to be really portable so why does the display need to be
  • you can pick up a 17in 1024x768 LCD on eBay for about $25 if you can't find one just lying around
ooooohh... what does this button do???
 

Offline sub

  • Regular Contributor
  • *
  • Posts: 107
  • Country: au
Re: Open Source "Lab instrument system" Idea
« Reply #47 on: March 07, 2013, 07:27:37 am »
While the device may not be portable it still takes up bench space, moreso because you now need your chassis to accommodate front panels for every instrument in addition to a 17in display.  For some applications USB bandwidth could be a limiting factor, but it seems that Android supports USB 2.0, so it might still be fine---I'll try it on my mobile when I eventually get some FTDI chips.

I agree that one would probably need to go down the route you suggest for graphically intensive things, but for a signal generator, function generator, power supply, etc. the requirements are not that high.

 

Offline Rerouter

  • Super Contributor
  • ***
  • Posts: 4694
  • Country: au
  • Question Everything... Except This Statement
Re: Open Source "Lab instrument system" Idea
« Reply #48 on: March 07, 2013, 08:23:07 am »
looking back to some of the earlier discussion on possible housing and mounting options, as a rough guideline, why not make use of what already accommodates for most of our connectors, even if not a professional case in the early implementations, i'm referring to PC expansion slot covers, its a simple layout for the blanks with only 1 bend, 2 filed or cut grooves at the top and 2 slanted edges at the bottom, just manipulating the direction of usage slightly,

flip it around so it curves back in towards your pcb, have either a screwed on flap at the bottom pin the grooved part or some clip arrangement, and have either another screw or a pin arrangement molded out of a single piece of plastic and pushed up with a spring for a lock in for the top hole (trying to cater for both people with tools and people without)

most of the modules, such as a volt meter, ammeter 4 wire ohm meter etc, could easily fit on a single blank, even very high density logic analyzer cards could be made with plenty of space left for additional processing like decode on them if required, and the required front piece can be stolen out of old pc cases, (always plenty around)

that leaves us with what card length, and is not an easy question, on one hand you have very cheap 5x5cm boards, but on the other it should probably cater to larger higher complexity boards, say a 5x10 or 10x10, personally i would say the 100mm long should be the standard, as double or triple wide blanks would not be that hard to make, though fitting into the casing may take better supporting the pc mounting side, and removing the middle strip supports :/

thoughts?

 

jucole

  • Guest
Re: Open Source "Lab instrument system" Idea
« Reply #49 on: March 07, 2013, 10:28:01 am »
I think a project like this needs to start with a standard bus and case design and work from there. One of the hardest things is going to be the user interface. An old 17" LCD monitor would work great as the screen. But the input is a lot harder. For some things like logic analysers a keyboard and mouse work fine but for a scope or function gen you need buttons and knobs...

.. In my experience open source projects generally need to be tightly managed or the user interface gets out of control.

Some good points in there!
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf