Page 1 of 1

Computer interfacing

Posted: Wed Aug 11, 2010 12:41 am
by Dustin
Here is a cheap way to log all that fusor data.

http://labjack.com/

(no affiliation)

Re: Computer interfacing

Posted: Wed Aug 11, 2010 12:43 am
by Tyler Christensen
Looks like it might be useful, although there are much cheaper ways to go to get the same functionality if you code it, such as Arduino. I'm nearly done making a full logging system for my reactor on an Arduino dev board, I'll post details when I'm done.

Re: Computer interfacing

Posted: Wed Aug 11, 2010 11:18 am
by Richard Hull
I have used pico computer instruments for years, but lab jack is a great extension of that idea.

Richard Hull

Re: Computer interfacing

Posted: Wed Aug 11, 2010 10:16 pm
by Jerry Biehler
I have one of the ones from Dataq, starts at $25.

http://www.dataq.com/data-acquisition-s ... r-kits.htm

Re: Computer interfacing

Posted: Wed Aug 11, 2010 11:16 pm
by Dustin
Wow that is cheap.
Would be better if it had some outputs so you could have some controlling capabilities
but for logging would be ideal.

Re: Computer interfacing

Posted: Thu Aug 12, 2010 12:47 pm
by Doug Coulter
Linux drivers!

I already built a up front end (PIC uP in my case) and am now adding the 10 fast inputs from an M Audio DAW board....but this is cool too --

Most of the work in my case is shoving it all into a well designed MySQL database and then mining the data -- keeping it all organized and time-aligned etc.

Re: Computer interfacing

Posted: Sun Aug 15, 2010 8:16 pm
by bpaddock
Watch the input impedance of the low end Dataq units, it can be as low as 50k.

Re: Computer interfacing

Posted: Mon Aug 16, 2010 4:15 am
by derekm
I know this may sound obvious,

Please take special care with computer interfacing for high voltage protection.
A particular thing I have found by hard experience is to have a separate ground for the protection circuitry.The protection can be quite simple, ferrites, capacitors, diodes etc but dont feed the protective ground into the house earth... A HV pulse into the house earth can blow up the appliances connected to it. This will have the largest effect on those electronic devices nearest in the house wiring to the source of the pulse.

Re: Computer interfacing

Posted: Thu Aug 19, 2010 7:44 pm
by Doug Coulter
Well, it may sound obvious, but it IS a real problem (and many people only see the obvious in hindsight after the magic smoke comes out of something); even the normal fixes for ground loops don't always solve it, nor do all the ones that try to protect things like sensor or data aq inputs from spikes.

Even what you'd consider a stray capacity (picofarads) in something (like a wire to a feedthrough) if arced to ground can make kiloamps of peak current that will really couple into anything nearby even if not directly connected -- an accidental transformer isn't hard to make at all. Takes awhile to shake all that down and get a good stable data aq system that isn't fooled by EMI. And these things can really make some EMI. I've had computers fried just being in the same room, not connected, but with energy that simply coupled into things like USB, VGA, serial port cables not connected to the fusor. It's a real situation if you're not pretty careful.

I did just get a LabJack U6-pro here, and it's in preliminary testing now. I've never seen anything so trouble free in Linux for scientific type gear -- it just works! I got a U6 pro -- way worth it even for someone who usually rolls their own in this area. Drivers work, the examples work, it's reasonably fast, it's accurate, what can I say? You do have to know how to follow directions and do the make, make install kinds of things you normally need to know anyway for Linux as soon as you do anything a little off the mainstream. No big deal.

Ethernet can give some limited galvanic isolation (500v rated but it has some capacitive coupling in the transformer), but they were out of stock on them, and there's an app note on their site that shows how to use a certain wireless dongle to put the USB model on the ehternet cheaper than their ethernet version (which was out of stock anyway). That gives far better isolation and flexibility than a wired connection. The computer doesn't even have to be in the same room then.

The add on connector boards seem pricey until you have them, then they seem pretty cheap. That's one heck of a lot of connectors for all these can do.

Re: Computer interfacing

Posted: Fri Aug 20, 2010 11:51 am
by derekm
Doug Coulter wrote:
> ... Takes awhile to shake all that down and get a good stable data aq system that isn't fooled by EMI. And these things can really make some EMI. I've had computers fried just being in the same room, not connected, ...
>

Done exactly the same. Although not with a Fusor, but an Electrostatics research projectt. A 40 Kv flash over zapped the Single board Computer sitting on the bench.2 metres away. Then an explanation to to the Prof...

After that, apart from the separate protective ground, it was careful routing of all cables so they would be sheilded by large earthed metallic parts of the apparatus from "seeing" the potential flash over. thenn The first level of protection circutry on was mounted on the experimental apparatus.itself.

Trying to count the nanoamps with Kv flying around meant lots of fun with filters and amplfiers and level shifting. The analogue section its self had another 2 levels of circuit protection to protect the DACs and the computer . it was separated physically from the apparatus, from the DACs, with circuit protection on both the entry and exit.

In a phrase - Layered in depth protection. All of which helped it survive a number of incidents

Re: Computer interfacing

Posted: Tue Aug 24, 2010 7:04 pm
by SkyPuppy
Is there any way to turn these into optically isolated circuits? How would one change varying voltage, for example, into something that you could connect optically? I'm way out of my experience here.

Re: Computer interfacing

Posted: Tue Aug 24, 2010 7:36 pm
by Tyler Christensen
There wouldn't be a reliable way to optically couple the analog data (you could, but you would lose significant accuracy), however you could take a number of approaches at coupling a digital stream. One method could be to run the analog data through an ADC, couple this digital datastream optically to the microcontroller or data acquisition device, then decode the digital stream in software. Alternatively, you could have a microcontroller that is sacrificial in the case of a fault and then optically couple a serial datastream to the computer that is doing analysis and logging on the data. This would protect the computer, but not the data acquisition unit.

But ultimately, even if it is optically coupled, you still need to properly shield microcontrollers and computers. These sorts of devices can be destroyed by a nearby arc or discharge even without being connected to the same circuit. Proper isolated grounded chassis and shielded cables leading to appropriate non-looped grounds is the only sure way to protect stuff even if it's not directly connected to the fusor, just nearby.

Re: Computer interfacing

Posted: Sun Sep 12, 2010 9:51 pm
by walter_b_marvin
Full Linux boxes with wireless and ethernet are selling for under $100. Some of these have up to 7 RS232 inputs, for say, a microcontroller board with A/D capability. This gives you

1) Remote Monitoring
2) Positive remote control

I should be able to write a C# gui application to read all meters, and apply voltages, while sitting safely behind a wire saftey cage. My PC application will talk to the LInux box, witch in turn will drive up to 7 controllers, each of which might have 2 or three functions. At least that is my approach.

Re: Computer interfacing

Posted: Wed Sep 15, 2010 6:21 pm
by walter_b_marvin
I one converts rs232 to an optical level, transmission loss on the optical line is irrelivant