Replacing the Oscillloscope with a singel chip

[scionEE]

I'm designing a test system for our product that uses an oscilloscope to capture a pulse. I would like to replace the function of the oscilloscope with a micro-controller or DSP chip. The pulse occurs every half second to second and is a minimum of .5ms-1.5ms wide and has 3-5V amplitude. We will have 3 pulses per device. If you could give me direction in which MC to choose it would be much appreciated.

Also, I'm an intern at the company so if there is a simple solution I'm probably over looking it.

 

[MikeHalloran]

Choose a chip for which local expertise is plentiful.

For a hint, look at the bookshelves around you.



Mike Halloran
Pembroke Pines, FL, USA

 

[PNachtwey]

Buy a starter or evaluation kit! Don't buy just the CPU. Buy the whole kit! There are many kits that will do the job. As stated above you should prefer to buy from a kit that your company is likely to support in the future.

I like TI DSP evaluation kits. Many TI DSPs have analog inputs built in.

Peter Nachtwey
Delta Computer Systems

http://www.deltamotion.com

 

[IRstuff]

Nothing is quite that trivial. It depends on what you're supposed to do when you "capture" a pulse. Do you timetag it, do you simply note it, do you measure it? Who's going to be writing the software to make this work? Who's going to be qualifying and certifying that it does exactly what the oscilloscope did?

TTFN

FAQ731-376

 

[Skogsgurra]

For that, you do not need much. A PIC with a built-in AD will easily resolve down to 0.03 milliseconds if one channel is needed and it has 12 bit resolution which, to everyone's surprise, actually is very close to 12 bits. It is used in our ARCUS transient recorder and does exactly what you describe - plus a lot of other things. We power the 'system' via the USB bus and also control and get data over the USB.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[Skogsgurra]

Sorry, the chip itself will resolve much better than the 30 microseconds. The USB bus sets the speed limit in our case. The chip can do much better.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[itsmoked]

I'd grab one of these:

http://www.parallax.com/tabid/295/Default.aspx

The fastest is:

http://www.parallax.com/Store/Micro...1/Default.aspx?SortField=UnitCost,ProductName

Your timing doesn't appear to be rigorous. If you only need a hand full of units there's no way to justify the cost and hassle of laying out a board, purchasing parts, purchasing a compiler, and getting the tool chain setup as compared to buying a few of these, hooking any terminal program to it and be writing code for the hardware you'd have in your hand a day or two later.

These are all interpreted so they contain the operating system on board.

You need to refine your timing needs a little better and what the results of monitoring these pulses are.

None of us could solve for the rather vague spec you've given.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[scionEE]

I'll try to narrow down the specs. The pulse will be 1ms wide, 4V in amplitude, and repeat every half second. The signal line is pulled to 5V. When a pulse is triggered it pulls the line down by 4V. While pulled low the signal has some RC charging towards 5V and at the end of the signal the line is pulled back up to 5V. The default settings are 0.5ms and 5V, which occurs on reset aka powerloss. So, if the pulse is not one of the two states listed it should be failed. Are the parameters listed enough information?

What the microcontroller needs to do is be able to trigger on the falling edge, store and pass on the event, and then close when the pulse is completed. The triggering of the event is what I don't know how to program into a MCU.

What would ideally be done with the signal is it would be compared and given a pass/fail/reset tag, which would later be reviewed by a engineer, as long as the event can be identified by the MC and passed onto a PC the processing can occur there.

The current system uses Agilent VEE 8.5 to interface with an i2c microcontroller which communicates with i2c chips (muxes and ADCs).

I will look though the basic stamp and the PIC devices.

 

[IRstuff]

So, why can't you look at the setup of the oscilloscope? Whatever the oscilloscope is programmed for, you duplicate.

TTFN

FAQ731-376

 

[scionEE]

I'm not sure what you mean? Do you mean the settings I apply to the oscilloscope to capture the pulse? Or, do you mean the code on the DSP chip on the oscilloscope?

The former is done using the remote control commands through a GPIB interface. Where do I begin to find the latter?

The idea is to save $10,000/system by having a chip doing just what is needed, capturing a pulse and passing it to a PC.

 

[MikeHalloran]

I think you have to really understand what capturing a pulse and passing it to a PC really means, e.g., how the pulse's image is stored, with what resolution in time and amplitude, how exactly it's stored, and exactly what is sent to the PC, on the wire.

Then you can work backwards from there to derive a minimal capture system.

I don't intend to minimize the difficulty of those two separate tasks. Much reading and learning awaits you. Enjoy.



Mike Halloran
Pembroke Pines, FL, USA

 

[IRstuff]

Well, it's both, then. Some oscilloscopes have the capability of actually checking the entire shape of a pulse against a template. SOMETHING, either in the oscilloscope, or elsewhere, is making a decision to pass or fail the pulse. Whatever processing is required for that SOMETHING to make the decision is what you need to duplicate. You appear to suggest that there is a DSP within the oscilloscope, so you need to ask, or find out, about where that code resides and how to get a copy of it. You need to find out if the remote controller does any processing to make the decision. These are all the things you should have already investigated and detailed BEFORE you asked the question about a single chip solution.

I think that the other responders have already been quite generous with their responses to a student posting that would otherwise have been deleted by now.

TTFN

FAQ731-376

 

[scionEE]

You know what, your absoultly correct(not that you need to know that). I posted under the assumption that there exsist a solution with a already made product.

 

[IRstuff]

No doubt others have already trodden down this road, but it's not a done deal, simply because, as we've all indicated, each situation is unique, requiring unique solutions.

TTFN

FAQ731-376

 

[itsmoked]

You can take MikeHalloran's points about looking at the scope, understanding the actual "test", and then understanding what's being sent to the PC as The Required Preliminary job.

You need to draw little sketches of the various timings and perhaps shapes of the waveforms under test. Also draw or describe clearly, (to yourself), the limits of acceptable results.

Then understand the numbers that will come from analyzing the waveforms up to their allowed limits.

Next, you need to understand the signaling protocol the scope is sending to where ever, because this will define the communications hardware that will have to be provided.

Once you have this information in your head, this preparation, you can look effectively at the various solutions you can apply to see which falls out as the best solution. This will be based on time-to-implement, cost, and any potential hurtles with a particular solution. If one exists some immediate investigation is needed to prevent a project failure when you reach that point in the project.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[Skogsgurra]

We may need to get a few basic facts right before telling you what to do.

1. Shall the new function do exactly what the old one did? I.e. show a picture of the pulses that someone can examine or shall it analyze the pulses all by itself? In the former case, there are lots of very low cost solutions like USB scopes that can replace your unnecessarily costly scope.

2. Will the device do the same thing year after year or is there a need to change limits from time to time?

3. Is this for a few units, for ten units, hundreds, thousands? If you need hundreds or thousands, you can throw in some development money and get exactly what you need. Else, development cost may become a major problem.



Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...