fpga4fun.com

Anyone looking for one of these?

I haven't found one that was reasonably inexpensive around, and I'm planning on doing one for me in the next month or so. Just the basics, a BNC connector, the interface circuitry to do coupling with a scope probe (or two), the ADC, and either a ribbon cable or pin-style connector.

Just wondering if anyone here is looking, or if more flashy boards are coming soon - rather use one of those.

This PCB can include the instrumentation input amplifier, resistor divider net and an AC/DC selector? I'm planning to do something like this one day also, but I wanted an electronic selection of the divider net. Is this feasible?

Best regards,

Nelson Sicuro

You should look into Digital Variable Gain Amplifiers like this one http://www.national.com/pf/CL/CLC5526.html and see if any fits your needs.

It would be very hard to include everything and keep an accuracy close to one LSB no matter what method you use. Some comercial circuits use special relays for switching. Normal relays will not work because they are not reliable at low currents/voltages.

My advice is to look for integrated solutions to save you a lot of hairpulling.

In talking to Jean, he tells me that the thing that keeps the Flashy board from really serving as a practical oscilloscope is that its input stage is too simple. A real scope can probe signals from a few mV to 50V, while Flashy's input is limited to one range (about -3V to +7V). That's fine for troubleshooting digital logic, but you will always need a more robust input. I know a lot about digital logic design but nothing about analog stuff. Can anyone come up with a better input stage so that this can serve as a practical tool? That's about the only way that I will ever be able to afford a good digital scope, and what we have here comes so close.

So, what do you say, people? Is anyone up to the challenge? When I first found this site, I was very excited about the possibilities it opened up. I still am. Let's keep going and take this to the logical conclusion.

Quickly, glossing over details:

Input stages on DSOs (assuming they are the 1MOhm type, not 50 Ohm) are typically a resistive divider with an equivalen impedance of 1Mohm plus compensation capacitors on each leg to keep the bandpass level through the input frequency bandwidth (ie 100MHz, 200MHz,). The equivalent capacitance as seen looking "into" the scope should be around 11pf to 18pf.

The input resistors should be of high enough wattage so that they are not damaged by input voltages. The current flowing through a resistor will be bounded on one side by the input voltage and on the other by either ground or the internal protection clamp voltages. Like wise the capacitor should be able to withstand the voltage. Don't use wire wound resistors, low inductance metal film are pretty common. Need low temp. co resistors and capacitors.

Some scopes have multiple attenuators, switched by relays.

And the AC/DC coupling capacitor can be right at the input or in some cases after the attenuator.

There should be a pair of clamping voltages immedately after the input attenuator. The simplest clamps are done with diodes and zeners. Keep in mind that they can cause signal offset and also influence the node capacitance (variably I might add).

After the clamps there is typically a follower of some kind. In lower frequency designs it can be an op-amp buffer but higher frequency equipment often has a fet-follower pair. The FETs are bootstrapped to help improve offset accuracy but you sill may have some thermal drift.

Now you can begin your signal chain.

If you can get your hands on older Tektronix scope manuals they often have complete schematics. You can see the "state of the art" of 1980 in all its gorey detail. The concepts are still the same in more recent equipment but with improvements in bandwidth, speed, power dissipation, component size, etc.

Ebay is a good source for old manuals. Also a good source for used equipment in general.

Have fun.

Rob,
Well, your reply just confirmed what I've always known: as an analog circuit designer, I'd make a good plumber. Since you obviously know what your doing, why don't you just throw that circuit together, debug it, get it laid out, debug it again, stress test it, limit test it, and make it available to the rest of us?(darn! there doesn't seem to be an emoticon here for wide eyed innocence )

Anyway, I wish somebody could do that. I certainly know that I can't. Oh well. (Heavy sigh)

Thanks for the electronics lesson.
Mark

mickelsen wrote:Rob,
and make it available to the rest of us

To a small degree it is available to you. Building a balanced attenuator isn't that hard to do. And if you promise not to try and measure lightning strikes or voltages higher than about 100V, you can limp along without the clamping circuit after the attenuator. Likewise, if you can live with bandwidth below 100MHz, use an op-amp follower instead of a FET-follower.

If you would be satisfied with 10MHz bandwidth (as a first cut, just for playing around) your layout can be pretty sloppy. Sometime between 10MHz and 100MHz the layout becomes important and after 100MHz the layout is definately important.

The other item you will need is some dual- rail power, at least for a little while near the input stage until you can shift the signal up so that single supply op-amps and ADCs can deal with it.

Rob,

You write that stuff as if I had even the faintest idea of how to do any of it. I don't. I truly wish I did. It's one of my goals to learn it before I die. Meanwhile, I'm hoping someone else will do the analog front end for this project. I'm much more able to design the logic to download to the FPGA and the program to run on the PC to emulate an oscilloscope front panel. I hope someone else will create the robust front end for the Flashy-D, or perhaps lay out a Flashy-E (sounds sort of like the Enterprise E, doesn't it), so that we could have a real tool to use on other projects. I have a sneaking hunch that you could, or perhaps xtal, or Jean, or someone. Actually, as it stands, the Flashy-D can help debug digital logic with its allowable input swing of -3v to +7v. There just isn't any really good protection there, if I understand it correctly (which I probably don't).

Anyway, I continue to dream of a good, modern, RECONFIGURABLE, storage scope that I understand (because I did a lot of the programming for the FPGA to make it go).

Is there any hope out there?

Longingly,
Mark

There are MANY inexpensive (I'm avoiding the word "cheap") DSOs available. In recent issues of Embedded Desgin there were some reviews of USB interfaced DSOs.

I'm not saying I won't do a front end design (I've done them before at a job way a back in the 20th century) but honestly if you are intending to use this as a tool for debugging other designs you will be better off buying a DSO (new or used) and learning how it works.

DSOs are notorious for telling lies. Often it is NOT the equipemnt's fault, it is the operator's fault because they did not understand the limitations of the equipemnt (sampling process, bandwidth vs. sample rate, trigger configuration and the list goes on and on). Things are much better in recent designs from Tek, Agilent (HP) and LeCroy. Much more analog-scope-like operation and variable persistance displays. Non-display DSOs (USB, PCI etc) tend to lag a bit in ease of use.

If you are serous about learning how to use a scope go to the Tektronix web site and search for "The XYZs of Oscilloscopes". There is a similar document for digital sampling oscilloscopes too. Agilent has a couple of docs along the same line. But you will need to do a little snooping around to find them. Trust me, knowing the difference between AUTO, AUTO LEVEL and NORMAL triggering will put you miles (well maybe kilometers) ahead of other hobbiests. Leanring how to use the dual-timebase feature will put you even further ahead.

Save your pennies and stalk ebay and prowl around DSO manufacturer web sites. Do you homework and BUY the best tool you can afford. Then get familiar with it. Then think about building your own. You will come out ahead in the long run.

One last thing, while you are saving your pennies, save enough to get a copy of Horrowitz and Hill's "Art of Electronics" 2nd edition. More than enough information there for you to do some real damage!

Master Robert,

Your words hold much wisdom, Sensei. Using an unproven tool to debug an unproven project would not be prudent. I must return to my quest for a commercially produced analog scope that is within my budget. Unfortunately, that probably means one that is very old or pretty hashed.
I used scopes a lot, in a previous lifetime, but they were all analog, so I should probably stick with what I know. That doesn't mean that getting a robust front end for the Flashy-D or something equivalent isn't a great idea for a project. I've always enjoyed creating tools; scopes, debuggers, logic analyzers, etc. Things that make the developer's job easier. And, since I was also the developer in many cases, I was just making my own job easier. It was usually worth the investment in time.

I hope you will decide to do a front end for a DSO, and you might also think about a multi channel front end for a logic analyzer using FPGA's.
Just a thought.

I hope to hear more from you.
Thanks again,
Mark

P.S. My copy of The Art of Electronics plus the student manual arrived two weeks ago. It must be the right one to get, since everybody seems to recommend the same one.

mickelsen wrote:Master Robert,

I must return to my quest for a commercially produced analog scope that is within my budget. Unfortunately, that probably means one that is very old or pretty hashed. Mark

Not necessarily.

Depending on what you expect to be doing with the scope, there are some very reasonable prices on ebay but you have to know when to walk away.

If you are content to see sinewaves below 100MHz and can deal with the fact that your square waves and pulses display with a rising edge no faster than 3ns to 5ns something like a Tek 2245 can be had for less than $500. These were workhorse scopes and there are lots of them out there. Two channels, good trigger system and dual timebase. Analog though.

The HP (Agilent) 54622D and a couple of its cousins make nice scopes to use around microcontroller projects with their two "analog" channels (these are DSOs) and 16 digital input channels which can be used for logic analyzier functions. New these were around $6000 but should be showing up used for $1500 to $3000. Yes you can get a used car for that much money but can you debug your projects with a used car? Good tools are worth what you pay for them.

The reviews in http://www.embedded.com//showArticle.jhtml?articleID=160902529 and there is a followup article. These are pretty reasonably priced for their performance and probably hold the most promise for playing at home. Not included are PCI bus cards like those from Protek, National Instruments, Gage and others. Like the USB pods, some expensive and some cheap. Also a few parallel port interfaced units still limping along out there.

But since you got your copy of Art of Electronics, maybe the best thing to do right now is put the pennies in the piggy bank and concentrate on your new book. I guarantee you will dog-ear that thing and have its spine cracked in no time, both signs of love and respect for a book, kinda like the Velvetean Rabbit...


Rob