Re: Oscilloscope delivers 25 GHz bandwith on 4 channels

Tek did a random sampling plugin for their 7000 series scopes, the

I want one, but the price is extreme. There's probably not much inside.

The 7T11 did a sawtooth scan to sweep a 7S11 sampling unit (with sampler plug-in) across some time interval. With an S-6 plug-in, it would do 30ps risetime, which was fast enough for my needs. This was 1970's equipment, pretty good for the time.

I was very happy when I realized it was possible to drive the

I loved that piece of equipment. Tektronix were the best in that era.

Jeroen Belleman

At a trade show, in the Tek booth, they had a little box that hosted two SD series heads and interfaced to a PC. They never offered that for sale.

It wouldn't be hard to do that now. There are lots of S and SD and 7S heads on ebay.

Very nice idea, but that will work only for sinusoidal signals, right?

There were some superhet oscilloscopes that split the input signal into bands with RF techniques, namely downconverting bands and digitizing them, then somehow putting that mess back together mathematically. Of course, one was a LeCroy.

Integrated shockline samplers killed that idea.

But 100 GHz electrical signals barely exist, so the market is small for those megabuck scopes.

I should be possible to abuse a cheap fast latched comparator as a sampler with ~10GHz bandwidth or so. Something like an ADCMP580.

Jeroen Belleman

I've done that and have a PCB, TDR actually. It seemed to work but I haven't had much time to play with it.

Does anyone want to take over and see how well it actually works? I guess it could become a product.

Simon and I are just finishing up a TDR gizmo for measuring soil moisture and salinity vs depth for an ag customer. It’s a 150-ps-class device, which is much better than good enough for the application, and we’re getting the first 20 fully-stuffed boards for $23 each from JLCPCB, including the data converters, MCU, voltage regulators, as well as the TDR proper.

It uses a two-diode sampler, which avoids the major pain of sampler design, the need to match diodes. Of course it has horrible kickout, but that’s perfectly okay in this situation.

Fun gizmo.

Cheers

Phil Hobbs

Yes, but you do also need to keep track of phase in order to reconstruct the waveform.

John

At an earlier employment a proposal was made to include a TDR into a product, to be able to preventive warn of cable faults or even motor winding shorts.

Then a RF engineer, one that I never liked much, took the brute force approach using a GHz sampling ADC, costing hundreds of dollars per product (would effectively kill the idea). He said it could not be done in any other way.

I then made a diode sampler, with a sliding picosecond STM32 timer, and made it for 10 USD instead :-)

Our gizmo is replacing something like that—a 250 MSa transient digitizer run in equivalent time mode. Its BOM cost was around $400, plus a lot of the parts were EOL.

Savings like that sure make the licensing conversation easier. ;)

Cheers

Phil Hobbs

So you were able to make a deal with the client that you part owned the IP, and could use it for other projects?

I am in a similar situation right now, working on a dedicated HW solution that I would like to begin to sell afterwards. Guessing either telling the client they get later improvements to the design for free, reducing my hours billed, or letting them get a percentage of the profits of my sales.

The conversation is still underway, but I expect that we’ll wind up with a win-win deal, as we have previously.

We position ourselves as a design consultancy with a lot of existing “background IP”, including full product designs, design studies, and general know-how.

In the present case, we’re looking at filing a patent for a new measurement principle, and charging the customer a combined patent/know-how royalty. Since the total cost is a good bit less than the BOM savings alone, the negotiations are pretty amicable. ;)

Cheers

Phil Hobbs

Sounds like an idea situation. Important to take that upfront

I have spend a lot of time on cost savings over the years, so when I take a consulting assignment, I often try to solve the task and reduce cost significantly at the same time. Ideally with a ROI of a year. That is a compelling selling argument against the customer.

Cheers

Klaus

Yup. It’s worth pitching a royalty,because the value you bring isn’t just a one-time payoff. That’s true whether you’re customizing an existing design or doing something new. You have a lot of background IP, including previous designs and

The most common objection is, “We’re paying you for the work—why should we pay twice?” The answer is that the royalty (5% of revenue or thereabouts) is for the background IP, and the hourly work is for customizing it to their application.

Try saying that you want to succeed together with them—that gets the point across pretty well, we find.

Cheers

Phil Hobbs

design approaches, circuit topologies, and general expertise, such as how to make an amplifier fast and accurate while avoiding oscillation.

Oh