RLC bridge

Looks like a good tool. Supposedly the Signalhound can also do scalar network analysis (to 4.4GHz) but I haven't looked into that yet.

I find spectrum analysers largely underused in R&D labs. Sometimes people almost tear their hair out trying to see some "fuzz" on a scope that realistically only delivers 6-7bit ENOB at speed. When I ask for a spectrum analyzer they bring out some real old box and then, minutes later, we have the mismatch and also its location. Those things have so much more dynamic range. You can switch them to 1dB/div and it's almost like an x-ray machine.

From talking to the marketing guys I know, for every engineering group they sell a scope to, they typically sell four or five to production groups -- and the production groups care a lot more about keeping instruments in cal than engineering groups usually do.

Yeah, I suppose that's not a bad way to go... at least if the "faceless box" doesn't already need tons of computing horsepower.

OK, but didn't we already have this discussion with Keith? :-) Just as IBM was too big to produce dirt-cheap PCs profitably, Tek/Agilent/LeCroy are seemingly too big to produce dirt-cheap test instrumentation. Heck, the low-end of many traditional test equipment companies today (Keithley, Agilent, etc.) is often just re-badged boxes designed & made in China.

Mmm... I'm not sure.

It does seem to me that if, e.g., Highland Technology wanted to produce, say, a 1GHz DSO, they could do so and make money for a lot less than what Tek asks for.

It's a significant part of the market, certainly... although I'd be curious to know what percentage of dollars it is. There's an awful lot of margin in the high-end test equipment business, although that does have to support an awful lot of R&D as well.

Another thing the Chinese companies often don't have is a lot of credibility when it comes to safety ratings. That's only a small part of the market, certainly, but if I were buying meters to hand to electricians to measure, say, 120/240/480V with decent energy behind them (i.e., many horsepower motor drives), I'd generally pony up for the nice Fluke meters and probes. even though it costs a lot more than getting the $3 Habor Freight meter. :-)

Fluke had a nice PowerPoint presentation somewhere with photos of various cheap meters that had self-destructed when subjected to overvoltages.

---Joel

I would *love* to see someone like Dave Jones go and measure ENOB on real scopes -- few people realize that buying an 8-bit or 12-bit scope doesn't really get you nearly that many ENOBs on a single-shot acquisition. The scope manufacturers at times seem to feel that, well, if none of their competitors have mentioned some of the ugly real-world warts of a particular piece of test equipment, perhaps it's best if they don't either and sully the well?

(At least these days you can usually get decent app notes on what a particular brand of scope is really doing in "noise enhancement/averaging" or similar modes. I.e., some average adjacent samples of the same acquisition, some average adjacent acquisitions, etc. With the exception of "smoothing" on VNAs, they're generally all "legitimate" approaches, but the outcomes are somewhat different based on the exact approach used.)

Wow, that's surprising. Yes, in production everything has to be calibrated, same in QC and service. When I ran part of a company we had a fairly large production. In the cleanroom there were no scopes at all, it was all mostly rack-mount custom stuff, machines, computer-based testers and so on. Systems production had a couple of scopes but not very high end ones.

The Signalhound does. Although I found it to run with minimal speed penalties even on a Samsung NC10 netbook. The really nice thing during off-site jobs is that you can write the report at the same time, keep the old EMC report from the lab open, pull datasheets off the Internet, et cetera. This way I could keep feeding my client part numbers to order so they'd be there next morning. All right there from my "spectrum analyzer". Their engineer was only across a hallway but since I sat tangled in the middle or myriad wires I sometimes used email for that, so my requests went to Kansas and then on to a computer at the other side of the hallway :-)

And we as a country keep losing our shirts over it :-(

Then it would be high time for Tek to get some business consultants in there that could find where the bottlenecks are.

If a company wants to only play in the luxury market segment and is ok with remaining a small company, nothing against that. But they have to constantly watch their back and I don't quite see them doing that. Just imagine, who would have thought 10 years ago that some day you can buy a pretty good GHz spectrum analyzer for around $1k? That day is now here.

Well ...

They are trying to buy time I guess. I remember when western auto mfgs touted their great crash test results and so on, and that the competition from Asia can't live up to that, and so now. Today they do live up to that.

I was quite surprised when Instek scope showed noticeably lower noise in a plot than a name-brand scope that was at least 30% more expensive. The knobs and the whole thing feel sturdier as well.

Order more!

Cheers

Phil Hobbs

I'll bet the engineering models have more bells and whistles on them. Generally, I don't care much about cal, as long as the thing works. If it's at the cal lab, it's not working.

I just got a note from the TestEquity (Agilent) sales engineer thanking me for my order. I guess the order for a couple of MSO3054Xs and stuff (power supplies, etc.) did go through. The boss said to order but he didn't have a "capital" budget. His boss told him to do it, so...

It's *exactly* the same. Large companies have large overhead, particularly those with large direct sales forces.

Badgeware. The next step is out of the business. Exactly the path IBM followed in the PC business.

They're more flexible and change faster. There are far more of them, so if

Perhaps, but not less than the Chinese. My bet is that T&A have the market for the high-end, all bells and whistles, pretty well locked up, though. There probably aren't enough sold to interest the bottom feeders.

Exactly, something the bottom feeders don't have to do.

Or a badgeware Agilent? ;-) They're giving them away with the purchase of a DSO. So? Who wants it?

I used to use that LCR meter some 25 to 35 years ago, i liked it a lot.

?-)

tho.

It has been on my "i want one" list since.

?-)

At defense companies the rules can be strict. I was in the middle of some tough measurements when a guy on a golf cart size electric truck pulled up, unhooked my probes and loaded the scope onto the bed. "Hey, I need that right now!" ... "Sorry, it's due for calibration by 12:00pm today".

Then they are often doing something wrong. I had a similar discussion with the CEO of my first employer over 20 years ago. Told him that some day medical ultrasound will be down to 5-10 major players and all the little ones are bought up. He said no way, their overhead and large sales force would be too much of a burden. It's mostly direct sales at the big players. Well, fast forward 20 years and it is exactly as I predicted back then. Even the place where he and I worked was not spared, it is now Philips Medical.

Bingo! That can be done in the computer world as well, and has been done (Dell, for example).

Huh? Where are the 100 auto manufacturers? Or even 100 TV manufacturers? It's almost only the big ones now.

I have personally seen the risk for the big players when they are thinking that market is locked up in their favor. Brought my scope (Instek GDS-2204) to a client because I don't really trust the lower end Tek scopes for analog stuff. Did the job, manager walks by, looks, is a bit surprised since he never heard of that brand. "Under $2k for all this? Wow! So what else do they sell?". I didn't really know so I gave him the web site. He sure must have done his research there. During my next visit I saw that they had bought a barrage of Instek programmable power supplies for their quality control. They saved many kilobucks.

I have half a dozen Harborfreight meters for the road, just to check grounding and stuff. If one dies in the action or is lost it's not a big deal. They are surprisingly accurate.

I don't think so. It would take millions to develop a scope, its software, and its packaging, and its certifications, and it would be difficult to market. There's already Rigol and Instek and others fighting for the business. Their production margins are probably pretty low, what with their low-end scopes retailing for around $350 with probes. I can sell one small PCB in a simple aluminum box for ten times that, more than a Tek DPO2024, admittedly in much smaller quantities.

I would like to do a USB TDR one of these days. We have a working prototype but we haven't had the time to pursue it as a real product.

I think a lot of that "custom stuff" has now turned into off-the-shelf test equipment -- they don't have nearly as many people like you around to build the custom stuff anymore. :-)

At least test & measurement is only a very small chunk of the economy?

But I agree, it still sucks.

I suppose there's a chance Tek already knows how to make things inexpensively, they're just reaping the profits until those other competitors force them to lower their prices. :-)

That doesn't look great for Tek, but you have to ask: Would Atten or Rigol or Instek even put out a recall notice if they had a similar problem? I'm kinda leaning towards thinking, "no."

Indeed, when companies like Agilent just go off and re-badge Rigols, what you're really paying for is *supposed* to be the assurance that Agilent went over the thing with a fine-toothed comb and made sure it really met its performance specs as well as being designed reliably (i.e., don't get too extreme with overclocking the ADCs!) and safely. For a commercial operation, those assurances are worth something... whereas for hobbyists (and perhaps consultants :-) ) I can see that it might not always be.

Good point -- there's no rocket science involved in making safe meters/probes/etc., it just requires the culture to want to do so and the willingness to spend the extra money that it requires. (Not a lot of extra money, but not the $3 meter level either.)

---Joel

OK, perhaps you would have to aim more into the, say, 3-6GHz ballpark to clearly get out of the Rigol/Instek low-end. But I suspect you have the technical know-how to do it... and of course Joerg is only an hour or so away if you need to cost-minimize the thing. :-)

Many people are unaware that Rohde & Schwarz launchaed a line of DSOs last year... to not a whole lot of fanfare (that's a German company for you, I guess). I've had the opportunity to play with them briefly, and for a 1st generation design, they're actually pretty nice -- up to 4GHz for their high-end model -- if still a bit "clunky" user-interface-wise.

Their pricing is in the usual Tek/Agilent/LeCroy ballpark, though, so at least in the U.S. their growth will likely be slow.. Still, you can bet that T/A/L are all going to be keeping a close eye on their popularity, particularly in Europe.

---Joel

I didn't mean custom-custom, I meant systems engineering. IOW where you piece a rack together from off-the-shelf chunks and lash it all into a standalone system with DAQ software or LabView.

It is, but it's a lucrative chunk. We have to occasionally take stock and look at what's left, and whether we need to do something against losing some of that. I believe we have to because it has the typical markings of a gradual erosion.

They are way past that point. How will they ever catch up with stuff like the Signalhound or this little box that John pointed out the other day:

If I was in need of sampling capabilities above 1GHz right now that's probably what I would buy, without even looking at the Tek offerings because I know they can't beat this anytime soon. Or maybe never.

Once the Federales find out they will have to, or they'll come after their domestic dealers. Of course I've had my Instek open and I must say that I am rather impressed. The power supply is all by itself on the upper side of a steel barrier, and it looks pretty sturdy. It's not a cheapo plastic lunchbox enclosure. They also use the good stuff for AD conversion, from Analog Devices. What was less enthusing was the fact that the screws and steel threads showed slight traces of rust after only three years in a non-humid environment. I can only imagine how this would have looked if I lived in places like Houston.

Even commercial that is no longer a guarantee of a continuing gravy train for the big manufacturers. Like I wrote in another thread, after a client saw my Instek they bought their whole automated test setup (consisting of lots of programmable power supplies) from Instek instead.

Not there, but take a look at a $20-$30 meter. They aren't bad.

Rohde&Schwarz bought Hameg and with that they acquired a ton of engineering know-how. For a while I was eyeing their analog/digital HM2008 Combiscope but then the bottom fell out of the Dollar and they didn't have a really good mainstream sales channel over here, so I bought the Instek instead.

They should keep an eye on Asia. That's where steamroller is going to come from.

Nowadays its very difficult to make an instrument that fullfills the customer's expectations. Ethernet, data exports, and so on. About 8 years ago I embarked on putting an USB scope together. I've abandoned that project because Rigol and so on came up with cheap oscilloscopes.

Things are improving. I recently bought a couple of dummy loads and every metal part has its own ground lead.

That's pretty cool.

Not for the price, but if I had the money around (and most of industry does), I'd just get an 8GHz *real time* scope from one of the big guys

-- far more convenient, and the user interfaces can do *much* more than FastSampling.Com's. (I mean... it doesn't even appear to be able to calculate eye openings for serial data analysis -- that's where a lot of the 6-12GHz DSO market is.)

It's really two separate markets, kinda like Amazon Kindle Fire ($199) and Apple iPads ($499-$829, depending on features): The Fire is hands-down a better value for the dollar, but iPads very clearly have more capabilities. Amazon sold six million Kindle Fires over the last Christmas season, Apple sold three million Kindles (pre-orders) last weekend. There's room in the marketplace for both types of devices, IMO.

Sounds like a smart client -- and I really hope more companies do realize that paying the big bucks for test equipment is not always necessary... but that change seems to be so slow that I doubt the big guys worry about it too much yet.

---Joel

...but I can't help thinking that for every company your "enlighten," the schools kick out 10 MBAs who -- having little or no technical background -- fall back on their "best practices" manual when it comes to purchasing test equipment for their technology companies, and that means buying from the big guys.

Granted, who the big buys are changes with time too -- back in 1990, no Fortune 500 company was buying much from Dell, whereas today of course Dell, Lenovo, and HP probably make up 90+% of such corporate sales.