Big scope

And then send a photo to my clients, with me on the couch :-)

It's the equipment. For example, the old logic analyzer has a very small CRT but one can use RS232. Goes into PC, PC display pics -> bingo.

Today in the pool I had the idea of projecting a pic there, have a small margarita mixer on the pool deck, and float around in the water. My wife thought that would be decadent :-)

I really don't trust Windows when it comes to realtime.

My experience with NI is that everything will cost at least $500. So I've largely stayed away from them.

Well, I prefer something that's already natively capable of realtime. Such as QNX.

I priced it out 15 years ago and licensing was quite favorable. Of course, there wasn't anything else comparable out there and as far as I can see there still isn't.

"Overheated and not survived" clearly points to serious design flaws. If there is a chance of overheating there should always be temp sense and graceful shutdown.

Those are the people you want in quality control :-)

I have a similar trait: Worked in medical ultrasound much of my life but I am one of the worst when it comes to ultrasound penetration and visibility.

Too late. This has already been invented and usually comes in the form of toner and ink cartridges.

You'd have to watch a great deal of TV to be exposed to subliminal advertizing. And there is the first problem marketeers have in our case. We have no cable or dish. Only an antenna and since that dreaded DTV was made law, reception falls apart so often that what little TV consumption is left has shrunk to a minimum. So the drill is:

Tape old movies we like, repeat recording as often as needed until you luck out and the station won't pixelate out in the middle of it, then watch it some time later while skipping all ads via the FFW button.

TV stations have shot themselves in the foot by two mistakes:

a. Agreeing to have a working system replaced by a flakey one with sub-par RF path error tolerance.

b. Giving up precious VHF space for less lucrative UHF space. That one I'll never understand. It has resulted in numerous stations not reaching suburbia anymore. Suburbia -> people with discretionary dough -> no longer reached -> drop in ad viewer number -> drop in ad revenue.

Well, yeah, but only the good ones stick. Volkswagen marketing is exceptionally bright there. But will it result in increased sales? Not with me and my wife. Or take the repetitive slogan everyone knows: "If it fits, it ships". So I took a peek, saw that the prices aren't favorable, and to this day I still pack my own box and ship for a buck or two less. Sometimes the difference is well over $5.

Or should work. They fail to work with many people who have lots of discretionary income. For example, you can't win much if you convince people who live on food stamps.

I have one, a HD-1250, probably the same one you have. But those easily go *PHUT* in a high energy system.

Au contraire. I had cases where people mumbled things like "That looks like voodoo" or "Now he's really off his rocker". Then at lunchtime I announced the problem to be found and some jaws dropped.

In engineering? Who cares?

No, because it frequently found stuff that their fancy schmantzy $10k+ analyzer simply did not see.

They often don't. Classic real-life example: I was called out to a client who had internal noise issues in a high-sensitivity chemical analysis machine they just developed. I dreaded it because it meant at least 3h each way on the freeways, can't fly there plus I needed to schlepp tons of gear. And my lower back doesn't like loading all that. So I got there, they had a luxurious Stanford Research analyzer set up, looked and looked and looked, couldn't find a thing. "Must be software, but the guys said it's impossible because the noise is clearly in the rwa data". So I tried. Shut the SRS analyzer down within 10 minutes. Pulled a laptop and a fairly ugly contraption out of my stuff, people cracked jokes about it. By early afternoon we had it, fan-blade induced noise that triggered air flow pulsing near a cap which in turn created wee microphonics. I showed them a fairly well distinguished 7-8Hz peak that was not very stable. On the SRS there was ... nothing. "No, you've got to be kidding, that's an artifact or something". Slowed the fan down with the palm of my hand, peak moved towards 4Hz. "I'll be darn!"

Even many boat anchors have never been reached in performance by "modern" gear. One is the HP-3577 network analyzer. Same with the HP-3585 baseband analyzer. Or the Tek 2565 scopes. And on and on. So if you face a difficult analog job chances are that you won't be able to deliver your best performance if you only have "modern" gear.

Nothing is real and there's never enough time. Windoze just makes it look nice.

If QNX is that good a fit, so perfect for the application, and has been around for many years, why aren't more instrument makers using it? There's usually an obscure reason, such as them irritating someone important in the industry, creative pricing, or their presidents belonging to the wrong country club.

One of my customers got a QNX runtime license quote in about 2002 of about $350 per device. Since the customer didn't plan to ship enough units to justify a license with better prices, the high license cost effectively killed the idea. I don't recall what OS or RTOS was eventually used.

The rack of equipment had no centralized power management or alarms. Each piece of equipment allegedly took care of its own power and alarm requirements. At least that was the plan. What usually happened is that the hot air would accumulate at the top of the rack, and heat up the top boxes. There were vents but convection wasn't enough to keep the rack cool. The top box also blocked air flow from the top of the rack. The customer refused to allow rack fans because of "reliability issues" or some such vague excuse. I was seeing overheating traps from equipment that hardly belched any heat but happened to be at the top of the rack. I suggested that it might be helpful if we inverted the rack, but that was rejected for aesthetic reasons.

In the end, we repackaged one of the boxes, and used the available rack space to improve the air flow. I put some effort into optimizing the air flow. However, the big improvement came when the customer was finally convinced that maybe a big fan would be helpful. As for reliability, I get the BS award by suggesting that TWO fans would be more reliable than one.

I believe that's called "thick skinned" which is roughly defined as: 3. insensitive or hardened to criticism; obtuse; callous.

Very true. I wonder if the warranty timer idea is patented:

Method and Apparatus for Validating Time-Limited Warranty

This one displays how many days are left on the warranty on a stun gun. I can only speculate what happens when the warranty expires:

Plenty of others that mostly deal with methods of providing a variable warranty period. I think there's room for my idea. Does anyone make a chip with a self destruct feature or explosive device inside? Can such a small bomb be reliably modeled with LTSPICE?

Nope. It's heavily used. At one point in the 1970's, public pressure was sufficient to get Congress to consider enacting laws against the use of subliminals. An investigation was ordered, which concluded that subliminals don't work and therefore no law was required. That was the signal for Madison Avenue to go berserk with subliminals. It's been like that ever since.

A good example is what happened when an employer decided to hire a professional advertising agency to promote a product that I had helped design. The agency put together a full page magazine advertisement. This was before electronic layout so everything was photographed and later overlaid on a light table or in a darkroom. The original proofs had some subliminals that ad agency added for no obvious reason. It then went to the photo lab for color separation, where the lab added subliminals using an overlay. The ad now had a very light mosaic of the word "sex" all over it. The magazine or the printers added their overlay of subliminals in the form of the words "good job" that I found in several places on the final magazine advert. The radio sold well.

So, why did literally everyone that could touch the advertisement add subliminals? For one thing, it's easy to do overlays. However, the real reason was that they work. Essentially, they're hidden advertisements for the ad agency, for the photo lab, for the printers, and for the magazine. By adding subliminals, they sell their services, and maybe a little will rub off on the customers product.

It's much more difficult to see subliminals on TV or in movies. However, if you can do stop frames, they're fairly obvious.

(major cut due to lack of time)

Yep, I've been there with a similar troubleshooting exercise. I tried (and failed) to help an agency deal with an interference problem. We could hear the interference but could not identify the source because the intermod mix order was sufficiently high that several expensive service monitors could not decode the PL, DCS, tone remote, or other control tones. Tweak and tune, but nothing would show. So, I plugged my Droid X cell phone audio into the discriminator output, waited for the IM to hit, and captured some audio. Played back on a spectrum analyzer app, I could see faint lines where the tones were expected. I repeated the capture with my laptop and used Spectrum Lab software to find the tones, which were something like 50 or 60dB below the peak audio levels. The expensive service monitors just weren't designed to pick out such low level modulation. It took a while, but the intermod sources were eventually identified (by someone else).

While this is not an example of the superiority of using antique test equipment instead of modern test equipment, it does illustrate that some pieces of equipment are better at doing some things than others. It may be that the instruments that you list are superior to modern equipment in some specific or specialize ways, but overall, the new stuff is generally better for the more common tasks.

I have no idea. It is in wide use but not in consumer applications. Well, except for smart phones.

When we made our decision for QNX and against Windows CE in the 90's the reasons were crystal-clear. The QNX support team came across as vastly more knowledgeable and the license fees were very modest. Open source was not an option because this was med tech where you need 100% trackability.

We were quoted way lower in 1997. $350 is a show stopper for just about any product.

But still, nothing is supposed to die in here. If it does -> back to the drawing board. It's the same with automotive, if the water-cooling system fails because a hose burst that is not supposed to take out the pricey electronics.

:-)

I usually try not to be but that seems to have no impact on the huge ultrasound signal losses in my body.

I sure hope not.

Ok, but if they worked, why is it that I pretty much never buy any products from ads? When I need something I browse for it, but only then. Maybe us guys are wired differently than women but I have never felt the urge to absolutely have to have something, or even go on a shopping tour with no particular needs.

The perfect companion for a $50k scope! (you know, the kind with Windows :p )

Tim

"It can take weeks to get a quote from a vendor, with lots of back-and-forth with a sales rep about specific requirements and features. I am not at liberty to divulge specific pricing for various RTOS development licenses, but you should expect a number in the tens of thousands of US dollars. QNX Momentics is among the least expensive for initial setup, and can come in under ten thousand depending on requirements. The trade-off is that there is also a runtime royalty, which can tip the balance depending on volume. The cost runtime licenses for QNX can vary greatly, especially since price will go down as volume goes up. The industry as a whole ranges from royalty-free to hundreds of dollars per unit." I also found some postings mentioning developers licenses for $4,500. I find it rather odd that QNX would not have a standard pricing schedule, and that they seem to have some kind of restriction on customers disclosing the price of their licenses. However, despite my suspicions, QNX apparently have been around a long time, which suggests that they may be doing something right, and have something worthwhile to offer. Unreliable vendors, with predatory pricing, just don't seem to last.

There's a bit of a complexicated story here. The original commercial rack full of equipment came with BITE (built in test equipment). This was about 1979 and such things were done with huge circular MIL-SPEC connectors and ugly bundles of spaghetti. The customer decided that they didn't want all that junk and removed it from the bid. Unfortunately, all the safety, shutdown, interlock, alarm, and flashing lights were in the BITE panel, where all the spaghetti came together. One of the other engineers threw together a minimal alarm, interlock, and shutdown panel, when it became obvious that the rack wasn't going to be safe without one.

If it's not dangerous, it's not fun. Fortunately, there's no need for a post-warranty explosion. The chip could simply short itself internally, leaving a charred crater on the PCB, or just open an important data line, resulting in a non-operational PCB. The opportunity to "license" the use of the design by the elapsed or running time, is tremendous. It's just like electric water heaters. You can get them with a 6, 9, or 12 year warranty. The difference is the size and number of sacrificial magnesium anodes, which are sized to fail according to the warranty. With my warranty timer circuit, a manufacturer would no longer need 3 different builds of water heater. Instead, they could make them identical and use the warranty timer to blow a hole in the PCB or the bottom of the tank to announce that a replacement is past due. Consumers are certain to appreciate my efforts on their behalf. In the past, they had no idea when their purchases were going to fail. Now, they know exactly when it will fail, and can plan accordingly.

Probably because you're only thinking of obviously obnoxious radio and TV advertisements. If you look carefully at the products that are being advertised on broadcast media, they all a products or services that are of marginal value, possibly a scam, or are sufficiently useless that only saturation advertising will produce any results. Insurance, reverse mortgages, strange kitchen gadgets, and cosmetics come to mine. Anyone with a reasonable IQ can easily through that crap.

However, what you don't see are the very short exposure to billboards, display advertising, magazine ads, and web pages. These are all designed to do their job in a few seconds at most. When grilled afterwards with a recollection or retention study, most people won't remember seeing the ads. It will seem like it never happened. Yet, these are the most effective types of ads. They work by either creating an association with the product, or by telling a story.

Association is the most common. The advert is full of sexual hints. If you buy the product, your sex life will improve, and stuff like that. Such methods only go so far, and the sheer number of such subliminals have made them somewhat ineffective. So, new associations have been contrived. Fear of death, religious connections, connections with famous people, cartoon characters for the kids, etc. Anything that associates something good (or bad) with the product. The connection is triggered most commonly at the store or supermarket. You are faced with a myriad of almost identical unfamiliar products on the shelves. You have a limited amount of time to make a selection. How do you decide? Well, there's price, heft, packaging, but most commonly it's by how the product "feels" to you, which means what images does it produce in your mind. In extreme cases, just looking at a product can make one salivate with the anticipation of eating or playing with it. Creating a positive association makes that happen. If all you think about is sex when you're shopping at the supermarket, you now know why.

I once went to a restaurant that had a printed menu with lots of photos of their various offerings. Across the chili pot was the word "HOT" in two places. I have a poster with a picture of the Matterhorn at home on the wall. After several years, I noticed that the glaciers down the side had been retouched into the letters "SEX". Morris, the tabby cat, has the word "SEX" all over his fur, which is why they used a tabby, instead of a solid colored cat. The stuff is everywhere.

Unfortunately, associations can also backfire. I don't recall the companies, but one of the canned soup companies decided that giving away panty hose with their soup was a good sales idea. They setup the promotional display at the markets and found that sales dropped like a rock. Buyers were associating the promotion as putting their feet into their soup. Yech.

The other way is to tell a story. An old example of that are the liquor ads that showed a party scene. You can still find those in some popular magazines. Look at the people in the picture and try to determine who is married to whom, who is playing around with whom, who is groping whom, who's advertising themselves, who's available, and what's really going on in the picture. If you pretend it's an orgy, it's easier to understand. The attraction is that if you buy the booze, you too can join in the orgy.

There's also story type advert that I found in a mechanical engineering magazine many years ago. If I find it, I'll scan it and post it. I won't go into the whole story, but basically it shows a murder and offers indemnity for the perpetrator if he buys the manufacturers ball bearings (or something). I suspect that almost everyone has thought about killing their boss or a coworker at some time in their career. Well, here's an advertiser that capitalizes on those thoughts.

I would have great difficulty proving that such things actually work. All I can offer is that advertisers spend about $48 each year, much of which is on subliminals. I think it's a fair assumption that businesses do not spend money on things that don't work.

It's highly likely that many readers are going to be on the lookout for subliminals in advertising. Here's what will happen. At first, most people have trouble seeing the figures and letters until after they are shown examples. Fairly soon thereafter, they see subliminals everywhere they look, in huge numbers, and with amazing clarity. Then, nothing. They can't see any. Don't panic, this is normal. What happens is that the brain simply cannot deal with the transition between something that was previously buried deep in the subconscious, and now has to be deal with in real time. After a few weeks, they begin seeing the subliminals again, but never in the huge numbers that they originally detected. Maybe it's better that way.

Do you ever go shopping without a shopping list? Do you buy more at the market when you're hungry, than when you're stuffed? Have you ever had a lousy day and then gone shopping for some toy to make you feel better? I'm really guilty of the latter, even if it's shopping on eBay. I have to force myself to delay my purchase for a few hours to avoid the splurge.

It works because you cannot afford the time to carefully analyze everything that you do. Nobody can do that. Important purchases require analysis, spreadsheets, calculations, research, expert advice, etc. Impulse purchase require nothing. If an advertiser can somehow convince you that their product is a splurge purchase, you're likely to buy it without even thinking. My father's company manufactured lingerie from 1953 to about 1988 which was sold as an impulse item in large department stores. The rack of garments had to be located near the checkout counter or they simply would not sell. Everything we did was designed to enhance the impulsive nature of the purchase, such as reducing the amount of thinking and handling required to conclude the purchase. I'll spare you the details and simply indicate that it worked.

huh? By "trackability" do you mean you need someone to hide behind if the s*it hts the fan, or do you mean something else, and if so what?

What do you think will happen to QNX if something happens to BBRY/RIM?

Best regards, Spehro Pefhany

Yes that seemed peculiar to me too. Typically an open source project has every line of every commit signed off and "trackable". As opposed to some opaque binary blob that you just have to hope works. (And the supplier likely explicitly disclaims all warranty on).

But I assume Joerg is talking more about something like "certified".

With software licensing, think about it like going to a car dealership :-)

Seriously, it's like with components. A good purchaser can get you that hot RF transistor for 23 cents while the 10,000 qty price on the distributor web site says 50 cents.

When negotiating, _always_ have an alternative and don't forget to mention that on occasion. "Well, we plan to do that with Windows CE but wanted to first have a look at other ...".

Ok, then chalk that one up to customer incompetence. It's rare but can happen. The challenge is how to diplomatically solve the pickle that this gets everyone into.

Don't do that in aerospace or medical :-)

And then the customer is unhappy, buys the next heater from the competition. And the one after that. This is what I'd do. Things like that go on a black list here and it is next to impossible to get off the list.

Again, doesn't seem to work with me. When I need something I nearly always start out on the web. Clean slate, look at specs, reviews and so on. At that point I usually have no clue what brands I am even looking at.

Simple, my wife taught me: By list of ingredients, price and coupon values. She also knows exactly what will go on sale when.

Well, I don't :-)

Who is Morris?

Doesn't say anything about subliminal ads though.

But all this requires that the person looks at ads and billboards in the first place. I generally don't.

No, never. Same with my wife.

Can't say that this ever happened to me. What did happen was that, for example, I went to a hardware store or a web site with my shopping list, found something else that I liked but that wasn't on the list, and bought it along with the other stuff.

They also do that with candy. Knowing that there will be lots of kids in tow, kids can't wait in line, but there are lines. So ...

What I mean (and what's required by law) is that you must be able to trace back everything in the production chain. All the way down to lot number, place of production, personnel involved in producing this lot, and so on. In med tech one gets audited a lot for that, it's normal.

It goes much farther than that. Take med tech as an example: If there is a case where something bad happened or even just in an audit, the auditors will trace back and drill down, including figuring out if all employees involved in a particular production step (the one that was found to be the culprit) had adequate training. Then they will want to see the actual training records. Try that with open source.

Exactly that is what has me concerned lately. Had a similar discussion with a software pro in Europe.

Well, can you really do that with any kind of software you did not write (or directly control the development of)? If I find a bug in QNX, how do I then trace back and drill down etc? Is there really anything behind it in practice other than a certificate?

It's been over 10 years since I dealt with QNX but yes, that was one of the key questions we had before taking the plunge. QNX still hols seminars on the topic:

I have been through many audits. They are just routine things in many fields but it still feels a bit like a root-canal every time. In the end the difference is this: Properly run corporations in medical, aerospace and other safety critical fields are run with a complete set of standard operating procedures (SOPs). You can hold them to that because once released they must follow these SOPs. I certainly do not want to belittle the open source environment (using Open Office myself) but you typically do not have such guarantees there. In med/aero your whole entrepreneurial future can depend on this.

Free-RTOS might have a useful model, there is a freely available version, and if you want certification you can pay for the certified version. AIUI it is the same API (the same product for all I know) so you can get everything working first for zero cost, make sure it does exactly what you want.

It was really just your "trackability" word that seemed wrong; this is the essence of modern open source projects, and it is the closed source products that are totally opaque. I am not really advocating dropping some random freeware students project into a pacemaker :)

It is really not that easy. To give you an example, the FDA requires all medical devices (and especially the software in there, after some not so good events) to have a complete design history file or DHF. In those there have to many many things, would go beyond the scope of this thread. But the key ones that are often missing with open source:

Documented design decisions: Why was a certain method chosen and not another? Were there any tests or experiments to determine this? If so, where are they documented? All involved stakeholders must attend, and this includes in particular the Regulatory and QC department. Most open source communities don't even have one of those.

Design reviews: Must be thoroughly recorded and file, including name of participants, results, action items, assignments for the follow-up process, and so on.

There are about a dozen other requirements for the DHF.

Then there is the hazard analysis. I have never seen one of those for open source software, but of course I am not a SW expert and some might very well exist.

Closed source products might seem opaque but only until you start discussions with the vendor. For some products there is no discussion, either they open up about it or no dice. I have refused or recommended to refuse cooperation for this reason in more than one instance. Another option can be if the respective party has their own audit trail and that is recognized and has been blessed by the respective federal agency. But then I'd like to see some paperwork comfirming this, plus an ironclad written assurance that they remain in compliance.

Europe is a bit easier in that respect but not by a lot.

Gosh, does this mean that i get a Kewpie prize for having a 2560 x 1440

?-)

No. But you do get a prize for making everyone with a smaller monitor feel bad. Mine is only 1920 x 1200. Sniff. Oh wait... is it an OLED screen? If not, it's obsolete.

You get a prize for a unique but expensive solution for nearsightedness.

Big screen TV's usually have miserable dot resolution when used as a monitor. 1377 x 768 and 1920 x 1080 are common, often at 60Hz fixed vertical scan rate. They're made for TV with computahs as an afterthought. A small part of the problem, which will eventually float to the top of the priority stack, is 3D capability. It's really difficult to do 3D with tiny pixels. As usual, technical progress has its critics:

Why 4K TVs are stupid