Prob. a very basic Q re: capacitors

Admittedly, caps don't wear out, but what's the energy density of one? OP is talking about using supercaps in lieu of batteries, and that's just stupid.

And if you have a battery life problem, maybe you should take a look at your _own_ energy consumption. ;-)

Cheers! Rich

Hi Joerg,

Thanks in advance for your advice here; as usual it's quite good.

MMICs... stuff like HMC349s. There are many of these paths and of course it's supposed to be on a tiny little box, so relays are just too big.

120-140dB, give or take -- upper end is typically at 10-20dBm, and you'd like to still be able to hear something down at -120dBm. 3-6dB, although the upper levels would probably require a lot of justification that it can't get any better without huge price/power/size/etc. increases. 1.5:1 ballpark -- 2:1 would probably still be OK.

Mainly price :-) -- hence the viewpoint of doing this more for the educational value than any return on investment.

Additionally, specifications are "typical" -- it's unlikely that anyone would complain if, e.g., we miss the noise figure or isolation or VSR, etc. over various narrow frequency ranges. (E.g., diplexer crossovers, perhaps...)

High hundreds, in all likelihood.

Yeah, but you need a different current for the leg of the T than the arms if you don't want large mismatches, right? (And the arms need to be at least reasonably matched...) That too would tend to suggest to me a microcontroller/DAC approach where you can calibrate the DAC output on a unit-by-unit basis to get the best SWR.

I really need to start building some test boards for this sort of thing (RF switches, attenuators, amplifiers, etc.) and posting the schematics/results on ABSE for commentary rather than just using someone's pre-packaged device all the time...

---Joel

Ok, sound like normal comms receive gear then, or a good scanner.

3dB almost requires a common preamp before the switch to overcome switch losses but that isn't too expensive.

But hopefully not during switching times?

That sure puts a discrete solution into the educational domain.

Why would a mismatch during switching matter? In the off state the receiver input see a pretty high-Z anyhow. Else you'd sacrifice quite some noise figure whether fully integrated or discrete.

Yes, but don't use a.b.s.e. The wonderful NY attorney general has caused most ISPs including mine to kill it. I wrote to him but, as expected, he did not care to answer. That guy must really feel like a hero now, having taken a sledgehammer and smashed one of our windows. I guess a.b.s.e. must be pretty low in US traffic these days.

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

They're still not great. Many use more power than is acceptable for battery power.

IIRC they had BOR but it didn't prevent corruption of the on-chip flash. Again, IIRC, this was clear from the datasheet, but for some reason a lot of people tended to make assumptions without reading the fine print.

Best regards, Spehro Pefhany

--
"it\'s the network..."                          "The Journey is the reward"
speff@interlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com

Hi Joerg,

Yeah, "mid- to high-end commercial receiver" range.

You wouldn't go the discrete route there for a 6GHz amp though, would you? ...when Hittite/Mini-Circuits/Watkins-Johnson/Macom/etc. will sell you a MMIC to do so for no more than a few bucks?

Correct. There is some spec about there aren't supposed to be any switching transients above -110dBm, but it's not clear to me if that's even a "real" spec -- since you're switching during a very brief period of time, wouldn't it make more sense to talk about energy? (Essentially charge injection, like the analog switch guys do down at audio/low RF.) I suppose what they really mean is that -110dBm corresponds to 31.6uV peak in a 50 ohm system, so they don't want to see any glitches above 31.6uV? Presumably you'd test this by terminating all the inputs, running the switch back and forth a bunch, and monitoring the output with a scope... that would need potentially some pretty high bandwidth?

You're right, it doesn't -- my idea of unit-to-unit calibration doesn't make sense.

I didn't realize your ISP was included; sorry to hear that. It really sucks!

---Joel

John Popelish wrote in news:zs6dnZeatLimPRLVnZ2dnUVZ snipped-for-privacy@comcast.com:

Thanks, good to knwo which direction I need to head in - I also ordered a couple more books. AS usual, the more I learn, the more I realize I don't know ;)

Good link, Thanks!

Yup, search terms are often a stumbling-block for me. This referrence is a good start, I'm printing it out (yeah, I'm old-fashioned that way ;) )

Makes sense but I hadn't thought of it those terms, so that helps.

Thanks again :)

- Kris

Rich Grise wrote in news: snipped-for-privacy@example.net:

I saw some solar LED lighting products (esp., "solar bricks" and road-lane- markers) that already use capacitors as batteries. Maybe you ought to find the products online and write to their manufacturers - I'm sure they'd greatly appreciate being enlightened about their stupidity. [snip]

For large quantities I would certainly look into a discrete solution first. A few bucks is a lot of money there. You can buy transistors like the BFP620 for around 30-40 Cents:

After all, a MMIC is usually nothing but an RF transistors or two with some resistors included. In the lower bands they may be cheap enough but one must make sure there is at least one alternative source. At higher frequencies the premiums you pay for the privilege to be able to plop down a pill and be done with the amplifier design become quite steep.

There will always be some glitch energy. You cannot tell all transmitters and stations in and around Oregon to hold it for a second because you intend to switch. If you muffle the ramps the bulk of that energy will just concentrate in the lower bands.

Even if it wasn't I heard that most other major US ISPs did. I thought about complaining but what good does it do if I and very few others would be the lone rangers out there on a.b.s.e. IMHO the politicos are now beginning to cause a competitive disadvantage for our country. Not good at all.

I did write to the NY AG. No response. A behavior which I anticipated. Just like with some AGW zealots who fall silent the millisecond you want hard data. Ok, ok, I am getting off my soap box now.

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

OK, compare that to this part:

(which is actually $3 in quantities of 100). The gain is quite flat -- within a couple dB from VLF to 6GHz -- with a return loss better than 10dB (not spectacular, but OK). What do I do with that BFP620 to get anything close to that sort of gain flatness? The data sheet is showing a 25dB gain variation for the test circuit they used...!

Hey, I thought you were on the "alternative sources don't really matter that much anymore" side of that discussion last week... guess not? :-)

Certainly, but how are you supposed to measure it? The way I suggested -- terminate the inputs and (inside a screen room) look at pulses coming from the output as you switch with a scope?

Thanks for the help,

---Joel

Well, of course you shouldn't shoot for much more than 10dB of gain per stage which is plenty (or almost too much) in front of a mixer. ft is

65GHz, less at currents away from ideal. You can get hotter transistors but for a 6GHz preamp this looks sufficient.

My usual RF stages have a feedback resistor in the emitter with a series RC "peaker" across it if I really have to push it.

Nope. Only for small production runs. Even then I am careful, don't want my clients sitting dry when they want to order another run a few years down the road. With some bad luck that ritzy RF chip may be gone because a certain cell phone variety fell from grace with the consumers.

You could do that, or use a spectrum analyzer with defined averaging. In the end I'd look at the application, try it out, increase the glitch energy and find out where the pain threshold is.

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

Here is something to play around with the parameters and all kinds of SMPS architectures:

It's simplified, no losses take into consideration, but very good to find out whether your design roughly hits the ballpark or not.

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

And that can change a 25dB gain variation down to, say, 3dB over the span over four decades? I'll have to try it...!

Hey, back when you were mentioning that ultrasound machines you had helped to design twety years ago were still in use... how many ultrasound machines does a good manufacturer sell in a year? Thousands? Tens of thousands? If you make them last 20 years, it doesn't seem as though it would be up in the millions like, e.g., big screen TVs are... :-)

What are the "consumables" with an ultrasound machine anyway? Do consumables from one manufacturer work on most other manufacturers' machines?

---Joel

Yep, that would be the objective. Transistors are just like opamps in that their open loop gain rolls off towards higher frequencies. Feedback flattens that until you reach the point where it loses steam.

It's more like in the thousands. The smaller and lower in cost the more they make.

The scanheads will mechanically wear off over time. Face rubs though, cable jacket stiffens and cracks, that kind of thing. But that takes many years of serious every-day use.

On this one I did the analog front end parts:

Here you have sterile catheters that can only be used once on one patient so there is large sales volume in disposables. If you have access to IEEEplore (I don't) this describes the technology in a nutshell:

?arnumber=921550

Looks like this, just imagine, there are five ICs inside a tip of under

1.5mm diameter:

So, if you ever develop a complicated blood vessel occlusion where they need an ultrasound from the inside out to place a stent you'd be in good hands :-)

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

Yeah, although I've always thought of the resistor in the emitter approach as being more about stabilizing the bias point of a transistor rather than the actual gain -- I've done that on simulated (SPICE'd) op-amps, where the overall feedback/gain determination was the usual explicit negative feedback from output to inverting input.

I'll play around with it some and see what thappens...

Looks impressive; very neat.

Have you done any work on automated defibrillators? The prices ($1-2k) still seem a little high, but definitely within the reach of even small companies today. (Philips must be using some awfully fancy metal for their wall-mount bracket:

-

- $78!? For a piece of bent sheet metal with a few punched depressions/holes and a sticker on the front!?)

What does the doctor use the ultrasound results for -- to determine the size/thickness/etc. of the stent? (To tell you just how many cheeseburgers too many you ate before coming to see him? :-) )

I would imagine there's a certain sense of pride and happiness that comes with working on devices that pretty directly help people live better...

---Joel

The total Z present in the collector circuitry versus the total Z down at the emitter is what determines the gain of a transistors stage. So if Z at the collector is four times the emitter-Z you'll get 12dB. Of course, at some point internal emitter path resistance, capacitances and so on will play into the equation. Plus you'll usually need to get back down to 50 or 75 ohms.

-

Well, there is all the liability insurance. Joe Sixpack mounts it to drywall. Crumbles off, the whole thing falls onto his foot. He sees a lawyer of the ambulance chaser type run an ad, calls the number, lawyer sees medical = deep pockets, well, there it goes. What Americans often do not realize is that in the end it's not the "big fat deep-pocketed medical manufacturer who can afford it" will pay. The consumer will.

Many things: See whether the plaque is about to flake off (VERY dangerous), what size stent would be best, whether a stent is applied correctly or needs some additional balloon inflations to become nice and snug with the artery wall.

Yeah, there were some real heart-warming stories. Such as a clinical trial of a not yet FDA approved device in Europe. Not allowed to name the client but according to the medical folks it went like this: Guy got told by doc that it's over, meds don't work anymore and surgery is impossible. Only pain meds, eventually followed by death. He was so weak that he couldn't do the simplest chores like walking over to the sink to wash a tea cup. But there'd be this high-falluting device from the US, so if he'd signed his life away on that waiver sheet ... he did sign. Weeks later he was back on his racing bike.

Then the not so good remembrances. When we were done with an ultrasound machine one of the sales guys tested it, by coincidence happened to look at his thyroid and said "Oh s..t!". He found his own thyroid cancer. Didn't make it :-(

Unfortunately the liability situation in the US and the fact that all underwriters declined to quote because of med design on my apps I am slowly migrating to areas outside medical. But I'll always have one foot in medical, grew up there. Also met my wife there. She was the marketeer assigned to test the very first machine I was involved in.

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

Joerg wrote in news:1Ppkk.20096$ snipped-for-privacy@nlpi068.nbdc.sbc.com:

Hey, thanks! :)

I've given a quick look at seceral of teh explanations, and they look pretty decent - seom intermediate explanation, plus, for experts, soem of the physics. Nicely-done. It's also interesting that the page has teh example circuits organized taht way, becasue it means I can compare others with them and, if there are parallels, that should help me better understand them, or at elast the context so to speak.

My books should be here by 4 Aug at the latest, so I'm also looking forward to their arrival.

ANy recommendations for toher good "self-learnign" or "intermediate" books are always welcome as well ;)

I'm sure that I'll have to modify my opriginal idea about the brightness- level, but I'm also certain that I can do it, if I keep at this, rather than giving up just because it's difficult...which it is: I'm a visual- spatial thinker, quite right-brained, so my strengths are arts, information analysis, and learning languages, so things related to math and computer programming are *very* difficult for me. OTOH, once I got my AARP card, I found that the older I get, the more cantankerously pigheaded I become and I've made up my mind that I *will* manage to figure this out.

I very much appreciate all of the links, references, and other positive input :)

Thanks again,

- Kris

[...]

Which brings up a point: What advantages are there to becoming an AARP member? I've heard that there'd be an increase in junk mail, plus it costs membership fees, plus they are sometimes politically a bit liberal for my taste. Took one look at their health insurance stuff, not much to write home about that would work in our country IMHO. So I wonder whether it's worth it. Maybe Jim and a few other "AARP-age" folks could also chime in ;-)

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

Ah, gotcha. (As you're undoubtedly well aware, with op-amps the deal was that Z at the collector (drain) was as high as possible -- generally a current source -- in order to maximize gain -- "there's no such thing as too much gain in an op-amp?")

They declined to quote libability insurance for you? I.e., if some machine you designed blows up and injures me, I can perhaps successfully sue you -- even though you were working as a consultant to the equipment manufacturer -- and if I win I get your pool and the chance to fix your sprinkler system myself? Hmm...

Thanks again for the help and the stories!

---Joel

(Recalling op-amps again...)

I had a professor who said you should design op-amp stages for ~30dB per stage -- hence the typical two-stage op-amp gets you the not-so-at ypical 60dB gain op-amp. (At low frequencies, of course.)

The guy had previously worked at Analog Devices, although if he ever designed any real product there he never mentioned it. :-)

He always claimed that our design assignments were "something [he] could design in a few hours" but then with "the specifications reduced maybe 10%."

I was never certain whether or not that was true, especially since he never provided his own solution to the assignments -- he always chose one of ours that he liked best and made that the "official" solution.

This and his rather brusk personality caused some students to think that he might have been a case of, "those who can do, those who can't, teach."