God yes! tell me about it, when the reception is good its very good, but its so temperamental and the glitches are extreme! during times of atmospheric disturbance, even just strong winds, they can create transients that could easily damage speakers, continuous dropouts, gaps in transmission that make it impossible to follow a sentence etc.
It is so much worse than analogue TV, I can only imagine what its like in more remote areas. i'll take a slightly imperfect image, some noise, a shadow etc well before these abhorrent issues that make me scared to connect my DC coupled system up in the living room.
As a result I tend to watch things on replay on the internet archive, rather than live if I can, and only watch news or sport live.
Anemometers are interesting but probably too difficult. However MEMs microphones are getting very good. because of the processes they tend to all track within less than 1 dB. They are also very small and available with digital outputs. They don't have the SNR of the best measurement mikes but may be good enough. Conceivably a sound intensity system could be built with 2 (or more ) MEMs mikes and something like an ADI ADAU17xx DSP processor (that can connect directly to the digital MEMs mikes) and make a sound intensity system for less than $20.
QUOTE=jan.didden;3434588]... or a super-duper car that just decides to stop because of a softwarebug that went undetected
Parts of the US west are known as big sky country because the land is flat there are few trees or other obstructions. So when you are standing and looking around you are the tallest thing and most of what you see is the sky
Unfortunately my car has a Micosoft logo on the dash for the Bluetooth hands free cellphone. So when I drive in big sky country as soon as my car gets on any rise and all you can see through the windshield is clear blue sky, it stops running. When you try to restart it unlike normal cold starts it takes two to three minutes.
[/QUOTE]To be honest, my audi has a fail-safe mode that allows me to drive it home without all the sophisticated engine management. Consumption will be terrible, and so will accelaration. The management chip reputedly also has a watchdog-timer that is periodically reset by the software. When the reset doesn't happen and the watchdog runs out, it generates an interrupt that resets the software in an attempt to pull it out of its rut.
It's a nice system, I've used it with success in some of my own embedded projects.
jan
Yes Demian. That’s promising.
As with freq response inherent good matching, so may be with phase response with MEM mics.
In any case, differences will be consistent across the lot making probably phase calibration obsolete or turning it into a delay issue which can easily be adjusted by the DSP processor
As for MEM anemometers, if you ever come across one (two) try them. I don’t think I will have the chance to meet them.
George
All Engineers I work with have a vested interest in the layout of their circuits.
Gee I'm a bit disappointed you cant share some of the sophisticated layout techniques with us amateurs, I only do simple using designs techniques like this quite often these days:
http://www.dnu.no/arkiv3/HDI -IPC presentation_Norge 090924-2.pdf
oh and as you like quoting stuff from the past, I also did the PCB's mentioned here many years ago:
http://accelconf.web.cern.ch/accelconf/e94/PDF/EPAC1994_2333.PDF
Just for your interest (and as you love op-amps) here's some 21st century SMD analogue next to an 0.1" pitch 8 pin DIL, the opportunities for small compact sensitive analogue layout with minimal parasitics and minimal EMC target area is readily available these days, with advanced thermal engineering you can also keep the whole assembly cool, minimising thermal distortion,
So I actually have a professional interest in the Blowtorch layout, and have a curiosity as you all went to the expense of using a Teflon based substrate, what else you did...
Attachments
Last edited:
It what I do for a living John, and have done for over 27 years. In fact when it comes to PCB design I am probably a bit of a geek, as I have a deep interest in all aspects of design, manufacture and assembly issues, and I am always interested in how others do their designs and the reasons why. As to analogue layout, most small signal layouts involve ground planes, sometimes one contiguous plane, sometimes separate GND pours for different sections with a starpoint, and often with outer layers GND poured and used as a dirty EMC ground on the outer layers, often routed directly to the chassis. Low noise circuitry requires careful layout, that's one reason I like the new generation of packaging, HDI PCB layout etc, the new packaging is very thermal efficient, a necessity with the amount of stuff that gets crammed on some boards these days.
Elektroj, ha ha wait till I put my 3D mechanical head on, I'll be after the machining program for case, so I can reverse engineer it.
Elektroj, ha ha wait till I put my 3D mechanical head on, I'll be after the machining program for case, so I can reverse engineer it.
That ain't crowded. Positively as open and with as much space between the components as the Serengeti plains. You want to take a look at one of the latest smart phones, or an Ultrabook notebook PCB to see what crowded is.
Nice color representation,
Marce if you been doing these for 27 years does Johns attachment remind you of triple prints?
Thanks
-Antonio
The latest technology is really crowded with 6 and 8 layer PCB and chip scale packages. But they have very little exposed analog. I have been working with a PCB with a Bluetooth chip and 2 full DSP's plus antenna in less than 1 square inch. We could shrink it more if we put parts on both sides but that has consequences. Despite all of the connections (6 microphones, USB and RF) there is very little analog that is not in a chip. Which also means little opportunity to make it better.
The constellation board is much denser than other discrete boards using that generation technology (really, late 80's and early 90's). Surface mount makes for much smaller footprints, but the "good" parts are unavailable in surface mount if at all.
The IC argument will pop up and it's true, the functionality of that board could be accomplished in smaller space with IC's but one real limitation on IC's is dynamic range. Even really low noise IC's are still limited in output swing. Going discrete can permit much higher voltage rails and more max out.
The constellation board is much denser than other discrete boards using that generation technology (really, late 80's and early 90's). Surface mount makes for much smaller footprints, but the "good" parts are unavailable in surface mount if at all.
The IC argument will pop up and it's true, the functionality of that board could be accomplished in smaller space with IC's but one real limitation on IC's is dynamic range. Even really low noise IC's are still limited in output swing. Going discrete can permit much higher voltage rails and more max out.
Thanks Demian for your input. All I know is, NEVER AGAIN! Any board that I have to troubleshoot in future will be more optimally layed out for symmetry and ease of identification of the individual components.
Don't worry Demian, I won't show the other layers or the bottom of the board, even though it has a bit of surface mount.
Demian has more experience in 'troubleshooting' this board than I do. Trust me, it is a 'headache' not just because of the schematic, but finding the individual parts and measuring voltages, as well as replacing them.
This board has been a hard lesson for me, as I only approximate the specific resistances in a design and once a prototype board is made, I optimize the entire circuit. The difficulty of changing the resistor values in this schematic, as well as the surface mount changes has caused an embarrassingly high amount of trouble.
I guess, in future, I will have to computer SPICE modeling, before making anything. I never had to, before.
Don't worry Demian, I won't show the other layers or the bottom of the board, even though it has a bit of surface mount.
Demian has more experience in 'troubleshooting' this board than I do. Trust me, it is a 'headache' not just because of the schematic, but finding the individual parts and measuring voltages, as well as replacing them.
This board has been a hard lesson for me, as I only approximate the specific resistances in a design and once a prototype board is made, I optimize the entire circuit. The difficulty of changing the resistor values in this schematic, as well as the surface mount changes has caused an embarrassingly high amount of trouble.
I guess, in future, I will have to computer SPICE modeling, before making anything. I never had to, before.
Indeed. It's almost as if a cottage industry could spring up around discrete circuits for augmenting the output voltage and current swings of ICs, without fatally affecting bandwidth, slew rate, settling time, and so on.
Of course the notion of resorting to such is not a negative commentary on the skills and ingenuity of IC designers. It's just that to date at least there are no compelling applications to provide an adequate market and thus an incentive to develop such. It was interesting to see somewhere in here the query about the evaluation of the AD797 being based on sound quality, as we tend to suppose it was developed for audio, which Scott pointed out was not the case.
The pernicious effects of obsession with power consumption, coupled with the convenience of integrating analog circuitry with lower-and-lower-voltage digital hardware, have also limited our choices.
I do welcome the ubiquity of surface mount as it allows us to do low-volume designs that come near to integrated performance by reducing parasitic inductances etc. And even changing out parts, once one gets the hang of it, is in some ways easier than working with leaded parts and plated-through holes --- at least with 0805 or thereabouts. Much below that gets tough for manual assembly without specialized equipment.
Magnoman, it reminds me of the tape ups I started on, firstly simple single sided boards with black tape, then up to red and blue tape ups (which I presume are similar or the same to triple prints, photographic techniques to drop one colour or another), then after a year slaving at theses and drafting the schematics (A0 plus drawing board), I was allowed on the CAD system (Racal Redacs Redboard, now Cadstar), heaven. I don't miss the old days though, a lot of exciting developments (though in some aspects negative see #37616 above) brought on a lot by the sheer market volume of hand help devices has shrunk devices to often silly levels. I now do boards where 0603 chip components look huge. The reduction in both size and voltages does add its own problems, recently did an ADC with a 1.2V swing!
A problem also for redundancy, for equipment that has to run 15 years or more, trying to get a circuit to work the same today with new parts causes engineers headaches, on the same note, see a lot of older designs coming in to have EMC components added to I/O specially to cables.
Well, layout is NOT my thing. I leave it to others, people who have a proven record. We only miniaturize circuit boards when it is absolutely necessary.
Surface mount has a VERY LIMITED selection, these days, for my circuit designs, at least.
My biggest headache is CASES. It is difficult, if not impossible, to make a case that is attractive (at a world class level), cost effective, and useful as a shield, enclosure, heatsink, and stable platform for switches.
Most of the time, these days, I leave the CASES to others who I consult for. They do a pretty good job at making the tradeoffs, if necessary.
I will stick to fundamental analog design, with or without IC's, leave the board layouts to Carl, and hope to improve my troubleshooting skills, that have declined since I have aged and lost an eye. I sure wish there were some skilled techs or jr engineers who could do the things that I once had to do, when I was younger, and save myself as an overseer of the circuit design, rather than have to do the mechanics of it. I can still read test results easily enough to optimize the circuitry.
Unfortunately, there are many here, especially the vocal ones, who believe that design efforts, like my own, are redundant, and a waste of time. Fortunately, I do hear from others, who think differently, and I still hear differences in IC's, preamps, power amps and EVEN cables (at times), that tell me I am still on the right track. I'll just 'keep on truckin'. '-)
Surface mount has a VERY LIMITED selection, these days, for my circuit designs, at least.
My biggest headache is CASES. It is difficult, if not impossible, to make a case that is attractive (at a world class level), cost effective, and useful as a shield, enclosure, heatsink, and stable platform for switches.
Most of the time, these days, I leave the CASES to others who I consult for. They do a pretty good job at making the tradeoffs, if necessary.
I will stick to fundamental analog design, with or without IC's, leave the board layouts to Carl, and hope to improve my troubleshooting skills, that have declined since I have aged and lost an eye. I sure wish there were some skilled techs or jr engineers who could do the things that I once had to do, when I was younger, and save myself as an overseer of the circuit design, rather than have to do the mechanics of it. I can still read test results easily enough to optimize the circuitry.
Unfortunately, there are many here, especially the vocal ones, who believe that design efforts, like my own, are redundant, and a waste of time. Fortunately, I do hear from others, who think differently, and I still hear differences in IC's, preamps, power amps and EVEN cables (at times), that tell me I am still on the right track. I'll just 'keep on truckin'. '-)
This is what we call a 'crowded' analog board.
Hi Guys,
Layout is fun for me, especially if I am not under any schedule pressures.
Layout brings lots of challenges to a designer. It really sucks when you have to do it again, for one reason or another, kinda have to allow for at least two rev's before Mfg Release. I remember back at HP, the average was a between 1-2 pcb rev's, which is darn good. One design that we did had a brain dead ASIC on it, I did 6-8? pcb rev's before MR release, made lots of OTime$ on that project. A mass of FPGA's on it to fix the prob.This was done using Mentor BoardStation.
In this latest design, "Portable Media Player" I have done, as shown, a mixed signal design. Tuner(Si4735), MP3/WMA/AAC memory stick player(BU9458), USB CODEC(PCM2902B), headphone amp(MAX9729), 25W Class "D" amp(TPA3100D2). Amtel ATxmega MCU. This done in ORCAD.
It blows me a way, that I can do the same & a whole lot more in this small form factor than an ole receiver. I look at my old Pioneer SX-950, still selling on eBay for almost the same, I paid for it in 1978 and get a chuckle.Okay this portable ain't a 100W/ch Class "AB" amp, but for the majority of your listening it is fine.
Right now I am listening to Van M, playing the CD in my computer, through the USB to the portable, on to my Dynaudio Gemini's, the sound is great and this all that really matters. I am drawing about 3W idle. Maybe double if I crank it.
I am an advocate of jamming as much functionality into a small package as possible, considering this is a design for my own use.
This is what I call AudioDIY for myself at least, say compared to what I did for a college project 30+ years ago.
Anybody interested in one? Assembled or bare pcb? Just kidding, I know this ain't the commercial section. Actually I need more help with SW than HW, of late.
Rick
Hi Guys,
Maybe in the days of old, but not by today's standards.
Layout is fun for me, especially if I am not under any schedule pressures.
Layout brings lots of challenges to a designer. It really sucks when you have to do it again, for one reason or another, kinda have to allow for at least two rev's before Mfg Release. I remember back at HP, the average was a between 1-2 pcb rev's, which is darn good. One design that we did had a brain dead ASIC on it, I did 6-8? pcb rev's before MR release, made lots of OTime$ on that project. A mass of FPGA's on it to fix the prob.This was done using Mentor BoardStation.
In this latest design, "Portable Media Player" I have done, as shown, a mixed signal design. Tuner(Si4735), MP3/WMA/AAC memory stick player(BU9458), USB CODEC(PCM2902B), headphone amp(MAX9729), 25W Class "D" amp(TPA3100D2). Amtel ATxmega MCU. This done in ORCAD.
It blows me a way, that I can do the same & a whole lot more in this small form factor than an ole receiver. I look at my old Pioneer SX-950, still selling on eBay for almost the same, I paid for it in 1978 and get a chuckle.Okay this portable ain't a 100W/ch Class "AB" amp, but for the majority of your listening it is fine.
Right now I am listening to Van M, playing the CD in my computer, through the USB to the portable, on to my Dynaudio Gemini's, the sound is great and this all that really matters. I am drawing about 3W idle. Maybe double if I crank it.
I am an advocate of jamming as much functionality into a small package as possible, considering this is a design for my own use.
This is what I call AudioDIY for myself at least, say compared to what I did for a college project 30+ years ago.
Anybody interested in one?
Assembled or bare pcb? Just kidding, I know this ain't the commercial section. Actually I need more help with SW than HW, of late.Rick
Attachments
- Status
- Not open for further replies.