There is a large amount of literature on the design of old-style radio circuits (MF and HF, eg shortwave), but surprisingly little available on the design of modern VHF/UHF consumer electronics.
In particular I found it interesting how the manufactures coped with the crazy proliferations of different TV broadcast standards around the world.
The authors home page at https://www.maximus-randd.com/ looks like being a huge source of interesting reading for RF enthusiasts.
I found this page when searching for the ICs in (what I now know is) a tuner module. Surprised that this site has never been posted to HN before. I'm working on FMCW radar related stuff so it was super interesting to me too.
I found it sometime last year, looking for data on a Philips video IF processing chip, I thought about posting it. Great site from an interesting guy, a rather sad history of the decline of electronics manufacturing in western europe though, death by bean counters.
What was the IC you were looking for, PLL synthesizer or Demodulator?
I was just curious about the chips inside one of these tuner modules, so I searched up their label and went to images. The two things are TDA5737 and TSA5522, which seems to be mixer and synthesizer.
I think it's because old radio circuits were topologically fairly simple but designed using priceless experience and big budgets (e.g. tangentially Tektronix had a scope with synthetic sampling rates in the THz in the 60s or so), whereas modern stuff is becoming - even in simple designs - harder and harder to analyse as margins are being chased.
You can usually work out the basics by finding out the main working chips in a product and reading their application notes - as long as it's not Chinese it should be well documented.
I assume GP is talking about equivalent-time sampling. ETS is a technique for extending sampled system bandwidth at the expense of requiring a perfectly repetitive waveform. Consequently, it's useless for single-shot captures, where you only get the native sample rate of the system.
ETS used to be much more common in the days before (affordable) digital oscilloscopes could directly sample at high enough frequencies to be useful.
Tektronix 7T11 system - up to 10 ps/div (at ~15 GHz or so analog bandwidth with the right sampling head), ~250 samples per trace (~10 divs wide), so one sample every 100/250 ps = 0.4 ps = about 2.5 THz ETS rate. Whether this really says so much about the system ... I'm pretty sure that the sampling jitter exceeds 0.4 ps by quite some margin. In a system constructed like this it is actually really easy to turn the ETS rate up to incredibly high values; whether the data that generates is useful, is another question.
"(e.g. tangentially Tektronix had a scope with synthetic sampling rates in the THz in the 60s or so), whereas modern stuff is becoming - even in simple designs - harder and harder to analyze as margins are being chased."
Yeah, yet another sad engineering story is modern documentation—or more precisely the almost complete lack thereof. One of the great things about Tektronix was that it was an electronics instrument company run by engineers—engineers who were truly committed to both technical excellence and the highest quality of manufacturing [at least that was my experience of the company back some decades ago when I was buying its CROs and analyzers—I'm unsure of the current situation].
Rigorously complying with technical standards was a Tektronix mantra—and that included supplying customers/users with top-notch, high quality documentation. Tektronix user and service manuals were legendary for not only being both comprehensive and clear to understand but also for their ergonomics—you could find any info needed about an instrument within seconds (the manuals were ergonomic design masterpieces).
Unlike most of today's equipment where most of the technical information about a device is considered proprietary and held locked up by corporations, Tektronix manuals provided the customer with just about everything a customer wanted to know: detailed technical descriptions of its electronics operation, clear and comprehensive circuit diagrams with layouts including parts cross references together with properly indexed parts lists for purchasing spares. [Right, back then, there was none of this current fight about the 'Right to Repair' as that was the assumed norm by both users and manufacturers alike - yes, we had the 'Right to Repair' and it was stolen from us.]
That Tektronix published such detailed information for everyone to read is proof that patents, copyrights and manufacturing trade secrets alone could protect the company's IP. It demonstrates the fact that it is unnecessary to hide almost everything about equipment except its bare operational details.
Tragically, the problem about the paucity of information about equipment coming from manufacturers these days has now been around so long that both younger users and engineers simply do not know what decent equipment documentation looks like or entails. For those who don't then I'd suggest you have a look by downloading the Tektronix 7834 oscilloscope manual from the Internet Archive (this is a device that I'm very familiar with):
(Note: for a better quality presentation, download the PDF version without the embedded text; alternatively download both versions.)
INCIDENTALLY, NOTE THAT 'TROUBLESHOOTING', P-4.3; 'CORRECTIVE MAINTENANCE', P4.12; AND THE SPARE PARTS LIST, P7.1, ARE ALL COMPLETE AND WITHIN THIS ONE SINGLE MANUAL. IT'S PROOF THAT BACK IN THE 1980S AND BEFORE THAT THE 'RIGHT TO REPAIR' WAS AUTOMATICALLY ASSUMED BY EVERYBODY!
Finally, I'd add that today's practice of locking up equipment information behind corporations' firewalls—information that only a few decades ago was freely available from many tech manufacturers—has the negative effect of deskilling* the population. Not only this but also it wastes considerable time and effort as users have to try and discover (or reverse engineer) info that rightfully ought to be in equipment manuals (consider the total productivity loss when taken across the full population of users).
* For example, Tektronix manuals were wonderful training aids for those wishing to understand how an oscilloscope works. With the aid of a second oscilloscope and the manual of the first, one, say, could trace the passage of signal from the vertical input amplifier to the defection pla...
Although the kind of training-documentation a la Tek is dead, Analog have some rather nice docs for the things they care about - for their SDR-on-chip chipsets they have long form (600page) books to read.
There's also the problem that a modern scope is just a maze of ASICs, FPGA blobs, etc.
Right, I know. That's another serious issue much too big to discuss here except to mention two well-known examples that highlight the problems:
1. There's are problem with having compiled, thus obfuscated, code in embedded devices such as you mention. The high profile case being the Volkswagen emissions affair. Given the potential for vehicles (normal or EV) to kill people, there's a strong case that all such embedded code in mission-critical applications ought to be open whether the manufacturer likes it or not.
2. Backdoors such as Intel's Active Management Technology/AMT/IME should not be allowed by law unless all aspects of their APIs are published openly.
It's not only high profile stuff such as I've mentioned either. Several weeks ago I bought a NETGEAR LB2120 4G/LTE modem. I know nothing about the information it sends back to NETGEAR other than it actually does so by virtue of the fact that it advised me of an update (no other information about the update was supplied). It seems to me that, as user, I not only have the right to know what information is sent back but also that I should have the option to block it—and that this right should be enshrined in law.
Guess he is more familiar with old tuners than with old computers :) Also, a quick Wikipedia search could have easily prevented this: https://en.wikipedia.org/wiki/Amiga_4000
from: https://en.wikipedia.org/wiki/Amiga_4000
The three ISA slots can be activated by use of a bridgeboard, which connects the Zorro and ISA buses. Such bridgeboards typically feature on-board IBM-PC-compatible hardware, including Intel 80286, 80386, or 80486 microprocessors allowing emulation of an entire IBM-PC system in hardware.
from: https://en.wikipedia.org/wiki/Motorola_68040
The 68040 offered the same features as the Intel 80486, but on a clock-for-clock basis could significantly outperform the Intel chip in integer and floating point instructions. However, the 80486 had the ability to be clocked significantly faster without suffering from overheating problems.
Amazing site -- safe to say this info doesn't exist anywhere else, at least not online. It covers some of the earliest TV development work in the 1930s all the way to satellite/DBS in the 2000s.
Worth looking at the older stuff, just for the examples of European industrial design in the WWII and later eras.
I remember the analog era. That lasted quite a while and it worked well (bad signal just meant a slightly more snow in a picture). Then on digital era I have already transitioned from DVB-T to DVB-T2. When changing a channel in those 70ies or 80ies televisions, it happened instantly. Now we have huge amounts of computing power, but every press on remote feels like the command is queued somewhere for a second or two. I hate using today's televisions and tuners. While the picture and sound is considerably better, the usability is horror. Also Teletext worked always fast and was instant. Today, if you try to open simple program guide, it feels like it is compiling a kernel on a background.
This is covered somewhere in the ~middle of the article:
>Because a block of 12 pictures are jointly interpolated and coded this method is also called Block Coding. A block takes roughly 0,5s, which is then also the minimal lock time of a receiver.
>Intra refresh periods of a half-second are common on such applications as digital television broadcast
Cable TV providers are totally screwing the pooch here. Most modern Cable Boxes employ reverse channel to spy^^^report telemetry and offer additional features. This means all the hardware and software infrastructure is already in place to request dynamically generated Inter frame from the backend on the fly and skip this up to 12 frames wait time.
Of course the other part is "modern" stupid software stacks, with TVs running Android, and probably rendering UI using web technologies.
I have the same feelings when it comes to the analog VGA connections, especially on projectors. You could have a mangled connector with poor quality cabling and in the worst case, the image would degrade, have ghosting and maybe miss a color channel or two, but you could still get your message on screen since the analog receiver will try to lock on to any signal it gets regardless of how poor the quality is, while with digital, it either works or it doesn't, there's no middle ground, and with connectors that have seen much usage it tends to not work more often than it does. If the protocol handshake or something else fails from the start it's game over. Analog degrades much more gracefully.
Yeah it's crazy how much worse call quality has become relative to landlines. It's so bad that it's a borderline accessibility issue at this point. Would be interesting to know what public input there is when it comes to bandwidth optimization. There also isn't a good quantitative metric for measuring call quality in a way that could spur competition between carriers. It seems to mostly be a race to the bottom.
Digital video compression uses several frame types. Most frequent ones are inter-frames, and those encode the difference from the previous frame — so if you don't have those previous frames, you can't decode a usable picture. You have to wait for the next intra-frame to be transmitted, which contains a complete picture by itself. Analog video doesn't have that restriction, every single frame is self-sufficient.
> Also Teletext worked always fast and was instant.
That's not at all how I remember it. I remember it taking anywhere from instant to 15 seconds to load a page.
"...but every press on remote feels like the command is queued somewhere for a second or two. I hate using today's televisions and tuners.
I agree completely, I find this delay more than annoying—in fact, it's infuriating! I consider this delay not just an ergonomic limitation but also a definite design fault (i.e.: the DVB standard allows the delay to occur). The fact that regulators authorized DVB/TV digital services with this fault speaks more to me about the commercial pressure brought to bear on the regulators (to accept the limitation) than it does about any proper consultative/evaluation process (user ergonomics evaluation) having been properly undertaken before these services were introduced.
(…But then, I suppose we can't have expected much better from DVB standards given that in the 1980s many governments around the world essentially nuked their spectrum management departments and outsourced much of the work to commercial organizations that had vested interests—thus, effectively they were committed to lower standards. [This is a long sad story which, unfortunately, is much too involved to discuss here.])
Back before the RTL-SDR era, there were a number of writeups on building homebrew spectrum analyzers out of TV tuners. Example: https://www.qsl.net/n9zia/spec/w6hph.pdf
This is a wonderful site and I would highly recommend it to anyone in TV or RF engineering. By that, I mean the complete site not just the TV tuner history pages (i.e.: https://www.maximus-randd.com/). This is how the internet should be used to convey information. Not only is the information therein both comprehensive and very detailed but also the 'cleanliness' and clear layout of the pages across the site is second to none. It's a long while since I've seen such a well presented and well laid out site (even JavaScript is used sparingly and not abused as is usually the case across much of the web). When I create my next website, I'll certainly keep the presentation of this site in mind.
The information about tuners is remarkably comprehensive (it's certainly the best I've seen in a single documented collection on the subject). I don't work with TV tuners on a day-to-day basis but I've had a lot to do with them in the past—and that goes back to the vacuum tube/turret type well before we had varicap* tuning (so I feel competent to comment on the site).
That said, there are stacks of valuable information on this site about tuners and other stuff that I was either unaware of or not very familiar about. There's nothing more I can add as to do so would be superfluous—as the excellence of this site speaks for itself.
_
* [I recall initially hating varicap tuning due to its propensity to suffer cross modulation (which meant that one had to ensure that TV channels had roughly the same RF input levels and that strong out-of-band services were kept out of such tuners by installing filters/traps in the antenna line before the tuner input—lest herringbone patterns result).]
Fascinating view into the history of Philips RF developments in 1990-2000 timeframe. Absolutely loved it!
WSP (World Standard Pinning) "standardization" story with management naive dreams of selling own products in Asia without cannibalizing home European market while simultaneously moving Consumer Hardware Development to Asia .. where all the hardware was thereafter designed with Asian parts first was priceless.
Philips TV division deciding to second source, skipping Philips Tuners division thus eroding own company market share and profit. Then in turn Philips Tuners division commissioning a cheaper clone of Philips MOPLL from Siemens with a huge minimal volume quota, consequently eroding Philips Semiconductors market share by 25% and forcing Tuner division to manufacture inferior products with already obsolete part. Just beautiful, <Chef's Kiss>, Pure genius!
> The most devastating development, however, was purely internal and of severe structural impact.
'The history of a pioneering tube that was developed by Philips Research and that was, next to the magnetron, “The most important tube from World War II” - Ronald Dekker'
One of the problems Bowen was facing was that while early 1939 all components of the Airborn Radar seemed to have fallen in place, he still had only one receiver which up to that moment no supplier had been able to duplicate. But then quite unexpected his old professor comes up with a golden tip:
"Quite by chance in april or May of 1939, I heard some encouraging news from Edward Appleton, my old professor at King’s Colledge and now Jacksonian Professor of Physics at Cambridge. He told me that the Pye Radio Compagny, still hoping that there would be a television industry in Britain, had set up a production line for 45 Mc/s TRF chassis and had actually made a trial run. I went hot-foot to Cambridge to see B.J. Edwards, the Technical Director of Pye Radio, and was rewarded with a remarkable sight – he had scores [!] of TRF chassis just the type we were looking for. These used a new valve with an octal [sic] base, which had not yet appeared on the market. It was the EF50, a valve which was destined to play almost as important a part in Radar war as the magnetron itself."
Bowen took a few samples of the Pye chassis back to Bawdsey and quickly verified that it was significantly better than their old EMI chassis; it was also smaller and lighter. Touch and his men quickly put a 200 MHz mixer (at that time still based on an acorn tube, but later replaced by an EF54) in front of it and the receive problem was as good as solved.
29 comments
[ 3.5 ms ] story [ 24.9 ms ] threadThere is a large amount of literature on the design of old-style radio circuits (MF and HF, eg shortwave), but surprisingly little available on the design of modern VHF/UHF consumer electronics.
In particular I found it interesting how the manufactures coped with the crazy proliferations of different TV broadcast standards around the world.
The authors home page at https://www.maximus-randd.com/ looks like being a huge source of interesting reading for RF enthusiasts.
What was the IC you were looking for, PLL synthesizer or Demodulator?
You can usually work out the basics by finding out the main working chips in a product and reading their application notes - as long as it's not Chinese it should be well documented.
ETS used to be much more common in the days before (affordable) digital oscilloscopes could directly sample at high enough frequencies to be useful.
This was a fully analog system hence the ~.
Yeah, yet another sad engineering story is modern documentation—or more precisely the almost complete lack thereof. One of the great things about Tektronix was that it was an electronics instrument company run by engineers—engineers who were truly committed to both technical excellence and the highest quality of manufacturing [at least that was my experience of the company back some decades ago when I was buying its CROs and analyzers—I'm unsure of the current situation].
Rigorously complying with technical standards was a Tektronix mantra—and that included supplying customers/users with top-notch, high quality documentation. Tektronix user and service manuals were legendary for not only being both comprehensive and clear to understand but also for their ergonomics—you could find any info needed about an instrument within seconds (the manuals were ergonomic design masterpieces).
Unlike most of today's equipment where most of the technical information about a device is considered proprietary and held locked up by corporations, Tektronix manuals provided the customer with just about everything a customer wanted to know: detailed technical descriptions of its electronics operation, clear and comprehensive circuit diagrams with layouts including parts cross references together with properly indexed parts lists for purchasing spares. [Right, back then, there was none of this current fight about the 'Right to Repair' as that was the assumed norm by both users and manufacturers alike - yes, we had the 'Right to Repair' and it was stolen from us.]
That Tektronix published such detailed information for everyone to read is proof that patents, copyrights and manufacturing trade secrets alone could protect the company's IP. It demonstrates the fact that it is unnecessary to hide almost everything about equipment except its bare operational details.
Tragically, the problem about the paucity of information about equipment coming from manufacturers these days has now been around so long that both younger users and engineers simply do not know what decent equipment documentation looks like or entails. For those who don't then I'd suggest you have a look by downloading the Tektronix 7834 oscilloscope manual from the Internet Archive (this is a device that I'm very familiar with):
https://archive.org/details/tektronix_7834
(Note: for a better quality presentation, download the PDF version without the embedded text; alternatively download both versions.)
INCIDENTALLY, NOTE THAT 'TROUBLESHOOTING', P-4.3; 'CORRECTIVE MAINTENANCE', P4.12; AND THE SPARE PARTS LIST, P7.1, ARE ALL COMPLETE AND WITHIN THIS ONE SINGLE MANUAL. IT'S PROOF THAT BACK IN THE 1980S AND BEFORE THAT THE 'RIGHT TO REPAIR' WAS AUTOMATICALLY ASSUMED BY EVERYBODY!
Finally, I'd add that today's practice of locking up equipment information behind corporations' firewalls—information that only a few decades ago was freely available from many tech manufacturers—has the negative effect of deskilling* the population. Not only this but also it wastes considerable time and effort as users have to try and discover (or reverse engineer) info that rightfully ought to be in equipment manuals (consider the total productivity loss when taken across the full population of users).
* For example, Tektronix manuals were wonderful training aids for those wishing to understand how an oscilloscope works. With the aid of a second oscilloscope and the manual of the first, one, say, could trace the passage of signal from the vertical input amplifier to the defection pla...
There's also the problem that a modern scope is just a maze of ASICs, FPGA blobs, etc.
Right, I know. That's another serious issue much too big to discuss here except to mention two well-known examples that highlight the problems:
1. There's are problem with having compiled, thus obfuscated, code in embedded devices such as you mention. The high profile case being the Volkswagen emissions affair. Given the potential for vehicles (normal or EV) to kill people, there's a strong case that all such embedded code in mission-critical applications ought to be open whether the manufacturer likes it or not.
2. Backdoors such as Intel's Active Management Technology/AMT/IME should not be allowed by law unless all aspects of their APIs are published openly.
It's not only high profile stuff such as I've mentioned either. Several weeks ago I bought a NETGEAR LB2120 4G/LTE modem. I know nothing about the information it sends back to NETGEAR other than it actually does so by virtue of the fact that it advised me of an update (no other information about the update was supplied). It seems to me that, as user, I not only have the right to know what information is sent back but also that I should have the option to block it—and that this right should be enshrined in law.
wait, what? :o :-)
Guess he is more familiar with old tuners than with old computers :) Also, a quick Wikipedia search could have easily prevented this: https://en.wikipedia.org/wiki/Amiga_4000
from: https://en.wikipedia.org/wiki/Amiga_4000 The three ISA slots can be activated by use of a bridgeboard, which connects the Zorro and ISA buses. Such bridgeboards typically feature on-board IBM-PC-compatible hardware, including Intel 80286, 80386, or 80486 microprocessors allowing emulation of an entire IBM-PC system in hardware.
from: https://en.wikipedia.org/wiki/Motorola_68040 The 68040 offered the same features as the Intel 80486, but on a clock-for-clock basis could significantly outperform the Intel chip in integer and floating point instructions. However, the 80486 had the ability to be clocked significantly faster without suffering from overheating problems.
Worth looking at the older stuff, just for the examples of European industrial design in the WWII and later eras.
>Because a block of 12 pictures are jointly interpolated and coded this method is also called Block Coding. A block takes roughly 0,5s, which is then also the minimal lock time of a receiver.
TLDR: you need to wait for the intra-frame https://en.wikipedia.org/wiki/Video_compression_picture_type...
>Intra refresh periods of a half-second are common on such applications as digital television broadcast
Cable TV providers are totally screwing the pooch here. Most modern Cable Boxes employ reverse channel to spy^^^report telemetry and offer additional features. This means all the hardware and software infrastructure is already in place to request dynamically generated Inter frame from the backend on the fly and skip this up to 12 frames wait time.
Of course the other part is "modern" stupid software stacks, with TVs running Android, and probably rendering UI using web technologies.
edit: heck, I can go on : I have inherited a cast iron 'tillet'? ( frying pan ). It makes better steakes, and it shows no signs of wearing.
> Also Teletext worked always fast and was instant.
That's not at all how I remember it. I remember it taking anywhere from instant to 15 seconds to load a page.
That is consistent with 'the protocol', however there might have been 'smart' TVs which cached or prefetched pages.
I agree completely, I find this delay more than annoying—in fact, it's infuriating! I consider this delay not just an ergonomic limitation but also a definite design fault (i.e.: the DVB standard allows the delay to occur). The fact that regulators authorized DVB/TV digital services with this fault speaks more to me about the commercial pressure brought to bear on the regulators (to accept the limitation) than it does about any proper consultative/evaluation process (user ergonomics evaluation) having been properly undertaken before these services were introduced.
(…But then, I suppose we can't have expected much better from DVB standards given that in the 1980s many governments around the world essentially nuked their spectrum management departments and outsourced much of the work to commercial organizations that had vested interests—thus, effectively they were committed to lower standards. [This is a long sad story which, unfortunately, is much too involved to discuss here.])
The information about tuners is remarkably comprehensive (it's certainly the best I've seen in a single documented collection on the subject). I don't work with TV tuners on a day-to-day basis but I've had a lot to do with them in the past—and that goes back to the vacuum tube/turret type well before we had varicap* tuning (so I feel competent to comment on the site).
That said, there are stacks of valuable information on this site about tuners and other stuff that I was either unaware of or not very familiar about. There's nothing more I can add as to do so would be superfluous—as the excellence of this site speaks for itself.
_
* [I recall initially hating varicap tuning due to its propensity to suffer cross modulation (which meant that one had to ensure that TV channels had roughly the same RF input levels and that strong out-of-band services were kept out of such tuners by installing filters/traps in the antenna line before the tuner input—lest herringbone patterns result).]
WSP (World Standard Pinning) "standardization" story with management naive dreams of selling own products in Asia without cannibalizing home European market while simultaneously moving Consumer Hardware Development to Asia .. where all the hardware was thereafter designed with Asian parts first was priceless.
Philips TV division deciding to second source, skipping Philips Tuners division thus eroding own company market share and profit. Then in turn Philips Tuners division commissioning a cheaper clone of Philips MOPLL from Siemens with a huge minimal volume quota, consequently eroding Philips Semiconductors market share by 25% and forcing Tuner division to manufacture inferior products with already obsolete part. Just beautiful, <Chef's Kiss>, Pure genius!
> The most devastating development, however, was purely internal and of severe structural impact.
https://news.ycombinator.com/item?id=25735404
'The history of a pioneering tube that was developed by Philips Research and that was, next to the magnetron, “The most important tube from World War II” - Ronald Dekker'
One of the problems Bowen was facing was that while early 1939 all components of the Airborn Radar seemed to have fallen in place, he still had only one receiver which up to that moment no supplier had been able to duplicate. But then quite unexpected his old professor comes up with a golden tip:
"Quite by chance in april or May of 1939, I heard some encouraging news from Edward Appleton, my old professor at King’s Colledge and now Jacksonian Professor of Physics at Cambridge. He told me that the Pye Radio Compagny, still hoping that there would be a television industry in Britain, had set up a production line for 45 Mc/s TRF chassis and had actually made a trial run. I went hot-foot to Cambridge to see B.J. Edwards, the Technical Director of Pye Radio, and was rewarded with a remarkable sight – he had scores [!] of TRF chassis just the type we were looking for. These used a new valve with an octal [sic] base, which had not yet appeared on the market. It was the EF50, a valve which was destined to play almost as important a part in Radar war as the magnetron itself."
Bowen took a few samples of the Pye chassis back to Bawdsey and quickly verified that it was significantly better than their old EMI chassis; it was also smaller and lighter. Touch and his men quickly put a 200 MHz mixer (at that time still based on an acorn tube, but later replaced by an EF54) in front of it and the receive problem was as good as solved.