First things first: define what you mean by "measuring the effects of RF."
Is this for a product design that will be used in many different environments? If so, there are lots of standards out there which give a guide to quantifying RF immunity.
Is this just for your own use? Then try to be realistic about the RF challenges you face, and apply sound engineering prophylaxis.
Is this troubleshooting or characterizing an existing piece of kit? If so, what are the issues? Random noise from time to time? Continuous degradation? Buzz? AM detection?
With a target, it's obvious where to aim things, even with those very misunderstood conventional test tools.
Is this for a product design that will be used in many different environments? If so, there are lots of standards out there which give a guide to quantifying RF immunity.
Is this just for your own use? Then try to be realistic about the RF challenges you face, and apply sound engineering prophylaxis.
Is this troubleshooting or characterizing an existing piece of kit? If so, what are the issues? Random noise from time to time? Continuous degradation? Buzz? AM detection?
With a target, it's obvious where to aim things, even with those very misunderstood conventional test tools.
I have been working in audio electronics for a long time 50+ years. My full time job is an RF engineer (38 years) doing cell phone and two way radio design work. I have seen effects caused by RF devices affect the sound in subtle ways that are not easilly measured by typical steady state measurements in the audio band. I have also seen and chased "ghosts" where it looks like RF was affecting the audio equipment, but in reality the RF source was whacking the test equipment.
The worst offender is the cell phone. Here in the US there are several different cellular standards each with its own characteristics that affect things differently. All cell phones operate with 0.6 watts of RF power or less. Two way radios can also affect consumer electrical equipment.
Background information:
The most popular world wide cell phone standard is GSM. GSM is used in several frequency bands worldwide from 380 MHz to 2.4 GHZ. 900 MHz and 1.8 GHz are the most common in the US and used by older T-mobile and AT&T phones. GSM transmits a constant envelope carrier, meaning that the RF carrier is held at a constant amplitude during transmission. The information is carried by phase modulating the carrier. GSM uses Time Division Multiplexing (TDMA) which means that the transmitter is pulsed on and off at a base rate of about 8 Hz with multiple timeslots available, making the apparent pulse rate faster. In effect the carrier is 100% AM modulated with narrow pulses. This makes a GSM phone a strong interference generator whose effects are easilly heard. Many consumer devices like computer speakers, TV sets, and audio equipment will buzz in the presence of a GSM cell phone. It's effects on audio equipment are easilly measured in the audio band.
iDEN is an older standard used by Nextel (Sprint) in the US and Telus in Canada. It is also used in South America and Japan. It transmits at 800 MHz and 900 MHz. It uses TDMA with an 11 or 22 Hz rate AND the RF carrier is amplitude modulated. iDEN will get into anything remotely succeptible to RF and its effects are easilly heard and measured.
CDMA evolved from military secure communications. (Sprint and Verizon in the US) The system uses Code Division Multiplexing to spread a fairly low data rate across a wide frequency range. The transmitter does not pulse, it operates continuously during transmission. The information is modulated onto the RF carrier using a form of Amplitude Modulation, meaning that the RF carrier power is changing with the applied data. The data is combined with the spreading code to create a data rate in the 1.5 MHz range. The key here is that the RF frequency AND the modulation are both far above the audio range.
3G and 4G standards also use Amplitude Modulation with a higher data rate in the 1 to 10 MHz range. Again the RF frequency AND the modulation are both far above the audio range.
Hand held two way radios like carried by police and ham radio operators use Frequency Modulation but the RF power output can be as much as 10 watts. The RF frequency can range from 30 MHz to 1.3 GHz.
Automobile based two way radio equipment can cover 1.8 MHz to 2.4 GHz and the RF power level can be as high as 100 watts. Some ham radio equipment can produce over 1 killowatt of RF power. Both AM and FM modulation modes are used.
When RF radiation impacts consumer electrical equipment some of the RF energy will be picked up on the external wiring and even the internal circuitry. This RF energy can be rectified by ANY nonlinear device that it reaches. The rectified RF will create a DC voltage across this nonlinear element. The most common nonlinear element is a semiconductor junction. Ordinarily the rectified RF will be in the microvolt range and present no obvious effect other than a small bias shift. If the RF field is constantly changing in strength (ANY amplitude modulated or pulsed signal) there will be a constantly changing DC bias shift. Frequency Modulation does NOT vary the RF field. If the rate of change is in the audio range the effect can be plainly audible. This is why you hear a GSM or iDEN cell phone in your cheap unshielded computer speakers. This effect can be heard and measured, although since scope probes and test leads are good antennas the mere connection of a scope probe can turn a clean circuit into a succeptible one.
The area where difficulty lies is when the rate of change in the RF signal strength varies at a rate above (or far above) the audio range. This is the case with CDMA, 3G, and 4G cell phones. There will be some change in the operating point of internal circuitry but the rate of change can be in the 1 to 10 MHz range. You will not hear this directly, nor will most audio test equipment. In some cases its effects can be plainly heard and measured as an increase in distortion. In more subtle cases the effects can cause mild distortion effects or compressing of dynamic range that can not be easilly measured without disturbing the effect.
So, if you have a piece of equipment that can "hear" a GSM cell phone, you can bet it is also affected by CDMA and 3G or 4G phones. The effect may not be as obvious but you can probably find a spot where the phone does something to the sound that doesn't quite sound right, but doesn't cause blatant distortion. You can't always measure this with a sine wave and distortion meter.
I have seen one case where a 1KW ham radio transmitter blew the tweeters in a neighbors speakers without the stereo even turned on (SS pioneer receiver in the 1970's).
The worst offender is the cell phone. Here in the US there are several different cellular standards each with its own characteristics that affect things differently. All cell phones operate with 0.6 watts of RF power or less. Two way radios can also affect consumer electrical equipment.
Background information:
The most popular world wide cell phone standard is GSM. GSM is used in several frequency bands worldwide from 380 MHz to 2.4 GHZ. 900 MHz and 1.8 GHz are the most common in the US and used by older T-mobile and AT&T phones. GSM transmits a constant envelope carrier, meaning that the RF carrier is held at a constant amplitude during transmission. The information is carried by phase modulating the carrier. GSM uses Time Division Multiplexing (TDMA) which means that the transmitter is pulsed on and off at a base rate of about 8 Hz with multiple timeslots available, making the apparent pulse rate faster. In effect the carrier is 100% AM modulated with narrow pulses. This makes a GSM phone a strong interference generator whose effects are easilly heard. Many consumer devices like computer speakers, TV sets, and audio equipment will buzz in the presence of a GSM cell phone. It's effects on audio equipment are easilly measured in the audio band.
iDEN is an older standard used by Nextel (Sprint) in the US and Telus in Canada. It is also used in South America and Japan. It transmits at 800 MHz and 900 MHz. It uses TDMA with an 11 or 22 Hz rate AND the RF carrier is amplitude modulated. iDEN will get into anything remotely succeptible to RF and its effects are easilly heard and measured.
CDMA evolved from military secure communications. (Sprint and Verizon in the US) The system uses Code Division Multiplexing to spread a fairly low data rate across a wide frequency range. The transmitter does not pulse, it operates continuously during transmission. The information is modulated onto the RF carrier using a form of Amplitude Modulation, meaning that the RF carrier power is changing with the applied data. The data is combined with the spreading code to create a data rate in the 1.5 MHz range. The key here is that the RF frequency AND the modulation are both far above the audio range.
3G and 4G standards also use Amplitude Modulation with a higher data rate in the 1 to 10 MHz range. Again the RF frequency AND the modulation are both far above the audio range.
Hand held two way radios like carried by police and ham radio operators use Frequency Modulation but the RF power output can be as much as 10 watts. The RF frequency can range from 30 MHz to 1.3 GHz.
Automobile based two way radio equipment can cover 1.8 MHz to 2.4 GHz and the RF power level can be as high as 100 watts. Some ham radio equipment can produce over 1 killowatt of RF power. Both AM and FM modulation modes are used.
When RF radiation impacts consumer electrical equipment some of the RF energy will be picked up on the external wiring and even the internal circuitry. This RF energy can be rectified by ANY nonlinear device that it reaches. The rectified RF will create a DC voltage across this nonlinear element. The most common nonlinear element is a semiconductor junction. Ordinarily the rectified RF will be in the microvolt range and present no obvious effect other than a small bias shift. If the RF field is constantly changing in strength (ANY amplitude modulated or pulsed signal) there will be a constantly changing DC bias shift. Frequency Modulation does NOT vary the RF field. If the rate of change is in the audio range the effect can be plainly audible. This is why you hear a GSM or iDEN cell phone in your cheap unshielded computer speakers. This effect can be heard and measured, although since scope probes and test leads are good antennas the mere connection of a scope probe can turn a clean circuit into a succeptible one.
The area where difficulty lies is when the rate of change in the RF signal strength varies at a rate above (or far above) the audio range. This is the case with CDMA, 3G, and 4G cell phones. There will be some change in the operating point of internal circuitry but the rate of change can be in the 1 to 10 MHz range. You will not hear this directly, nor will most audio test equipment. In some cases its effects can be plainly heard and measured as an increase in distortion. In more subtle cases the effects can cause mild distortion effects or compressing of dynamic range that can not be easilly measured without disturbing the effect.
So, if you have a piece of equipment that can "hear" a GSM cell phone, you can bet it is also affected by CDMA and 3G or 4G phones. The effect may not be as obvious but you can probably find a spot where the phone does something to the sound that doesn't quite sound right, but doesn't cause blatant distortion. You can't always measure this with a sine wave and distortion meter.
I have seen one case where a 1KW ham radio transmitter blew the tweeters in a neighbors speakers without the stereo even turned on (SS pioneer receiver in the 1970's).
Agreed, I presented scope shot measurements of RF attenuators & their reduction of RF signal reflections by twice the amount that the signal is reduced. i.e a 10dB attenuator reduces reflections by 20dB approx. Two other people confirmed this with their own tests & measurements. Now a third, Pano, has also confirmed this double attenuation. SY could not measure this doubling effect for whatever reason.
It is an example of measuring signals in the RF domain & the practical difficulties that might be encountered. How do we know how RF might effect audio if we can't even measure it accurately, consistently? I believe that it is educational to learn how to do this & this example throws up an anomalie that the RF experts may well be able to shed some light on for the benefit of all. I believe that it is very much on topic!
It's by studying & explaining anomalies that we can learn! If you don't want to discuss your test conditions that's fine but Pano also experienced this as stated. The 1M was his test condition, not mine (probably a scope input impedance?).
Last edited:
A cable itself (Bybeed or not) is unlikely to convert RF to audio, so if the audio effects of a cable are due to RF attenuation then this will only be seen via RF measurements. (Cable terminations can convert RF to audio via the rusty bolt effect.)
If the effect of stray RF is to shift device bias points, then this might or might not show up in distortion tests. The problem is that distortion tests usually assume that the effect is constant (being caused by the circuit), so may smear the result and just give an average. It is conceivable that rapidly varying distortion could sound worse than constant distortion with the same average level.
If the effect of stray RF is to shift device bias points, then this might or might not show up in distortion tests. The problem is that distortion tests usually assume that the effect is constant (being caused by the circuit), so may smear the result and just give an average. It is conceivable that rapidly varying distortion could sound worse than constant distortion with the same average level.
I got this fancy 3G cell phone from AT&T (Samsung Galaxy) It makes calls in the 3G mode unless the coverage is bad or the network is busy, then it reverts to GSM. I can tell the difference if I am sitting here in front of the computer by the sound coming from the computer speakers. At home the Sony TV will make the GSM buzz if the phone is in GSM mode, but the audio will subtly distort as I walk around the TV if the phone is using 3G.
I routinely test my electronic creations with a cell phone. It is expected that some equipment will misbehave when approached with an iDEN or GSM phone. I look for bad stuff. I was a design engineer on many of the early iDEN phones (1997) and we had a lot of fun in the local Circuit City store before people knew what a Nextel phone was. I found that leaving the phone on the passenger floorboard of a 1984 Dodge Daytona can cause unexpected car stalling when the phone starts transmitting. Its best to find this stuff early in the design cycle.
I realise this is a bit OT, but if a line is properly terminated there are no reverse reflections to be attenuated. I don't know why audio people would want to measure things like this. If it does what it should, then you have merely confirmed that the textbooks are correct. If it appears not to, then you have merely confirmed that you are measuring the wrong thing. This assumes you live in a rational universe where RF transmission lines propagate signals according to well-known principles.
Please can we kill off this ghost? Then talk about the actual topic.
Please can we kill off this ghost? Then talk about the actual topic.
Originally posted by abraxalito
Some links rel. to grounding, impedance consideration on lines, return currents ect, all in relation to unwanted generation and unwanted reception of HF signals. *PS
The last one may be applicable for the "Groundside Electrons" thread too.
http://www.hottconsultants.com/pdf_files/june2001pcd_mixedsignal.pdf
http://www.hottconsultants.com/pdf_files/ground.pdf
The Return Path: Impedance Control on Printed Circuit Boards
The Return Path: Impedance Control on Printed Circuit Boards
http://www.hottconsultants.com/pdf_files/image_plane.pdf
Regards
George
Edit:
*PS Every circuit element that can efficiently radiate HF energy due to it’s dimensional nature , it can equally efficiently receive HF energy
Some links rel. to grounding, impedance consideration on lines, return currents ect, all in relation to unwanted generation and unwanted reception of HF signals. *PS
The last one may be applicable for the "Groundside Electrons" thread too.
http://www.hottconsultants.com/pdf_files/june2001pcd_mixedsignal.pdf
http://www.hottconsultants.com/pdf_files/ground.pdf
The Return Path: Impedance Control on Printed Circuit Boards
The Return Path: Impedance Control on Printed Circuit Boards
http://www.hottconsultants.com/pdf_files/image_plane.pdf
Regards
George
Edit:
*PS Every circuit element that can efficiently radiate HF energy due to it’s dimensional nature , it can equally efficiently receive HF energy
Last edited:
Grounding has a huge effect on RF. You just have to stop thinking about circuits and start thinking about waves. Circuit theory is just a low frequency approximation to reality, where "low frequency" means roughly under 100MHz or thereabouts (depending on circuit size).
I once read part of the groundside electrons thread - they seem to live in a different universe from me, where random loops of wire attached to ground terminals magically improve sound. If it works at all, which I severely doubt, it is simply shifting RF resonances away from interfering frequencies so no two people will find the same results. Much more likely is the combination of ignorance and the placebo effect.
I once read part of the groundside electrons thread - they seem to live in a different universe from me, where random loops of wire attached to ground terminals magically improve sound. If it works at all, which I severely doubt, it is simply shifting RF resonances away from interfering frequencies so no two people will find the same results. Much more likely is the combination of ignorance and the placebo effect.
The devil's in the detail here - 'properly terminated' is a common mantra, rather like 'competently designed'. If the rise time of a digital signal is 1nS then 'proper termination' must be effective at least over a bandwidth of 350MHz and probably higher. Its not so trivial, especially when non-RF connectors such as RCA phonos are in use.
I see SPDIF discussions as falling under the remit of 'RF and Audio' myself. What's the problem? DACs are susceptible to stray RF on their digital inputs in my experience.
OT posts removed.The signal reflection refered to by H.Ott is based on the TEM wave coupling, it has to be a plane under the signals.
A lot of what you want to investigate is covered by EMC and Signal Integrity. With the speed of clocks
and signal rise times on even the simplest digital PCB these days, you have to follow the guidlines.
One of the main causes of problems is fast (>1ns) rise times of signals even when they are not necessary,
but most modern chips can achieve these rise times with no problems, adding harmonics into the GHzs
When you start adding DDR ram interfaces, gigabit ethernet interface, and controlled inpedance PCB's,
simultaneous switching noise from FPGA's etc and you are realy having fun. Most high speed PCB design and where cost is secondary to
reliability is now done using stripline (all signals burried on inner layers) with each signal layer referenced to
a contigous ground plane, so boards have multiple ground planes. When adding powerplane pairs (you always add a power layer and a ground layer)
you go for the minimum distance (>0.1mm) to increase plane capacitance.
Its all about minimising the current loops, and above 1MHz this is the inductive loop, making sure that the energy
injected into a line ALL reached the load by minimising impedance discontinuities, and understanding some of how the TEM wave propogates down a line
and where its e and h field are coupling. This is why LVDS is becoming more prevelant, even though the diff pairs are coupled to the return plane their
close coupling effectively cancels the current in the ground, whilts making the signals less prone to localy coupled noise.
We quite often have clocks in the 100MHz + range, whith double data rate switching, ie on the rising and falling edge you have 200MHz
data buses whizzing around.
All comercial equipement has to conform to certain EMC standards CE, FCC, when its Aero, Millitary or Medical these requirements are even more
onerous, and has had to for years so is a branch of electronics that is both well documented and researched. That is why engineers and PCB designers
go to great lengths to design systems so that EMC is not a problem and does not effect the circuitry.
RF interference in its simplest form can be seen on Audio outputs as what we call "hairy waveforms", to a
greator or lesser extent there will be some RF artifacts on the audio waveform where you have digital
and analogue on the same board, and or local RF coupling.
Some interesting reading and links.
"Electromagnetic Compatability Engineering" Henry Ott ISBN798-470-18930-6
"High-Speed Digital Design" Howard Johnson & Martin Graham ISBN 0-13-395724-1
"Right The First Time" Lee W. Ritchey Free Download Right the First Time
Henry Ott Consultants
Tech Tips
The EMC Compliance Club
A source of some excellent articles by Keith Armstrong and others, including a few articles on shielding audio equipement
EMC Information Centre - The EMC Journal (Free in the UK)
Signal Integrity Sites
beTheSignal.com
High-Speed Digital Design Publications Archive by Keyword
Speeding Edge consultants specialize in high-speed PCB and system design disciplines
A simple presentation of some of the Issues.
http://www.x2y.com/filters/TechDay0...log_Designs_Demand_GoodPCBLayouts _JohnWu.pdf
And one of my favorites:
Especially number one, regarding clocks, which will apeal to all those that instead of using the
clock or crystal next to a device on a board, add a seperate board with long leads to achieve low jitter.
The 10 Best Ways to Maximize Emission from Your Product
A lot of what you want to investigate is covered by EMC and Signal Integrity. With the speed of clocks
and signal rise times on even the simplest digital PCB these days, you have to follow the guidlines.
One of the main causes of problems is fast (>1ns) rise times of signals even when they are not necessary,
but most modern chips can achieve these rise times with no problems, adding harmonics into the GHzs
When you start adding DDR ram interfaces, gigabit ethernet interface, and controlled inpedance PCB's,
simultaneous switching noise from FPGA's etc and you are realy having fun. Most high speed PCB design and where cost is secondary to
reliability is now done using stripline (all signals burried on inner layers) with each signal layer referenced to
a contigous ground plane, so boards have multiple ground planes. When adding powerplane pairs (you always add a power layer and a ground layer)
you go for the minimum distance (>0.1mm) to increase plane capacitance.
Its all about minimising the current loops, and above 1MHz this is the inductive loop, making sure that the energy
injected into a line ALL reached the load by minimising impedance discontinuities, and understanding some of how the TEM wave propogates down a line
and where its e and h field are coupling. This is why LVDS is becoming more prevelant, even though the diff pairs are coupled to the return plane their
close coupling effectively cancels the current in the ground, whilts making the signals less prone to localy coupled noise.
We quite often have clocks in the 100MHz + range, whith double data rate switching, ie on the rising and falling edge you have 200MHz
data buses whizzing around.
All comercial equipement has to conform to certain EMC standards CE, FCC, when its Aero, Millitary or Medical these requirements are even more
onerous, and has had to for years so is a branch of electronics that is both well documented and researched. That is why engineers and PCB designers
go to great lengths to design systems so that EMC is not a problem and does not effect the circuitry.
RF interference in its simplest form can be seen on Audio outputs as what we call "hairy waveforms", to a
greator or lesser extent there will be some RF artifacts on the audio waveform where you have digital
and analogue on the same board, and or local RF coupling.
Some interesting reading and links.
"Electromagnetic Compatability Engineering" Henry Ott ISBN798-470-18930-6
"High-Speed Digital Design" Howard Johnson & Martin Graham ISBN 0-13-395724-1
"Right The First Time" Lee W. Ritchey Free Download Right the First Time
Henry Ott Consultants
Tech Tips
The EMC Compliance Club
A source of some excellent articles by Keith Armstrong and others, including a few articles on shielding audio equipement
EMC Information Centre - The EMC Journal (Free in the UK)
Signal Integrity Sites
beTheSignal.com
High-Speed Digital Design Publications Archive by Keyword
Speeding Edge consultants specialize in high-speed PCB and system design disciplines
A simple presentation of some of the Issues.
http://www.x2y.com/filters/TechDay0...log_Designs_Demand_GoodPCBLayouts _JohnWu.pdf
And one of my favorites:
Especially number one, regarding clocks, which will apeal to all those that instead of using the
clock or crystal next to a device on a board, add a seperate board with long leads to achieve low jitter.
The 10 Best Ways to Maximize Emission from Your Product
I'm not sure this thread is supposed to be about designing reliable digital circuits, interesting though that is. I thought it was about the effects of RF interference on analogue audio circuits, and how to measure this.
In some cases modulated RF will cause LF and subsonic things to happen (below where we usually look), but it might not be easy to distinguish this from mains voltage variations.
In some cases modulated RF will cause LF and subsonic things to happen (below where we usually look), but it might not be easy to distinguish this from mains voltage variations.
Hi
DF96, the reason for all the references to digital is that any digital components in your system, or in your house could be a cause of "rf" interferance. The clock rates and RISE TIMES of all digital equipement is the part of the cause of "RF" interferance, and probably one of the main ones. Your TV if its a modern flat panel, your DVD, CD, all have digital circuitry. More modern ones will quite often use modern intergrated circuitry. Where there is a market you will find IC manufacturers have a minimum chip solution, these chips are becoming more complex, reducing component count and improving functionality, but at a cost. And that cost is more and more digital circuitry, SMPS's etc etc. Any noise not controlled will polute your mains, get a cable length near a 20th of a clock and you have a quite effective antenna.
The point of all the links is to show some of the engineering that does go on outside of Audio for problems that effect all electronic and to show that "RF" is common place. To explain what I do at work to the uninitiated is to put on the radio (Radio 1 in the UK) and then show them the memory block I am routing; the clock on the memory is running at more than twice what the radio station is broadcasting at. It puts RF into perspective these days.
And if you use power line communications your just adding to the problem.
RISE TIMES.
A 1ns second rise time adds harmonics into the GHz range. Sy (and Henry Ott, Eric Bogatin, Howard Johnson) are right about rise times, the slower the better, but the rest are right, it mustn't be too slow. Modern FPGA's and Processors have controllable current drive (rise time), that you match to both the characteristic impedance of the layout and the clock period, you choose the slowest rise time within tolerences, to minimise both reflections and Delta I noise (ground bounce). We recently observed a 16Mhz clock, (not fast by todays standards) that was being problematic. The problem was a severe impedance mismatch in its tracking, a PTH test point, and a sub 1ns rise time, adding a series terminating resistor and selecting the value through observation, reduced the rise time to over 1.5ns and reduced the over and indershoot (ringing) by a considerable amount. (SPDIF) The series resistor I belive attenuates the second mode reflection (the one that has bounced back from the reflection point, impedance mismatch), parallel resistors terminate the first mode, ie absorbe the excess energy the source cannot handle. This is why impedance matching is so mportant, when a device switches it sees the immediate impedance, the destination is a few pico seconds away, so it drives the current accordingly, a ps or so later it sees a change in impedance and the current has to back up, kerboom noise.
Sorry for rattling on, but I do see these problems at work every day, as a PCB designer, doing some fun stuff. There is a lot of interesting work out there that we can use as a background for some real Audio research and dicussion
DF96, the reason for all the references to digital is that any digital components in your system, or in your house could be a cause of "rf" interferance. The clock rates and RISE TIMES of all digital equipement is the part of the cause of "RF" interferance, and probably one of the main ones. Your TV if its a modern flat panel, your DVD, CD, all have digital circuitry. More modern ones will quite often use modern intergrated circuitry. Where there is a market you will find IC manufacturers have a minimum chip solution, these chips are becoming more complex, reducing component count and improving functionality, but at a cost. And that cost is more and more digital circuitry, SMPS's etc etc. Any noise not controlled will polute your mains, get a cable length near a 20th of a clock and you have a quite effective antenna.
The point of all the links is to show some of the engineering that does go on outside of Audio for problems that effect all electronic and to show that "RF" is common place. To explain what I do at work to the uninitiated is to put on the radio (Radio 1 in the UK) and then show them the memory block I am routing; the clock on the memory is running at more than twice what the radio station is broadcasting at. It puts RF into perspective these days.
And if you use power line communications your just adding to the problem.
RISE TIMES.
A 1ns second rise time adds harmonics into the GHz range. Sy (and Henry Ott, Eric Bogatin, Howard Johnson) are right about rise times, the slower the better, but the rest are right, it mustn't be too slow. Modern FPGA's and Processors have controllable current drive (rise time), that you match to both the characteristic impedance of the layout and the clock period, you choose the slowest rise time within tolerences, to minimise both reflections and Delta I noise (ground bounce). We recently observed a 16Mhz clock, (not fast by todays standards) that was being problematic. The problem was a severe impedance mismatch in its tracking, a PTH test point, and a sub 1ns rise time, adding a series terminating resistor and selecting the value through observation, reduced the rise time to over 1.5ns and reduced the over and indershoot (ringing) by a considerable amount. (SPDIF) The series resistor I belive attenuates the second mode reflection (the one that has bounced back from the reflection point, impedance mismatch), parallel resistors terminate the first mode, ie absorbe the excess energy the source cannot handle. This is why impedance matching is so mportant, when a device switches it sees the immediate impedance, the destination is a few pico seconds away, so it drives the current accordingly, a ps or so later it sees a change in impedance and the current has to back up, kerboom noise.
Sorry for rattling on, but I do see these problems at work every day, as a PCB designer, doing some fun stuff. There is a lot of interesting work out there that we can use as a background for some real Audio research and dicussion
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.