Hi John,
I wanted to deal with this on it's own.
Since the degeneration resistor is in series with the active device and collector (or drain) circuit, by definition there should not be any phase change. Not until you get to extremely high frequencies anyway. Are you referring to a feedback loop? In this case, I completely agree that phase might be a problem IFF the circuit has phase problems into the passband of the signal. The third harmonic of 1 KHz (for example) is only 3 KHz. I would not expect large phase shifts in that area.
-Chris
I wanted to deal with this on it's own.
Since the degeneration resistor is in series with the active device and collector (or drain) circuit, by definition there should not be any phase change. Not until you get to extremely high frequencies anyway. Are you referring to a feedback loop? In this case, I completely agree that phase might be a problem IFF the circuit has phase problems into the passband of the signal. The third harmonic of 1 KHz (for example) is only 3 KHz. I would not expect large phase shifts in that area.
-Chris
Chris, we don't know EXACTLY why local feedback may be less problematic than LOOP feedback, BUT it is our design experience that it is so. Perhaps, IF a circuit has an almost infinite bandwidth (and that might be possible with new technologies in the GHz or TeraHertz region in the far future). However, today, we are still limited to about 100MHz or much less for audio circuits. You would think it enough, but only when used carefully, in my experience.
Also, looking at Pederson's book more carefully, it is even MORE complex than I first estimated. Wow! I hope you will get something important from this book, if even just how complex it really is (physicist's view) vs. what do we need to know and do to make the specifications required (engineer's view). Both are important.
Also, looking at Pederson's book more carefully, it is even MORE complex than I first estimated. Wow! I hope you will get something important from this book, if even just how complex it really is (physicist's view) vs. what do we need to know and do to make the specifications required (engineer's view). Both are important.
Dashing through after a "late Valentine's" dinner and still drinking the wine (Chateau Camensac '95--holding up well at 12 years)...
On a less mathematical plane, it occurs to me that resistance in series with the Source will slow the charging of Cgs and Cds. This will be a distortion mechanism in and of itself at high enough frequencies and/or large voltage swings, i.e. slew rate situations. No, I haven't scratched out the math, don't have time. It may be that it's only important at 100Vrms and 100MHz for all I know...in which case we're back to RF.
If it's RF, I've found that there's a 50/50 chance that it will be a problem at AF. You can view this from a glass half-full or glass half-empty perspective. I try to take it on a case by case basis.
(Yes, obviously it depends on the value of the resistor and the capacitance of the device in question...JFETs will be waaaaay out there, MOSFETs more problematical, bipolars closer to JFETs...)
Grey
On a less mathematical plane, it occurs to me that resistance in series with the Source will slow the charging of Cgs and Cds. This will be a distortion mechanism in and of itself at high enough frequencies and/or large voltage swings, i.e. slew rate situations. No, I haven't scratched out the math, don't have time. It may be that it's only important at 100Vrms and 100MHz for all I know...in which case we're back to RF.
If it's RF, I've found that there's a 50/50 chance that it will be a problem at AF. You can view this from a glass half-full or glass half-empty perspective. I try to take it on a case by case basis.
(Yes, obviously it depends on the value of the resistor and the capacitance of the device in question...JFETs will be waaaaay out there, MOSFETs more problematical, bipolars closer to JFETs...)
Grey
In the Microwave industry RF usually considered GHz. or higher. Anything under GHz. is considered I.F.
Most on the amplifiers are designed for an octave. Better performance with Lange couplers.
Most on the amplifiers are designed for an octave. Better performance with Lange couplers.
To stay on track, IF the open loop bandwith was just a few hundred KHz, then feedback should be OK in any amount, but then the CLOSED LOOP BANDWIDTH would have to extend to 1GHz or more with high feedback designs. It is not impossible, but it is impractical at this time.
Hi Mike,
I can barely live with calling those frequencies IF. Often it is considered RF at MW and above, but I'd call IF frequencies around 10 MHz and RF a little above that and up.
We have microcontrollers running merrily at 40 MHz and uP's running into your RF region. Most audio hobbyists are only dimly aware of what goes on at 1 MHz, they seem to thing that most things that we see above 20 KHz is not their problem.
Hi John,
I'm looking forward to getting that book. I'm sure there is a lot in there for me. I may need some help with it.
I have to hit the hay now.
-Chris
Edit: Hi John, yes. Closed loop gains might be "doable" with current feedback topolgies. I have some op amps that are good for a couple GHz. PCB layout is too critical for me to do.
I can barely live with calling those frequencies IF. Often it is considered RF at MW and above, but I'd call IF frequencies around 10 MHz and RF a little above that and up.
We have microcontrollers running merrily at 40 MHz and uP's running into your RF region. Most audio hobbyists are only dimly aware of what goes on at 1 MHz, they seem to thing that most things that we see above 20 KHz is not their problem.
Hi John,
I'm looking forward to getting that book. I'm sure there is a lot in there for me. I may need some help with it.
That seems to be on the hairy edge for most of us. At RF frequencies, all those benign parasitic issues have grown a full set of fangs. RF engineering is beyond my experience, and many others I'll bet. It does explain your choice of PCB substrates nicely. Understand that getting many people to consider what goes on at 1 MHz is a job.
I have to hit the hay now.
-Chris
Edit: Hi John, yes. Closed loop gains might be "doable" with current feedback topolgies. I have some op amps that are good for a couple GHz. PCB layout is too critical for me to do.
Hi John,
I guess the question is: Are they interesting enough to warrant you investigating what you can do with them?
Me? I have a lot of studying to do, but then, that is the beauty of this hobby / profession. Having access to the proper equipment sometimes will hold people back from developing further. Designing by ear alone is not what I would call enough to produce a good design.
Hi MikeW,
Your points are well taken. RF design guidelines are required for really good audio work.
-Chris
I guess the question is: Are they interesting enough to warrant you investigating what you can do with them?
Me? I have a lot of studying to do, but then, that is the beauty of this hobby / profession. Having access to the proper equipment sometimes will hold people back from developing further. Designing by ear alone is not what I would call enough to produce a good design.
Hi MikeW,
Your points are well taken. RF design guidelines are required for really good audio work.
-Chris
RF and EMI proof design guidelines are THE KEY FACTOR for really good audio. Almost ALL designs do not fulfill them.
PMA that is not what we are referring to. We are discussing increased gain bandwidth considerations. Of course, RFI is important, but it is another subject entirely.
MikeW said:In the Microwave industry RF usually considered GHz. or higher. Anything under GHz. is considered I.F.
Most on the amplifiers are designed for an octave. Better performance with Lange couplers.
"RF" is neatly divided into a number of frequency bands from 3Hz to 300GHz, based on wavelength (WL):
ELF extremely low frequency 3Hz to 30Hz 100,000km to 10,000 km WL
SLF superlow frequency 30Hz to 300Hz 10,000km to 1,000km WL
ULF ultralow frequency 300Hz to 3000Hz 1,000km to 100km WL
VLF very low frequency 3kHz to 30kHz 100km to 10km WL
LF low frequency 30kHz to 300kHz 10km to 1km WL
MF medium frequency 300kHz to 3MHz 1km to 100m WL
HF high frequency 3MHz to 30MHz 100m to 10m WL
VHF very high frequency 30MHz to 300MHz 10m to 1m WL
UHF ultrahigh frequency 300MHz to 3000MHz 1m to 10cm WL
SHF superhigh frequency 3GHz to 30GHz 10cm to 1cm WL
EHF extremely high frequency 30GHz to 300GHz 1cm to 1mm WL
For aerospace EMI testing purposes some of those tests set limits as low as 15kHz E-field and H-field (audio frequency interference and susceptibility).
Sorry if I'm causing this to get off topic.
Best Regards, Chuck Hansen
Thanks Chuck. So expressed another way: In the beginning, the GBW of op amps was in the MF region. Today, cheap IC's are usually in the HF region, but the better and more expensive linear IC's are now in the VHF or UHF region which looks like today's limit. I will get more interested in classical op amp configured IC's when their GBW moves to the EHF region, perhaps not within my lifetime. However, conceptually I can see many of my concerns evaporating at that time.
I have some video op-amps good to 10 MHz. I'll send them to you if you want them John. Only +/- 5 volt.
Chuck come up for lunch and I will take you for a tour of the plant.
I'll buy.😀
Chuck come up for lunch and I will take you for a tour of the plant.
I'll buy.😀
Mike, I have access to most IC op amps, but I usually prefer discrete designs. However, I am designing at this time, a phono stage with IC's. It will not replace my better designs, but it should work OK.
+/-5V is useless. You can return to something like AD811, AD817, AD845, they are fast enough and have enough headroom.
MikeW said:I have some video op-amps good to 10 MHz. I'll send them to you if you want them John. Only +/- 5 volt.
Chuck come up for lunch and I will take you for a tour of the plant.
I'll buy.😀
Thanks for the offer Mike, but I have a bum back that really limits my travel radius. Thank goodness work is only 3 miles away, and audioXpress writing is via email/UPS.
Best Regards, Chuck Hansen
Hi Pavel,
The 5 VDC supplies are used for signal conditioning. Normally with A/D and D/A converters. I have some of those beasties to try as well, but they are all smt. A couple get up into the GHz range! What is really cool are the differential in and out op amps.
Hi John,
All that speed can really bite you in the backside though. High gain with extremely high cut off frequencies are just begging for a reason to oscillate or ring. Things are getting hairy for a poor audio guy.
-Chris
The 5 VDC supplies are used for signal conditioning. Normally with A/D and D/A converters. I have some of those beasties to try as well, but they are all smt. A couple get up into the GHz range! What is really cool are the differential in and out op amps.
Hi John,
All that speed can really bite you in the backside though. High gain with extremely high cut off frequencies are just begging for a reason to oscillate or ring. Things are getting hairy for a poor audio guy.
-Chris
Hi Chris,
many of CD/SACD players have analog output 2.3Vrms, i.e. +/-3.25Vp. It is much better then to have Vs = +/-12V at least.
many of CD/SACD players have analog output 2.3Vrms, i.e. +/-3.25Vp. It is much better then to have Vs = +/-12V at least.
- Status
- Not open for further replies.