| Home |
| Search |
| Today's Posts |
|
#1
Posted to rec.audio.tubes
|
|||
|
|||
Next Project
Well, Gort was a Christmas day success so why not FM next? Maybe even
gasp stereo. So far I have a crude, but working, reactance tube FM Transmitter breadboarded and some 'ideas' for the MPX. http://flipperhome.dyndns.org/FM%20Stereo.htm I was trying to make it a "dollar days" transmitter, meaning the tubes cost $1/ea from ABCvacuumtubes, but the 6DT8 is 'full price' at $3.00. |
|
#2
Posted to rec.audio.tubes
|
|||
|
|||
|
Next Project
In article ,
flipper wrote: Well, Gort was a Christmas day success so why not FM next? Maybe even gasp stereo. So far I have a crude, but working, reactance tube FM Transmitter breadboarded and some 'ideas' for the MPX. http://flipperhome.dyndns.org/FM%20Stereo.htm I was trying to make it a "dollar days" transmitter, meaning the tubes cost $1/ea from ABCvacuumtubes, but the 6DT8 is 'full price' at $3.00. Hi Flipper, The 6ME8, and its similar cousins, has always frustrated me because I have never been able to figure out a reasonable way to combine the outputs from the two plates without using a transformer. I would like to use two 6ME8s, one for each stereo channel to generate the complete stereo composite baseband signal without using the L+R L-R matrix scheme. Unfortunately this requires an anode combining circuit that is phase linear from about 3 Hz to 53 kHz, a tall order for a transformer! One thing that I hadn't realized about the 6ME8 circuit, until your version got me thinking about it, is that it looks like the 6ME8 can function as a linear multiplier if the grid is not over driven, eliminating the need for a filter on the output to avoid subcarrier harmonics. As we all know from the Optimod this is the proper way to build a stereo coder. A separate 38 kHz oscillator would probably be necessary to avoid subcarrier harmonics, I assume that your self oscillating scheme probably has too much harmonic content, although I am just guessing there. This scheme also suffers from phase shift in the output transformer, even though it only has to handle the subcarrier and its sidebands. Still it is food for thought. I would probably just go with a 38 kHz oscillator driving the four diodes of a 6JU8 tube through a transformer to generate the stereo composite signal in the reverse of the way a typical FM stereo receiver demodulates it. The down side with this approach is that it requires a phase linear low pass filter to get rid of the subcarrier harmonics in the composite signal. -- Regards, John Byrns Surf my web pages at, http://fmamradios.com/ |
|
#3
Posted to rec.audio.tubes
|
|||
|
|||
|
Next Project
On Mon, 09 Jan 2012 22:43:06 -0600, John Byrns
wrote: In article , flipper wrote: Well, Gort was a Christmas day success so why not FM next? Maybe even gasp stereo. So far I have a crude, but working, reactance tube FM Transmitter breadboarded and some 'ideas' for the MPX. http://flipperhome.dyndns.org/FM%20Stereo.htm I was trying to make it a "dollar days" transmitter, meaning the tubes cost $1/ea from ABCvacuumtubes, but the 6DT8 is 'full price' at $3.00. Hi Flipper, The 6ME8, and its similar cousins, has always frustrated me because I have never been able to figure out a reasonable way to combine the outputs from the two plates without using a transformer. I would like to use two 6ME8s, one for each stereo channel to generate the complete stereo composite baseband signal without using the L+R L-R matrix scheme. Unfortunately this requires an anode combining circuit that is phase linear from about 3 Hz to 53 kHz, a tall order for a transformer! If you mean using two 6ME8s switching the deflectors at 38KHz for a 'chopper' style matrix generator, I thought about that too but I don't think you need a transformer as you can simply add the outputs. You need a large symmetrical square wave, though, and that's not 'trivial'. I was trying to come up with something 'simple' but the word doesn't seem to apply. One thing that I hadn't realized about the 6ME8 circuit, until your version got me thinking about it, is that it looks like the 6ME8 can function as a linear multiplier if the grid is not over driven, eliminating the need for a filter on the output to avoid subcarrier harmonics. As we all know from the Optimod this is the proper way to build a stereo coder. A separate 38 kHz oscillator would probably be necessary to avoid subcarrier harmonics, I assume that your self oscillating scheme probably has too much harmonic content, although I am just guessing there. This scheme also suffers from phase shift in the output transformer, even though it only has to handle the subcarrier and its sidebands. Still it is food for thought. Yeah, I'm not sure about harmonic content. For one, I have no model to simulate with, just a drawing. Shouldn't it be out of the passband though? I would probably just go with a 38 kHz oscillator driving the four diodes of a 6JU8 tube through a transformer to generate the stereo composite signal in the reverse of the way a typical FM stereo receiver demodulates it. The down side with this approach is that it requires a phase linear low pass filter to get rid of the subcarrier harmonics in the composite signal. I may go with a diode modulator but in that case I'd probably use solid state. Another problem is balancing it and that's one thing about the 6ME8, you can balance it with deflector bias and don't need tunable cores. I really haven't gone past the speculation phase at this point and I appreciate the suggestions. |
|
#4
|
|||
|
|||
Quote:
The whole thing is still here in my workshop. Cheers, John |
|
#5
Posted to rec.audio.tubes
|
|||
|
|||
|
Next Project
In article ,
flipper wrote: On Mon, 09 Jan 2012 22:43:06 -0600, John Byrns wrote: In article , flipper wrote: Well, Gort was a Christmas day success so why not FM next? Maybe even gasp stereo. So far I have a crude, but working, reactance tube FM Transmitter breadboarded and some 'ideas' for the MPX. http://flipperhome.dyndns.org/FM%20Stereo.htm I was trying to make it a "dollar days" transmitter, meaning the tubes cost $1/ea from ABCvacuumtubes, but the 6DT8 is 'full price' at $3.00. Hi Flipper, The 6ME8, and its similar cousins, has always frustrated me because I have never been able to figure out a reasonable way to combine the outputs from the two plates without using a transformer. I would like to use two 6ME8s, one for each stereo channel to generate the complete stereo composite baseband signal without using the L+R L-R matrix scheme. Unfortunately this requires an anode combining circuit that is phase linear from about 3 Hz to 53 kHz, a tall order for a transformer! If you mean using two 6ME8s switching the deflectors at 38KHz for a 'chopper' style matrix generator, I thought about that too but I don't think you need a transformer as you can simply add the outputs. You need a large symmetrical square wave, though, and that's not 'trivial'. I'll have to think about it a bit to be sure it actually works, but I think you have the right idea, with two tubes, you don't need the transformer to combine the plates, the subcarrier cancels without one. The second pair of plates could drive the other input to a push pull reactance tube modulator. If I am not missing something this is really cool, two 6ME8s make a linear multiplier type stereo coder that doesn't need a composite filter! As I am conceiving it the audio would drive the deflectors as in your design, and out of phase subcarrier signals would drive the grids of the two tubes. What could be simpler! I was trying to come up with something 'simple' but the word doesn't seem to apply. If the two 6ME8 scheme really works, that would be about as simple as you can get, what with no complex composite filter required. One thing that I hadn't realized about the 6ME8 circuit, until your version got me thinking about it, is that it looks like the 6ME8 can function as a linear multiplier if the grid is not over driven, eliminating the need for a filter on the output to avoid subcarrier harmonics. As we all know from the Optimod this is the proper way to build a stereo coder. A separate 38 kHz oscillator would probably be necessary to avoid subcarrier harmonics, I assume that your self oscillating scheme probably has too much harmonic content, although I am just guessing there. This scheme also suffers from phase shift in the output transformer, even though it only has to handle the subcarrier and its sidebands. Still it is food for thought. Yeah, I'm not sure about harmonic content. For one, I have no model to simulate with, just a drawing. Shouldn't it be out of the passband though? What determines the pass band? I would think anything that hits the reactance tube(s) is going to be transmitted, and the receiver is going to determine the actual passband, with some receivers demodulating the subcarrier harmonic degrading separation. I would probably just go with a 38 kHz oscillator driving the four diodes of a 6JU8 tube through a transformer to generate the stereo composite signal in the reverse of the way a typical FM stereo receiver demodulates it. The down side with this approach is that it requires a phase linear low pass filter to get rid of the subcarrier harmonics in the composite signal. I may go with a diode modulator but in that case I'd probably use solid state. Another problem is balancing it and that's one thing about the 6ME8, you can balance it with deflector bias and don't need tunable cores. The diode modulator can probably be balanced without too much difficulty, a lot of old broadcast transmitters managed it. I don't follow the bit about the tunable cores? Why would a diode modulator require tunable cores, while a 6ME8 wouldn't? The RCA BTS-1A Stereo Coder described on my web pages did the job without tunable cores and with only two tubes. I really haven't gone past the speculation phase at this point and I appreciate the suggestions. I wasn't really making suggestions, just musing about various ways to build a stereo coder that I hadn't really thought about before. -- Regards, John Byrns Surf my web pages at, http://fmamradios.com/ |
|
#6
Posted to rec.audio.tubes
|
|||
|
|||
|
Next Project
On Tue, 10 Jan 2012 22:37:32 -0600, John Byrns
wrote: In article , flipper wrote: On Mon, 09 Jan 2012 22:43:06 -0600, John Byrns wrote: In article , flipper wrote: Well, Gort was a Christmas day success so why not FM next? Maybe even gasp stereo. So far I have a crude, but working, reactance tube FM Transmitter breadboarded and some 'ideas' for the MPX. http://flipperhome.dyndns.org/FM%20Stereo.htm I was trying to make it a "dollar days" transmitter, meaning the tubes cost $1/ea from ABCvacuumtubes, but the 6DT8 is 'full price' at $3.00. Hi Flipper, The 6ME8, and its similar cousins, has always frustrated me because I have never been able to figure out a reasonable way to combine the outputs from the two plates without using a transformer. I would like to use two 6ME8s, one for each stereo channel to generate the complete stereo composite baseband signal without using the L+R L-R matrix scheme. Unfortunately this requires an anode combining circuit that is phase linear from about 3 Hz to 53 kHz, a tall order for a transformer! If you mean using two 6ME8s switching the deflectors at 38KHz for a 'chopper' style matrix generator, I thought about that too but I don't think you need a transformer as you can simply add the outputs. You need a large symmetrical square wave, though, and that's not 'trivial'. I'll have to think about it a bit to be sure it actually works, but I think you have the right idea, with two tubes, you don't need the transformer to combine the plates, the subcarrier cancels without one. The second pair of plates could drive the other input to a push pull reactance tube modulator. If I am not missing something this is really cool, two 6ME8s make a linear multiplier type stereo coder that doesn't need a composite filter! As I am conceiving it the audio would drive the deflectors as in your design, and out of phase subcarrier signals would drive the grids of the two tubes. What could be simpler! I think we're talking different approaches, although the net result is the same (or else it wouldn't be FM stereo MPX). I'm talking about time division multiplexing, ala what's described here http://web.archive.org/web/200602140...com/stereo.htm Using the 6ME8 gets you the 38 KHz inversion. I.E. One plate is the 'inverse' of the other so you could feed a common 38KHz square wave, large enough to swing cutoff, to the deflectors on each, pre-emphasis audio to the grid, and use the L 'when high' plate on one to sum with the R 'when low' plate on the other. Just tie the two plates together through a common load resistor and whichever one is 'on' passes its channel's signal. The sum then, after a harmonics filter, has the pilot added and goes straight to the modulator. The other two plates go to B+ but aren't used. The square wave needs to be symmetrical and constant amplitude. Actually, ideal would be a sine, rather than square, perfectly aligned between 'on' and cutoff but cutoff isn't all that sharp so I doubt it would work. I was trying to come up with something 'simple' but the word doesn't seem to apply. If the two 6ME8 scheme really works, that would be about as simple as you can get, what with no complex composite filter required. Well, in the TDM circuit you need to filter out the harmonics but that should be a relatively simple low pass. One thing that I hadn't realized about the 6ME8 circuit, until your version got me thinking about it, is that it looks like the 6ME8 can function as a linear multiplier if the grid is not over driven, eliminating the need for a filter on the output to avoid subcarrier harmonics. As we all know from the Optimod this is the proper way to build a stereo coder. A separate 38 kHz oscillator would probably be necessary to avoid subcarrier harmonics, I assume that your self oscillating scheme probably has too much harmonic content, although I am just guessing there. This scheme also suffers from phase shift in the output transformer, even though it only has to handle the subcarrier and its sidebands. Still it is food for thought. Yeah, I'm not sure about harmonic content. For one, I have no model to simulate with, just a drawing. Shouldn't it be out of the passband though? What determines the pass band? The "unknown" LP filter after the modulator. Second harmonics should cancel in the transformer leaving 3'rd and up to filter. I would think anything that hits the reactance tube(s) is going to be transmitted, and the receiver is going to determine the actual passband, with some receivers demodulating the subcarrier harmonic degrading separation. I would probably just go with a 38 kHz oscillator driving the four diodes of a 6JU8 tube through a transformer to generate the stereo composite signal in the reverse of the way a typical FM stereo receiver demodulates it. The down side with this approach is that it requires a phase linear low pass filter to get rid of the subcarrier harmonics in the composite signal. I may go with a diode modulator but in that case I'd probably use solid state. Another problem is balancing it and that's one thing about the 6ME8, you can balance it with deflector bias and don't need tunable cores. The diode modulator can probably be balanced without too much difficulty, a lot of old broadcast transmitters managed it. I don't follow the bit about the tunable cores? Why would a diode modulator require tunable cores, while a 6ME8 wouldn't? The RCA BTS-1A Stereo Coder described on my web pages did the job without tunable cores and with only two tubes. Oh that's right. They used a couple of pots to balance. I really haven't gone past the speculation phase at this point and I appreciate the suggestions. I wasn't really making suggestions, just musing about various ways to build a stereo coder that I hadn't really thought about before. Well, musings help too  |
|
#7
Posted to rec.audio.tubes
|
|||
|
|||
|
Next Project
In article ,
flipper wrote: On Tue, 10 Jan 2012 22:37:32 -0600, John Byrns wrote: In article , flipper wrote: On Mon, 09 Jan 2012 22:43:06 -0600, John Byrns wrote: In article , flipper wrote: Well, Gort was a Christmas day success so why not FM next? Maybe even gasp stereo. So far I have a crude, but working, reactance tube FM Transmitter breadboarded and some 'ideas' for the MPX. http://flipperhome.dyndns.org/FM%20Stereo.htm I was trying to make it a "dollar days" transmitter, meaning the tubes cost $1/ea from ABCvacuumtubes, but the 6DT8 is 'full price' at $3.00. Hi Flipper, The 6ME8, and its similar cousins, has always frustrated me because I have never been able to figure out a reasonable way to combine the outputs from the two plates without using a transformer. I would like to use two 6ME8s, one for each stereo channel to generate the complete stereo composite baseband signal without using the L+R L-R matrix scheme. Unfortunately this requires an anode combining circuit that is phase linear from about 3 Hz to 53 kHz, a tall order for a transformer! If you mean using two 6ME8s switching the deflectors at 38KHz for a 'chopper' style matrix generator, I thought about that too but I don't think you need a transformer as you can simply add the outputs. You need a large symmetrical square wave, though, and that's not 'trivial'. I'll have to think about it a bit to be sure it actually works, but I think you have the right idea, with two tubes, you don't need the transformer to combine the plates, the subcarrier cancels without one. The second pair of plates could drive the other input to a push pull reactance tube modulator. If I am not missing something this is really cool, two 6ME8s make a linear multiplier type stereo coder that doesn't need a composite filter! As I am conceiving it the audio would drive the deflectors as in your design, and out of phase subcarrier signals would drive the grids of the two tubes. What could be simpler! I think we're talking different approaches, although the net result is the same (or else it wouldn't be FM stereo MPX). I'm talking about time division multiplexing, ala what's described here http://web.archive.org/web/200602140...ipod.com/stere o.htm Using time division multiplexing to generate the FM stereo signal became obsolete in the 1970s when the optimod was introduced. The optimod used a linear multiplier scheme to generate the L+R subcarrier signal, allowing higher modulation levels to be achieved without violating the FCC rules. Using the 6ME8 gets you the 38 KHz inversion. I.E. One plate is the 'inverse' of the other so you could feed a common 38KHz square wave, large enough to swing cutoff, to the deflectors on each, pre-emphasis audio to the grid, and use the L 'when high' plate on one to sum with the R 'when low' plate on the other. Just tie the two plates together through a common load resistor and whichever one is 'on' passes its channel's signal. The sum then, after a harmonics filter, has the pilot added and goes straight to the modulator. The other two plates go to B+ but aren't used. The square wave needs to be symmetrical and constant amplitude. Using a square wave requires low pass filtering the composite signal to remove the subcarrier harmonics. The 38 kHz subcarrier component at the output of the time division multiplexer is too large relative to the L+R signal, so a cross blend circuit is also required to reduce the effective L-R level feeding the time division multiplex circuit, so that the output matches the smaller L+R signal. Actually, ideal would be a sine, rather than square, perfectly aligned between 'on' and cutoff but cutoff isn't all that sharp so I doubt it would work. After further though, my idea is to use two 6ME8s, with the Left channel audio feed to the grid of one and the Right channel audio feed to the grid of the other. A pure 38 kHz sine wave would feed the deflectors of both tubes so that the tubes would operate as linear multipliers. The plates of the two tubes would be connected together, with opposite plates relative to the deflectors connected, and output taken from one pair of plates. DC bias would be applied to the deflectors both to balance out the subcarrier, and to provide the correct level of L+R relative to the L-R subcarrier sidebands. The relative L-R level is controlled by the level of the 38 kHz sine wave applied to the deflectors, and the relative L+R level is controlled by the DC bias applied between the deflectors. Trying to work this out on the 6ME8 transfer characteristics I noticed what appears to be an anomaly in the data used to draw the curves. Worse the plate dissipation is only 2 Watts per plate, so that you must operate in the lower part of the transfer characteristics which are difficult to read, it would have been better if they had expanded the curves vertically, and cut off the high current parts of the curves. -- Regards, John Byrns Surf my web pages at, http://fmamradios.com/ |
|
#8
Posted to rec.audio.tubes
|
|||
|
|||
|
Next Project
On Thu, 12 Jan 2012 00:27:30 -0600, John Byrns
wrote: In article , flipper wrote: On Tue, 10 Jan 2012 22:37:32 -0600, John Byrns wrote: In article , flipper wrote: On Mon, 09 Jan 2012 22:43:06 -0600, John Byrns wrote: In article , flipper wrote: Well, Gort was a Christmas day success so why not FM next? Maybe even gasp stereo. So far I have a crude, but working, reactance tube FM Transmitter breadboarded and some 'ideas' for the MPX. http://flipperhome.dyndns.org/FM%20Stereo.htm I was trying to make it a "dollar days" transmitter, meaning the tubes cost $1/ea from ABCvacuumtubes, but the 6DT8 is 'full price' at $3.00. Hi Flipper, The 6ME8, and its similar cousins, has always frustrated me because I have never been able to figure out a reasonable way to combine the outputs from the two plates without using a transformer. I would like to use two 6ME8s, one for each stereo channel to generate the complete stereo composite baseband signal without using the L+R L-R matrix scheme. Unfortunately this requires an anode combining circuit that is phase linear from about 3 Hz to 53 kHz, a tall order for a transformer! If you mean using two 6ME8s switching the deflectors at 38KHz for a 'chopper' style matrix generator, I thought about that too but I don't think you need a transformer as you can simply add the outputs. You need a large symmetrical square wave, though, and that's not 'trivial'. I'll have to think about it a bit to be sure it actually works, but I think you have the right idea, with two tubes, you don't need the transformer to combine the plates, the subcarrier cancels without one. The second pair of plates could drive the other input to a push pull reactance tube modulator. If I am not missing something this is really cool, two 6ME8s make a linear multiplier type stereo coder that doesn't need a composite filter! As I am conceiving it the audio would drive the deflectors as in your design, and out of phase subcarrier signals would drive the grids of the two tubes. What could be simpler! I think we're talking different approaches, although the net result is the same (or else it wouldn't be FM stereo MPX). I'm talking about time division multiplexing, ala what's described here http://web.archive.org/web/200602140...ipod.com/stere o.htm Using time division multiplexing to generate the FM stereo signal became obsolete in the 1970s when the optimod was introduced. The optimod used a linear multiplier scheme to generate the L+R subcarrier signal, allowing higher modulation levels to be achieved without violating the FCC rules. Hard to believe, but true, you actually used the word "obsolete" in a TUBE newsgroup. LOL Heck, *analog* is 'obsolete' Using the 6ME8 gets you the 38 KHz inversion. I.E. One plate is the 'inverse' of the other so you could feed a common 38KHz square wave, large enough to swing cutoff, to the deflectors on each, pre-emphasis audio to the grid, and use the L 'when high' plate on one to sum with the R 'when low' plate on the other. Just tie the two plates together through a common load resistor and whichever one is 'on' passes its channel's signal. The sum then, after a harmonics filter, has the pilot added and goes straight to the modulator. The other two plates go to B+ but aren't used. The square wave needs to be symmetrical and constant amplitude. Using a square wave requires low pass filtering the composite signal to remove the subcarrier harmonics. Yeah, except a lot of the BA1404 'kits' don't bother. The 38 kHz subcarrier component at the output of the time division multiplexer is too large relative to the L+R signal, so a cross blend circuit is also required to reduce the effective L-R level feeding the time division multiplex circuit, so that the output matches the smaller L+R signal. How can a TDM signal be 'bigger' than the signals it's chopping between? Actually, ideal would be a sine, rather than square, perfectly aligned between 'on' and cutoff but cutoff isn't all that sharp so I doubt it would work. After further though, my idea is to use two 6ME8s, with the Left channel audio feed to the grid of one and the Right channel audio feed to the grid of the other. A pure 38 kHz sine wave would feed the deflectors of both tubes so that the tubes would operate as linear multipliers. The plates of the two tubes would be connected together, with opposite plates relative to the deflectors connected, and output taken from one pair of plates. That's the topology I described with the 'ideal' being a sine wave, which, in TDM lingo, amounts to 'infinite' over sampling. Makes sense because TDM is essentially an 'approximation' of the linear case, which is where all the harmonics come from. DC bias would be applied to the deflectors both to balance out the subcarrier, and to provide the correct level of L+R relative to the L-R subcarrier sidebands. The relative L-R level is controlled by the level of the 38 kHz sine wave applied to the deflectors, and the relative L+R level is controlled by the DC bias applied between the deflectors. It's the sloping cutoff that bothers me. I'd think that's got to introduce distortion or bleed through. Trying to work this out on the 6ME8 transfer characteristics I noticed what appears to be an anomaly in the data used to draw the curves. Worse the plate dissipation is only 2 Watts per plate, so that you must operate in the lower part of the transfer characteristics which are difficult to read, it would have been better if they had expanded the curves vertically, and cut off the high current parts of the curves. I imagine they didn't expect that to be DC so the higher curves are sine peak. Still looks a bit 'too much' to me but there was probably some reason. |
|
#9
Posted to rec.audio.tubes
|
|||
|
|||
|
Next Project
In article ,
flipper wrote: On Thu, 12 Jan 2012 00:27:30 -0600, John Byrns wrote: In article , flipper wrote: I think we're talking different approaches, although the net result is the same (or else it wouldn't be FM stereo MPX). I'm talking about time division multiplexing, ala what's described here http://web.archive.org/web/200602140....tripod.com/st ere o.htm Using time division multiplexing to generate the FM stereo signal became obsolete in the 1970s when the optimod was introduced. The optimod used a linear multiplier scheme to generate the L+R subcarrier signal, allowing higher modulation levels to be achieved without violating the FCC rules. Hard to believe, but true, you actually used the word "obsolete" in a TUBE newsgroup. LOL Most, if not all, of the tube era broadcast FM stereo generators, as opposed to test bench equipment, used the subcarrier approach to FM stereo generation. I suspect this had nothing to do with tubes but had to do with the dominance of the serasoid modulator in the monophonic FM era. Transistors brought in new FM exciters and "time division multiplex" stereo generators. But what was old became new again when the increasing importance of FM brought in the optimod which reverted back to the subcarrier technique. Heck, *analog* is 'obsolete' There seems to be some opinion about that to the contrary. Using the 6ME8 gets you the 38 KHz inversion. I.E. One plate is the 'inverse' of the other so you could feed a common 38KHz square wave, large enough to swing cutoff, to the deflectors on each, pre-emphasis audio to the grid, and use the L 'when high' plate on one to sum with the R 'when low' plate on the other. Just tie the two plates together through a common load resistor and whichever one is 'on' passes its channel's signal. The sum then, after a harmonics filter, has the pilot added and goes straight to the modulator. The other two plates go to B+ but aren't used. The square wave needs to be symmetrical and constant amplitude. Using a square wave requires low pass filtering the composite signal to remove the subcarrier harmonics. Yeah, except a lot of the BA1404 'kits' don't bother. The 38 kHz subcarrier component at the output of the time division multiplexer is too large relative to the L+R signal, so a cross blend circuit is also required to reduce the effective L-R level feeding the time division multiplex circuit, so that the output matches the smaller L+R signal. How can a TDM signal be 'bigger' than the signals it's chopping between? That's easy, to illustrate it in the easiest possible way, think of the transmitter being modulated by a 50 Hz signal out of phase in the left and right channels. This produces only a DSBSC 38 kHz signal, no L +R signal. The period of the 50 Hz tone is so long with respect to the 38 kHz subcarrier signal, that if we look at a handful of 38 kHz cycles around the peak of the 50 Hz signal, it will appear pretty much as an unmodulated 38 kHz square wave. If you analyze the amplitude of the various frequency components of a square wave with an amplitude of unity, you will find that the amplitude of the fundamental is about 27% greater than the amplitude of the square wave. I could be slightly off on the on 27% number as I did the math in my head, but you get the idea. This extra 27% of subcarrier level that the "time division multiplex" system produces needs to be reduced, this can be accomplished by in phase cross blending between the two stereo channels to knock down the L-R level. -- Regards, John Byrns Surf my web pages at, http://fmamradios.com/ |
|
#10
Posted to rec.audio.tubes
|
|||
|
|||
|
Next Project
On Thu, 12 Jan 2012 22:10:58 -0600, John Byrns
wrote: In article , flipper wrote: On Thu, 12 Jan 2012 00:27:30 -0600, John Byrns wrote: In article , flipper wrote: I think we're talking different approaches, although the net result is the same (or else it wouldn't be FM stereo MPX). I'm talking about time division multiplexing, ala what's described here http://web.archive.org/web/200602140....tripod.com/st ere o.htm Using time division multiplexing to generate the FM stereo signal became obsolete in the 1970s when the optimod was introduced. The optimod used a linear multiplier scheme to generate the L+R subcarrier signal, allowing higher modulation levels to be achieved without violating the FCC rules. Hard to believe, but true, you actually used the word "obsolete" in a TUBE newsgroup. LOL Most, if not all, of the tube era broadcast FM stereo generators, as opposed to test bench equipment, used the subcarrier approach to FM stereo generation. I suspect this had nothing to do with tubes but had to do with the dominance of the serasoid modulator in the monophonic FM era. Transistors brought in new FM exciters and "time division multiplex" stereo generators. But what was old became new again when the increasing importance of FM brought in the optimod which reverted back to the subcarrier technique. Heck, *analog* is 'obsolete' There seems to be some opinion about that to the contrary. No doubt, but the Optimod 8600 is DSP. It's moot, really, because I wasn't intending to 'compete' with Optimod and the goal was something 'simple'. I'd be happy, at least to begin with, if something simply 'worked'. Using the 6ME8 gets you the 38 KHz inversion. I.E. One plate is the 'inverse' of the other so you could feed a common 38KHz square wave, large enough to swing cutoff, to the deflectors on each, pre-emphasis audio to the grid, and use the L 'when high' plate on one to sum with the R 'when low' plate on the other. Just tie the two plates together through a common load resistor and whichever one is 'on' passes its channel's signal. The sum then, after a harmonics filter, has the pilot added and goes straight to the modulator. The other two plates go to B+ but aren't used. The square wave needs to be symmetrical and constant amplitude. Using a square wave requires low pass filtering the composite signal to remove the subcarrier harmonics. Yeah, except a lot of the BA1404 'kits' don't bother. The 38 kHz subcarrier component at the output of the time division multiplexer is too large relative to the L+R signal, so a cross blend circuit is also required to reduce the effective L-R level feeding the time division multiplex circuit, so that the output matches the smaller L+R signal. How can a TDM signal be 'bigger' than the signals it's chopping between? That's easy, to illustrate it in the easiest possible way, think of the transmitter being modulated by a 50 Hz signal out of phase in the left and right channels. This produces only a DSBSC 38 kHz signal, no L +R signal. The period of the 50 Hz tone is so long with respect to the 38 kHz subcarrier signal, that if we look at a handful of 38 kHz cycles around the peak of the 50 Hz signal, it will appear pretty much as an unmodulated 38 kHz square wave. If you analyze the amplitude of the various frequency components of a square wave with an amplitude of unity, you will find that the amplitude of the fundamental is about 27% greater than the amplitude of the square wave. I could be slightly off on the on 27% number as I did the math in my head, but you get the idea. This extra 27% of subcarrier level that the "time division multiplex" system produces needs to be reduced, this can be accomplished by in phase cross blending between the two stereo channels to knock down the L-R level. How come none of those 'simple' circuits in the above link bother with this? |
|
#11
Posted to rec.audio.tubes
|
|||
|
|||
|
Next Project
In article ,
flipper wrote: On Thu, 12 Jan 2012 22:10:58 -0600, John Byrns wrote: In article , flipper wrote: On Thu, 12 Jan 2012 00:27:30 -0600, John Byrns wrote: In article , flipper wrote: I think we're talking different approaches, although the net result is the same (or else it wouldn't be FM stereo MPX). I'm talking about time division multiplexing, ala what's described here http://web.archive.org/web/200602140...ers.tripod.com /st ere o.htm Using time division multiplexing to generate the FM stereo signal became obsolete in the 1970s when the optimod was introduced. The optimod used a linear multiplier scheme to generate the L+R subcarrier signal, allowing higher modulation levels to be achieved without violating the FCC rules. Hard to believe, but true, you actually used the word "obsolete" in a TUBE newsgroup. LOL Most, if not all, of the tube era broadcast FM stereo generators, as opposed to test bench equipment, used the subcarrier approach to FM stereo generation. I suspect this had nothing to do with tubes but had to do with the dominance of the serasoid modulator in the monophonic FM era. Transistors brought in new FM exciters and "time division multiplex" stereo generators. But what was old became new again when the increasing importance of FM brought in the optimod which reverted back to the subcarrier technique. Heck, *analog* is 'obsolete' There seems to be some opinion about that to the contrary. No doubt, but the Optimod 8600 is DSP. Which raises the interesting question of whether the Optimod 8600 uses a subcarrier, or a TDM algorithm in its firmware? It's moot, really, because I wasn't intending to 'compete' with Optimod and the goal was something 'simple'. I'd be happy, at least to begin with, if something simply 'worked'. Using the 6ME8 gets you the 38 KHz inversion. I.E. One plate is the 'inverse' of the other so you could feed a common 38KHz square wave, large enough to swing cutoff, to the deflectors on each, pre-emphasis audio to the grid, and use the L 'when high' plate on one to sum with the R 'when low' plate on the other. Just tie the two plates together through a common load resistor and whichever one is 'on' passes its channel's signal. The sum then, after a harmonics filter, has the pilot added and goes straight to the modulator. The other two plates go to B+ but aren't used. The square wave needs to be symmetrical and constant amplitude. Using a square wave requires low pass filtering the composite signal to remove the subcarrier harmonics. Yeah, except a lot of the BA1404 'kits' don't bother. The 38 kHz subcarrier component at the output of the time division multiplexer is too large relative to the L+R signal, so a cross blend circuit is also required to reduce the effective L-R level feeding the time division multiplex circuit, so that the output matches the smaller L+R signal. How can a TDM signal be 'bigger' than the signals it's chopping between? That's easy, to illustrate it in the easiest possible way, think of the transmitter being modulated by a 50 Hz signal out of phase in the left and right channels. This produces only a DSBSC 38 kHz signal, no L +R signal. The period of the 50 Hz tone is so long with respect to the 38 kHz subcarrier signal, that if we look at a handful of 38 kHz cycles around the peak of the 50 Hz signal, it will appear pretty much as an unmodulated 38 kHz square wave. If you analyze the amplitude of the various frequency components of a square wave with an amplitude of unity, you will find that the amplitude of the fundamental is about 27% greater than the amplitude of the square wave. I could be slightly off on the on 27% number as I did the math in my head, but you get the idea. This extra 27% of subcarrier level that the "time division multiplex" system produces needs to be reduced, this can be accomplished by in phase cross blending between the two stereo channels to knock down the L-R level. How come none of those 'simple' circuits in the above link bother with this? I don't see any circuits, "simple" or otherwise at the "above link"? By the way a more direct link to that page is: http://transmitters.tripod.com/stereo.htm I assume that it is the cross blending that you are saying that the "circuits in the above link" don't bother with? How do you know they don't bother with it? IIRC the final "circuits" on that web page use an oversampling technique to generate the composite stereo signal rather than a 38 kHz square wave to do the switching. Oversampling greatly reduces the need for filtering, eliminating it in the limit. I'm not sure why so many people like that web page? I got into a lengthy discussion about that page 10 years ago, after another person posted it as recommended reading on an WWW Discussion Group. I have 50 some odd messages from the discussion group, and emails I eventually exchanged with the author of the page. While the web page has a superficial gloss to it, it is riddled with subtle errors, and half truths, that have the potential of leading an unwary reader astray, although I suppose that if you keep its title firmly in mind, and don't take everything it says as gospel, it sort of fulfills the promise of its title, In additions to the errors, there are also some unexplained artifacts in some of the plots. It isn't clear if these artifacts were somehow caused by the simulator program the web page author used generate the plots, or if there was something unusual about the "circuits" the author used. One good thing about the web page is that I did learn from it how the stereo coder kits that used multiple analog switches work, which in retrospect should have been obvious, but wasn't to me. I was familiar with the harmonic cancellation concept summing the outputs of separate modulators running at one or more harmonics of 38 kHz, with the output of the 38 kHz modulator. Zenith had done that in one of their early vacuum tube coders, using a 114 kHz modulator, in addition to the 38 kHz modulator, all vacuum tubes of course. However I wasn't familiar with the multiple analog switch oversampling technique, which sort of combines an analog multiplier and analog table lookup all in one simple circuit. -- Regards, John Byrns Surf my web pages at, http://fmamradios.com/ |
|
#12
Posted to rec.audio.tubes
|
|||
|
|||
|
Next Project
On Fri, 13 Jan 2012 15:13:04 -0600, John Byrns
wrote: In article , flipper wrote: On Thu, 12 Jan 2012 22:10:58 -0600, John Byrns wrote: In article , flipper wrote: On Thu, 12 Jan 2012 00:27:30 -0600, John Byrns wrote: In article , flipper wrote: I think we're talking different approaches, although the net result is the same (or else it wouldn't be FM stereo MPX). I'm talking about time division multiplexing, ala what's described here http://web.archive.org/web/200602140...ers.tripod.com /st ere o.htm Using time division multiplexing to generate the FM stereo signal became obsolete in the 1970s when the optimod was introduced. The optimod used a linear multiplier scheme to generate the L+R subcarrier signal, allowing higher modulation levels to be achieved without violating the FCC rules. Hard to believe, but true, you actually used the word "obsolete" in a TUBE newsgroup. LOL Most, if not all, of the tube era broadcast FM stereo generators, as opposed to test bench equipment, used the subcarrier approach to FM stereo generation. I suspect this had nothing to do with tubes but had to do with the dominance of the serasoid modulator in the monophonic FM era. Transistors brought in new FM exciters and "time division multiplex" stereo generators. But what was old became new again when the increasing importance of FM brought in the optimod which reverted back to the subcarrier technique. Heck, *analog* is 'obsolete' There seems to be some opinion about that to the contrary. No doubt, but the Optimod 8600 is DSP. Which raises the interesting question of whether the Optimod 8600 uses a subcarrier, or a TDM algorithm in its firmware? I don't know. Maybe something DSP 'specific' because they do mention in one spot 'can't do without DSP', although, that might have been audio processing related. It's moot, really, because I wasn't intending to 'compete' with Optimod and the goal was something 'simple'. I'd be happy, at least to begin with, if something simply 'worked'. Using the 6ME8 gets you the 38 KHz inversion. I.E. One plate is the 'inverse' of the other so you could feed a common 38KHz square wave, large enough to swing cutoff, to the deflectors on each, pre-emphasis audio to the grid, and use the L 'when high' plate on one to sum with the R 'when low' plate on the other. Just tie the two plates together through a common load resistor and whichever one is 'on' passes its channel's signal. The sum then, after a harmonics filter, has the pilot added and goes straight to the modulator. The other two plates go to B+ but aren't used. The square wave needs to be symmetrical and constant amplitude. Using a square wave requires low pass filtering the composite signal to remove the subcarrier harmonics. Yeah, except a lot of the BA1404 'kits' don't bother. The 38 kHz subcarrier component at the output of the time division multiplexer is too large relative to the L+R signal, so a cross blend circuit is also required to reduce the effective L-R level feeding the time division multiplex circuit, so that the output matches the smaller L+R signal. How can a TDM signal be 'bigger' than the signals it's chopping between? That's easy, to illustrate it in the easiest possible way, think of the transmitter being modulated by a 50 Hz signal out of phase in the left and right channels. This produces only a DSBSC 38 kHz signal, no L +R signal. The period of the 50 Hz tone is so long with respect to the 38 kHz subcarrier signal, that if we look at a handful of 38 kHz cycles around the peak of the 50 Hz signal, it will appear pretty much as an unmodulated 38 kHz square wave. If you analyze the amplitude of the various frequency components of a square wave with an amplitude of unity, you will find that the amplitude of the fundamental is about 27% greater than the amplitude of the square wave. I could be slightly off on the on 27% number as I did the math in my head, but you get the idea. This extra 27% of subcarrier level that the "time division multiplex" system produces needs to be reduced, this can be accomplished by in phase cross blending between the two stereo channels to knock down the L-R level. How come none of those 'simple' circuits in the above link bother with this? I don't see any circuits, "simple" or otherwise at the "above link"? By the way a more direct link to that page is: http://transmitters.tripod.com/stereo.htm I assume that it is the cross blending that you are saying that the "circuits in the above link" don't bother with? How do you know they don't bother with it? IIRC the final "circuits" on that web page use an oversampling technique to generate the composite stereo signal rather than a 38 kHz square wave to do the switching. Oversampling greatly reduces the need for filtering, eliminating it in the limit. I'm not sure why so many people like that web page? I got into a lengthy discussion about that page 10 years ago, after another person posted it as recommended reading on an WWW Discussion Group. I have 50 some odd messages from the discussion group, and emails I eventually exchanged with the author of the page. While the web page has a superficial gloss to it, it is riddled with subtle errors, and half truths, that have the potential of leading an unwary reader astray, although I suppose that if you keep its title firmly in mind, and don't take everything it says as gospel, it sort of fulfills the promise of its title, In additions to the errors, there are also some unexplained artifacts in some of the plots. It isn't clear if these artifacts were somehow caused by the simulator program the web page author used generate the plots, or if there was something unusual about the "circuits" the author used. One good thing about the web page is that I did learn from it how the stereo coder kits that used multiple analog switches work, which in retrospect should have been obvious, but wasn't to me. I was familiar with the harmonic cancellation concept summing the outputs of separate modulators running at one or more harmonics of 38 kHz, with the output of the 38 kHz modulator. Zenith had done that in one of their early vacuum tube coders, using a 114 kHz modulator, in addition to the 38 kHz modulator, all vacuum tubes of course. However I wasn't familiar with the multiple analog switch oversampling technique, which sort of combines an analog multiplier and analog table lookup all in one simple circuit. Oh, sorry. I though I had posted the first link but that was one he linked to. http://cappels.org/dproj/FM_MPX_STER...CIRCU IT.html He uses a PIC to simply switch the channels on and off, which is his 'simpler' version of one he also links to. http://www.sm0vpo.com/audio/stereo_enc.htm Other than a simple, single, RC low pass filter there's no 'post processing' of any kind. The page you don't like claims the BA1404 chip uses the same 38 KHz chopper method. |
|
#13
Posted to rec.audio.tubes
|
|||
|
|||
|
Next Project
In article ,
flipper wrote: On Fri, 13 Jan 2012 15:13:04 -0600, John Byrns wrote: In article , flipper wrote: On Thu, 12 Jan 2012 22:10:58 -0600, John Byrns wrote: In article , flipper wrote: Heck, *analog* is 'obsolete' There seems to be some opinion about that to the contrary. No doubt, but the Optimod 8600 is DSP. Which raises the interesting question of whether the Optimod 8600 uses a subcarrier, or a TDM algorithm in its firmware? I don't know. Maybe something DSP 'specific' because they do mention in one spot 'can't do without DSP', although, that might have been audio processing related. It could have been the stereo coder too, I gave some thought to how I would do a stereo coder using DSP, what techniques I would use, and one of them, while not theoretically impossible with analog, would certainly be nearly impossible in a practical sense. In DSP I would use several transversal filters which would be practically impossible to do in analog. If you analyze the amplitude of the various frequency components of a square wave with an amplitude of unity, you will find that the amplitude of the fundamental is about 27% greater than the amplitude of the square wave. I could be slightly off on the on 27% number as I did the math in my head, but you get the idea. This extra 27% of subcarrier level that the "time division multiplex" system produces needs to be reduced, this can be accomplished by in phase cross blending between the two stereo channels to knock down the L-R level. How come none of those 'simple' circuits in the above link bother with this? I don't see any circuits, "simple" or otherwise at the "above link"? Oh, sorry. I though I had posted the first link but that was one he linked to. http://cappels.org/dproj/FM_MPX_STER...LTIPLEX%20ENOC DER%20CIRCUIT.html Are you saying that the web page you referred to at http://web.archive.org/web/200602140...ers.tripod.com has a link to the "cappels" page? He uses a PIC to simply switch the channels on and off, which is his 'simpler' version of one he also links to. This circuit doesn't appear to include pre-emphasis, I wonder why? Without pre-emphasis the transmitter is going to be mighty dull sounding. While there is no overt cross blending in this circuit, the series resistance of the shunt switches in the "PIC" will provide some cross blending. I wonder if he picked the resistor values in the "resistor matrix" to optimize this? I would put pots in series with the "PIC" switches to make the switch resistance equal in the two channels, and to optimize the stereo separation. IIRC the maximum stereo separation possible from square wave TDM without cross blending is something on the order of only 18 dB, perhaps that is enough for a project like this? http://www.sm0vpo.com/audio/stereo_enc.htm Other than a simple, single, RC low pass filter there's no 'post processing' of any kind. This one at least includes pre-emphasis, but again I don't see any overt cross blending, and unlike the shunt switches used in the first circuit, the series switches used in this circuit wouldn't provide any covert cross blending. The page you don't like claims the BA1404 chip uses the same 38 KHz chopper method. But I'll bet that the BA1404 chip includes cross blending, without which the stereo separation is abysmal as noted above. -- Regards, John Byrns Surf my web pages at, http://fmamradios.com/ |
|
#14
Posted to rec.audio.tubes
|
|||
|
|||
|
Next Project
On Sat, 14 Jan 2012 17:11:36 -0600, John Byrns
wrote: In article , flipper wrote: On Fri, 13 Jan 2012 15:13:04 -0600, John Byrns wrote: In article , flipper wrote: On Thu, 12 Jan 2012 22:10:58 -0600, John Byrns wrote: In article , flipper wrote: Heck, *analog* is 'obsolete' There seems to be some opinion about that to the contrary. No doubt, but the Optimod 8600 is DSP. Which raises the interesting question of whether the Optimod 8600 uses a subcarrier, or a TDM algorithm in its firmware? I don't know. Maybe something DSP 'specific' because they do mention in one spot 'can't do without DSP', although, that might have been audio processing related. It could have been the stereo coder too, I gave some thought to how I would do a stereo coder using DSP, what techniques I would use, and one of them, while not theoretically impossible with analog, would certainly be nearly impossible in a practical sense. In DSP I would use several transversal filters which would be practically impossible to do in analog. That's the great thing about DSP, you don't have to program in 'reality' mode. hehe You only have to eventually 'interface' to reality If you analyze the amplitude of the various frequency components of a square wave with an amplitude of unity, you will find that the amplitude of the fundamental is about 27% greater than the amplitude of the square wave. I could be slightly off on the on 27% number as I did the math in my head, but you get the idea. This extra 27% of subcarrier level that the "time division multiplex" system produces needs to be reduced, this can be accomplished by in phase cross blending between the two stereo channels to knock down the L-R level. How come none of those 'simple' circuits in the above link bother with this? I don't see any circuits, "simple" or otherwise at the "above link"? Oh, sorry. I though I had posted the first link but that was one he linked to. http://cappels.org/dproj/FM_MPX_STER...LTIPLEX%20ENOC DER%20CIRCUIT.html Are you saying that the web page you referred to at http://web.archive.org/web/200602140...ers.tripod.com has a link to the "cappels" page? No, cappels is the 'first link' and his page references the other. He uses a PIC to simply switch the channels on and off, which is his 'simpler' version of one he also links to. This circuit doesn't appear to include pre-emphasis, I wonder why? Without pre-emphasis the transmitter is going to be mighty dull sounding. He's doing pre-emphasis in whatever is driving the inputs, just like I did at first before adding the preamp. He mentions adding pre-emphasis near the bottom. While there is no overt cross blending in this circuit, the series resistance of the shunt switches in the "PIC" will provide some cross blending. I wonder if he picked the resistor values in the "resistor matrix" to optimize this? I doubt it because that would be 'clever' enough to mention. The ATTINY12 datasheet indicates the switch on resistance is 50 Ohms or less, which doesn't seem significant. I would put pots in series with the "PIC" switches to make the switch resistance equal in the two channels, and to optimize the stereo separation. Wouldn't that also affect L+R? IIRC the maximum stereo separation possible from square wave TDM without cross blending is something on the order of only 18 dB, perhaps that is enough for a project like this? Unfortunately, neither he nor the other project gives any performance numbers. http://www.sm0vpo.com/audio/stereo_enc.htm Other than a simple, single, RC low pass filter there's no 'post processing' of any kind. This one at least includes pre-emphasis, but again I don't see any overt cross blending, and unlike the shunt switches used in the first circuit, the series switches used in this circuit wouldn't provide any covert cross blending. Yes, and the other is just doing 'the same thing' but 'simpler'. The page you don't like claims the BA1404 chip uses the same 38 KHz chopper method. But I'll bet that the BA1404 chip includes cross blending, without which the stereo separation is abysmal as noted above. Unfortunately, the BA1404 datasheet is entirely mysterious about it, just drawing a 'MPX' box. They spec separation as 45 dB "typical" with 25 dB minimum. |
|
#15
Posted to rec.audio.tubes
|
|||
|
|||
|
Next Project
In article ,
flipper wrote: On Sat, 14 Jan 2012 17:11:36 -0600, John Byrns wrote: In article , flipper wrote: On Fri, 13 Jan 2012 15:13:04 -0600, John Byrns wrote: In article , flipper wrote: How come none of those 'simple' circuits in the above link bother with this? I don't see any circuits, "simple" or otherwise at the "above link"? Oh, sorry. I though I had posted the first link but that was one he linked to. http://cappels.org/dproj/FM_MPX_STER...0MULTIPLEX%20E NOC DER%20CIRCUIT.html Are you saying that the web page you referred to at http://web.archive.org/web/200602140...ers.tripod.com has a link to the "cappels" page? No, cappels is the 'first link' and his page references the other. Yeah, as I was pressing the send button I was beginning to suspect that might be the case, as I was familiar with the "transmitters.tripod" site in 2002, and as I pressed the send button I remembered seeing a 2007 date mentioned on the "cappels" site. I just checked and you are indeed correct about which site references which. He uses a PIC to simply switch the channels on and off, which is his 'simpler' version of one he also links to. This circuit doesn't appear to include pre-emphasis, I wonder why? Without pre-emphasis the transmitter is going to be mighty dull sounding. He's doing pre-emphasis in whatever is driving the inputs, just like I did at first before adding the preamp. He mentions adding pre-emphasis near the bottom. I must confess that I haven't yet gotten around to reading either of the two sites you referenced, I have only looked at the "circuits" so far, but I will soon get around to reading the entire pages. While there is no overt cross blending in this circuit, the series resistance of the shunt switches in the "PIC" will provide some cross blending. I wonder if he picked the resistor values in the "resistor matrix" to optimize this? I doubt it because that would be 'clever' enough to mention. The ATTINY12 datasheet indicates the switch on resistance is 50 Ohms or less, which doesn't seem significant. Agreed, 50 Ohms wouldn't have a large effect. I would put pots in series with the "PIC" switches to make the switch resistance equal in the two channels, and to optimize the stereo separation. Wouldn't that also affect L+R? Yes, but in the opposite direction, so at some point the ratio of L+R to L-R subcarrier should be correct. When the switch resistance is infinite the L+R is unaffected, i.e. at its maximum, and when the switch resistance is zero Ohms, the L+R is attenuated 6 dB. When the switch resistance is infinite no subcarrier is produced, and when the switch resistance is zero Ohms, the subcarrier is at its maximum level. A switch resistance somewhere between zero Ohms and infinity should produce a signal with the proper ratio. IIRC the maximum stereo separation possible from square wave TDM without cross blending is something on the order of only 18 dB, perhaps that is enough for a project like this? Unfortunately, neither he nor the other project gives any performance numbers. http://www.sm0vpo.com/audio/stereo_enc.htm Other than a simple, single, RC low pass filter there's no 'post processing' of any kind. This one at least includes pre-emphasis, but again I don't see any overt cross blending, and unlike the shunt switches used in the first circuit, the series switches used in this circuit wouldn't provide any covert cross blending. Yes, and the other is just doing 'the same thing' but 'simpler'. I wonder what the simplest tube circuit is? My entry would be three tubes, a 38 kHz oscillator, 19 kHz divider, and a quad diode tube operating as a pair of shunt switches driven by a 38 kHz transformer. The page you don't like claims the BA1404 chip uses the same 38 KHz chopper method. But I'll bet that the BA1404 chip includes cross blending, without which the stereo separation is abysmal as noted above. Unfortunately, the BA1404 datasheet is entirely mysterious about it, just drawing a 'MPX' box. They spec separation as 45 dB "typical" with 25 dB minimum. Assuming my 18 dB separation guess is anywhere near correct, the BA1404 must be using cross blending to get a typical separation of 45 dB! Or alternatively they could be using a narrower sampling function, rather than a square wave. IIRC A more impulsive sampling function reduces the need for cross blending, but increases the level of the sucarrier harmonics considerably, necessitating better filtering of the output. -- Regards, John Byrns Surf my web pages at, http://fmamradios.com/ |
| Reply |
| Thread Tools | |
| Display Modes | |
|
|
Similar Threads
|
||||
| Thread | Forum | |||
| DIY project... | Tech | |||
| Looking for some D.I Y project | Tech | |||
| FS: UAD-1 Project Pak | Pro Audio | |||
| FS: UAD-1 Project Pak | Pro Audio | |||
Linear Mode
