|
MW Receiver progress report
Hi folks,
I finally got the AM sync detector running. The hardest part of it is the loop filter. I haven't been able to find any information on loop filter design, when there is no divider in the feedback loop. I have discovered that with a 4046, you set your lock range by setting the Fmin, Fmax of your VCO, while the capture range is set by the loop filter. With a wideband loop filter, the demodulated audio sounds fine, but with a + or - 3kHz capture range, there is a chirping type of distortion on the modulation peaks. In this case, settling time is 50mSec, while with the broadband loop, settling time is 100uSec. Overall, these things aren't too bad to build, except for that loop filter. Once I get that optimized, I will have a sync detector that is worthy of being sold as a stand alone unit. Projected price of this unit will be around the 100 dollar mark. Pete |
Behold, Pete KE9OA signalled from keyed 4-1000A filament:
Projected price of this unit will be around the 100 dollar mark. Amazing! -- Gregg *It's probably useful, even if it can't be SPICE'd* http://geek.scorpiorising.ca |
Pete KE9OA wrote:
Once I get that optimized, I will have a sync detector that is worthy of being sold as a stand alone unit. Projected price of this unit will be around the 100 dollar mark. Do you mean the MW receiver or outboard sync' detector for $100? -----= Posted via Newsfeeds.Com, Uncensored Usenet News =----- http://www.newsfeeds.com - The #1 Newsgroup Service in the World! -----== Over 100,000 Newsgroups - 19 Different Servers! =----- |
"starman" wrote in message ... Pete KE9OA wrote: Once I get that optimized, I will have a sync detector that is worthy of being sold as a stand alone unit. Projected price of this unit will be around the 100 dollar mark. Do you mean the MW receiver or outboard sync' detector for $100? I had a nice phone conversation with Pete yesterday- his sync detector sounds wonderful- can't wait to integrate it into my version of his HF all mode RX It is the sync detector that will be priced in the $100 range. Dale W4OP |
Dale Parfitt wrote:
"starman" wrote in message ... Pete KE9OA wrote: Once I get that optimized, I will have a sync detector that is worthy of being sold as a stand alone unit. Projected price of this unit will be around the 100 dollar mark. Do you mean the MW receiver or outboard sync' detector for $100? I had a nice phone conversation with Pete yesterday- his sync detector sounds wonderful- can't wait to integrate it into my version of his HF all mode RX It is the sync detector that will be priced in the $100 range. Dale W4OP Will it be similar to the Sherwood SE-3 with both double and single sideband selection? -----= Posted via Newsfeeds.Com, Uncensored Usenet News =----- http://www.newsfeeds.com - The #1 Newsgroup Service in the World! -----== Over 100,000 Newsgroups - 19 Different Servers! =----- |
|
I mean the outboard detector. It will have a choice of quasi-sync detection
and true PLL based sync detection. Pete starman wrote in message ... Pete KE9OA wrote: Once I get that optimized, I will have a sync detector that is worthy of being sold as a stand alone unit. Projected price of this unit will be around the 100 dollar mark. Do you mean the MW receiver or outboard sync' detector for $100? -----= Posted via Newsfeeds.Com, Uncensored Usenet News =----- http://www.newsfeeds.com - The #1 Newsgroup Service in the World! -----== Over 100,000 Newsgroups - 19 Different Servers! =----- |
Thanks John! I did finish the MC1350-NE602 design, but took it a step
further, integrating that circuit into an SA637 chip. The SA637 does an even better job than the first circuit, eliminating the need for that coupling transformer that is needed for the MC1350 for best performance. Once I got that circuit successfully developed, I decided to go with a true PLL type of sync detector. I just got through with the board layout yesterday, using an SA637 as a limiting stage, feeding the limited I.F. signal into the phase detector input of a 4046 PLL chip. The I.F. signal is split into two branches............the 1st branch is applied to the I.F. input of the SA637, while the 2nd branch is applied to the RF port of an NE602. The VCO output of the 4046 is applied to the LO input of the NE602, and the audio output is taken from the I.F. port of the NE602. I decided to upgrade the circuit a bit, including a ceramic bandpass filter between the I.F. output and the limiter input of the SA637. It should work pretty well, but I still want to be able to vary the phase of the 4046's VCO signal. This way, I will end up with zero degrees phase shift through the system, affording good performance. It is important to keep the absolute system phase shift with plus or minus a few degrees. It has been quite a bit of fun so far. It will be cool, once our software guy writes the code for the spectrum analyzer display. We are still a few months off, but things are looking good. The active loopstick circuit is working well enough to put on the market as a stand-alone unit. The gain is quite high, so I may need to put a variable attenuator on the output. Because I am using a source follower ahead of the final RF amplifier, the unloaded Q of the tuned circuit can be taken advantage of. I am not sure it I will put any regeneration into the active loopstick design yet. None of the regeneration circuits that I have found on the Internet have worked very well. Pete John Crabtree wrote in message ... "Pete KE9OA" on 12/20/03 wrote: Hi folks, I finally got the AM sync detector running. The hardest part of it is the loop filter. I haven't been able to find any information on loop filter design, when there is no divider in the feedback loop. I have discovered that with a 4046, you set your lock range by setting the Fmin, Fmax of your VCO, while the capture range is set by the loop filter. With a wideband loop filter, the demodulated audio sounds fine, but with a + or - 3kHz capture range, there is a chirping type of distortion on the modulation peaks. In this case, settling time is 50mSec, while with the broadband loop, settling time is 100uSec. Overall, these things aren't too bad to build, except for that loop filter. Once I get that optimized, I will have a sync detector that is worthy of being sold as a stand alone unit. Projected price of this unit will be around the 100 dollar mark. Pete Are we to understand that you have moved away from using the MC1350 and NE602 in a homodyne configuration (your update of 2003/12/05) to the synchrodyne approach with a PLL oscillator ? I have just dug out an article: Trevor Wheatley (of Surrey Electronics), "AM synchronous demodulator", Electronics and Wireless World, Sept 1989, pp858-860 It offers SSB (LSB or USB), ISB, envelope, DSB and quadrature detection. Reading the article closely shows that a primitive form of passband tuning is available as well. The article has one of the most comprehensive sync detector circuits which I have seen. It also has some interesting comments on the effects of phase modulated carriers (used on some LW stations in Europe) and allowable loop bandwidths. Keep up the good work. 73 John KC0GGH |
I've got the latest version of the sync detector completed. When I called
Future Electronics for some quotes on the SA637, I discovered that this chip had been replaced by the SA647. Both of the chips are discontinued. I've got about 1400 of the 637s in my private stock, but that won't be enough to put this into production, so the next step is to design a sync detector, using an Analog Devices AD607. The folks at AD were nice enough to send me a workable application circuit, but unfortunately, this is going to require a four layer board, so I am not going to be able to etch the boards at home with this one. Not a bad thing in itself, but I am going to have a board house do the prototype boards. Pete Pete KE9OA wrote in message ... Thanks John! I did finish the MC1350-NE602 design, but took it a step further, integrating that circuit into an SA637 chip. The SA637 does an even better job than the first circuit, eliminating the need for that coupling transformer that is needed for the MC1350 for best performance. Once I got that circuit successfully developed, I decided to go with a true PLL type of sync detector. I just got through with the board layout yesterday, using an SA637 as a limiting stage, feeding the limited I.F. signal into the phase detector input of a 4046 PLL chip. The I.F. signal is split into two branches............the 1st branch is applied to the I.F. input of the SA637, while the 2nd branch is applied to the RF port of an NE602. The VCO output of the 4046 is applied to the LO input of the NE602, and the audio output is taken from the I.F. port of the NE602. I decided to upgrade the circuit a bit, including a ceramic bandpass filter between the I.F. output and the limiter input of the SA637. It should work pretty well, but I still want to be able to vary the phase of the 4046's VCO signal. This way, I will end up with zero degrees phase shift through the system, affording good performance. It is important to keep the absolute system phase shift with plus or minus a few degrees. It has been quite a bit of fun so far. It will be cool, once our software guy writes the code for the spectrum analyzer display. We are still a few months off, but things are looking good. The active loopstick circuit is working well enough to put on the market as a stand-alone unit. The gain is quite high, so I may need to put a variable attenuator on the output. Because I am using a source follower ahead of the final RF amplifier, the unloaded Q of the tuned circuit can be taken advantage of. I am not sure it I will put any regeneration into the active loopstick design yet. None of the regeneration circuits that I have found on the Internet have worked very well. Pete John Crabtree wrote in message ... "Pete KE9OA" on 12/20/03 wrote: Hi folks, I finally got the AM sync detector running. The hardest part of it is the loop filter. I haven't been able to find any information on loop filter design, when there is no divider in the feedback loop. I have discovered that with a 4046, you set your lock range by setting the Fmin, Fmax of your VCO, while the capture range is set by the loop filter. With a wideband loop filter, the demodulated audio sounds fine, but with a + or - 3kHz capture range, there is a chirping type of distortion on the modulation peaks. In this case, settling time is 50mSec, while with the broadband loop, settling time is 100uSec. Overall, these things aren't too bad to build, except for that loop filter. Once I get that optimized, I will have a sync detector that is worthy of being sold as a stand alone unit. Projected price of this unit will be around the 100 dollar mark. Pete Are we to understand that you have moved away from using the MC1350 and NE602 in a homodyne configuration (your update of 2003/12/05) to the synchrodyne approach with a PLL oscillator ? I have just dug out an article: Trevor Wheatley (of Surrey Electronics), "AM synchronous demodulator", Electronics and Wireless World, Sept 1989, pp858-860 It offers SSB (LSB or USB), ISB, envelope, DSB and quadrature detection. Reading the article closely shows that a primitive form of passband tuning is available as well. The article has one of the most comprehensive sync detector circuits which I have seen. It also has some interesting comments on the effects of phase modulated carriers (used on some LW stations in Europe) and allowable loop bandwidths. Keep up the good work. 73 John KC0GGH |
"Pete KE9OA" wrote in message [snip]
this into production, so the next step is to design a sync detector, using an Analog Devices AD607. The folks at AD were nice enough to send me a workable application circuit, but unfortunately, this is going to require a four layer board, so I am not going to be able to etch the boards at home with this one. Not a bad thing in itself, but I am going to have a board house do the prototype boards. [snip] Pete, here's an article on using the AD607 as a sync demod: http://home.att.net/~wa1sov/technical/sync_det.html 73, Tom |
Thanks Tom,
I do have that article, and that circuit is similar to the quasi-sync detector that I developed a couple of weeks ago. The appnote that i have describes a design that I believe was created by one of the application engineers at AD. This is a really cool circuit, because it is optimized for 455kHz. I've got the board layout almost done...........maybe later this week, I will have it built up. I've bumped up the overload point on the receiver to 80,000uv, so it isn't too bad now. Measured distortion with either the envelope detector or the current version of the sync detector is .4% @ 300uV, but there is still something I don't like about the sound of that sync detector. If I were to try to market the receiver with the current sync detector, it would get shot down by anybody who was unfortunate enough to buy it. Still more work to do in this area. One fellow from Norway suggested that I have an I.F. output on the radio, so that folks can use their own sync detector. I could do that, but my goal is to design a sync detector that is better than anything else on the market, regardless of price. Hopefully, I will be able to do it. Pete tom Holden wrote in message om... "Pete KE9OA" wrote in message [snip] this into production, so the next step is to design a sync detector, using an Analog Devices AD607. The folks at AD were nice enough to send me a workable application circuit, but unfortunately, this is going to require a four layer board, so I am not going to be able to etch the boards at home with this one. Not a bad thing in itself, but I am going to have a board house do the prototype boards. [snip] Pete, here's an article on using the AD607 as a sync demod: http://home.att.net/~wa1sov/technical/sync_det.html 73, Tom |
Pete KE9OA wrote:
...my goal is to design a sync detector that is better than anything else on the market, regardless of price. Hopefully, I will be able to do it. I appreciate you're striving for the best Pete but I think most everyone on this group would be very satisfied if the sync' works like the R8B or AOR-7030. BTW- The sync' should be double sideband in addition to sideband selectable. -----= Posted via Newsfeeds.Com, Uncensored Usenet News =----- http://www.newsfeeds.com - The #1 Newsgroup Service in the World! -----== Over 100,000 Newsgroups - 19 Different Servers! =----- |
I am not sure how to do the selectable sideband function, but I can figure
it out, I will sure throw that function in. I just ordered some samples of the high speed op-amp that is used for squaring up the signal to the phase detector. They should be in on Wednesday. I am doing two versions of the detector...........one of them will use an AD op-amp, while the other will use a Burr-Brown device. This sync detector will have two ceramic bandpass filters in the signal chain.........one of them will be between the mixer output and the I.F. input, while the other will be between the I.F. output and the demodulator input. It shoud be a low-noise system. AD specifies this configuration as having an MDS of -90dBm. I will be feeding in a -20dBm signal, so the earlier stages of the receiver will have more than enough takeover gain. If there is enough interest in the circuit design, I will post the AD application note up on my website. Thanks for encouragement! Pete starman wrote in message ... Pete KE9OA wrote: ...my goal is to design a sync detector that is better than anything else on the market, regardless of price. Hopefully, I will be able to do it. I appreciate you're striving for the best Pete but I think most everyone on this group would be very satisfied if the sync' works like the R8B or AOR-7030. BTW- The sync' should be double sideband in addition to sideband selectable. -----= Posted via Newsfeeds.Com, Uncensored Usenet News =----- http://www.newsfeeds.com - The #1 Newsgroup Service in the World! -----== Over 100,000 Newsgroups - 19 Different Servers! =----- |
|
|
Thanks John...........that will be good.
Pete John Crabtree wrote in message ... "Pete KE9OA" on 12/30/03 wrote: I am not sure how to do the selectable sideband function, but I can figure it out, I will sure throw that function in. I just ordered some samples of the high speed op-amp that is used for squaring up the signal to the phase detector. They should be in on Wednesday. I am doing two versions of the detector...........one of them will use an AD op-amp, while the other will use a Burr-Brown device. This sync detector will have two ceramic bandpass filters in the signal chain.........one of them will be between the mixer output and the I.F. input, while the other will be between the I.F. output and the demodulator input. It shoud be a low-noise system. AD specifies this configuration as having an MDS of -90dBm. I will be feeding in a -20dBm signal, so the earlier stages of the receiver will have more than enough takeover gain. If there is enough interest in the circuit design, I will post the AD application note up on my website. Thanks for encouragement! Pete You have some options as to how to provide selectable sideband in the synchronous detector: 1. You can do it with passband tuning and move the signal carrier to the edge of the IF filter passband. IIRC this is the way that you can select the sideband in the AOR7030. I suspect that doing this will cause some phase distortion to the carrier if you place it on the edge of the passband. To what extent this matters I do not know. 2. Once you have the I and Q signals from a quadrature detector, you can use all-pass networks to phase the outputs and then add or subtract as necessary. Sony did this with the ICF-2010. Trevor Brook in his Electronics and Wireless World article did this as well. One issue with this approach is that the level of opposite sideband reduction is dependent upon the quality and number of stages in the phasing networks. IIRC Sony only achieve ca. 25db of opposite sideband reduction. The quality of phasing networks has been discussed in the ham radio literature re. the generation of SSB signals. It is possible to design simple networks which are 'OK' over the range 300 to 3000Hz. Once you have the Q output, it then should be very straightforward to offfer quadrature detection as well, where you null out the strongest station on the frequency to which the detector has locked. Which ever way you might choose, there are inevitable compromises. Another issue is what audio bandwidth is necessary in an AM receiver. Some time ago, in Short Wave Magazine (UK), John Wilson showed the spectrum analyser display of a BBC MW broadcast station. It was tightly filtered above 4+ kHz to stay with the allowed channel. Many AM stations also process their signals.. I like your idea of having two ceramic filters before the synchronous detector. I see that the oscillator in the AD607 has a wide tuning range, and it would be very unhelpful if it locked to the 'wrong' signal. As a contrast the synch oscillator in the ICF-2010, which is also used as the BFO, can only be moved a small frequency eg. 3 kHz (please do not quote me on this - I did measure it once) either side of 455kHz. I must admit that I am still intrigued by the idea in Brook's article of having using the sync. oscillator as a BFO and restricting it to a very narrow (+/- 20Hz ?) locking range. One could then listen normally with envelope detection, turn on the BFO to get ECSS detection, and then once one has obtained a zero beat, turn it to sync mode. However it is not something which I would wish to put into a general purpose radio. Let me register my interest in seeing the AD application note on your web site. If you want a copy of the Brook article, please contact me off list. 73 John KC0GGH |
It looks pretty good............I should take a look at my Drake schematics.
I hope they aren't using those impossible to get chips, you know, the ones that all of the radio manufacturers have bought up! Anyway, I got the AD607 board layout completed, but it required four layers. I am going to try to reduce it to two layers, at least for the 1st article, so I can build it up this weekend. Too bad that Analog Devices doesn't internally bias their chips, the way that Philips does. With the quagmire if external biasing resistors, this chip reminds me more of something like the OLD,OLD MC1496, which came out in the last century, not the modern day chip that it is touted as being. I do have the AD607 Eval Board, but that thing is harder to use than just doing my own board layout. I can understand why the manufacturers of those external sync boards charge so much money. Quite a bit of development time. I am still shooting for the 100 to 125 dollar range for this board, if I can get it into existance! Pete starman wrote in message ... John Crabtree wrote: "Pete KE9OA" on 12/30/03 wrote: I am not sure how to do the selectable sideband function, but I can figure it out, I will sure throw that function in. I just ordered some samples of the high speed op-amp that is used for squaring up the signal to the phase detector. They should be in on Wednesday. I am doing two versions of the detector...........one of them will use an AD op-amp, while the other will use a Burr-Brown device. This sync detector will have two ceramic bandpass filters in the signal chain.........one of them will be between the mixer output and the I.F. input, while the other will be between the I.F. output and the demodulator input. It shoud be a low-noise system. AD specifies this configuration as having an MDS of -90dBm. I will be feeding in a -20dBm signal, so the earlier stages of the receiver will have more than enough takeover gain. If there is enough interest in the circuit design, I will post the AD application note up on my website. Thanks for encouragement! Pete You have some options as to how to provide selectable sideband in the synchronous detector: 1. You can do it with passband tuning and move the signal carrier to the edge of the IF filter passband. IIRC this is the way that you can select the sideband in the AOR7030. I suspect that doing this will cause some phase distortion to the carrier if you place it on the edge of the passband. To what extent this matters I do not know. 2. Once you have the I and Q signals from a quadrature detector, you can use all-pass networks to phase the outputs and then add or subtract as necessary. Sony did this with the ICF-2010. Trevor Brook in his Electronics and Wireless World article did this as well. One issue with this approach is that the level of opposite sideband reduction is dependent upon the quality and number of stages in the phasing networks. IIRC Sony only achieve ca. 25db of opposite sideband reduction. The quality of phasing networks has been discussed in the ham radio literature re. the generation of SSB signals. It is possible to design simple networks which are 'OK' over the range 300 to 3000Hz. Once you have the Q output, it then should be very straightforward to offfer quadrature detection as well, where you null out the strongest station on the frequency to which the detector has locked. Which ever way you might choose, there are inevitable compromises. snipped Pete, I'm not sure if you mentioned it but I assume you're working with an I.F. of 455-Khz? The Drake-SW8 and it's clone the Sat-800, implement the sync' detector in the second I.F. at 455-Khz while the Drake-R8x does it at 50-Khz. It's basically the same circuit in both receivers, just operating at different I.F.'s. FWIW- I vote for the I/Q method for selectable sideband sync' detection. It seems like the elegant solution to me. -----= Posted via Newsfeeds.Com, Uncensored Usenet News =----- http://www.newsfeeds.com - The #1 Newsgroup Service in the World! -----== Over 100,000 Newsgroups - 19 Different Servers! =----- |
On Tue, 30 Dec 2003 07:08:46 GMT, "Pete KE9OA"
wrote: I am not sure how to do the selectable sideband function, but I can figure it out, I will sure throw that function in. On my bench as I type this is a nearly completed audio phase shift network for KK7B's R2pro. It is designed for handling the 90 degree I and Q phase shifts for audio from below 300 Hz to at least 4 kHz. Yes, it uses 1% components for the resistors and capacitors. However, it's built with commonplace operational amplifiers. With just an amplitude balance pot, one should be able to achieve opposite sideband rejection ratios of at least 50 dB. If you really want to get the very best out of it, one could reach rejection ratios of 60 dB, though KK7B says that maintaining this performance over typical room temperature shifts is probably more trouble than it's worth. To put this sort of performance in perspective, most consumer SSB receivers audio noise floors would mask the opposite sideband. Jake Brodsky, AB3A "Beware of the massive impossible!" |
I remember doing something with one of AD's quadrature modulators, using an
AD9854 as the LO source. I came up with a circuit that would give 50dB opposite sideband suppression from 4MHz to 1.5GHz. I don't know if Motorola ever applied for a patent for that circuit. I guess that circuit could also work. It sounds like you have taken a sound approach. Pete Jake Brodsky wrote in message ... On Tue, 30 Dec 2003 07:08:46 GMT, "Pete KE9OA" wrote: I am not sure how to do the selectable sideband function, but I can figure it out, I will sure throw that function in. On my bench as I type this is a nearly completed audio phase shift network for KK7B's R2pro. It is designed for handling the 90 degree I and Q phase shifts for audio from below 300 Hz to at least 4 kHz. Yes, it uses 1% components for the resistors and capacitors. However, it's built with commonplace operational amplifiers. With just an amplitude balance pot, one should be able to achieve opposite sideband rejection ratios of at least 50 dB. If you really want to get the very best out of it, one could reach rejection ratios of 60 dB, though KK7B says that maintaining this performance over typical room temperature shifts is probably more trouble than it's worth. To put this sort of performance in perspective, most consumer SSB receivers audio noise floors would mask the opposite sideband. Jake Brodsky, AB3A "Beware of the massive impossible!" |
| All times are GMT +1. The time now is 04:10 AM. |
Powered by vBulletin® Copyright ©2000 - 2025, Jelsoft Enterprises Ltd.
RadioBanter.com