Receive Antenna & Impedance "Matching"
Re: Receive Antenna & Impedance "Matching"
Have started a new thread on pre-selection in "GENERAL".
Gianfranco
Gianfranco
Reason: No reason
Re: Receive Antenna & Impedance "Matching"
Well done and looking good Gianfranco.
No further construction progress here, been setting up a Pi on WSPR (but it's only a 2B and groaning under the load) so as to free up the Windows VM for better testing, plus I'm on a short leash doing chores.
The current preselector arrangement as posted using the two vertical coils has now been running for several days on 30m WSPR and appears to be performing well, now I'm slowly working towards setting up two identical receivers to do a simultaneous WSPR comparison using a single antenna via a splitter as that would give definitive SNR and received spot count comparisons.
I really need to do that and assess the results with broader V narrower passbands before deciding what to do next with the filter.
Looking forward to your next post.
Regards, Phil
No further construction progress here, been setting up a Pi on WSPR (but it's only a 2B and groaning under the load) so as to free up the Windows VM for better testing, plus I'm on a short leash doing chores.
The current preselector arrangement as posted using the two vertical coils has now been running for several days on 30m WSPR and appears to be performing well, now I'm slowly working towards setting up two identical receivers to do a simultaneous WSPR comparison using a single antenna via a splitter as that would give definitive SNR and received spot count comparisons.
I really need to do that and assess the results with broader V narrower passbands before deciding what to do next with the filter.
Looking forward to your next post.
Regards, Phil
Reason: No reason
Re: Receive Antenna & Impedance "Matching"
Hi Phil, good for you you are in wintertime, here we are roasting at 38 °C, going to 40° and still some do not believe in climate change! Next week I am leaving for Sardinia: it is sailing time now and I will somehow be off the Forum for a while.
As for filters, I have managed to look at 3.8 MHz. Here powder iron toroids in N2 material easily provide measured Qu's in the order of 350 to 400 with inductance ranges from 0.5 to 45 uH, thus narrow band filters from 1.8 to 3.8 MHz are feasible with low insertion loss. Loss being percentage wise, at these frequencies significantly smaller bandwith is possible. Comparable Qu's are possible with air core coils, but diameters must grow, in order to keep coils "cubical" enough, so the filter dimensions become impractical.
Finally I have revived my old MAC PC & stored filter software from the nineties and now I can design crystal filters up to 4 elements. So if you have a crystal supplier there, I could provide the full filter data & specification.
Best regards,
Gianfranco
Please explain to me what you mean by the above, as I do not understand it.I really need to do that and assess the results with broader V narrower passbands (?)before deciding what to do next with the filter.
As for filters, I have managed to look at 3.8 MHz. Here powder iron toroids in N2 material easily provide measured Qu's in the order of 350 to 400 with inductance ranges from 0.5 to 45 uH, thus narrow band filters from 1.8 to 3.8 MHz are feasible with low insertion loss. Loss being percentage wise, at these frequencies significantly smaller bandwith is possible. Comparable Qu's are possible with air core coils, but diameters must grow, in order to keep coils "cubical" enough, so the filter dimensions become impractical.
Finally I have revived my old MAC PC & stored filter software from the nineties and now I can design crystal filters up to 4 elements. So if you have a crystal supplier there, I could provide the full filter data & specification.
Best regards,
Gianfranco
Reason: No reason
Re: Receive Antenna & Impedance "Matching"
Forgive me, badly expressed. I need to vary the inter-coil coupling using my first sliding coils and experiment with the trade off of insertion loss against selectivity.
eg. in the last hour I've received 21 stations on 30m WSPR with the SNRs as below:
-27, -26, -25, -25, -25, -24, -24, -24, -24, -21, -21, -20, -20, -19, -19, -18, -17, -15, -13, -12, -10
You can see how many I'd miss by losing a straight 6dB (my SNR floor is around -33dB) so the question that needs answering is would the RF gain and AGC improvements caused by removal of less or more band noise/strong stations make up for whatever the corresponding insertion loss was. Mmm, that's confusing too but you'll understand.
Understood with your temperatures, they are in the news here and yes, our 3 month outlook is forecast to be significantly drier and warmer. Our (small farm) dams are at record lows and have not recovered. And the bulk of our rain happens in the next three months...
OK on your 80m experimenting, that's a really good idea, not so low as to need specialist ferrite and large numbers of turns, plus its a popular band and one could do with a good narrow preselector to play with!
Enjoy Sardinia, our daughter in law was there three weeks ago but has returned to a Melbourne winter, and we were envious enough of her.
Great on the older Mac, I have several, though I've sold my genuine museum pieces to collectors, I wonder what yours is?
Phil
eg. in the last hour I've received 21 stations on 30m WSPR with the SNRs as below:
-27, -26, -25, -25, -25, -24, -24, -24, -24, -21, -21, -20, -20, -19, -19, -18, -17, -15, -13, -12, -10
You can see how many I'd miss by losing a straight 6dB (my SNR floor is around -33dB) so the question that needs answering is would the RF gain and AGC improvements caused by removal of less or more band noise/strong stations make up for whatever the corresponding insertion loss was. Mmm, that's confusing too but you'll understand.
Understood with your temperatures, they are in the news here and yes, our 3 month outlook is forecast to be significantly drier and warmer. Our (small farm) dams are at record lows and have not recovered. And the bulk of our rain happens in the next three months...
OK on your 80m experimenting, that's a really good idea, not so low as to need specialist ferrite and large numbers of turns, plus its a popular band and one could do with a good narrow preselector to play with!
Enjoy Sardinia, our daughter in law was there three weeks ago but has returned to a Melbourne winter, and we were envious enough of her.
Great on the older Mac, I have several, though I've sold my genuine museum pieces to collectors, I wonder what yours is?
Phil
Reason: No reason
Re: Receive Antenna & Impedance "Matching"
Phil, is this below the WSPR signal on 10,138.700 ? It has a kind of wobbly modulation.
The Mac in use is the SE30, but I have quite a collection!
You did not say if you wish me to do the xtal filter and where.
Cheers,
Gianfranco
Must switch off, or the laptop will blow up with heat!
The Mac in use is the SE30, but I have quite a collection!
You did not say if you wish me to do the xtal filter and where.
Cheers,
Gianfranco
Must switch off, or the laptop will blow up with heat!
- Attachments
-
- Looks too strong to me.....
- ScreenHunter 10.png (317.82 KiB) Viewed 49521 times
Reason: No reason
Re: Receive Antenna & Impedance "Matching"
Hi Phil,
The dilemma about selectivity Vs. insertion loss is not really a dilemma, if you carefully consider the data at your disposal.
1) The minimum detectable signal (MDS) of the RSP2 @ 10 MHz (given its rated Noise Figure) can be safely set at -120 dBm.
2) You now have to give a figure to your antenna noise level. Using my antenna data (yes, at at 14 MHz, but this is an example) shown below, the MINIMUM AVERAGE noise level you can hope for would be -112 dBm, if you used my antenna. But it is easy for you to measure yours.
3) You also have to consider a noise margin on S/N. Since you are receiving digital signals, I would set this @ -1 dB (-3 dB for SSB).
4) Thus your absolute sensitivity is -119 dBm and you can safely attenuate (signal & noise go down by the same amount) down to the level of - 112 dBm. Thus the allowable filter insertion loss is 119-112= 7 dB. E.g. with 7 dB insertion loss you do not lose any sensitivity, because you are just removing that much noise which would not allow to receive that signal anyway.
All the above holds with no strong adjacent interfering signals: but do you really have them so near? This is the entire point. If the strong overloading signals are far enough in frequency, then you can resort to a fixed frequency bandpass and forget about really high Qu components (always hammering on the same nail this Gianfranco!!)
Gianfranco
The dilemma about selectivity Vs. insertion loss is not really a dilemma, if you carefully consider the data at your disposal.
1) The minimum detectable signal (MDS) of the RSP2 @ 10 MHz (given its rated Noise Figure) can be safely set at -120 dBm.
2) You now have to give a figure to your antenna noise level. Using my antenna data (yes, at at 14 MHz, but this is an example) shown below, the MINIMUM AVERAGE noise level you can hope for would be -112 dBm, if you used my antenna. But it is easy for you to measure yours.
3) You also have to consider a noise margin on S/N. Since you are receiving digital signals, I would set this @ -1 dB (-3 dB for SSB).
4) Thus your absolute sensitivity is -119 dBm and you can safely attenuate (signal & noise go down by the same amount) down to the level of - 112 dBm. Thus the allowable filter insertion loss is 119-112= 7 dB. E.g. with 7 dB insertion loss you do not lose any sensitivity, because you are just removing that much noise which would not allow to receive that signal anyway.
All the above holds with no strong adjacent interfering signals: but do you really have them so near? This is the entire point. If the strong overloading signals are far enough in frequency, then you can resort to a fixed frequency bandpass and forget about really high Qu components (always hammering on the same nail this Gianfranco!!)
Gianfranco
- Attachments
-
- Data taken in june 2019 on my Beverage
- Noise floor compare 14.2 MHz.jpg (81.25 KiB) Viewed 49503 times
Reason: No reason
Re: Receive Antenna & Impedance "Matching"
Hi Phil,
We now have to consider how much strong is "strong" and " and how much "far" is far. If you look here:
https://www.sdrplay.com/community/viewt ... f=5&t=3926
you see that for the RSP Processor 2° order and 3° order effects start at a level of -35 /-40 dBm of the unwanted signal. These levels are only produced by broadcasters. Allowing a safety margin of 20 dB, we can be pretty sure that ham, commercial and broadcast signals reasonably far from our receive frequency will do no harm, if at levels below -60 / -65 dBm. We must now define what we mean for "reasonably far in frequency". Looking at the published block diagrams of the RSP's and for the 30 m band example, the protection is represented by a 2-12 Mhz filter (see below). This eliminates the detrimental effect of the MW broadcast band signals & below, but does nothing for the strong broadcast signals in the 5 MHz and 7 MHz segments. Let us leave this point for the moment.
We now have to look at unwanted AGC action. The best protection here is to use the MINIMUM possible AGC level in the EX control window. I have found that a level of 35 to 40 is optimal. If audio does not turn on at switch on, one has to "nudge" it up a little and then back off to the optimum setting. Using two RF generators it is easy to see that if the strong (-80 dBm) interfering signal is at least 200 KHz away (SSB LSB) it will not produce negative AGC action at all.
Now we can piece it all together. Carrying on the example at 30 m, two option (solving both the lower band overload danger and the spurious AGC action danger with the lowest possible Qu) are available:
1) A highpass filter with cutoff at 10,100 KHz, OR
2) A bandpass filter 10,100 to 10,200 KHz.
Best regards,
Gianfranco
We now have to consider how much strong is "strong" and " and how much "far" is far. If you look here:
https://www.sdrplay.com/community/viewt ... f=5&t=3926
you see that for the RSP Processor 2° order and 3° order effects start at a level of -35 /-40 dBm of the unwanted signal. These levels are only produced by broadcasters. Allowing a safety margin of 20 dB, we can be pretty sure that ham, commercial and broadcast signals reasonably far from our receive frequency will do no harm, if at levels below -60 / -65 dBm. We must now define what we mean for "reasonably far in frequency". Looking at the published block diagrams of the RSP's and for the 30 m band example, the protection is represented by a 2-12 Mhz filter (see below). This eliminates the detrimental effect of the MW broadcast band signals & below, but does nothing for the strong broadcast signals in the 5 MHz and 7 MHz segments. Let us leave this point for the moment.
We now have to look at unwanted AGC action. The best protection here is to use the MINIMUM possible AGC level in the EX control window. I have found that a level of 35 to 40 is optimal. If audio does not turn on at switch on, one has to "nudge" it up a little and then back off to the optimum setting. Using two RF generators it is easy to see that if the strong (-80 dBm) interfering signal is at least 200 KHz away (SSB LSB) it will not produce negative AGC action at all.
Now we can piece it all together. Carrying on the example at 30 m, two option (solving both the lower band overload danger and the spurious AGC action danger with the lowest possible Qu) are available:
1) A highpass filter with cutoff at 10,100 KHz, OR
2) A bandpass filter 10,100 to 10,200 KHz.
Best regards,
Gianfranco
- Attachments
-
- 50 Ohm in Blocks.png (51.65 KiB) Viewed 49474 times
Reason: No reason
Re: Receive Antenna & Impedance "Matching"
Sorry about failing to respond to your crystal offer and thanks for mentioning it. As per earlier comments I have tried using a single crystal. Given that the WSPR sub bands are 200Hz wide a single crystal has exactly the right bandwidth at around 500Hz according to my tests on 40m and 10m crystals I had lying round. I've not been able to find any affordable supplier and shelved the idea, moving on to the twin tank coils.
Fine on your calculations and thank you for the detailed example. Given my interest and large investment in WSPR and its ability to provide direct SNR readings at around 1dB comparative accuracy using real world / real time measurements from hundreds of transmitting stations all over the planet I've determined to go the empirical route.
So, today I turned off the RSP2 on 30m and set up an FT-857D and an FT-991, both on 30m fed from a single antenna via toroidal splitter to provide base line comparison data without the preselector.
The first dozen or so SNR repots on WSPR showed that they were within 1dB of each other but to my horror even though the FT-857D appeared to have a slight edge, it failed to decode some stations that were easily decoded by the other. That problem may well lie with the lack of processing power of the Raspberry Pi I mentioned earlier and I've temporarily run out of computers as I don't want to run multiple instances of the decoding software WSJT-X because of the nature of the audio feeds. I'll get there in the end though, currently awaiting the newest Pi4/4.
However, checking again a moment ago I find things have swung the other way, the two callsigns on WSPR are VK7JJ being the FT-991 with WSPR 2.0.1 on Linux Mint on a respectable laptop with 142 spots and VK7JJ2 being the FT-857D with the same WSPR 2.0.1 on the Pi2B with 159 spots.
A new 30m twin sliding coil preselector is all tuned up and ready to insert but I'll need to leave the baseline config running a while first.
--
Great on the SE30, a machine for the connoisseur of the time with its super extra speed. Pity it uses SCSI internally and externally making it harder to do a SSD transplant but the external connector would facilitate the use of an adapter if you wanted to really make it go! Running system 7? From memory that seemed the best on that machine. For a substantial period starting in 1986 I co-owned one of the original so called Apple Centres, with two of my best ham radio friends as the senior and next senior technician. We also did DOS/Windows/IBM etc etc on the side. We got up to all sorts of interesting stuff! Grand memories.
Fine on your calculations and thank you for the detailed example. Given my interest and large investment in WSPR and its ability to provide direct SNR readings at around 1dB comparative accuracy using real world / real time measurements from hundreds of transmitting stations all over the planet I've determined to go the empirical route.
So, today I turned off the RSP2 on 30m and set up an FT-857D and an FT-991, both on 30m fed from a single antenna via toroidal splitter to provide base line comparison data without the preselector.
The first dozen or so SNR repots on WSPR showed that they were within 1dB of each other but to my horror even though the FT-857D appeared to have a slight edge, it failed to decode some stations that were easily decoded by the other. That problem may well lie with the lack of processing power of the Raspberry Pi I mentioned earlier and I've temporarily run out of computers as I don't want to run multiple instances of the decoding software WSJT-X because of the nature of the audio feeds. I'll get there in the end though, currently awaiting the newest Pi4/4.
However, checking again a moment ago I find things have swung the other way, the two callsigns on WSPR are VK7JJ being the FT-991 with WSPR 2.0.1 on Linux Mint on a respectable laptop with 142 spots and VK7JJ2 being the FT-857D with the same WSPR 2.0.1 on the Pi2B with 159 spots.
A new 30m twin sliding coil preselector is all tuned up and ready to insert but I'll need to leave the baseline config running a while first.
--
Great on the SE30, a machine for the connoisseur of the time with its super extra speed. Pity it uses SCSI internally and externally making it harder to do a SSD transplant but the external connector would facilitate the use of an adapter if you wanted to really make it go! Running system 7? From memory that seemed the best on that machine. For a substantial period starting in 1986 I co-owned one of the original so called Apple Centres, with two of my best ham radio friends as the senior and next senior technician. We also did DOS/Windows/IBM etc etc on the side. We got up to all sorts of interesting stuff! Grand memories.
Reason: No reason
Re: Receive Antenna & Impedance "Matching"
Hi Gianfranco, as you might guess I just crossed posts with you, will catch up a bit later!
Regards, Phil
Regards, Phil
Reason: No reason
Re: Receive Antenna & Impedance "Matching"
Yes I still have the external SICSI drives and the Iomega storage chartridges with lots of data & design software, which can be useful from time to time. in the eighties and for all the nineties you had to use MAC for serious work. Windows software was still in toyland, with just about three crashes per hour! But only if you were lucky enough.
Phil, please help me: I do not wish to look conceited or spiteful at all, I just wish to understand. Nowadays you can buy a cartload of good used, but high quality, professional laptops for about $ 80 apiece. They run at 1.3 GHz and above, have on average 4 Gb RAMs and can work with WXP or with W7. They handle any version of SDRplay easily, are a sugar to operate, give good, reliable service ad let you devote your attention to the job at hand, e.g. RSP operation.
Having said all this, can you please explain in a few words why one should spend just about the same to get a Raspberry toy, with ridicolous performance, which needs hours of attention and setting up to get surely inferior results? Why one has to stop and wait for a cooling fan, a special housing box, etc.etc ... Is all this meant to take pleasure in difficulties? Something like hard mountain climbing or boat sailing in storms?
I hope to be forgiven for my sincerity, perhaps its the heat here that makes me difficult.
Gianfranco
Phil, please help me: I do not wish to look conceited or spiteful at all, I just wish to understand. Nowadays you can buy a cartload of good used, but high quality, professional laptops for about $ 80 apiece. They run at 1.3 GHz and above, have on average 4 Gb RAMs and can work with WXP or with W7. They handle any version of SDRplay easily, are a sugar to operate, give good, reliable service ad let you devote your attention to the job at hand, e.g. RSP operation.
Having said all this, can you please explain in a few words why one should spend just about the same to get a Raspberry toy, with ridicolous performance, which needs hours of attention and setting up to get surely inferior results? Why one has to stop and wait for a cooling fan, a special housing box, etc.etc ... Is all this meant to take pleasure in difficulties? Something like hard mountain climbing or boat sailing in storms?
I hope to be forgiven for my sincerity, perhaps its the heat here that makes me difficult.
Gianfranco
Reason: No reason