Results 1 to 19 of 19

Thread: Patch panel?

  1. #1

    Join Date
    May 2020
    Location
    Central Ohio
    Posts
    55

    Default Patch panel?

    I'm a new ham coming from an audio-production background, and one thing that audio pukes just loooove is patch bays.

    In case you're not familiar with the concept, it's a rack-mount device that has two horizontal rows of jacks on the front, say M jacks on top and N jacks on the bottom, where usually M = N, and another two horizontal rows of jacks on the back, again M jacks on top and N jacks on the bottom. Every front top jack is connected to the corresponding rear top jack, and the same for the bottom row.

    The idea is that you bring together a bunch of audio outputs of various kinds and plug them all into the back of the patch bay, into jacks in the top row; and you bring together a bunch of audio inputs and plug them all into rear jacks in the bottom row. Finally, you make or buy a bunch of short audio patch cables and, on the front panel, you patch outputs on the top row to inputs on the bottom row. This way, if you want to swap mixer channels for a couple of mics, you don't bother with repatching the mixer or rerouting cables on stage; all you do is go back to the patch bay and unplug one end each of two patch cables and swap them.

    (Some folks arrange outputs on the bottom and inputs on the top. They are heathens and blasphemers and should be shunned by all decent, right-thinking people everywhere.)

    So...with practically no experience in amateur radio, I'm beginning to see a fair amount of value in an RF patch bay, consisting of a couple of rows of dual-ended SO-239 all-threads poked through an aluminum or copper ground plane. Antenna feedlines and other RF sources would be connected to the back of the top row, and radios and other RF sinks would be connected to the back of the bottom row, and in front would be a series of one-foot PL-259-to-PL-259 patch cables connecting outputs to inputs.

    Then instead of having my 80-meter feedline getting in my way on the bench or falling down behind it while I'm using my 40/20/10/6-meter feedline, both feedlines and my radio would be plugged into the back of the patch bay, and switching from 40 meters to 80 meters would just be unscrewing a PL-259 from one SO-239 and screwing it into the one beside it.

    So...the question.

    This patch panel I've been describing: does it actually exist as an acknowledged product in the amateur-radio world? I've never seen one in real life or advertised. I'm planning to build one if A) I can't buy one and B) I can't be convinced that it's a stupid idea.

    Thanks,
    Dan

  2. #2

    Default

    Multi-coax switches are common. The patch bay would be great if you had a variety of radio gear as well as antennas, as most switches do not have enough isolation to keep RF from one transmitter out of the receiver of another radio. If you have multiple radios, go for it. If you have one radio and several antennas, just get a switch.

  3. #3

    Join Date
    May 2020
    Location
    Central Ohio
    Posts
    55

    Default

    Yes, I already have more than one radio, and I'll eventually have more than that. Interesting that the patch bay isn't already a thing.

    So thinking about building the patch bay already has me considering another project.

    None of my radios are fancy enough to have waterfall displays. My FT-891 tries to emulate one by sweeping its VFO really fast, but it's not the same, especially with only black and white. However, I do have a USB SDR dongle and gqrx, which isn't bad at waterfalling. So my idea is to hook both a transceiver and the SDR dongle to the same antenna, set gqrx to show me a whole ham band at once, and build some kind of device--there's probably already a name for it; it seems useful enough to be a thing--to detect transmit power on the feedline and instantly disconnect the SDR dongle to protect it until the transmit power goes away, then reconnect it. (I'm thinking about rectifying the signal from the antenna and feeding it to a transistor that opens a relay; but that'd probably end up distorting the signal to the radio somehow, and a mechanical relay might not open fast enough. Are there radio-frequency optoisolators?) I haven't looked at the gqrx code, but it might be worth working in some rig-control stuff so that I could control the transceiver's VFO from gqrx.

  4. #4

    Default

    Most transceivers have a diode in the receive chain that acts as a switch to shunt the RX path to ground during transmit (thankfully that DC decoupling cap also acts to limit the current). Relays and mechanical coax switches are bad at this because the contact leads inside the switching device couple enough power (even if shorted to ground a cm away) to give the receiver a big headache and possibly destroy crystals and other components. What I would do if I were attempting that is to study the schematic and locate this diode and bias it externally to prevent RF from entering the RX chain on the other radio's transmit. Long story short, I like your patch panel idea. Unless you intend to crack that radio open and do some modifications, I would resist the temptation to connect any additional receiver to a coax that will eventually carry power.
    Last edited by brandon lind; Fri 20th Nov 2020 at 06:51.

  5. #5

    Default

    Update.... Have you done an internet search for RF Patch Panel? I just did. Seems you are not the first one to have that thought. There is even one on ebay right now.

  6. #6

    Default

    I built a very simple patch panel to interconnect my two VHF/UHF radios to three different antennas, one a dual band 2M/70CM antenna, one a tri-band 6M/2M/70CM, and the third a wide-band discone. I was constructed out of a piece of white board faced masonite, six UHF barrel connectors and washers to support the connectors. The inputs from the antennas are on the top and the outputs to the radios (and a connection for my scanners) are on the bottom. It can be seen in this picture taken during a recent installation of a new desk in my shack.

    https://rightstuff.smugmug.com/Amate...el/i-P33FscL/A

  7. #7

  8. #8

    Join Date
    May 2020
    Location
    Central Ohio
    Posts
    55

    Default

    A diode. Hmm. A diode...really? How does that work? Does it somehow use the diode drop to not conduct tiny received signals, but then suddenly ground the snot out of the positive-going half of the transmitted signal? (Be gentle with me: I'm new to RF.)

    It's not what you think of when you think "radio." It's a USB dongle, about four times the volume of a standard flash drive, with a female SMA connector on the end opposite the USB, and no external controls or display. It came with a couple of telescoping dipoles, one about the right size for three meters and one about the right size for 80cm, but it works fine on the HF bands with my backyard antenna as well.

    Anyway, there's not much room inside the tiny case to fart around with adding circuitry, so anything I added would probably be external.

  9. #9

    Join Date
    May 2020
    Location
    Central Ohio
    Posts
    55

    Default

    Quote Originally Posted by brandon lind View Post
    Update.... Have you done an internet search for RF Patch Panel? I just did. Seems you are not the first one to have that thought. There is even one on ebay right now.
    No, but I did when I read your post. Most of the commercial ones seem to be 75-ohm, but yeah: like that.

  10. #10

    Join Date
    May 2020
    Location
    Central Ohio
    Posts
    55

    Default

    Quote Originally Posted by K6CPO View Post
    I built a very simple patch panel to interconnect my two VHF/UHF radios to three different antennas, one a dual band 2M/70CM antenna, one a tri-band 6M/2M/70CM, and the third a wide-band discone. I was constructed out of a piece of white board faced masonite, six UHF barrel connectors and washers to support the connectors. The inputs from the antennas are on the top and the outputs to the radios (and a connection for my scanners) are on the bottom. It can be seen in this picture taken during a recent installation of a new desk in my shack.

    https://rightstuff.smugmug.com/Amate...el/i-P33FscL/A
    That's kind of like what I had in mind, yes, although I was going to use an aluminum plate and bind it to my station ground, which would probably mean taking a length of copper strap, soldering one end with a blowtorch to a copper battery-cable eye terminal which I would bolt to the aluminum plate, and clamping the other end to my RF ground (two feet of copper pipe) with a hose clamp.

  11. #11

    Default

    Quote Originally Posted by dnwiebe View Post
    A diode. Hmm. A diode...really? How does that work? Does it somehow use the diode drop to not conduct tiny received signals, but then suddenly ground the snot out of the positive-going half of the transmitted signal? (Be gentle with me: I'm new to RF.)

    It's not what you think of when you think "radio." It's a USB dongle, about four times the volume of a standard flash drive, with a female SMA connector on the end opposite the USB, and no external controls or display. It came with a couple of telescoping dipoles, one about the right size for three meters and one about the right size for 80cm, but it works fine on the HF bands with my backyard antenna as well.

    Anyway, there's not much room inside the tiny case to fart around with adding circuitry, so anything I added would probably be external.
    I have an SRD, the SDRSmart something or another. The fact remains, it cannot handle high power levels on the antenna port.

    Attached is a drawing of how a diode can act as an RF switch for small signals. C1 and C2 are bypass capacitors that allow RF to pass but block DC. This prevents the DC from going beyond this circuit. RFC is a radio frequency choke, essentially an inductor with a high reactance at the frequency of use. This stops the RF from going into the power supply. The Diode only conducts if the voltage is over about 0.7v so if the RF is, for example, 0.5v peak to peak, the diode will never conduct. However, if we close the switch and add 2v bias, now the diode will conduct the entirety of the RF signal to ground along with the DC bias voltage. The resistors limit the bias current and keep the diode from burning up. Forgive the poor graph and artwork, i used excel and paint in a hurry.

    Diode as RF switch.jpg
    Attached Images Attached Images

  12. #12
    gnuuser's Avatar
    Join Date
    Apr 2020
    Location
    NW Pa,USA
    Posts
    101

    Default

    Patch panels are an interesting Idea If your equipment is buss grounded(as it should be for safety) your antenna switch might want to be able to isolate both conductors of the different transmission lines in order to prevent errant noise from induced signal voltages.
    anyone who has ever worked with early twin lead networking cables knows all about cross-talk from being too close to another active conductor.

    in the industrial world you definitely do not want to run a signal conductor in the same conduit as a power conductor ( Really messes up automation control systems)
    consider your power output also!
    qrp or low power radios do not produce a large enough induction current to interfere unless nearly touching another conductor, but they are extremely sensitive to it.
    and the more power you are putting out the greater the chance of spurious emissions unless your antenna is perfectly tuned and you have taken all steps to block common mode current.

    if you are going to use a patch panel, you may have to research the emissivity of your cables and make appropriate measures to block noise Either with spacing, shielding, or attenuating!
    all in all may turn out to be more trouble than its worth, but then again it may be a good setup
    too much blood in my caffeine system.

  13. #13

    Join Date
    May 2020
    Location
    Central Ohio
    Posts
    55

    Default

    Thanks for the drawing.

    Yeah, okay, that makes sense (I would not have thought of putting in the choke) as long as you have an extraneous signal from somewhere that says, "I'm about to transmit now."

    I was thinking of something that would key directly off the amount of power in the RF signal and act kind of like cornstarch paste. Treat it gently, and it lets you through; smack it hard, and it bounces you right off. That is, small signals pass through fine, but large signals find essentially infinite resistance.

    I have a vague picture in my head of a transistor or maybe a FET that's saturated on in receive mode so that it passes a biased version of the signal. It'd be held open by an envelope follower (I think you RF guys call them "detectors") that wasn't sensitive enough to detect a small signal; but when the enve--I mean the detector saw something big, it'd turn off the transistor (and maybe turn on another one to ground the SDR's input).

    But that's really just a semiconductor version of my relay that's held closed by a spring until the rectified energy from the transmitter pops it open. It might not be fast enough. I should play around with my oscilloscope and see if I can measure how fast the voltage on the feedline rises when I key the mic.

  14. #14

    Join Date
    May 2020
    Location
    Central Ohio
    Posts
    55

    Default

    I could look into it. I'm using mostly LMR-400 (or compatible) cables where I can, even though they're a little unwieldy at times.

    I bought 25' of a Flex version, expecting to get something nice and floppy to use for patch cables, but I really can't tell the difference in stiffness between the Flex stuff and the regular bury-me-in-the-ground stuff. Maybe I should have gotten the Super-Flex: I saw the manufacturer of a cutting-and-stripping tool complaining that its jacket was so rubbery that it was incompatible with the tool.

  15. #15

    Default

    Quote Originally Posted by dnwiebe View Post
    Thanks for the drawing.

    Yeah, okay, that makes sense (I would not have thought of putting in the choke) as long as you have an extraneous signal from somewhere that says, "I'm about to transmit now."

    I was thinking of something that would key directly off the amount of power in the RF signal and act kind of like cornstarch paste. Treat it gently, and it lets you through; smack it hard, and it bounces you right off. That is, small signals pass through fine, but large signals find essentially infinite resistance.

    I have a vague picture in my head of a transistor or maybe a FET that's saturated on in receive mode so that it passes a biased version of the signal. It'd be held open by an envelope follower (I think you RF guys call them "detectors") that wasn't sensitive enough to detect a small signal; but when the enve--I mean the detector saw something big, it'd turn off the transistor (and maybe turn on another one to ground the SDR's input).

    But that's really just a semiconductor version of my relay that's held closed by a spring until the rectified energy from the transmitter pops it open. It might not be fast enough. I should play around with my oscilloscope and see if I can measure how fast the voltage on the feedline rises when I key the mic.
    How fast is that oscilloscope? What is it's bandwidth? You would need to have a dual channel scope, one channel on PTT and another on the RF output to see the actual delay between the two. I'm guessing it would take a better scope than my 200MHz one to properly measure that delay. And I am certain no relay within a meter or ten could keep up without delay circuitry elsewhere in the radio. Electromagnetic relays are SLOW compared to even a tenth of the speed of light.

    Another issue you have is impedance. You are talking about running two 50ohm receivers on one antenna (presumably also 50 ohm to match the transmitter) and intermittently disconnecting one of them (thus changing the circuit impedance). When the two are in parallel during RX, you have 25ohms on receive, not 50ohm. 25ohms is a 2:1 SWR right from the start. Now you are switching matching networks too, not just antennas. The idea of running two simultaneous receivers on the same band with a common antenna and having one of them go into TX is a good way to make some expensive geranium-smelling smoke. Consider a repeater. It is a transmitter and receiver simultaneously using the same antenna with a small (but necessary) offset. The duplexer is a tuned filter that keeps the RF from the transmitter out of the receiver and the received signal out of the transmitter. The standard offset is 600kHz for VHF and most 4 to 6 can cavity duplexers struggle to obtain 90dB to 120dB of isolation at that close of frequency. You want to TX on the SAME frequency the secondary receiver is listening on and that is IMPOSSIBLE without serious attenuation, something you will not get with a cheap relay ~ never ever ever. That relay would need to be followed by a semiconductor to handle the RF the relay contact "arms" pick up, even if shorted to ground. Lets consider the math.... your receiver has a sensitivity of -120dBm, thats 120dB below a milliwatt to open the squelch. You key a 100w transmitter, thats 100,000mW or 50dBm for a difference of 170dB. Your switching device needs to have a minimum of 170dB of isolation (best of luck) just to keep the RX from lighting up. You won't do that with a relay any cheaper than you could buy a radio with a better waterfall display.

    Stick to one transceiver per antenna and hope like heck the other nearby antennas don't pick up enough RF from the transmitting one to cause problems in the other receivers. Look into contesting stations and how they operate. They never have a receiver operating on the same band (band, not just frequency) as the radio next to them. because it will get overloaded. Even when they are on different bands, they often use coax stubs to filter out the band used next to them so when the guy next to them is operating on one band, their receiver on another is not desensitized.

    I will leave it at that, if anyone else can tell me I am wrong, please correct me!

  16. #16

    Join Date
    May 2020
    Location
    Central Ohio
    Posts
    55

    Default

    Quote Originally Posted by brandon lind View Post
    How fast is that oscilloscope? What is it's bandwidth? You would need to have a dual channel scope, one channel on PTT and another on the RF output to see the actual delay between the two. I'm guessing it would take a better scope than my 200MHz one to properly measure that delay. And I am certain no relay within a meter or ten could keep up without delay circuitry elsewhere in the radio. Electromagnetic relays are SLOW compared to even a tenth of the speed of light.
    I think there are two misunderstandings here.

    First, I don't care what the delay is between PTT and RF output, because I never anticipated using any PTT signal. What I'm wondering about is, when I key the transmitter, how quickly does the voltage slew from zero to maximum? Is it practically instantaneous, or is there time in there for a circuit to detect it and shut off a transistor before enough current surges through the transistor to blow it and/or the equipment it's protecting?

    Which leads into the second: I understand that an electromechanical relay is not fast enough: that makes perfect sense. Is a detector followed by a BJT or FET fast enough? Dunno.

    Another issue you have is impedance. You are talking about running two 50ohm receivers on one antenna (presumably also 50 ohm to match the transmitter) and intermittently disconnecting one of them (thus changing the circuit impedance). When the two are in parallel during RX, you have 25ohms on receive, not 50ohm. 25ohms is a 2:1 SWR right from the start.
    Impedance is a question I hadn't even considered. Great point. It'd annoy both receivers. There might be a way to use active circuitry to raise the impedance the transceiver sees from the SDR, maybe to a few thousand ohms, but then convert it back to around 50 ohms for the SDR.

    Is SWR strictly applicable, though? The idea is to switch the SDR completely out of the circuit whenever the transceiver transmits; so the transmitter should never see it, except possibly for the barest flash of petticoats disappearing around a corner.

    The idea of running two simultaneous receivers on the same band with a common antenna and having one of them go into TX is a good way to make some expensive geranium-smelling smoke.
    Aha, I've caught you! It takes more than an idea to make expensive smoke, which is why I come here and talk to people like you. Even when the idea is stupid, I still learn from finding out the various ways in which it's stupid.

    Consider a repeater. It is a transmitter and receiver simultaneously using the same antenna with a small (but necessary) offset. The duplexer is a tuned filter that keeps the RF from the transmitter out of the receiver and the received signal out of the transmitter. The standard offset is 600kHz for VHF and most 4 to 6 can cavity duplexers struggle to obtain 90dB to 120dB of isolation at that close of frequency. You want to TX on the SAME frequency the secondary receiver is listening on and that is IMPOSSIBLE without serious attenuation, something you will not get with a cheap relay ~ never ever ever.
    Good, good...except that I don't want to transmit and receive simultaneously the way a repeater does. I hadn't even considered trying to leave the SDR listening to the whole band and just filter out the few kilohertz carrying the transmit power. I know you RF types work mostly with sine waves or near-sine waves, so you don't really care that much about the distortion introduced by steep filter slopes like audio guys have to; but the kind of filter you'd need for something like that would be insane anyway.

    That relay would need to be followed by a semiconductor to handle the RF the relay contact "arms" pick up, even if shorted to ground.
    No relay! No relay! You convinced me! But I have to say that I never would have considered RF leakage between the contact arms. RF is a whole new world.

    It surprises me, though. The inductance of a contact arm, or the capacitance of a pair of them, has to be pretty small. Stipulating in advance that I'm seriously no longer considering using a relay, I can see how contact arms might be a factor at SHF or above, but do they really cause problems at HF?

    Lets consider the math.... your receiver has a sensitivity of -120dBm, thats 120dB below a milliwatt to open the squelch. You key a 100w transmitter, thats 100,000mW or 50dBm for a difference of 170dB. Your switching device needs to have a minimum of 170dB of isolation (best of luck) just to keep the RX from lighting up. You won't do that with a relay any cheaper than you could buy a radio with a better waterfall display.
    170dB is indeed a daunting number. But say we had a switching transistor with an "off" resistance of 100 megohms. (For AF transistors, that's no big deal; are RF transistors different?) A 100W transmitter driving a 50-ohm system ought to be peaking at about 71V, right? --crap. Reverse voltage. That's the booger, isn't it? Argh. Biasing out a few millivolts is one thing; biasing out 35-40 volts is something else. At the very least, we're no longer talking about a little chip of perfboard with a few tiny components haphazardly soldered to it.

    Stick to one transceiver per antenna and hope like heck the other nearby antennas don't pick up enough RF from the transmitting one to cause problems in the other receivers. Look into contesting stations and how they operate.
    Don't know much about contesting stations. The pursuit hasn't appealed to me yet, so I haven't poked around in it. Do contesters usually gather in groups like that? Seems strange to me. If I wanted to brag that my station was better than your station, I'd want to make sure my station was operating in the best possible environment for it so that it turned in its best performance, rather than crowding it in with a bunch of other transmitters.

    But hey, there are people who run with the bulls in Spain, right? There are even people--so I hear, anyway--who run for political office. People do strange things.

    I will leave it at that, if anyone else can tell me I am wrong, please correct me!
    Well, thanks for putting up with me this long. I enjoy learning, and I certainly have a lot to learn about RF electronics--and about RF in general, and about electronics in general.

    Shalom,
    Dan

  17. #17

    Default

    Quote Originally Posted by dnwiebe View Post
    I think there are two misunderstandings here.

    First, I don't care what the delay is between PTT and RF output, because I never anticipated using any PTT signal. What I'm wondering about is, when I key the transmitter, how quickly does the voltage slew from zero to maximum? Is it practically instantaneous, or is there time in there for a circuit to detect it and shut off a transistor before enough current surges through the transistor to blow it and/or the equipment it's protecting?

    Which leads into the second: I understand that an electromechanical relay is not fast enough: that makes perfect sense. Is a detector followed by a BJT or FET fast enough? Dunno.

    Impedance is a question I hadn't even considered. Great point. It'd annoy both receivers. There might be a way to use active circuitry to raise the impedance the transceiver sees from the SDR, maybe to a few thousand ohms, but then convert it back to around 50 ohms for the SDR.

    Is SWR strictly applicable, though? The idea is to switch the SDR completely out of the circuit whenever the transceiver transmits; so the transmitter should never see it, except possibly for the barest flash of petticoats disappearing around a corner.

    Aha, I've caught you! It takes more than an idea to make expensive smoke, which is why I come here and talk to people like you. Even when the idea is stupid, I still learn from finding out the various ways in which it's stupid.

    Good, good...except that I don't want to transmit and receive simultaneously the way a repeater does. I hadn't even considered trying to leave the SDR listening to the whole band and just filter out the few kilohertz carrying the transmit power. I know you RF types work mostly with sine waves or near-sine waves, so you don't really care that much about the distortion introduced by steep filter slopes like audio guys have to; but the kind of filter you'd need for something like that would be insane anyway.

    No relay! No relay! You convinced me! But I have to say that I never would have considered RF leakage between the contact arms. RF is a whole new world.

    It surprises me, though. The inductance of a contact arm, or the capacitance of a pair of them, has to be pretty small. Stipulating in advance that I'm seriously no longer considering using a relay, I can see how contact arms might be a factor at SHF or above, but do they really cause problems at HF?

    170dB is indeed a daunting number. But say we had a switching transistor with an "off" resistance of 100 megohms. (For AF transistors, that's no big deal; are RF transistors different?) A 100W transmitter driving a 50-ohm system ought to be peaking at about 71V, right? --crap. Reverse voltage. That's the booger, isn't it? Argh. Biasing out a few millivolts is one thing; biasing out 35-40 volts is something else. At the very least, we're no longer talking about a little chip of perfboard with a few tiny components haphazardly soldered to it.

    Don't know much about contesting stations. The pursuit hasn't appealed to me yet, so I haven't poked around in it. Do contesters usually gather in groups like that? Seems strange to me. If I wanted to brag that my station was better than your station, I'd want to make sure my station was operating in the best possible environment for it so that it turned in its best performance, rather than crowding it in with a bunch of other transmitters.

    But hey, there are people who run with the bulls in Spain, right? There are even people--so I hear, anyway--who run for political office. People do strange things.

    Well, thanks for putting up with me this long. I enjoy learning, and I certainly have a lot to learn about RF electronics--and about RF in general, and about electronics in general.

    Shalom,
    Dan
    Got it, no relays. But you could with enough delay and a few other things.

    The rise time of the voltage is a function of frequency being what comes out is a sine wave. If you have 20MHz, the signal goes positive then negative and back to zero 20 million times in one second. This means the amount of time it takes for the voltage to go from zero to maximum is the reciprocal of 20 million divided by 4 (because we are only concerned with the first 90 of signal), or 12.5ns. You would need to detect the very beginning of that rise (way less than 12ns) and activate a semiconductor to disconnect the SDR antenna port. The circuitry also needs to hold that state until transmitting had finished. To do that, you would need a diode to rectify the signal and charge a capacitor to hold the high level state the switching device needs to stay active. That would likely require a high ohm resistor to limit the current through the diode detector and the diode will need to be a very fast one. Also, the capacitor will need to be small enough to not slow down the rising edge much. This means the switching device must have an extremely high input impedance to not drain the capacitor in between rectified voltage peaks. Turning it all off is the next issue. You could get really specific with your measurements such that the capacitor holding the gate of the switching device high has some high value resistance to ground such that the capacitor will lose voltage and shut off the switch after transmitting has stopped, but not a resistance so low that the capacitor runs out of juice between rectified peaks where the RF voltage dips to zero. I am sure there are other ways to sample the RF for controlling the switch, but that seems like the most obvious and easy to me. And just to be on the safe side, I would find a way to also switch the SDR port to ground at the same time the series switch disconnects it from RF just for good measure (like that diode switch I mentioned).

    The impedance matching thing will likely require a wide band toroid transformer to convert the 50 ohm to 25 ohm (which is what two 50ohm receivers look like paralleled). Being the two receivers are only paralleled during RX, now we need a way to switch in the matching network really fast too (unless a mismatch at your receivers and subsequent signal loss is acceptable to you). This is why I brought the PTT switch into the conversation. Sampling the fact the radio is about to transmit at the PTT switch as opposed to finding out at the antenna port give you a bit more time to pull all this off and greatly simplifies the circuitry. Now consider this... Lets say you add a few millisecond delay between your PTT button and the radio's PTT circuitry. Now you have bought yourself all the RF time in the world and your human senses will never even notice the delay in keying up. Using the PTT signal is greatly advantageous to your goal.

    As for what device to choose as a switch, mosfets are probably the way to go due to their extremely high input resistance the detector cap needs to not get pulled down to nothing between peaks. If, however, you bought yourself more time as I suggested, all of this would be greatly simplified.

    Brandon KE0KOY


    EDIT: If I drop off the map and stop replying, it is because of computer trouble. I do have a laptop if it totally craps the bed. There is a thermal stress fracture somewhere on my motherboard that has required me to preheat half of the motherboard with my solder station heat gun just to get the bugger to post and boot up. I used to be able to just let it sit there and think for a period of time (warming up) and a restart would do it, but the past couple days, it has required external assistance Not sure how long this bugger is going to hang in there. Thought it was my CPU at first, but the more it happens, the more I narrow it down with the heat gun. And when I do, its re-flow time. And by "ill leave it at that", I was referring to my current post, not this thread. I didn't abandon the conversation
    Last edited by brandon lind; Sat 21st Nov 2020 at 20:24.

  18. #18

    Join Date
    May 2020
    Location
    Central Ohio
    Posts
    55

    Default

    Quote Originally Posted by brandon lind View Post
    The rise time of the voltage is a function of frequency being what comes out is a sine wave. If you have 20MHz, the signal goes positive then negative and back to zero 20 million times in one second. This means the amount of time it takes for the voltage to go from zero to maximum is the reciprocal of 20 million divided by 4 (because we are only concerned with the first 90 of signal), or 12.5ns. You would need to detect the very beginning of that rise (way less than 12ns) and activate a semiconductor to disconnect the SDR antenna port.
    That's not exactly what I mean.

    First, you're assuming that the transmitter gets switched on precisely at the moment the carrier's instantaneous voltage makes a positive-going zero crossing, which might not be the case at all. If the transmitter is switched on when the carrier is at its negative maximum, you'll get a faster negative slew.

    Second, and this is the more important one...well, let me talk briefly about audio power amps. If you have an audio power amp running but set to standby (that is, ignoring the input signal, with the power transistors nice and relaxed), and you start an input--say pluck a guitar or bass string and let it sustain, or hold down a synth key--and then flip the switch from standby to on, you don't get a sharp attack on the output. The amp takes a bit of time--not long, maybe a few dozen milliseconds--to go from zero output to full power. I don't know why--probably capacitors capaciting somewhere--but that's been the way of every audio power amp I've ever encountered.

    I assumed--perhaps incorrectly--that the same was true of RF power amps, and if it was, maybe a semiconductor could be fast enough to sense the rising power level and holler, "High water's a-comin', Mabel! Git!" before it had to deal with the full 100W. That was the transition I was talking about capturing on my scope. Hook up a dummy load, put a tone on the mic, key the transmitter, and record a couple milliseconds before and a couple milliseconds after. Do we get a clean, sharp nanosecond-scale start to a full-power modulated sine wave, or do we get a quick ramp envelope? I haven't gotten around to it yet, because it'll require a little doing to get access to the feedline center conductor and attenuate it so that it doesn't blow up my scope or cause SWR problems.

    The circuitry also needs to hold that state until transmitting had finished. To do that, you would need a diode to rectify the signal and charge a capacitor to hold the high level state the switching device needs to stay active.
    Is that different in principle from the standard rectifier/filter capacitor arrangement you find in vanilla linear power supplies running at 50/60Hz?

    That would likely require a high ohm resistor to limit the current through the diode detector and the diode will need to be a very fast one. Also, the capacitor will need to be small enough to not slow down the rising edge much. This means the switching device must have an extremely high input impedance to not drain the capacitor in between rectified voltage peaks.
    Sounds like a MOSFET or a JFET. I don't have much experience with JFETs, but I saw a YouTube video where a guy set up a MOSFET to turn a motor, and then charged a 1uF capacitor and connected it across the gate, then went to bed. In the morning, the motor was still turning. He pulled the capacitor and the motor stopped. I don't have a feeling for whether MOSFETs are fast enough for RF.

    Turning it all off is the next issue. You could get really specific with your measurements such that the capacitor holding the gate of the switching device high has some high value resistance to ground such that the capacitor will lose voltage and shut off the switch after transmitting has stopped, but not a resistance so low that the capacitor runs out of juice between rectified peaks where the RF voltage dips to zero. I am sure there are other ways to sample the RF for controlling the switch, but that seems like the most obvious and easy to me. And just to be on the safe side, I would find a way to also switch the SDR port to ground at the same time the series switch disconnects it from RF just for good measure (like that diode switch I mentioned).
    Sure. I wouldn't care if the SDR didn't start waterfalling again until a quarter or half second after the transmitter shut up.

    The impedance matching thing will likely require a wide band toroid transformer to convert the 50 ohm to 25 ohm (which is what two 50ohm receivers look like paralleled). Being the two receivers are only paralleled during RX, now we need a way to switch in the matching network really fast too (unless a mismatch at your receivers and subsequent signal loss is acceptable to you).
    A wide-band transformer. Hmm. How do you make a wide-band transformer? I'm guessing not with a large inductance, because that'd produce problems at high frequencies. With a particularly low inductance, then?

    This is why I brought the PTT switch into the conversation. Sampling the fact the radio is about to transmit at the PTT switch as opposed to finding out at the antenna port give you a bit more time to pull all this off and greatly simplifies the circuitry. Now consider this... Lets say you add a few millisecond delay between your PTT button and the radio's PTT circuitry. Now you have bought yourself all the RF time in the world and your human senses will never even notice the delay in keying up. Using the PTT signal is greatly advantageous to your goal.
    Yep, makes sense. Feels a little hacky, but I should look inside my radio and see how hard it looks to fool with the PTT circuit.

    EDIT: If I drop off the map and stop replying, it is because of computer trouble. I do have a laptop if it totally craps the bed. There is a thermal stress fracture somewhere on my motherboard that has required me to preheat half of the motherboard with my solder station heat gun just to get the bugger to post and boot up. I used to be able to just let it sit there and think for a period of time (warming up) and a restart would do it, but the past couple days, it has required external assistance Not sure how long this bugger is going to hang in there. Thought it was my CPU at first, but the more it happens, the more I narrow it down with the heat gun. And when I do, its re-flow time.
    That is amazing. I had a motherboard go like that a few years back, and I just used it as an excuse to replace an aging machine. I never would have thought to use a heat gun on it.

  19. #19

    Default

    To the first quoted statement (forgot how to separate quotes on here), you are correct. My assumption that it will start with a positive-going signal is completely wrong. Good point.

    Apart from the uncertainty of where the carrier is when the signal starts, I have no idea how long it takes for the signal to go full strength. I assume it varies between radio designs but assume it is very fast. You may very well be correct in a small delay, no idea. Being you work with audio amplifiers, I assume you know that a 12vDC amplifier can make hundreds of volts AC on the output end. Be sure to do the math on the voltage based on the load you are hooking your scope to so that you don't cook the scope input. Other than that, I am not experienced enough to give you more information on that particular question.

    Different than a power supply filter, actually identical. But the component values will vary considerably from those used at 60Hz.

    The youtube video you watched where the capacitor holds the gate open all night is accurate. The MO in MOSFET stands for metal oxide, the insulative layer between the gate and the source/drain channel. It is a "field effect" device which means the charge on the gate influences the conductivity of the channel without directly making contact with it. Because of this oxide insulative layer, there is little to nowhere for the electrons to go once they are on the gate material (like those on the plate of a capacitor connected to nothing).

    Wide band RF transformers would be more effectively explained by youtube. There are different mixes with different permeabilities and power handling capabilities each suited for different things. Each manufacturer makes their own mixes too. You would need to take a crash course in that and buy ferrites specifically for what you want based on their website. That said, type 31 is great for HF.

    For the PTT thing, you dont need to open the radio. All you gotta do is tap into the mic cable with another set of connectors. I do not know what radio you wish to use and cannot give you specifics on how to do this, nor would I want to be responsible for the mishaps. Someone else may be more suited to walk you through specific steps if that is what you are looking for.

    As for the motherboard, this morning I have narrowed the problem down to about 4 square inches of board. I have just about figured our exactly where heating helps and I think one or two more cold mornings should pinpoint the issue. This is a common problem for graphics cards, especially for bitcoin miners or serious gamers. The contacts between the GPU and the board are usually responsible. Sometimes, reflowing works, but other times, removal of the GPU and reballing it is necessary. Because many people try to cheat and toss their graphics card in an oven (damaging electrolytic capacitors and not using flux) usually results in a temporary pressure contact that only lasts a few months. Some people make money buying damaged graphics cards online for mere dollars and resell the ones that can be fixed ~ although its a gamble with the abuse bitcoin mining comes with.
    Last edited by brandon lind; Mon 23rd Nov 2020 at 04:16.

Tags for this Thread

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •