Podcasting since 2005! Listen to Latest SolderSmoke

Wednesday, December 11, 2024

An Evening Bandscan on 40 Meters using the High-School Direct-Conversion Receiver


This video shows how useful this receiver really is.  Build one of these!
at 8:38 PM
Labels: , ,

9 comments:

  1. WN2ADecember 12, 2024 at 12:41 AM

    Hi Bill,
    I watched the full video, and it sounds spectacular, with great DCR presence! It also proves that in the busy evening 40 meter RF environment, the lack of single-signal capability didn't impair usefulness one iota.
    Two questions:
    1) Was the audio output to the HB speaker directly from the output transformer of the Third Audio Stage? Sounds like ample volume.
    2) Could you make a measurement of the LO leakage for me? Set the RF attenuation for minimum (max RF gain), tune the double-tuned filter in band center (no stagger tune) if you can.
    Take spectrum analyzer data across the band. That will give us an idea what LO rejection you are getting with no problems.
    Mucho TNX in Advance! 73de WN2A

    ReplyDelete
    Replies
    1. Bill MearaDecember 12, 2024 at 1:24 PM

      Mike: That is what you hear from an unamplified speaker. There is no additional amplification between the transformer and the speaker. I will try to do the test you mention in the next day or so. 73 Bill

      Delete
  2. AnonymousDecember 12, 2024 at 2:45 AM

    Hi Bill

    Congrats on your basic DC receiver operating from a 9v supply.

    RE: your post mixer network comment 1 video back:

    The shunt cap [ usually 100 nF in series with a 51 resistor ] provides a 50 Ω termination for a somewhat narrow band of frequencies from ~~ LF into VHF . It depends on the capacitor value + the value of the series inductor [ usually 10 or more turns on a FT37-43 ferrite toroid ] since the shunt R + cap --- and the series L form a network.

    The main function of the inductor is to block RF from getting into the audio amplifier input — RFC.
    Thus, this simple post mixer network does not offers a 50 Ω termination for audio frequencies at the product detector IF port.

    We normally must AC couple the audio amp input to the post mixer network. Typically this is a via a capacitor in the order of 6.8 - 220µF. This cap’s reactance means that only audio frequencies will get terminated in the input Z of the post mixer preamp.

    The audio amplifier itself provides some 50 Ω termination if it has a wide band 50 Ω input impedance. If the 50 Ω input Z audio preamp offers a strong input return loss -- say >=20 dB from ~ 60 Hz to up to maybe 10 KHz, you may get a reasonable 50 Ω return loss at AF at the product detector “IF” port.

    The lead to the popularity of the common base as a post product detector amplifier. Certainly, you might design a 50 Ω input common emitter type as well. I made a 2-part video about this particular amp only using discrete components.

    I did some post network experiments here:

    https://qrp-popcorn.blogspot.com/2024/01/audio-frequency-return-loss-bridge-50.html

    I added another shunt capacitor after the RFC to better filter RF + extend the network bandwidth. I chose a 4.7 nF instead of a 100 nF cap for 50 Ω RF termination. This gave a peak IF port return loss at 10 MHz, and a decent return loss up to almost 100 MHz , but, sadly only down to 5 MHz . You can only do so much with such a bog-simple network extending over a massive bandwidth that includes AF.

    Best to you!
    Todd--VE7BPO--

    ReplyDelete
    Replies
    1. Bill MearaDecember 12, 2024 at 1:29 PM

      Thanks Todd. Good to hear from you. We just lifted this network from Roy Lewellan's Optimized QRP transceiver. It seemed to help knock down the AM broadcast RFI from Radio Marti. So we kept it. Nick M0NTV also had good success with this. I think we do have an electolytic at the input to the AF amp. Your analysis and your filter are clearly better than ours, but we were just trying to keep it simple, and, frankly, don't have the tech chops to do what you did. Thanks Todd. 73 Bill

      Delete
    2. AnonymousDecember 12, 2024 at 4:32 PM

      Hi Bill. Thanks for your comment. Sorry for writing so unclearly.

      The W7EL network is fine – hundreds have used it over the decades – mine is not better.
      My main experimental conclusion was that the input impedance of the audio amplifier dominates what impedance is seen by the product detector IF port at audio frequencies with this, or a related, simple network.

      You mentioned that Q2 your post- detector audio amplifier does not have a 50 Ω input impedance and you hoped that this network would help that. Sadly, it does not. Your Q2 input Z would depend on factors such as the 2N4904 Beta, the actual DC voltage from the battery and other factors -- Q2’s input impedance may lie between 675 and 925 Ω doing guesstimate in-brain calculations. I’m not suggesting to re-build it, or that it is bad in any way --- just that the W7EL post mixer network mostly affects RF and not the AF input impedance seen by the IF port.

      Big thanks and Best! Todd

      Delete
    3. Bill MearaDecember 12, 2024 at 6:04 PM

      Thanks Todd. Helpful as always. 73 Bill

      Delete
    4. WN2ADecember 12, 2024 at 6:27 PM

      Todd is correct about the amplifier input Z dominating.. and the factors that make that calculation highly variable. I did some QS simulations of the first stage(only). Referring to the Hackaday schematic and @700Hz (unless otherwise noted), we get:

      1)Into "Audio" node Zin=1320 ohms, Gv=20.8dB . Gv=21.3dB @ 5KHz
      (a pretty flat audio response).

      2) At "AFOut"node into diplexer Zin=361 ohms Gv=20.8dB. Gv=21.7dB
      (....interesting Z did drop with the diplexer some, but not a match yet. Still flat audio response)

      3)So let's do something like W7ZOI plus W7EL...insert at C16/C17, a 88mH (series) and another 0.47uF (to ground), making a 'pi' network. We get
      Zin=139 ohms, Gv=29.6dB. Gv=-10dB @ 5KHz.
      So, if we find a "loading coil" and another cap, we get a better still match, and about 9dB more "gain" at 700. That's nice, but not the real payoff-- Now the >5KHz interference is down 30dB w.r.t 700Hz before this gets into the audio chain. That's the payoff!
      Suitable ~88mH coils might be a tough find for some.

      The QS 4.3.1 simulations are available, if there is interest.
      73! WN2A

      Delete
  3. AnonymousDecember 12, 2024 at 11:18 PM

    Thanks for your comments WN2A -- and to Bill for putting up with men rambling about whatnot and so forth.

    Yikes --- 88 mH inductors are straight out of the 60-70’s Handbooks. 88 mH telephone line loading coils were a surplus Hamfest staple back in the day. Fabulous coils – the 1 that I measured back in the early 90s had a DC resistance of only 4.5 ohms.

    Even the inductors for the 'proper' diplexer in Rick, KK7B’s fabulous R2 receiver are no longer in Mouser-Key catalogs. You probably can get them, but they cost dearly. I built 4 R2 variants back in the day with all the proper components-- and even then ,they seemed quite expensive to purchase + ship to Canada where our dollar is nearly worthless.

    It very difficult to model coils wound on the FT37-43. Back in 1996, I tried to measure the unloaded Q of a 10 turn inductor. I could not do it. So I asked Wes – he could not do it. All I could conjecture was that the Qul of this FT-37-43 coil was >= 1 and less than 9. ??

    I found that it behaved more like a 10 ohm resistor than anything else. So for Z transformation with the '43 at RF -- it’s turns ratio or ‘tapping a link’ since pi style network transformation is decidedly lossy and unpredictable.

    Back in the tube heydays, GW Miller offered 455 Khz slug tuned coils that gave ~~~100 to 280 uH and offered some Q. I had 1 and found the ferrite mix was somewhat similar to # 61 mix. So perhaps a Qul of 30 to 60 ??? These would be fabulous to use in a post mixer network.
    Cheers! T!!

    ReplyDelete
  4. WN2ADecember 13, 2024 at 8:56 AM

    I was impressed on how superb the audio sounded with just a Class A Stage. Ample volume with no crossover distortion. Q5 draws only about 15-16mA. Many Class AB's draw more current! Sometimes, others's projects make you re-think some of your own.
    How about a nice freq readout, Bill? Mechanical 10-turn counting knob, or a HB Freq Counter? See
    https://www.qsl.net/wn2a/FreqCount.html and do some "Muntzing" to only what you need (much,much less!)
    Possibly shield the counter for digital/display noise.
    The PIC code is GPL3 Open-Source; so make the "splash-screen" display your callsign, not mine!
    That retains the HB flavor of the project compared to just buying a module off Amazon.

    .....Yeah, Class A output stage! Dead-Simple!!!

    ReplyDelete

Subscribe to: Post Comments (Atom)
Designer: Douglas Bowman | Dimodifikasi oleh Abdul Munir Original Posting Rounders 3 Column