Wednesday, November 11, 2015

Simpleceiver Part 8

The Simpleceiver Direct Conversion Build is Working ~ Houston We Have Ignition!

 
We now have the J310 "Dual Gate MOSFET" Product Detector working as a 40M Direct Conversion Receiver! All of the upfront analysis we prepared  has paid off in big dividends. The test configuration consists on the J310's as the Product Detector, an Arduino Pro-Mini driving an AD9850 in the range 7 to 7.3 MHz supplying the LO signal and the audio amplifier is the NE5534 driving an LM-380 . There is no additional amplification nor Band Pass Filter in line at this time. There is a short video to show "proof of life".
 
Our basic J310 "Dual Gate MOSFET" schematic was used as shown below. There are but two additions: 1) A 3 turn primary to 20 turn secondary wound on a FT-37-43 core (#26 wire) is used to match the 50 Ohms on the antenna connected to Gate 1 (J2), where the match is forced to 2.2K. C3 is not used with the matching transformer. This is a 44:1 match and 2) A 10 NF is connected to Gate 2 (J1) and the other end is connected to the AD9850 DDS.
 
 
This circuit forms the basic Dual Gate MOSFET template that will be used throughout the total Transreceiver project. With some slight changes this circuit will morph into a RF amplifier that can be used on either the receive or transmit side. More on this in a later post. By the Way in the above configuration this Product Detector is good for 18 dB gain. Run the LT Spice if you are not convinced or you could just look at Part 7.
 
Below is the schematic of the Audio Amplifier stage consisting of the NE5534 and the LM-380 --good for two watts. The LM-380 is used in many commercial transceivers and should allay the fears of those who pan the LM386. [Yes to the question-- this is a GIF --still don't understand why jpgs are no good.]
 

This is the test set up with Pro-Mini/AD9850, the Product Detector and the NE5534/LM380

 
 
The Arduino software I am using is based on the code from AD7C. The Microcontroller is the Arduino Pro-Min and for driving the AD9850 I am using pins 4,5,6 and 7. You must use Pins 2 & 3 for the encoder interrupt, Pin A3 is used for the step increment. The LCD is a 16X2 with a blue background.
 
Stay Tuned --let the fun begin.
 
73's
Pete N6QW
 
 


11 comments:

  1. As always, great information, Pete. I'm going to download LT Spice and give it a try. I can definitely see how it can be a good tool to design circuits.

    ReplyDelete
  2. Pete, loving this series of yours. It's going to be a great one to try as an "upgrade" to my Neophyte DC RX. Question, though: Are the values of C1 and C7 correct on the LTSpice schematic? I read them as "milli", but I wonder if they ought to be "micro"? Doesn't make much difference in the simulation output, but I ask just so I can be clear on my understanding of your circuit.

    ReplyDelete
    Replies
    1. C1 and C7 are microfarads as in100 Ufd and 10 Ufd. Thank you for your feedback.

      73's
      Pete N6QW

      Delete
  3. Thanks for clearing that up...the "m" vs. "u" prefix for microFarad capacitors in LTSpice was something that tripped me up when I first started with the sim tool. Doesn't make a big difference in the bypass capacitors for your product detector here, but when I was trying to model frequency-sensitive circuits I was pulling out some hair when I couldn't get it to perform as expected.

    ReplyDelete
    Replies
    1. Hi Chris,

      See Part 9 for my attempt to clear up the mystery and a bit more info on the importance of the value of the bypass caps --the plots tell an interesting story.

      Thanks for input as that caused the initiation of Part 9.

      73's
      Pete

      Delete
  4. I don't know if it matters for your goals for this project, but TI is discontinuing the LM380. It's on "last time buy" status right now and no more well be sold by TI starting sometime next year. I don't know of a second source for the part.

    ReplyDelete
    Replies
    1. Hi Stephen,

      Thanks for the input --Bummer I thought the LM-380 would never go away. I have tried the TDA7520 and smoked 6 of them in the recommended test circuit. The other option is to build the discrete part amp and replace the 2N3904 and 2N3906 with "TAB" type PNP /NPN transistors such as the TIP30 and TIP31 and then it is a non issue. But that is really good info to know.

      Thanks Pete N6QW

      Delete
    2. Pete,

      Yeah, the LM380 is a good part and I'm sad to see it go. TI is discontinuing the DIP versions of the LM386, too. The SMT versions are surviving for now.

      73,
      Stephen KK7ZD

      Delete
    3. That is sad --guess I should stock up with a few. But that begs the question -- there must be some sort of replacement in the pipeline any clue what that is?

      Pete

      Delete
  5. This is a great series, Pete, I'm really enjoying reading through it (even if I'm late to the party).

    To address your repeated comments on not understanding why JPEG is no good for these schematics, perhaps I can help. :-) It has to do with frequency analysis! JPEG is a compression standard designed to provide very small file sizes for very large pixel count images of natural things. At some point, some smart cookie noticed that if you consider the changes from pixel to pixel of an image, "real" photographs tend to have very low high-frequency components compared to the low-frequency components in these changes.

    (Warning: hand-waving and gross simplifications follow!)

    By performing what is called a Discrete Cosine Transformation (DCT) of the image pixels, their color space values can be transformed into frequency space values. The high-frequency components of this data are then discarded (reducing the size of the matrix) and the low-frequency components are retained as the "compressed" image. When this tranformation is reversed at the other end, the image is visually nearly identical to the original image, because the colors in natural images change slowly from pixel to pixel and the loss of the high-frequency data is relatively unimportant. Think of this like passing spoken human voice through a moderate BPF (say 100 Hz to 5 kHz); you lose information, but the resulting signal is probably not noticeably distorted unless you compare it to the original side-by-side.

    The problem is that, for text and line art such as schematics, the high-frequency components of the image are very important! Consider that a black pixel next to a white pixel is a 100% change in "signal level" across a single pixel boundary. This means that, when the image is compressed and then restored, the lost high-frequency components cause artifacts in the signal. This is exactly analogous to passing a square wave through a LPF that cuts off higher-order harmonics -- you get rounded edges, ringing, and oscillations across the "flats" of the wave.

    GIF, on the other hand, uses a "lossless" compression that produces exactly the same image data on decompression as was fed in at compression time. Its compression algorithm makes different assumptions, and causes significant damage to photographs (as its color space is quantized to no more than 256 values) and provides only modest compression in that use, but it provides very high compression with zero image distortion when applied to low-color-count line drawings.

    Computerphile has a few YouTube videos on JPEG compression that, provided the above did not bore you to tears, you may find interesting:

    https://www.youtube.com/watch?v=n_uNPbdenRs
    https://www.youtube.com/watch?v=Q2aEzeMDHMA

    And why this is problematic for line art:

    https://www.youtube.com/watch?v=yBX8GFqt6GA


    More than you ever wanted to know about JPG versus GIF, but that's why people gripe! For my own part, as long as the image is readable, I don't get too cranked up.

    73,
    Ethan KB8OJH (which you may recall from SS #182!)

    ReplyDelete
  6. Hi Ethan,

    Thank you for the explanation -- I have been trying hard to make them all GIF's so I won't receive further emails from the person in Spain. I vote in your column --if I can read it I don't get too cranked up too. I guess I am a bit more practical --most of the schematics I publish are small portions of an over all circuit so they should be fairly well readable in jpg format.

    Thank you for the enlightenment --yes SS 182 --where I was trying to wake up the dead!

    73's
    Pete N6QW

    ReplyDelete