Monday, November 16, 2015

Simpleceiver ~ Part 11

Bits and Pieces Needed for the Transition from the Direct Conversion Receiver to the Super-heterodyne  Receiver version of the Project.

Addendum: Oscilloscope plots of the BPF when driven by the Arduino + AD9850 and the filter connected to 50 Ohms. Amazingly what was theoretically shown in the simulation is in fact what is seen when connected to the real hardware. Let's hear it for SSDRA and LT Spice!
 
If you have followed along this far by building the Audio Amplifier stage, the J310 DGM Product Detector and the Arduino/AD9850 LO then we are ready to proceed with the next bits and pieces.
 
I struggled with having the prospective builder include the 40 Meter Band Pass Filter for the initial build and weighed carefully the balance between building yet another module or firing up the radio with a smoke test.
 
In my initial testing I saw that the radio worked with just these three modules thus proving the design concept. But the Direct Conversion Receiver should have a Band Pass Filter in the loop. Keep in mind with the antenna connected to Gate 1 --every bit of RF at all frequencies is being picked up by the antenna and sent to the J310's. As I have found out many times, builders are using  just a chunk of wire laid out on the floor of the shack versus an installed antenna, up in the air and cut for the band in use. 
 
We even had the individual who built the LBS Direct Conversion Receiver hook his receiver to the closest rain gutter for an anetenna and we got the usual "Your Receiver Isn't Working" email. If the builder intends to stop the project with just the direct conversion receiver, you still need to build the Band Pass Filter! YOU WILL ALSO NEED A PROPER ANTENNA!!!!!
 
So now we will take up next the Band Pass Filter (BPF) and its purpose in the circuit. Just as its name implies the BPF will pass certain bands of frequencies while rejecting others. These filters are used both on receive and especially on transmit. Not all filters are created equal as some have a very narrow band pass with steep signal rejections off of either side of the desired bandwidth, while others may be a broad as the proverbial barn door. Designing a filter that has the necessary bandwidth and covers the desired band while rejecting out of band signals can be done by hand!
 
W7ZOI in his Solid State Design for the Radio Amateur provides the necessary equations and detail to do this [located in the Appendix]. The same procedure is available in his EMRFD book and of course there is the Internet. I prefer to use the SSDRA in conjunction with the LT Spice Simulation. For this design I went into the SSDRA table of stock BPF constants and started with those values for the LT Spice simulation.
 
The simulation approach I find is most useful for fine tuning a filter and then almost in real time seeing the impact of the changed constants in the plot functionality. Another nice feature is that with LT Spice you can find a value of inductance that will work in the circuit while at the same time have a value that can be built with a specific number of turns. I haven't quite figured out how to wind an inductor on a toroid with 19.7568 turns --19 turns or 20 turns works fine business. But trust me that decimal value of inductance CAN be important.
 
Below is a simulation schematic of a 40 Meter Band Pass Filter that can be used with project. I call your attention to the two inductors whose value is 2.28 Uhy (Micro Henry). This value was tweaked so that with 20 Turns of #20 will fit on a Amidon T-68-2  (Red) Core and provide the needed inductance. A note here about values for the capacitors C3, 4 and C5 are standard capacitor values but an effort should be made to get close tolerance values (1% would be ideal but NPO or COG are a must!). The capacitors marked C1 and C2 at 164 PF is nothing more than a 150 PF (NPO or COG) in parallel with a 0-50 PF NPO Trimmer capacitor. Using this combination enables "fine tuning" the tank networks and by adjusting the trimmer "sneak up" on the required additional 14 PF.

Also noteworthy is that the SSDRA shows you how to treat "small values" of coupling capacitors. Some designs show a value of the center coupling capacitor like in the 1 PF range --it can be done with larger values of capacitor through a series/paralleling of the capacitors. For this build I actually had a stock of 7 PF caps but the bin was empty -- luckily long ago I purchased various values of small NPO trimmer caps [ 1-3 PF, 5 -13 PF, 3 -12 PF] where I used the 5 to 13 PF trimmer.

Important note: When using the simulation tool or measuring a built filter IT MUST BE TERMINATED WITH A 50 OHM NON-INDUCTIVE LOAD. In the simulations I used a 50 Ohm resistor in series with the source and the filter is terminated in 50 Ohms! Should a signal source be applied to the filter  and measurements taken with a scope--it must be taken independent of being in the circuit across a terminated filter with a 50 Ohm termination!!!!
 
 
 
Below is a plot of this filter from what I call 10,000 feet over the spectrum 1 to 100 MHz. It has a narrow pass at our desired 40 Meter range. Note that any harmonics are about 40 dB down from this fundamental.
 
 
 
Now we have a more detailed view of our band pass window and the variation over most of the band is less than 1/2 dB. Being that this will be a SSB transceiver I was unconcerned that the lower 40 kHz was somewhat attenuated. But if the intent is to use the Simpleceiver on die hard low band edge work then using the simulation a bit of tweaking will reduce the attenuation on the lower 40 kHz. I will leave that exercise to the reader.

 
So round up your parts and starting soldering. I usually place the Band Pass Filter ahead of the product detector in the Direct Conversion Receiver and if you happen to be using a DBM as a detector where it is a passive device (loss in the conversion process) then you might need an RF amplifier ahead of the detector. Thus my configuration for the is case would be the antenna > RF Amp> Band Pass Filter> Detector. But since our Product Detector is an active mixer the configuration would be Antenna > Band Pass Filter > Detector.

  A photo of the prototype 40M Band Pass Filter -- it works!!!! I couldn't find my bag of 7 PF caps so I substituted a 5.5 to 13 PF trimmer cap in its place.

 
  Addendum:


While I have said this Band Pass Filter works pretty good I ran some tests (we do have readers of this blog who believe nothing) so here are a bunch of screen shots of the response of the filter. For an RF Source I simply to the Arduino + AD9850 and ran that into one side and then on the other side I terminated the filter in 50 Ohms and the measured across the 50 Ohms with my DSO.

 


Our original simulation plot showed that we had a fairly flat response from about 7.040 MHz (who cares about low end CW) to about 7.280 MHz. Look on the scope face and you will see the frequencies that are being measured and sure enough our output plot follows that curve. Houston we have qa working Band Pass Filter!
 

 










 
 




 
73's
Pete N6QW

4 comments:

  1. Pete, just to confirm what I think I understand: This BPF has a 50 ohm output impedance and the product detector antenna port has a 2.2k ohm input impedance, therefore we still need to use the impedance transformer (3 turn primary to 20 turn secondary wound on a FT-37-43 core (#26 wire)) which you mentioned in Part 8 of this series.

    I married up this BPF design to the product detector design in LTSpice and without the matching transformer, the bandpass has about 8 dB of ripple. When I add in the transformer the response is flat (<0.5 dB ripple), so I suspect I'm on the right track.

    now, to get to ordering the toroids, trimmers, and melting some more solder.

    ReplyDelete
  2. Hi Christopher,

    That is the problem with being too close to this stuff --kind of hard to read my brain. Yes that was the whole reason for building the matching transformer on the J310 DGM PD so that the match would be 50 Ohms "in" and you simply could connect to the BPF and call it a day. I am working on the Crystal Filter Module that has a J310 DGM ahead and following the 4 pole filter. There are several sets of matching transformers needed --so order extra cores.

    Out of the 12 MHz 1st IF amp is a 2 dB Pad and the match will be 50 Ohms out to 300 ohms in for the filter. On the out side of the filter is 300 Ohms to 50 ohms in to the 2nd IF J310 DGM and it's out is again 50 Ohms to the 50 Ohms on the Product Detector.

    The FT-37-43 cores have become quite expensive but let me suggest you buy two dozen ~ about $20 --but you will save on the shipping.

    73's
    Pete N6QW

    ReplyDelete
  3. As we always said "two is one, one is none" so I'll of course overbuy (good way to establish a junk box). Not sure of your toroid source, but i see the for $6/25 or $18/100 at Kits and Parts dot Com for that core. Can you offer any good pointers for trimmer caps other than eBay?

    /r
    chris

    ReplyDelete
    Replies
    1. Hi Chris,

      Try Dan's Small Parts in Missoula Montana (yes that is correct) and you can find some on this link

      http://danscloseoutsandspecialdeals.com/specialdealsandleftovers.html

      RF Parts is another place. Check also HSC (Halted Specialties) in Santa Clara, CA. Ken's Electronics is another source (in Ohio)
      Pete

      Delete