A New Noise Blanker for the Drake R-4C

Overview of the Project

At my noisy QTH, a decent noise blanker is a must.  I had the Drake 4NB noise blanker in my R-4C for a while, but it didn't work at all well on my QRN.  I decided to sell it and put together something that would work.  I'm actually not very fond of soldering circuits together, so I decided to try and find an easy way to get the noise amplifier and maybe the pulse shaping circuits I'd need without having to build my own.  I found that Ten-Tec had just what I wanted in the noise blanker for their Scout model 555.  The noise blanker is known as model 297 and this link will take you to Ten-Tec's price page. (Ten Tec has since discontinued this blanker. Here is a link to a more readable schematic diagram of the blanker, so that the more adventurous can build their own.)  I purchased one for $19.50, plus postage.  Then I designed and built the circuitry to interface it to the R-4C.  This page describes how you can build it too.  The results are really good.  Even though I put the blanker in the first IF before any filtering, nearby signals have much less effect than they do on my Kenwood TS-850SAT.  There is no noticeable effect from nearby strong signals if blanking is not taking place (i.e. when there is no QRN).  In other words, strong nearby signals don't cause any noticeable blanking action.  There is a small amount of intermodulation distortion noticeable when blanking is taking place and strong signals are close by.  The blanking action is good.  It removed the QRN that the 4NB wouldn't touch.  I can now use the venerable R-4C without having to listen to that terrible buzz from the local power line arcing.  This is an ongoing project.  There are some things I'd like to try that should provide some improvement in performance.  Most of them are noted on the page.  You might want to check the page for further changes in the future.

Making the Modifications

There are three parts to the modification.  The first is to modify the model 297 blanker board to work in the R-4C.  The second is to construct and install a noise gate in the R-4C.  The third is to install the two boards and tune it up.  Total cost should be about $25, less if you have a well stocked junk box.  This compares favorably with obtaining a used 4NB which didn't even work well for me.  (I sold my 4NB for $125.)  The parts list is given below.
 

• One Ten-Tec model 297 noise blanker.
• Two FT37-43 toroid.
• Four fast switching diodes (1N914 or 1N4148).  (If you're a perfectionist, you will want some hot carrier switching diodes, often used in mixers.  HP and others make these.  A good one would be the Hewlett Packard HSMS-2822, diode pair.  I have now tried using the HSMS-2822 diodes, and notice on the scope that more of the noise pulse is being taken out.  It is hard to say if there is a big difference just listening though, but I think it is worth using the HSMS-2822 diode pairs.  As it stands, there is still about 1 uS before the gate shuts down.  It would be nice to find the source of this delay.)
• One 13 uH indictor.
• One 4.7 uF electrolytic capacitor.
• Two 4.7 Kilo ohm resistors.
• Two .001 uF bypass capacitors.
• Some wire.
• One perfboard.

  
   

Modifications to the 297

Get your soldering iron out and remove R1 the 220 ohm resistor across the input of U1.  The purpose of this is to raise the input impedance of the noise amplifier so that it doesn't load down the IF.  The next thing to do to the 297 board is to cut the trace that connects both pins of connector 8 together, and cut the emitter of Q5 lose from ground.  Run a jumper wire from the emitter of Q5 to the now free pin on connector 8.  See the schematic of the 297 below.

Remove the redundant wire from the connector that snaps on to connector 42 on the circuit board.  Lace the other wire through the hole in the center of noise blanker socket on the R-4C, and solder it to the +14 volt pin on the under side of the R-4C chassis.  Now lace the black wire on the wire connector 8 (which will connect to the emitter of Q5 of the 297) through the same center hole on the R-4C noise blanker socket, and solder it to the pin that connects to the function switch on the underside of the chassis.  This pin is right next to the +14 volt pin.  (This measured closer to +12 volts on my R-4C.)  The connection points shown below on the R-4C schematic.  If you would like to see a larger schematic of the R-4C, click here.

The cuts and jumps on the circuit board can be seen below if you have a good eye.  A slightly closer view is available here.

The IF frequency of the Ten-Tec Scout is 6.124 MHz, and the first IF of the R-4C is 5.645 MHz.  I thought I'd have to add a little capacitance to the resonant circuits in the noise amplifier, but it turned out that L1 and L2 had enough tuning range to resonate at the Drake's first IF.  What a nice serendipity!
 
 

Constructing the Noise Gate

At this point you will need to construct your noise gate.  Use the small perfboard, and dead bug wiring.  First wind the transformers, T1 and T2  They are identical and have 21 turns trifilar wound on an FT37-43 ferrite core.  Use two of the leads as the secondary, with the place where they are soldered together being the center tap.  See the figure below.

The circuit is shown below.  I believe it would work a bit better if a delay line was incorporated in the gate circuit just before the present input.  I intent to experiment with this a little.  It will depend on the reverse recovery time of the diodes somewhat.  I'd like to try some hot carrier (Schottky) diodes as well, but I don't have any at the moment.  Presently the pulse is around 10 uS long, and I when I set the scope to see that pretty well, I couldn't notice that any delay was necessary assuming the diodes turn off instantly.  They don't, but the time that it takes to turn them off is much greater than I could get with a reasonable lenght of coax, so I'm just living with it at present.  Using hot carrier diodes would solve this problem.  Anyway, keep leads short so they don't couple the signals around the gate.

Installing the Noise Blanker

It is important to keep the high gain noise amplifier away from the stray RF under the chassis or the R-4C.  I mounted it on the top side of the chassis.  I found it handy to put the noise gate beneath the chassis.  The delay from a transmission line is about 1 nS per foot.  See the photos below.

This photo above shows the overview of the position of the model 297 noise blanker board.  The one below shows a closer view of how I mounted it to the AM filter bracket.

The photo below shows how I mounted the noise gate under the chassis.  The noise gate is the little brown circuit board with the two toroids, etc. mounted on it.

Tuning up the Blanker

To tune the blanker, you can either adjust L1 and L2 for maximum blanking effectiveness on your noise, or if you have an oscilloscope available, you can put it on the collector of Q1 and tune for maximum AGC action as seen by the thinnest line on the scope.  (If there is noise present, disregard those peaks, and concentrate on the rest of the signal.)

Theory of Operation

The first IF of the R-4C is sampled by the high impedance (roughly 5 kilo ohms ) input of the Motorola MC1350 IF amplifier IC.  This amplifier has an automatic gain control on pin 5.  With 5 volts on this pin you have zero db attenuation; with 7 volts you get about 70 db.  The link above will get you the Adobe Acrobat PDF file data sheet.  There are two of these amplifiers in cascade giving a possible peak gain of almost 120 db.  This amplified signal is fed into Q1 and Q2 which are the AGC amplifiers.  Q1 is an envelope detector.  Q1, R11 and C12 form a fast attack, slower decay (56 uS).  R12, R13 and C13 provide a very slow attack (19 mS) and very slow decay (33 uS).  This is to keep noise pulses from having any affect on the AGC.  Q2 level shifts and amplifies the AGC to the correct level.  The AGC is then fed back to U1 and U2.

Q3, Q4 and Q5 provide the switching pulse.  Q3 is an envelope detector similar to Q1 only it takes slightly more to trigger it.  I'm considering putting a resistor similar to R10 in this circuit and seeing if I can adjust the blanking level a bit.  (Another thing to play with.)  Q3, R19 and C14 form a fast attack, and a decay time of 4.7 mS.  This is about what you want for a blanking pulse, because the bandwidth of the IF and noise amplifiers previous to this stage is about 100 kHz (the pulse is spread to about 10 uS).  It might be worth a little to widen this by increasing C14 to say 100 pf, but so far I haven't tried this.  Q4 amplifies the pulse and drives the final driver Q5 which again acts as a fast attack, slow decay with the .001 uF and the 4.7 kilo ohm giving a decay time of again 4.7 mS.  The two decays roughly add to give almost the 10 mS desired.

When Q5 grounds the anodes of the diodes in the gate, they don't conduct, and so the signal is blanked for about 10 mS.  The noise gate is a singly balanced one to eliminate harmonics of the blanking pulse from the output.  The 4.7 uF capacitor keeps the cathodes of the diode strings at 6 VDC potential during the switching operation, and serves as a reservoir of charge.