You will find a lot of very useful "tribal knowledge" here:
Tuesday, January 31, 2017
Tuesday, January 17, 2017
Heat Sinking The PA To The Cabinet
by Don Cantrell, ND6T
by Don Cantrell, ND6T
I am a real advocate for keeping components cool. They last longer and are generally more stable. The heat sink provided with the BITX40 is adequate for lower power levels but bigger and better heat sinks provide for longer component lives and peace of mind. If you are going to install the transceiver in a metal box then you might consider using that enclosure as a heat sink. Whenever I have done this it has always worked very well.
In the early stages of installation I positioned the board as close to the rear of the cabinet as was convenient but it still left about 1/4” (6mm) gap between the Power Amplifier transistor (Q15) and the back panel when the original heat sink was removed. I found a small block of aluminum that fit nicely. I drilled a hole toward one end to match the hole in the tab of that IRF510.
When the mounting spacers had been attached to the BITX board (so I could tell where the transistor was located in respect to the panel) I marked and drilled the transistor's attachment hole first. I then placed the mounting bolt through the transistor, spacer block, and rear panel to temporarily hold it before marking and drilling the four holes to the attached mounting spacers of the BITX.
When mounting stuff in cabinets a bit of extra time and planning goes a long way toward preventing mistakes (and forehead slaps with cries of “OY!”). Go slow and save big washers and hole plugs.
After the board is mounted use a shouldered nylon or ceramic washer to insulate the mounting bolt from the transistor tab which is connected to the collector. Don't let it short circuit to anything. A mica or silicone pad between the spacing block and the transistor completes the insulating job while still conducting heat. If you use a mica insulator be sure to use heat sink compound on both sides of the insulator. A silicone pad does not need it.
If you cannot find a suitable aluminum block then look for any heat conductive metal. Copper, brass, even steel. I haven't tried it but I would assume that a stack of coins (with any corrosion sanded off) would work. Considering the cost of new metal stock in the hardware stores, it might be cheaper to just use money!
My BITX output transistor now runs so cool that I cannot detect any temperature increase. Just DO NOT TOUCH THE TRANSISTOR WHILE TRANSMITTING! RF burns are worse than the damage you feel from picking up the wrong end of a soldering iron. Or so I'm told.
Posted by Bill Meara at 5:08 PM
Monday, January 16, 2017
BITX Voltage Monitor
by Don Cantrell, ND6T
With eight pins available for reading analog signals, the little Nano® has some enticing capabilities. More useful than an “S” meter reading, a measurement of the supply voltage is another display item that you can easily include. This will allow you to keep track of your battery status if you are portable. It could also be adapted to check on any other voltage (or voltages) that you might be concerned with. Like your PA supply if you are going to feed it with a separate higher voltage.
Looking at the wiring diagram of my VFO build you will notice a couple of resistors that are in series and placed between the +12 Volt input and ground. There is a 1.8 Kohm from ground and a 5.6 Kohm on the power feed side. Power rating is unimportant since they are going to dissipate only about 25 milliwatts. I chose these values so that it could measure up to 20 Volts without subjecting the Nano® to more than 5 volts. A safe Nano® is a happy Nano®.
Whenever measuring any voltage with the Nano®, endeavor to keep the maximum voltage to the analog pin below 5 Volts but compromise to retain the best dynamic range. The measurement involves converting the range into 1024 steps and then dividing the result by whatever number will display a reading that will be closest to what the measured voltage actually is. If you have too much range then those steps will be larger and more inaccurate.
Software implementation is wonderfully simple. Here's how to read the voltage on pin A2:
// Read supply voltageint supply = analogRead(A1);
The first line is just a comment, reminding you what the next code is about. This line is optional. The second line establishes a local variable (that I named “supply”) and specifies it to be an integer type, saving a little storage space and keeping handling simple. The “=” sign assigns the result of the “analogRead” function on port A1. If the voltage applied to that pin was 2.5 Volts, for instance, the “supply” variable is now worth 512. That's because 2.5 volts is half the maximum 5 volts and, therefore, half of the maximum 1024. Make sense?
To display that as a voltage on our LCD screen we would use this line:
Why divide the supply variable by 47.7? Because that would make it correspond to the voltage that we were feeding the resistive scaling divider with. Find it by experiment. In this case you would see 10.73 on the display. Apply 10.73 volts to the top of the divider and you will see 2.5 volts on the tap where you have pin A1 attached and you will read 10.73 (or so) on the display.
Easy. Two resistors, two lines of code. You can use similar schemes to monitor any voltage or current. Rectify the voltage from a winding on a toroid that you run the antenna lead through and read output power (antenna current). Set up a VSWR bridge. Ahhh! The possibilities!
Posted by Bill Meara at 3:06 PM
Sunday, January 15, 2017
Adding an “S” meter function
Don Cantrell, ND6T
Aside from dressing up the display (taking up room is more like it) there is little use for a signal strength indicator (“S” meter) unless it is accurate. Admittedly this implementation is not accurate, merely a nod to the notion. That said, it does work as well as a few “store-bought” rigs that I have used. Since it is simple and easy, I will pass it on for what it's worth.
Just three components, none critical. I ran a lead from the “hot” side of the volume control through a diode to an available analog port on the Nano® and paralleled a 0.1uF capacitor with a 4.7 Megohm resistor from that port to ground. The diode can be a silicon switching diode like a 1N914 or 1N4148 but a germanium like a 1N32 will work better for the lower strength levels.
The reason for the inaccuracy is primarily due to lack of the usual AGC circuitry. Normally one just measures the AGC voltage and scales it to suit. This mod is simply using the detected audio and then using a short-period peak hold circuit to feed the micro-controller.
The code is poorly done at this point but it met my low expectations. It currently resides in lines90 through 94 in Version 1.0 of the BITX40_VFO.ino sketch. Please feel free to change it and pass it along.
Posted by Bill Meara at 7:06 AM
Saturday, January 14, 2017
SECOND HARMONIC FIX
by Don Cantrell, ND6T
No matter how I tried, I could not drop the second harmonic to make 42 dBc. Best I could get was about 39. That's not far out of FCC regulations and is probably well within the suppression required in other parts of the world. Probably it is just my particular board, the minor variations of the capacitors and inductors combining to just miss the mark. I haven't heard of anyone else reporting it but I will pass this along just in case. If you are using an external amplifier then I am certain that the filtering on that amplifier will solve it easily. You could also use an external low-pass filter.
I ran a simulation of the filter bandpass using the AADE software. The results are shown in Figure 1(Simulation) as the upper trace. You will notice that the suppression provided at 14.4 MHz (the second harmonic) is around 26 dB. The lower trace is around 50 dB at that frequency and is the result of paralleling L7 with a 100 pF capacitor. This extra capacitor acts like a trap for the second harmonic. I find myself installing this on most of the transmitters that I design, just to save an additional filter section.
I just soldered the 100 pF disk across that inductor on the underside of the board. I don't notice any difference in operation but I am now legal. Figures 2 (Before) and 3 (After) show the before and after measurements on my spectrum analyzer.
Has anyone else experienced this problem? If so, here is a cheap and easy fix.de ND6T
Posted by Bill Meara at 4:51 AM
Friday, January 13, 2017
RF Gain Control
by Don Cantrell, ND6T
I've grown used to RF gain controls, especially on my home brewed receivers. Often it is the only gain control and serves to adjust audio output as well. When I first put the BITX on the air I called a friend who lives within shouting distance (literally!) for a test. That's when I began to miss that RF gain control on the BITX.
During my experiments with the analog VFO I installed an additional pot on the front panel that provided a precision 1 kHz tuning. The new DDS system, of course, didn't need that knob and it became available for re-purposing. What better use than a RF gain control?
My plan was to insert a 10K ohm variable resistor (rheostat) in series with the power feed to Q1, the receive RF amplifier. Original design integrity is maintained by placing it between R15 and R16. This was accomplished by “tombstoning ” R15 (10 ohms) on the pad next to the supply via. I attached one lead from the control to the top of the (now upright) R15 and the other lead to the (now vacated) solder pad. Done! That's it!
I now have my beloved RF control with about 17 dB of adjustment range, and the feng shui of the front panel is restored.
 “ Tombstoning” is the term for tipping up a surface mount component. It now stands like a headstone over the site of its original position. This is the normal method of opening a circuit during trouble shooting. Very handy and one of the many reasons that surface mount construction is so much easier to work on and modify. Thanks Farhan! The BITX40 is beautifully laid out large, logical, and well-spaced. A true joy to work with. Those relatively large components make it easy compared to through-hole.
Posted by Bill Meara at 1:51 AM
Thursday, January 12, 2017
by Don Cantrell, ND6T
As much as I love the native BITX40 VFO with it's clean signal and simple design, our rural mountain home has room temperatures that vary widely and the VCO just proved too much for me to stabilize to my taste. I caved and raided the junk box for a Direct Digital Synthesizer solution.
Finding an Arduino Nano®, an AD9850 board, and a generic I2C 1602 LCD display, I cut a slot for the display in the front panel of my Radio Shack® #270-235 cabinet that I found in a swap meet (5 ¼”W X 3” H X 5 7/8” D) and mounted it with 4 short stand-offs.
I used a small individual-hole-plated piece of vector board to mount the Nano® and DDS board using IC sockets. A 7805 regulator was then soldered to small piece of un-etched printed circuit board and glued to one edge. A hole was drilled through both for a support bracket. I also mounted the filter capacitors along with a couple of resistors for a voltage monitor feature. This daughter board is now bolted to the rear panel of the cabinet with an “L” bracket, therefore using that previously wasted space and also heat sinking the regulator (as if it needed it!).
Not having an encoder, I decided to use the original tuning pot in a “shuttle” arrangement. (That was what we called the control on video editing boards back in the 1960's). Instead of spinning the knob, you twist it one way or the other. In the center position, it is idle: No tuning, the frequency is stable. Turn it clockwise slightly and the frequency starts to slowly increase. Counterclockwise, the frequency decreases. As you twist farther it tunes progressively faster. No buttons, no pressing. Tune any frequency in mere seconds and in single Hertz steps when you near your target.
It's like having a power tuner. The amount of twist is like the amount of throttle. I have mine set up so that it takes a leisurely 15 seconds to tune from 7.0 to 7.3 full throttle but I can still tune down to the single Hz frequency. That's actually much faster than tuning any commercial rig when set to single Hz resolution. Turn to the 11 o'clock or 1 o'clock position and slowly scan for signals. When I get around to it, I intend to install spring loading to provide haptic feedback on the knob. In the meantime I display a tuning indicator after the frequency reading.
One big knob. No clicking, no pushing, no range switch. Best of all, no boggle like you get if your knob is not-quite-in-the-middle of a stepped value. It is one big range with the tuning speed adjusted at an exponential differential rate. If you don't like the rate or the idle zone size, just plug in the USB cord from your computer and change it to suit. Create band limits. It's an Arduino®!
Although Shuttle Tuning requires just a few lines of code, this blog isn't probably the place for it. If anyone is interested in reading my poor programming style they can email me at email@example.com or at ND6T@ARRL.net and I will gladly attach a copy of the simple software and answer questions.
Yesterday I ordered a second BITX40, this one with the Raduino. I can hardly wait!
Wednesday, January 11, 2017
Hacking an earlier version of the Bitx40 to include the Si5351 LO and an OLED Display.
I purchased my Bitx40 early in November 2016 before the "Digi" VFO was available from hfsigs.com. This hack replaces my original hack that used the AD9850 and Pro-Mini I initially installed with a Nano/Si5351 plus Black and White OLED Display. The code for my OLED/Si5351 can be found at http://www.n6qw.com/Phase7.html The current configuration of the Bitx40 being offered would obviate the need to build this hardware.
Monday, January 2, 2017
Adding a Linear Amp Hack to the Bitx40.
The Bitx40 is an amazing platform and VU2ESE has even included an option to select the power output level simply by changing what voltage is fed to the IRF510. This easily done using the on board power connector that supplies "juice" to the IRF510. At 12 VDC you have the typical 5+ watts (more like 7 on this end) or by upping the voltage at this connector to 24 VDC the Pout is now 20 watts. However at the higher power level you will need to increase the IRF510 heat sink size (significantly). Bill, N2CQR has done this and you can see this at his blog http://soldersmoke.blogspot.com
Here is an option to keep the Bitx40 at the normal power level (12 VDC) but switch in line an external linear amplifier. It takes but few parts and the modification is totally external to the Bitx40 board. I have an SB200 Linear Amp (heathkit workhorse) and with the Bitx40 as a driver the Pout is 100 watts. More than QRP but really great when 40 Meters is somewhat marginal.
The theory behind this hack is the panel mounted Push To Talk (PTT) line is modified to include two isolating diodes. The diodes essentially separates two functions, where one line, which connects to the PTT1 connector on the board puts the Bitx40 into the transmit mode and the second line engages a small reed relay to provide a closure to ground to turn on your external amplifier. The parts required include two 1N4007 isolating diodes (yes a bit of overkill but in the junk box) a 1N4148 "snubber" diode across the reed relay coil, a terminal strip, the 12 VDC reed relay (low current draw on the coil ~ about 10 ma) and a panel mounted RCA jack. You will need about two feet of insulated wire to make the connections.
The actual hardware installation part includes using a small piece of scrap PC board that serves a base for the Reed Relay. Essentially I super glued the relay upside down to the PC board and then mounted the board to my case (10X6X2 metal chassis). For the ground side of the linear amp contact closure, I simply soldered a wire to the copper PC and the other end to the reed relay. The other end of the relay has a single wire (blue) running to the rear mounted RCA jack. In my initial install of the Bitx40 I used a 5 lug terminal strip for the power connection/power switch. That only used up 3 of the terminals so the remaining two lugs were perfect for the linear amp switching hack.
Whether the linear amp is in line or not is controlled at my linear amp itself. An additional hack might be to include a panel mounted toggle switch at the Bitx40 that would "open" the amp control line. Since I had a bypass capability at my amp I did not install the additional toggle switch.
One advantage to this hack is that does not require you to have 24 VDC plus bigger heat sink to get more Pout. I guess another advantage is that it keeps the rig pure at QRP levels especially when you feel guilty about being QRO. There is also the advantage that with a bit of padding on the Bitx40 Pout, you could drive an LDMOS amp to 600 watts out. It just plain gives you a few more range of options.