Saturday, March 9, 2019

So How To Do CW on a Homebrew SSB Rig?

3/16/2019 ~ Remote Antenna Tuning

3/14/2019 ~ Calling All Junk Boxes????

Calling all Junk Boxes, N6QW is looking for a part!

Recently I bought an SBE-33 SSB Transceiver for $52.30 and have been doing some major restorations, which principally involved a great deal of capacitor replacements. I did have to replace the speaker which was blown; but had an exact replacement taken from a Junker unit. 

I have had the rig on the air and so far I am impressed at the 1963 technology --even with an analog VFO. Yesterday running my SB-200 I worked Croatia, 9A5W, Nikola on 40 Meters. That is a big first for me and I got a notable report of how good it sounded. 

A short story --- the dial markings are at every 5 kHz and a ham running a FLEX5000 contacted me to tell me I sounded bad and was not on frequency (7.208 MHz). I asked that he give me some chatter wherein I zero beat him. Then when I transmitted back at him inquired about the signal. He said I don't know what you did but you fixed it. You are now on frequency and the audio is clear. Wow --he has been a ham for 5 years and well I will stop there. 

The Junker SBE-33 has provided many mechanical parts that were missing (the speaker of course and mostly shields although I did have to fabricate one of the shields). Should mention all the transistors are original save for one and I think I have that one too. Look up the specs on a Philco 2N1727 just for fun.   It is now mostly stock and I am down to one item -- The Panel Meter. 

It is blown and I attempted to repair it but can't get it loose from the case. Thus the only option is to find a replacement meter. Not any meter but one that came from an SBE-33. I guess blowing the meter was a problem, as the Junker is minus the meter too.

So if you have an SBE-33 sitting gathering dust in your garage, here is your chance to move it elsewhere. I need just the meter but will take a whole rig. Please advise price and shipping to 91320. I can be reached at

Pete N6QW


Arduino: How to add the CW capability!

It is always best to turn off the soldering iron first (and to also know which is the "hot" end) and go through a bit of a noodling process to actually think about what you want to do.

  •  Firstly we need to take a look at the Pin Count. If you are using one of the really neat Color TFT displays as I have in many of my rigs you have taken up quite a few of the Arduino Pins. At this point you might want to start looking at an Arduino Mega 2560. Having 54 digital pins will not constrain you on inputs/outputs and the 16 Analog Pins is also a bonus. If you will be using say a 20X4 LCD then the use of the I2C buss frees up a lot of digital pins and opens up the possibilities for using a Pro-mini or a Nano. Overall size of your rig may force a smaller physical size microcontroller. Depending on how much you want to display you might even get by with an 8X2 LCD. When I documented my Bitx40 build which was fabricated just before the Raduino was added, I used an 8X2. You can see that on my website at
  • So from a Pin Count and for the normal SSB operation we have the encoder and step select ( 4 pins including ground). The I2C Buss (4 pins, but two are  5 VDC and Gnd). USB/LSB Select (one pin), VFO A/B Select ( one pin) and TUNE (one pin). True not all are digital pins, as five of those use Analog Pins 1 through 5. So we still have lots of Digital Pins to work with using  the Uno, Nano or Pro-Mini. 
  • Functions we are looking to add to the Arduino: the very first is to have a switch ( 1 pin) that says check if this pin is active and if so go to the CW routine. The very 1st item on the list is to shift the BFO to USB for CW Receive.
  •  Because we are using the Arduino we can accurately generate audio tones and thus we have the option as described earlier of using the tones to do CW. In one of my transceivers when I engage the TUNE function at the end of the pulsed tone I sent my call sign at 988 Hz. It sounded pretty decent. BUT we will have to do a lot of RC filtering as the TONE output is a Square Wave it would be nice to pump a sine wave down the throat of the Balanced Modulator. We will need to designate one pin as the TONE output that we later smooth up.
  • For our CW application there are tones closer to 700 Hz and I will find the appropriate one for the code. By using two pins on the Arduino we can accurately generate timed dits and dahs. By using one of the analog pins (we have three spare ones A0, A6, A7) we could include a speed control so that even though you physically send slower the output can be at a blistering 100 WPM. That will show those SSB only guys!
  • In the code for turning on the Si5351 LO we may need to include a constant, call it c,  in addition to the standard generated frequency and for SSB the constant "c" = 0. But for CW the constant is the amount of the offset. This may not be required but is put on the list so it can be added or not included depending on further noodling.
  • A timed output pin, (only active on CW) would key the PTT and hold the transmitter on for some time period. This output responds to when either the dit or dah key is engaged. Another pin could actually signal different delays --or two pins four delays. 

This is just some initial noodling on my part. But mind you a real Arduino programmer would have all of this as menus where you could select timing levels output speed and delays. But I do think this is a way of adding the capability albeit a bit klunky.
I will have to dust off the QRP Quarterly articles that AI6YR and I wrote to evaluate adding keyboard sending to the mix as that would change some of the functionality I described.
Now the thought occurred to me (and assuming unattended operation is OK for this application) --using canned code and reading the CW, in the morning turn on your rig and all day long the rig is having QSO's while you are at work. How cool is that?

Pete N6QW

The Answer is Not a Flippant: Carefully!

[Author's note: A friend in VK4 land made an inquiry about CW operation. I find that 99.99% of my operating time is SSB. But others spend a greater time on the air using CW so why not share some info and data that I have stashed on my computer where a SSB rig can be made to work CW. This also open the possibility of filter switching for a more narrow pass band. With Arduino anything may be  possible.]

My shack is full of homebrew SSB transceivers but only two have CW as an option as well as the normal SSB functions. A design of having both modes must be done with great care and forethought. The rig which does have the best CW functionality is my KWM-4 which hit the airwaves in 2013. 

I spent a great deal of time looking at CW schemes and it became obvious that many commercial rigs included CW as an after thought to aid in the marketing and did not really address the desires and wants of the ardent CW operator.

Here are some of the results of my research. 

  • One method was to simply unbalance the carrier balance control, turn off the mic circuit, crank back  the power and flip a switch often called MOX (manually operated transmit). Thus change over always involved the MOX so no break-in. You essentially keyed the PTT. The offset was poor and involved a lot of leap frogging to copy you. Often there was no side tone oscillator.
  • Another method involved shifting one of the BFO frequencies straight down the throat of the Filter Center Frequency (Cf) and then key a buffer stage somewhere in the loop. This method was OK; but often limited you to sending CW on one side band and receiving on the other. I think the convention now is to receive on USB. Some later commercial rigs let you pick CWL or CWU.
  • Some rigs used a clever approach that employed a tone being generated on "key down" that did three things: 1) the first is you now had a side tone; 2) you now were able to use that tone to key the VOX so you had break-in operation and the third the tone was sent to the microphone input circuit. I think this is covered under Part 97 as tone modulated CW. I think this was the system employed in the Collins KWM-2.
  • For the most part the "value engineering equation" included only one filter often (in early rigs) 2.7 kHz wide. For a CW signal 500 Hz wide that is a lot of signals snaking through the filter pass band.
  • But one homebrew rig in particular from the Dec89/Jan90 QST article authored by the Wizard himself Wes, W7ZOI used a separate crystal oscillator and keyed buffer to generate CW. (The rig was a 20M QRP SSB/CW transceiver.) But even this design had shortcomings in that to do CW you had to hold down the PTT switch. I fixed that problem when I built my version of this rig. This the 2nd rig I have with SSB and CW and predates the KWM-4.

Thus when I built the KWM-4, I looked to address those issues and decided on the W7ZOI approach; but having some refinements.

  • I would use a separate oscillator to generate CW. Somewhere in my travels I found a 455 kHz crystal and this is what I used.
  • My 455 kHz oscillator was turned on for short periods of time but not keyed. A downstream buffer stage was keyed. This is an old-timers technique as this method reduced chirp and key click as you would have in directly keying an oscillator. The short time I spoke of could be accurately set using a timing capacitor. I like a long time between letter so the code sounded distinct. Many CW operators are plain and simple crappy senders using a straight key. 
  • Next the keyed buffer stage was fitted to the circuit so that the CW signal was injected into a bilateral stage ahead of the crystal filter. This was done for two reasons: the first being I could tweak the 455 kHz signal so its offset was more in line with 700 Hz and secondly I did not have to worry about any effects of the filter and downstream circuitry. A small board mounted relay engaged only in CW on transmit disconnected the Mechanical Filter and the preceding bilateral stage fed the 455kHz RF.
  • I then developed a control circuit that did several tasks. One task was to provide break-in operation with a second to have functionality for both tuning the rig with output that required no "key down" or regular CW operation. As a bonus this circuit also providing the switching of voltages to go from Receive to Transmit and even included the usual outboard linear amplifier control. From a block diagram standpoint this is what it looked like.

This is a schematic of how it was done.

I spoke of the relay bypass on the IF Mechanical Filter

Some Snaps of the Control Board

Yes Virginia, all of the boards were made on my CNC machine. I think it turned out pretty cool. But then again I tend to think all that I do borders on genius.

Now with the advent of the Arduino and the Si5351 a lot of what I did six years ago will be just a few pieces of hardware aside from the Arduino. To wit the 1N914's coming off a the 7400 IC could now connected to pins on the Arduino. The timed voltage off of the NE555 is just a Pin Output to a DC switch that is in an "OFF" state until keyed. 

The 4.7 UF was a timing cap. But suppose we use one more pin with a switch that would give you a default time value and engaging the switch selects a different value. Two switches could give you four time values. 

I would consider the use of the third clock on the Si5351 that would provide the CW oscillator frequency that would operate in the same way the standalone did in the KWM-4 and that would feed a keyed buffer.

All of my sketches include a TUNE function that generates a pulsed 988 Hz tone that is fed into the Balanced Modulator. That same code can be modified to provide the timed voltage to the buffer as well as the side tone and show on the display you are in the CW mode. 

I just may have to build what I suggested and then this could be used with other homebrew SSB transceivers. At this stage it is always best to noddle a bit more but I think I have sketched out a possible roadmap to add the Arduino to the mix. If you used a Mega 2560 (more pins) then it would be possible to add my CW sender software so you could add a keyboard for sending CW --and even call up canned messages. With the Arduino anything maybe possible! (For those still subscribing to QRP Quarterly several years ago, Ben, AI6YR and myself did a series of articles on a CW Sender using the keyboard.)

Rock On, Pete!

Pete N6QW

Thursday, March 7, 2019

You Know You Are Gettin Old When ...

Stop Reading this has nothing to do with Ham Radio Stuff!

In a bit of wistful reflection, these days I am taking a more measured approach in what I do.

Take this morning when I was responding to renewing my AARP membership and checked the one year renewal period. I was thinking hey in five years I may not need this --maybe not even three.

Then I thought about what happened in the grocery checkout line yesterday (I was wearing my beret) and the rather attractive lady in front of me had a few items and her total amount was $19.56 wherein I blurted out --that was the year I was in the 9th grade. She turned to me  and said well you don't look that old. (I just know it was the beret).

Then I had a conversation with Go Daddy yesterday about the renewal of my hosting account and the five websites. They were trying to encourage me to sign up for three years. Again my thoughts about who knows if I will still be here in two let alone three years.

That gave cause to my penchant these days to NOT buy a new high technology radio; but instead find the $50 boat anchor specials. In part it is the idea of spending $6K on a FLEX6700 and only get to use it a short time. But maybe buying the old boat anchors is a way to try to relive the past as a way of avoiding the future. 

So if you have similar thoughts --maybe you are getting old too? But we can still learn, still enjoy and still have fun. but we must always think nothing is forever … even the emperor (with a small e).

Pete N6QW

Monday, March 4, 2019

The $52.30 SBE33 Boatanchor

1963 ~ Lots of new SSB Transceivers!

3/5 Late Update ~ On the Air w/SBE33

Well what I posted this morning was a good exercise but it turned up  different problems. The driver (12DQ7) was bad and the reason no indication --- the meter is broken. Upon close examination I saw some very thin wires floating around. One option is to find a company that rebuilds meters. I just set the bias for maximum smoke and we get > 60 watts on tune. Made a contact on 40M with a station in the SF Bay area. Got a good report. We are cooking with gas!

3/5/2019~ Happy Birthday To The US Navy Seabees!

On 5 March 1942 the US Navy Seabees were founded. So to all who "Can Do" hoist a few and Happy Birthday!

In yesterday's post and in the video I mention that I saw no current and no output after fixing the HV supply problem. I guess I was so involved in the many fixes that I didn't take time for a bit of noodling. My brain works while I sleep and so this morning we now have a theory as to the why.

First a graphic and some troubleshooting analysis. In the circuit on the left the Ammeter reads "0" for we do not have a complete circuit. In the right hand side we see the Ammeter (by virtue of Ohms Law) will read "1 Amp". That is the result of 100 volts being impressed on 100 Ohms ---100V/100R = 1 amp

Now that takes us to an analysis of just how is the "Transmit" function activated in the SBE33. There is a very small "ice cube" sized 4DPDT relay that when energized makes the transition from Receive to Transmit. In the receive mode the Cathodes of the driver tube (12DQ7) and the finals (PL-500's) are lifted above ground. Thus no current flows and no output! When the relay is activated then the cathodes are grounded and current flows and output should appear to the antenna. 

Shown below is a partial schematic of how the Three Cathodes in "Receive" are connected through a 270K resistor to the 460 VDC buss; but in Transmit the Relay engages and the Cathodes  go to a pseudo Ground. The Relay may be the issue! I can check this without the high voltage connected as I can simply place an Ohm meter on the Cathode buss and activate the PTT. If it doesn't go to "0" then that will confirm the relay as being the issue. OR that there is an open connection from the cathodes circuit to the relay. Just like the graphic no connection no current, no output!

Note that in the second photo below essentially when that relay is energized it creates a ground point above the chassis ground where the bias supply and negative side of the HV and Screen supply are common to the cathodes of the three tubes. That is why I plan on finding an isolation transformer to prevent me from getting shocked.

Noodling add lots of clarity. I will be checking this out this morning.

Pete, N6QW

March 4, 2019

In February of 1963 Faust Gonsett announced the introduction of the SBE33, which was a hybrid transceiver using mostly solid state devices and three tubes. It covered 4 bands, SSB only and produced a whopping 60 watt PEP. It also used a Collins Mechanical Filter --quite advanced for the time. This would have been a better choice to take to Midway Island versus the NCX-3

Over the years I have had several of these rigs and now added another one to my collection. My winning bid of $52.30 put in my hands. But many of these "bargains" come with a price. Let me explain …

The front panel cosmetics is what caught my eye --really great for something that is 56 years old. So let me explain what I had to do to fix this jewel. I should say I have the receiver working but the final amp stage still has some problems.

The list of issues.

  • First I put the SBE-33 on a variac and slowly brought up the voltages to reform the capacitors. But there was absolutely no sound coming out of the speaker. I put my scope on the audio output transistor and varied the "volume" control. I could see the static pattern but no sound. There is a speaker jack on the back panel and plugged in a speaker --boom sound but no signals. I have a replacement speaker but it is a major disassembly to replace the speaker.
  • I did a crude alignment and finally heard some signals; But there was this terrible popping static noise. I had replaced all of the electrolytic caps so I knew that was not the problem and that caused me to do some noodling. Below is a partial schematic and luckily the transistors are in sockets. So starting way upstream I pulled transistors in the receive chain. Finally I was down to the audio stage where Q1 and Q2 are the audio amplifiers. Q1 is the pre-amp and hung right in the base circuitry is a 0.1 Uf to ground. If that was shorting to ground that would cause the popping noise. I unsoldered one lead and noise went away. I replaced the cap with a 0.1 Uf at 100 Volts

  • I did find the tuning range was down in the CW portion of the band and found that a very small inconspicuous capacitor in the VFO tank network was unsoldered. reconnecting that made the tuning range proper. I also found some rewiring that do not match the schematic and that was corrected. In later production unit every solder joint had a dab of red paint. If the paint was missing then you know that some one had been "diddling" with the wiring. This unit is an early production unit as the decal on the back indicates it was made in Rancho Santa Fe (near San Diego). I can also tell by the fact the fuse is located on a terminal strip on the top side of the chassis. Most later units had the fuse on the bottom side.
  • Next I noted that no matter where I put the USB/LSB switch it receives LSB. My thoughts immediately went to a bad filter. The I looked closely at the filter and noted some really rank connections with bad solder joints. The SBE uses a unique scheme for shifting sidebands. The basic BFO is on 456 kHz. That signal is doubled to 912 kHz and then either doubles or tripled so you get outputs at 1824 and 2736 kHz, (I rounded the numbers). The signal is then injected into a mixing diode following the filter so the signal is added or subtracted to the main IF signal of 456 kHz. Thus the add is 2280 kHz and the subtract is 2280. kHz but with a sideband inversion.  The  VFO operates at 5 MHz and there are injection frequencies of 7, 10 , 16 and 24 Mhz. getting this all to sync is a bit of a challenge. Initially I thought there was a bad mixing diode. --it wasn't. then my attention turned to the filter. Yep the filter was connected wrong! A rewire made the USB LSB correct.
  • There are quite a few alignments and 20M still needs to be touched up but 40 and 80 seem to work OK.
  • I had disconnected the fuse assembly which is in the HV part of the power supply and that brought me to now firing up the HV. After installing a fuse, the good news is the fuse did not blow the bad news no HV. In looking at the schematic following the fuse is a 5 ohm 10 watt resistor. This is so there is some resistance in the mains ahead of the voltage quadrupling circuit. It also acts much like a step start circuit for when you turn on the juice the caps look like a dead short. The resistor looked fine with no burn marks but a ohmmeter across the resistor showed it was open. A replacement 5 Ohm 10W resistor from the junker radio now provides high voltage. 
  • The final amp problem -- no indication of plate current and no output --those are new tubes. So even with a bad driver tube you should see idling current. More super sleuthing work is required
  • It was a bargain price and with much trouble shooting and a few bucks in parts should get this puppy on the air. I hope to check into the vintage sideband net with the SBE33.

If you listened to Podcast #210 I mentioned that an ARC-5 BC453 could be made into a SSB transceiver using the 85 kHz IF. Here is where you can find it:

Pete N6QW

Saturday, March 2, 2019

My First SSB Transceiver ~ 1963.

I was asked a question...

A ham in the Mid-West with a "1" call frequents 40 Meters making contacts and always asks during the QSO "What was your first rig?"

I responded with my 80 Meter ARC-5 and the one tube Tx (6V6). But then I thought about my first SSB rig. It was the summer of 1963 and had just graduated from college and simultaneously entered the US Navy as an Ensign. My temporary duty station in Port Hueneme, CA was a three month school that focused on overseas construction and a dose of how to harden structures to survive a nuclear blast. (It was 1963). 

There was quite an active ham population on the base given that each Seabee Battalion had a complete Collins station that was activated during overseas deployments. KR6SB was the Seabee station on Okinawa. 

I happened to mention to one of the chiefs who ran Special Services that I was headed to Midway Island for a one year tour. He said "you need to take a rig with you and we have a deal with National Radio". So for $200 I had an NCX3 and the matching supply. 

Well I soon found out that you get what you pay for and besides the main ham station on Midway, KM6BI had a Collins KWM-2 and a 30L-1 and a Tri-bander up a 120 foot tower. Guess what I used?

When I got back stateside I sold the National gear and never looked back. That is until last week when I found the NCX-3 and NCXA on eBay. My winning bid was $69. Imagine my surprise that after fixing a loose HV Electrolytic that it works. Then I remembered why I liked the Collins. 

Actually the Cosmetics are astounding and no one has done any "tinkering". There may be a few marginal tubes and some "peaking and tweaking needed" but all in all not a lot needed.

Hers is a video I shot today and even made a contest contact on 20M.

Pete N6QW

Thursday, February 28, 2019

March 1, 2019 ~ A Digitally Controlled Analog VFO

A Digitally Controlled Analog VFO.

So OK here is the great reveal on some of the earlier posts on this blog. I built a prototype of a stepper motor controlling the main tuning capacitor of an analog VFO. 

We now have a digitally controlled Analog VFO. This is the first time that I have actually used an Arduino to control a stepper motor. Old hat undoubtedly to many of you; but a first for me with the Arduino. About 10 years ago I dabbled with a project using a PIC16F84. I was totally lost in the Assembly Language.

We will now, some 10 years later, circle back to that old project. It involved the controlling of a remote antenna tuner. This effort with the controlling of the capacitor with a stepper motor is a stepping stone to the Antenna Tuner.

See my website for more details on the VFO, the code to control the stepper motor including the LCD readout and the code for automatic zeroing of the capacitor on power shutdown. It also describes the Antenna Tuner Project approach and how the LCD will provide the readout for setting the capacitors and the tap points for the inductor using shorting relays.

Be sure to visit Yvan's website for the many excellent  tutorials. It all started with his you tube video.

But for now as you will see the whole thrust of digitally controlled Analog VFO was to give deference to my friend Bill in recognizing the relevance of the analog VFO in today's rigs. At the same time we have the ability to add a touch of "digi" stuff to simulate the HRO dial mechanism.

Pete N6QW

Wednesday, February 27, 2019

2019 ~ Big DX Contest Coming Up this Weekend

Big DX Contest This Weekend!

In preparation for the contest I reconfigured one of my 40 Meter rigs so it will work on 20 Meters. It was a pretty simple change over. Unsolder two caps in the BPF and retune the Trimmers and then swap in a pre-made 20M LPF. 

This is the two Filter, two VFO rig with a wooden front panel and of course the cool 20X4 display.

There are quite a few stations in the Caribbean that were testing their stations today. I picked a PJ2, and an HR8 using this rig with an outboard amp at 80 watts. I had my beam in line but can tell it is at times intermittent. I think it is a coax/connector problem at the beam itself. On the list to get fixed ASAP.

What I didn't mention was three lines of code changes, 5 minutes on the computer and I reset the tuning range to 20 Meters. Kind of hard to do with  analog systems unless you have a bunch of heterodyne crystals. 

VFO B defaults to FT8 on startup -- need to test FT8 on 20 Meters. Tomorrow.

The Red knob on the lower right hand side is the TUNE button. When the MOX switch is flipped on and the button momentarily pushed the word TUNE appears on the LCD right after the word Rig. After the 7 second cycle the word disappears and the rig returned to normal operation. During the TUNE a 988 Hz tone is generated and then pumped into the microphone circuit (following some filtering and attenuation). 

Pete N6QW

Friday, February 22, 2019

2019 ~ A Tale Of Two Hams

Two variations on the same theme.

2/27/2019 ~ Simple is Better!

 Several off and on blog comments has caused me to think (something I should have done more of sooner). 

My main issue was to "assure an automatic "ZEROING" of a circuit which led me down a path of trying to store a register value in EEPROM should I forget to manually Zero a control before power turn off. This quickly got complex. Then N8MN described to me how he used something I developed for an LDMOS amp control circuit to essentially do the same thing. Other commenters suggested a similar approach.

But then in a lightning moment at 3;00 AM this morning my brain said STOP!. You don't need to screw with the EEPROM! What I am trying to do is ZERO a control before the power goes off (either manually or automatically). When you power off the circuit logic takes over to automatically introduce a delay while the control is Zeroed and then power shut off. A bonus --some logic in there to provide reverse voltage protection.
Soon all will become apparent..

Pete N6QW

2/25/2019 ~ The Arduino Answer.

Some times the answer is at the end of our nose. Steve N8MN, contacted me off the blog and said he did what I wanted to do and he got the idea from me. Wow that is scary.

Several years ago I wanted to build an LDMOS amp but reasoned that the control system had to be extra "tight" so I wrote some control code using a Mega2560. I started off with the classical motor stop start circuit where a relay is latched across the contact of a momentary NO Push Button. The latching of the relay across the contact starts the motor. A similar in series NC PB once engaged drops the relay latch and juice is cut to the motor.

Soon all will be apparent...

Pete N6QW

Thus Steve suggests the same approach where you start the Arduino with a PB and then Break the circuit as with the Motor start only you add some delays to write once to the EEPROM before total power off. On start up passing through the setup lets you "0" the control. So I had the answer just didn't realize.

Joe, W3JDR suggested essentially the same thing in his comment. So now a bit of testing with hardware.

Pete N6QW

2/25/2019 ~ Follow On Arduino Question

Rather than try to respond to the several comments I will post some additional information here. Thanks to those who made inputs.

At least for me, I know I want to do something but often do not know how to articulate something well enough to describe it. So here is an example: not actually what I am trying to do but to illustrate the issue.

You have a digital VFO and you are tuning around and the band is dead. So you turn off the rig and turn on Fox News to get the latest real news. About two hours later you go back and turn on the rig hoping band conditions improved. In the code I use there is always a default frequency and that is where the rig starts up not where I last left it. I had an ICOM 730 a long time ago and it did the same thing only the start up was somewhere in the CW band and I wanted to operate SSB. So every time I turned it on I had to move it to the phone band. I found that annoying. wouldn't be nice to start up where you last left off? Yes most rigs today now do that...

My real application has a "0" point only this time if you do not intervene where you last left the circuit now becomes the new "0" point not the real "0" point. There is a way to mechanically "0" the circuit before turn off but if you forget to do that -- then you are stuck with a new "0" point. This results in a recalibration of hardware and software, which is not easily done. What I was asking is as follows.

  • The latest reading (now being stored in a register) is always stored in the EEPROM device (or some other storage device)
  • You turn off the rig
  • When you turn on the rig and before the loop starts you read the EEPROM and if it is not "0" but some other value then a correction process takes over and the hardware is physically placed in the "0" state. This is not a LO but physical hardware that must be changed much like a potentiometer that must always be dead center when you start the loop. 
  • My question involves just the EEPROM and how to store the value of a register, how to read that register in the setup and then make a correction. The rest I know how to do as it is now being done manually. I am trying to cure a my most common problem -- I forget to do it before I turn off the power. 

Now if some of the suggested hardware can do this then I am on the right path. My other alternative is to flash  the screen every 5 seconds. "Hey Dummy, Before Power Turn Off, Zero The Circuit". [Now That I know how to do!]

Pete N6QW

2/24/2019 ~ An Arduino Question

For those who have the Arduino skill set here is a question/task for you:

Before powering down the Arduino I want to read some value being stored and if that value is not "0", I want some action to occur which would cause that value being stored to become "0" upon reapplying power. It is important the code to do the actual zeroing process in the setup phase. This is like "zeroing" the meter before you take any measurements. I was thinking EEPROM but that is only good for maybe 100,000 cycles and if you constantly read and write to EEPROM that makes it only 50K times. So that is a consideration as well. The bottom line is that whatever is done once the loop starts, automatically it is starting at "0".

Pete N6QW

2/23/2019 ~ Capping off that magnificent Analog VFO.

Yes that is a bit of a play on words. However stability and lack of drift are not the only two factors of import in an Analog VFO. There is the nagging issue of how to set the frequency and do this on a repeatable basis. In the late 1930's enter "James Millen" a mechanical engineer of note who designed essentially an oversized micrometer dial mechanism mated to a gear box so that you could actually read small frequency change increments and do so on a repeatable basis. 

Shown below is the world famous HRO Dial Mechanism. Essentially the large dial was a Vernier dial mechanism that had numbers appear in windows cut out on the dial face and the movement of the dial was coupled to a gear box (Boston Gear Works) where a large gear ratio required many "cranks" of the main dial to achieve one full rotation out of the back end of the gear box. The variable capacitor connected to the gear box was the type so each degree of rotation resulted in a capacitance that when plotted resulted in a straight line of frequency change. Move it one increment and it was a 1 kHz chnage. I find it funny that at times I am being told I am not on frequency as my signal is 20 Hz low -- which today  is 1/50 of the resolution possible on the HRO dial. We have come a long way; but it was indeed advanced technology for the 1930's

I grabbed this photo off of a current eBay listing and you can have the above on a "buy it now" basis for a mere $275. Ouch. If any one has one of these either sell it for $$$ or you can donate it to the N6QW Laboratories where I will sell it for cash and buy more radio parts so I can keep making stuff so it can be posted on the blog and website. 

Bill, N2CQR has one of these very mechanisms on a homebrew receiver that he calls HRO-ish. In a recent podcast he mentioned that he was having repeatability problems with his Analog VFO coupled with this mechanical delight. It was not a stability or drift issue; but a problem that you could not come back to the same frequency for the same dial reading. The suggestion of many regarding the problem was that it was a 'mechanical problem". After all, the mechanism was long in the tooth perhaps nearing 80 years old. Bill has fixed the problem and that he will cover in our next Podcast (soon).

But I leave you with this tantalizing thought … what if there was another way to have a mechanically tuned Analog VFO that has high accuracy and repeatable frequency changes without resorting to one of those $13 SanJan frequency counters?

Pete N6QW

Many of the readers of this blog are aware that I participate in a podcast (almost monthly) along with the host and cruise director, Bill N2CQR. Our subject areas are wide ranging and frequently we travel a path where we use different approaches to arrive at the same end state.

One area in particular where there is great divergence is our methods of frequency generation for LO's and BFO's. Bill  typically takes the analog approach using coils and capacitors. With due credit to Bill's skill and patience he has been very successful in building highly stable analog VFO's. 

N2CQR as a student of Doug DeMaw's practices you often see the multiple capacitors in his tank circuits along with physical isolation of the coils (air wound no less) and tuning capacitor which are far removed from any heat generating circuits. The fact that he has many rigs using this method is a testament to his skills and understanding of the analog VFO topology. Analog VFO's can be used with today's rigs. So a tip of the cap to Bill.

But for Bill it is not just the idea of using the analog approach; but more of the technical understanding of each and every component in his VFO. He can see, in depth, the impact of component selection, physical isolation, rigidity of construction all with a measure of attention to detail. He will frequently say --you can stare at the Adafruit Si5351 all day long and learn nothing about how all of those bits and pieces work to generate an RF signal.

I on the other hand prefer the newer technologies and perhaps for a different set of reasons. I like building lots of stuff and using digital methods for frequency generation supports the rapid prototyping paradigm. In fact there is a bit of laziness as well. Many of the critical factors that affect the successful analog build are not so much of a concern with an Arduino controlling a Si5351. Frankly for me it comes down to a time factor --a really high quality analog VFO is not a simple output of spontaneous construction. I can build a Digital VFO in about an hour. For the average homebrewer the high quality analog VFO may take a whole day to fabricate. 

Bill and I are not in any sort of competition and that is where I think our divergent views makes for a better podcast in that the listener has an array of options at his disposal when undertaking a homebrew rig. 

In fact building an analog VFO is in itself a terrific learning tool as it addresses the electronic as well as mechanical aspects of rig fabrication. Build one, learn the details and then stand back and enjoy the wonders of those coils and caps.

I would appreciate hearing any comments on the subject of analog versus digital frequency generation.

Pete N6QW

Friday, February 8, 2019

2019 ~ Building SSB Transceivers

The Building of a SSB Transceiver ~ You can do it!

At Christmas my 3rd son gave me a transducer speaker that mounts to a piece of foam board and I sit here amazed at the sound reproduction. The amp driving the speaker is the 2N3904 and LM386. So it will provide more than just headphone volume (with a 100 Ohm resistor installed). I received a comment about the 100 Ohm resistor and just for fun substituted a 10 Ohm in the LT Spice simulation and saw no difference in the output curves. The short video tells much.

Pete, N6QW

2/21/2019 ~ Go Daddy Now Fixed! The audio Amp Schematic is on the website.

2/20/2019 ~ Problems with Go Daddy!!

For over 24 hours I have been trying to upload updated schematics and information to my n6qw website. It seems like Go Daddy who hosts my website has some sort of issue with a windows server. One solution that offered was for me to spend more money and move over to a Linux server OR just wait until they get around to fixing the problem. Not only is it a cost issue but also I would have to start from scratch and rebuild all five of my websites. This makes about as much sense as declaring an emergency so the emperor can have his wall. 

My alternative is to wait a bit longer to see if they can resolve the problem or just post the info here. Here is one of the updated schematics. Perhaps if any of you have a Go Daddy account you may have the same problem.

Pete N6QW

2/18/2019 ~ Toroids are not self shielding!

I noted that after installing the shield on the left hand side there appeared to be a detuning of the band pass filter. Despite what many have said toroids are not necessarily self shielding. 

When I installed the side piece, I thought there was a sufficient space gap between the steel based side panel and the T-68-2 cores. There was not and so the BPF was being detuned. There just was no way with the side panel in place to re-peak the BPF to the proper value unless I followed the practice espoused by Ian Roussel (Full Custom Garage on the Motor Trend Channel) which is make a hole.. The "A" change was to cut about a 1 and 1/8 inch round hole in the side panel stiffener right over the BPF. This accomplished two things. Firstly it reduced the effects of the steel on the BPF and secondly to provide access to the trimmer caps so that the BPF can be tuned with the steel side piece in place. Worked perfectly -- the hole does not affect the stiff properties of the side bracket. We are good to go. The 2nd photo below shows the side piece without the access hole. You have got to love when a plan comes together!

Pete N6QW

2/17/2019 ~ Be sure to visit where the project is documented  and new schematics have been added over the past two days. 

Learning to Hem ~ No we are not sewing!!!

In keeping with trying to use stuff I already have knocking about the shack we now address a common problem with homebrew construction: making or buying a case enclosure. The problem is compounded by the fact the front panel is  from the former 30M CW transceiver project that was a QRP Quarterly article. But the back panel came from a transceiver that was rebuilt at least three times. 

Now for the disconnect. The front panel is 4 inches high and the back panel a mere two inches. Standing alone the front panel flexes as you spin the encoder --not a good plan. Thus to stiffen things we need to connect the front panel to the back panel to add rigidity. We essentially have a trapezoid and how to fabricate those two pieces, one for each side, and not spend a fortune on material.

One of my currently favorite TV shows is the Velocity Channel (now Motor Trend) and my new hero is Ian Roussel of Full Custom Garage. This guy uses what is at hand and often comes up with very creative solutions. In one episode he needed a round metal structure to house the odometer that was being affixed to a new dashboard. He spotted his dog's water bowl which it turned out was stainless steel and just the right size. The poor dog now drinks out of a plastic dish. But it is that sort of thinking we need to employ when we fabricate our homebrew rigs.

A trip to Home Depot yielded a piece of galvanized metal flashing that was about 8 X 11 inches. The price was $1.14 and was enough material to make two sides. But wait you are thinking that stuff is awful flimsy. True; but if you bend it over along one of the seams --it becomes rigid. A pair of Tin Snips, a metal straight edge with scribe and a bench vise is all that is needed to make these very durable side braces. The galvanized metal can be soldered as well as painted. Thus the bending over and doubling is known in the fabrication world as "hemming".

The first item on the list should be a manila folder that will serve as a pattern to overlay on the sheet metal and provides a template to scribe the outline. Keep in mind you have right and left hand pieces and thus you must flip the pattern over to draw the scribe lines for the different sides. The 1st photo below shows the template of the sides and you can see the bend line for the hemming. On either end add 1/2 inch of material that is bent over and will provide the anchor points to the front and back panel.

You might have to "play a bit" with the metal bending so get two sheets of the material. The finished rig is a far cry from the boards screwed down to work bench as it was initially built. Nice looking Rig, Pete!

Ian Roussel would be proud.

Pete N6QW



[2/11/2019 I have started publishing schematics for the Simple SSB on my website at Although it is doubtful that many (if any) would take up the fabricating of this rig, there just may be one or two who have an interest.] 

Just the other day I was having a very nice QSO with Gene W6QFU and we exchanged our station line up. W6QFU is lucky in that he just recently acquired a very nice ICOM 7300 and has joined the ranks of perhaps over 100,000 other hams world wide who now have that rig sitting at the operating position. Some would suggest it is an appliance rig; but at the same time it is a great sounding rig that undoubtedly is very popular and the current price point puts it in reach of many hams. Guess there is a link to price point and the numbers in shacks around the world.

On my end --yes, yet another new rig just finished and W6QFU was the third contact on that rig. So I was anxious to see how it held up. Well I might add...

ARRL as previously reported will start a new podcast one month from now called "So What Now?" A possible answer to that question (or perhaps it is a challenge)  is to suggest you build your own SSB transceiver. Trust me what I am suggesting you build is not in the ICOM 7300 class; BUT it puts out a respectable signal with a rock solid frequency stability and it is something you built and can be built for about $100. [YMMV depending on how much junk you have in the junk box.]

But let us go back in time to review SSB transceivers starting in the late 1950's and early 1960's. BOOM there was an explosion of technology in that time frame starting with the Collins KWM-1 and KWM-2 (the cost --read a years salary for most middle class working folk) to more modest rigs that were less costly. Heathkit did much for the hobby in offering single band transceivers that could be had for about $100. I remember being on Midway island (actually two island Sand and Eastern Island) in 1963 where a fellow Navy ham on Eastern ran one of those rigs into a 13 dB gain Rhombic antenna pointed at the US. Everyone wanted to talk to him--firstly he was DX and secondly that gain antenna took that 100 watts and made it look like more than a KW (like 2KW and you can check the math [10*log(2000/100) = 13dB]) But even the simple Heathkit rigs involved some pretty complex circuitry to receive and transmit a signal using common circuit elements. 

There were many complicating factors in building a homebrew SSB rig in the 1950's and 1960's

  • Socketry --yes you had to do all of that metal bashing and mechanical engineering to layout all of the tubes and components that had to be affixed to a large aluminum chassis
  • Power Supplies. You often required Filament, Bias, LV, HV and control  power to operate the rig. We are indeed fortunate --12 VDC and we are there today
  • Front panel controls abound as does a space reservation for the Analog Dial Readout
So there was much involved aside from just a few schematics.

In todays rigs none of those issues exist as many circuit elements simply solder to a few pads. Socketry is minimized so you can feel secure in selling your set of Greenlee Chassis Punches at the next block yard sale. You won't need them. A 12 VDC well regulated supply or even batteries can run your rig. You only need to cut out a rectangular hole in the front panel for your favorite display (LCD or Color TFT). If it is a touch screen then that will reduce further the number of controls needed

Fast forward to today some 60 years later and building a transceiver is much easier and uses far less components thanks to IC's, PLL's and inexpensive commercial crystal filters. Let me take a moment to discourage the use of homebrew crystal filters especially if you have not previously built one, and two if you lack some pretty decent test gear. Sure you can purchase crystals inexpensively but that is only 5% of the task. So reject all of those claims that you only have to buy 5 crystals at 40 cents a piece and you are there --you aren't! INRAD sells a nice 4 pole 9 MHz experimenter's filter for about $30 and you can also buy 9 MHz filters from the GQRP club for about the same shipped to you. Yes I do have several rigs with homebrew crystal filters so I do speak with experience. But it is a simple decision if you do not have that experience --buy the filter!

So lets talk a minute about the SSB transceiver architecture. and  we'll proceed first to the  block diagram. 

Many rigs today are bilateral or as some prefer bi-directional in that a) signals can be routed in two directions through the circuit element and b) because of (a) you can use that circuit element in both transmit and receive.

Here is an example of one circuit element that I frequently use... a bidirectional IF Amplifier Module as shown above. This module has four transistors, a bunch of caps and resistors, two matching broadband transformers, a crystal filter and a relay to bias parts of the circuit to change the signal direction. The common 2N3904 and 2N3906 make up one amplifier stage and following the crystal is a second identical stage. The amplifier circuit was invented by Plessy and appears in EMRFD (just so it has street creds). Here is what is happening .. with one bias condition the two 2N3904's take on the amplifier role. With the alternate bias condition the 2N3906's are doing the heavy lifting. This circuit is good up to about the 6 Meter band and a single amp pair is good for 15 to 17 dB of gain. 

Here are some notable items. This is a broad band circuit --there are no IF transformers and there is nothing to peak or tweak. The resistor compliment for a single amp stage is 22 Ohms, 2 X 100 Ohms,. 2X 680 and a 1K. For caps -- six each 100 Nano Farad. The power source is actually 6 VDC but I run mine on 8 VDC through a three terminal regulator. Add in two FT-37-43 matching transformers [19 Turns solenoid wound tapped a 6 turns. The 19 turns squared is 361 and the 6 turns squared is 36 --thus 361/36 = 10. So this transformer matches the 50 Ohms of the Plessy to the 500 ohms on the GQRP filter.]

Including the Linear Amp control ( another 2N3904) there are eleven transistors and FET's along with two packaged ADE-1 Double Balanced Mixer's an Arduino Uno R3 (because it was in the junk box) a 16X4 Seasick Green LCD (also a junk box refugee). BTW I was looking at some 16x4 Blue White LCD's --$8 shipped from the USA. Lest I forget the GQRP 9 MHz crystal filter. If you use the INRAD it is a 4:1 match as the INRAD's are 200 Ohms.

This may be a good place to stop for now… but to whet your whistle:

Interesting note this front panel was cut out of the front panel that was used for the 30 Meter CW transceiver that originally was a QRP Quarterly article. I should tell you that I was encouraged by QRP Quarterly to design and build this CW rig so it would appeal to those QRPer's who dominate CW. I did as was suggested and did make one contact to prove it works . The rig then went into the never to use again and possible use in other projects bin. There it has sat for five years. Now is the time to use it!

The panel was reversed as it was double sided PCB. The Key is now the Mic jack and where the analog dial was --it was cut out for the LCD. The audio amp stage was cut out of the original main board. There are three empty holes in the new panel. Above the red knob the audio gain control will be installed and the hole in the lower right hand corner will house the audio output jack. and the small hole next to the mic jack will be filled with a bolt and nut. The original panel was 8 inches long and with a bit of juggling and cutting material off of each end the size is now 6 inches wide. 

Thank you Velocity Channel for giving me some really great ideas! Yes Virginia having a CNC sure makes life a lot easier when you are fabricating radios. The part that was cut off included the former part containing the volume control and headphone jack. The two bolt holes nearest those were enlarged to house one of the toggle switches and the new headphone jack. 

New panel controls include MOX and TUNE button (red button), Volume control, Main Tuning, USB / LSB Select and VFOA / VFOB Select, Microphone Jack and Headphone Jack. Lest I forget the 16X4 LCD.

Below is the rig screwed down to the top of the work bench. In the very bottom is the IF module, Audio amp and Microphone amp. To the left of the IF module is the BPF and above the IF the bidirectional J310 amp and the single 2N2219 feeding the IRF510. To the right of the IRF510 and the microphone amp is the LPF. The junk in the center is my power buss and some relays used for TR. The RCA plug/cord above the J310's is the linear amp control. The "T" line feeds a 1K resistor into the Base of a 2N3904 with the Emitter grounded at the collector goes to the control line in the follow on linear amp. 

Take a good look --aside from looking like crap --there are not many parts. I did use it on WSPR and FT8 as well as a few SSB contacts. the future work will involve some packaging to sanitize the look. 

It is a shame that many hams think building a rig is too difficult. Drop me a comment if you would like to see more detail on this project. Oh should mention portions of the design are LT Spice simulations so there is solid science behind the modules. These include the BPF, LPF, the J310 Bi-Directional amp, the 2N2219 driver stage and the microphone amp circuit. Oh --the mic amp design is now the same design used in the audio pre-amp stage. The circuits not simulated are the Plessy amps (from EMRFD) and audio final stage LM386 and the IRF510. 

I was asked a question about how the Simple Transceiver compared to the Sudden Transceiver Project. There are two areas of comparison the first of which is performance and the second the circuit elements themselves. Performance is equal and thus I am pleased; but the circuit elements do differ and the following explains the differences.

There is some circuits which are alike and some not so. Physically the Sudden is a smaller footprint. Some circuit differences include:

  • The IF Module uses the Plessy amps and external DBM’s. The Sudden using the NE602 provides the mixer stage and the balanced modulator on the input side and the mixer and product detector on the output side. The LO & BFO are switched whereas in the  Simple SSB rig they feed the same DBM regardless of T or R. The Sudden is a single pass whereas the Simple is bidirectional.
  • The simple SSB takes the pre-driver used in the Sudden and makes it bi-directional so it is the Rx RF Amp on receive and Tx pre-driver on transmit. The Sudden has a separate Rx RF Amplifier stage
  • Only one BPF versus two in the Sudden
  • The Driver and Linear Amp are the same
  • The Mic Amp is the same but the Audio Amp different
  • The Sudden uses a Color TFT and Nano and the Simple the Uno R3 and the 16X4 LCD
  • The LPF’s are the same.

Bottom line ~ not a lot of parts! For those who wonder this is not a Bitx40 repackaged. The Dual J310bi-directional  stage configured as a Dual Gate MOSFET has some interesting possibilities. The stage has a manual gain control pot built into the circuit board so that stage gain can be adjusted. This opens up some potential for adding AGC and ALC to this stage so now we can add more additional refinements. One comment I made arising from the Sudden Transceiver fabrication  was the opportunity for experimentation. With a change in the BPF and LPF the Simple SSB Transceiver can traverse to other bands. A simple DPDT toggle switch on the front panel could with 1/2 the switch  controlling the VFO range and the second half controlling 4 SPDT relays would switch in the proper BPF's and LPF's for each band. One switch, four relays and a few more toroids and caps and you are on two bands. 

Pete N6QW

So How To Do CW on a Homebrew SSB Rig?

3/16/2019 ~ Remote Antenna Tuning 3/14/2019 ~ Calling All Junk Boxes???? Calling all Junk Boxes, N6QW is looking for a part! ...