Monday, August 14, 2017

A New Line of Transceivers ~ DifX

A New Definition of HAM!


HAM = Homebrew Avoidance Methodologies.


This posting undoubtedly will create a stir in the greater ham community and may give pause that  my advanced age is now affecting my mental faculties. What I am about to post is based on over 60 years of homebrew experience dating back to the early 1950's when I first dabbled with the CK722 in solid state circuits. Or perhaps there will be questions from some quarters is 60 years experience, one year of experience 60 times over? Or is it truly a progressive building of an experience base spanning 60 years that started with vacuum tubes and now involves amazing digital electronic building blocks packed into boards the size of postage stamps?
Many would think given my lengthy time at the bench, would I resist moving from the buggy whip to the new fangled automobile. What I will be sharing, while solely my view, is fully supported with the "why" I am taking the various positions.
A week doesn't go by (or maybe just a few days) where I get an email requesting help or asking questions by those who are new to homebrewing or maybe from homebrewer's not so experienced at soldering their fingers together.

I am aiming this post at a specific audience and as mentioned, to those newer to homebrewing and the boiled down "why" is that what I am describing of what not to do, makes it an uneven playing field against having early on successes. These HAM Tips are being provided perhaps as a substitute approach, in essence, saying you will save yourself a lot of grief by not undertaking certain methods to homebrewing a project. The journey as the Reverend Dobbs of GQRP SPRAT states is the important part --but the starting place is not the top of Mt. Everest!

Homebrew Avoidance Methodologies (HAM)

Since we now live in the "age of impatience" below are the avoidance subjects I will be covering. This is provided now, so you won't have to read any further and thus you can simply move on and revel in the view that my mind is totally gone.
  1. Why not to build and use analog VFO's?
  2. Why not to build and use the discrete (2N3904/2N3906) audio amplifier?
  3. Why not to use the "prototype" bread boards?
  4. Why not to build and use the Dishal Filters?
  5. Why not to build "that really big project" without having built something small initially?

The Dreaded Analog VFO!


So OK for those who stuck with me this far let us take up subject Number 1, with me, and that is the non-use of Analog VFO's. Here come the arguments that with an analog VFO you have a link to the past and that there is so much satisfaction in finally getting one to exhale sine waves. Yes, there is the nostalgia trip and the idea of a connection with the past in how things used to be done. But for today's state of the art rigs they are barely "good enough".
It has even been suggested that I never have built an analog VFO or that I have never really built a VFO that was rock solid, didn't have warm up drift and didn't drift at all.  To those who think that, shown below is an example of an Analog VFO built by me. This did not happen in a span of two hours; but I started first by looking at all of the aspects that need to be covered to have a really solid VFO. Yes, I consulted DeMaw and Hayward and Hantchett (for those who don't know -- he was the guy from RCA, and his call sign W2YM. He built one of the first DGM VFO's). I read and reread the ARRL info on the mechanical side of building solid VFO's and even selected a VFO type that would afford the very best frequency generation characteristics as to stability and even considered the phase noise that incidentally could be generated. There was real engineering behind this project!
The VFO topology selected was the Hartley and the tank coil was wound on a ceramic form without any ferrite core, which has been installed in a captive cradle assembly. This was done so that the tank coil was solidly held but floating in air away from the any metal. Liberal use of coil dope [Gorilla Glue] was applied so that there was no movement of the windings. The enclosure was made of double sided PC Board and the method of construction had a metal cover that was solidly affixed to each corner as well as a center post. Thus you could flex the top cover and there was NO frequency change!!!! That said the bottom actual circuit board is single sided copper board so that there is no unwanted coupling through the small capacitance that would exist in the board given that you have two plates separated by a dielectric.
N6QW Homebrew Analog VFO
Hartley Oscillator


Multi-capacitors in parallel (NPO/COG and NPO Trimmers) were used to set the base frequency range of the VFO and a double bearing type variable capacitor was used for the incremental tuning. The main tuning cap was external to the box. [The black wire in the lower corner of the box connects to the variable capacitor.]  The multi-capacitor approach is to limit the circulating currents through any one capacitor so that there was no "heating" of individual capacitors which is a cause of drift.

To avoid loading of the VFO circuit I used a buffer/booster amp on the output  of the oscillator transistor and there were multiple outputs so that the main output drives a mixer circuit and one of the auxiliary outputs could furnish output to the frequency readout LCD. Still another output could feed a huff and puff stabilizer. The output stage was/is adjustable so that the appropriate drive level is provided to the mixer stage.

Voltage regulation is internal to the box for a reason. There would be some heat generated with the regulator and when the box is sealed up, in effect, you have a temperature controlled oven with a large external thermal mass (the enclosure). One approach for super critical applications is to have the VFO "hot" at all times even when the main rig is powered down. This eliminates the "turn on" drift.
Did this VFO work and what was the performance? The VFO was set up to tune several hundred kHz around 5 MHz for use with a 20 Meter transceiver and 9 MHz IF. It did work and the drift (both turn on and long term) were typical for a 1960's/70's well built VFO. But here are the issues concerning this approach:
  • It took several days to fabricate the parts and build the unit (no rapid prototyping here). McDonald's plastic straw coil forms are cute; but no where near what is needed for high end stability such as the process that was used. My tank coil took nearly 1/2 day to build!
  • The tuning range was limited to several hundred kHz around 5 MHz. Frequency changing say to a 16 MHz VFO for use on 40 Meters would have been tenuous as I doubt the  level of frequency stability at 16 MHz would not be the same as 5 MHz. Taking that to 30 MHz for 15 Meter operation would be out of the question. The alternative is to have dual conversion and leave the VFO on 5 MHz. Now you have added an order of magnitude of complexity.
  • Readout was a problem unless you employed some sort of counter that had offset functionality and if you went this far, then you are already 2/3 the way to a digital VFO. Mechanical Analog dials may get you 5 kHz or at best 1 kHz resolution
  • While I used an Oren Elliot Products reduction drive on the double bearing variable capacitor (6:1) that still limited tuning in precise steps such as you would have digitally in going down to 10 Hz or 100 Hz.
  • The physical size was almost larger than the rest of the rig components. So this type of VFO is not suited for a compact designs.
  • The drift problems even though slight were still apparent.
An Arduino based digital VFO using the Si5351 can be built and functioning in two hours time and none of the problems described above are issues. The total cost is around $20 and I get a color display, wide frequency range coverage, selectable step rate and ultra super stable frequency generation of both the LO and BFO. My upper frequency limit is 200 MHz --try that with your LC VFO! The actual construction methodology can be less stringent which is another bonus. The physical size lends itself well to compact construction techniques.
Building a stable analog VFO only comes with experience, a well stocked junk box  and having test and measuring tools to peak and tweak the VFO for maximum performance. A digital VFO on the other hand using the Color TFT display has built in test features and functionality. This is a critical point -- a person new to homebrewing an analog VFO may be stuck when things are not going well. Whereas the digital VFO has less issues with parts, drift, setting ranges, circulating currents and on and on. So just move on and learn how to use the Arduino and Si5351. If I can do it so can you!

The Case of the Wimpy Audio Amp!


Shown above is the 2N3904/2N3906 complimentary pair audio amplifier as originally built for the Let's Build Something Project. This amp sucked in so far has having a room filling output. It was later modified to use the TIP31C/TIP32C combo in direct replacement of the 2N3904/2N3906. This dramatically changed the output level. The same arguments for not building this amp applies as with the Analog VFO. Yes it is retro cool and gives one a feeling of being connected to the soul of the rig BUT great care must be exercised in the build and having little experience building discrete audio amplifiers ties one hand behind your back.
This was the one circuit board aside from builders not following the three videos on how to build Double Balanced Mixers that caused the most heartburn in the project. Typically the builders ignored the layout suggested and just went off on their own and ended up with an amp suffering from squealing feedback. This amp can be a high gain circuit and thus circuit layout was critical. More often than not the haphazard builds had lead lengths that were too long. This puny amp also consumed too many parts. Being able to say you built it and it involved no IC's is not a substitute for a squealing amp with marginal performance.
A really major shortcoming is the output connections. In the design we used (from the Internet) one of the output connections is taken off of the 12VDC rail. This means you must use an isolated output connector or you risk shorting the 12 VDC rail to ground -- that will smoke things nicely. It is a bad design!!!!! There are other designs that are similar but avoid the 12 VDC rail as one of the output connections.
One might ask why we would use such a design and the answer goes back to the premise of the Let's Build Something project -- common parts, no (or very minimal) IC's and homebrew to the max.
My current favorite amp is the 2N3904 driving either an LM386 or LM380. There will be the hue and cry that the LM386 is too noisy --so is the 2N3904/2N3906 so that doesn't hold water. There are other IC's that will give around 1/2 watt out and there are even higher power complete amplifier boards  that can be had for a few bucks on eBay. Now we have room filling volume, no solder burns, no squeals and a much smaller footprint so important if you want a compact rig. The photo below shows such an amp from eBay with a single supply voltage of anywhere between 5 and 12 VDC and good for 18 Watts. The price is $1.99 and free shipping. Why would you want to build a tepid 2N3904/2N3906 audio amp that most likely will be noisy and subject to feedback? It is hard to make a case to build the wimpy amp.

Just Say No To Prototype Boards!


Wow N6QW how can you be bashing prototype boards as shown below, when the very use of the boards suggests that their utility is in fact the ability to hook up a circuit without using any solder and have something working in minutes versus hours? C'mon Pete what is up?
Recently I received one of those emails that "your audio amp circuit doesn't work". As it turns out the amp was built on one these very same prototype boards. I then was asked to provide my voltage readings at various locations on my circuit.

I guess my first thought is "why am I doing this"; but in the true ham spirit I did. The reason I say this is that the homebrewer should have done this as a 1st step. As it turns out there was "0" volts on his collector of the 2N3904 and thus the reason for my circuit not working for him!!!!!!! 
There is fault with the homebrewer for not having a disciplined troubleshooting process to know what should be the voltages at various places in the circuit and the second problem is the fault of the prototype board in that it is easy to overlook wiring that must be supplied. The two rows along the right and left sides are not continuous but yet are intended to be the voltage rails. You have to "jumper" each grouping to have a continuity --an easy rookie mistake (or is it misteak). 
But more importantly here is my POX list for why not to use prototype boards:
  1. While you have lots of holes at times it is not easy to make all of the connections.
  2. It is hard to flow a circuit using the boards (ie lots of cross overs and opportunities for mistakes).
  3. With repeated use, the contacts become flaky and that leads to intermittent circuits which are difficult to identify and you are left wondering what is the problem.
  4. These boards do not provide a bottom ground plane so important for RF circuits. The boards are probably OK for making NE555 Flasher circuits but avoid them for RF circuits. So now I will get the hate mail that Mike, AA1TJ develops all of his amazing RF circuits using these boards. To that I would answer --Mike knows what he is doing and is well experienced. Most new to homebrewing do not have AA1TJ's skill set
  5. In use for audio circuits the cross over mentioned in item 2 provides ripe conditions for hum and feedback.
  6. Access points for voltage measures are not readily built in -- hard to fit a probe in the little holes.
In further response it is hard to beat Manhattan construction on a piece of single sided copper PC Board. You can flow the circuit, have an excellent ground plane, make connections short and direct and there is no fuzz that you have NOT connected voltage to the collector of the 2N3904. A ground plane is mandatory for RF circuits and the solid ground plane coupled with short direct connections mitigates against hum and feedback in audio circuits. So steel up your courage and throw out the prototype boards or give them to your grandkids with an NE555 flasher circuit. Unfortunately the grandkids will only be amused for about 30 seconds if that long.

The Dishappointing Dishal!


Several recent posts by me reflect my disfavor with the Dishal Crystal filter. What I got from the software resulted in an unsatisfactory filter. There just is no other way to sugar coat it. I actually built two Dishal Filters with one being a 6 pole and the second being a 4 pole. The six pole was a total bust and the plotted response curve looked nothing like the predicted curve.

The second filter was a four pole configuration and that seemed to be more successful --at least initially. The formal computer generated plot was based on the data that was supplied as input to the program. So if there was a problem with the input, the plot would have reflected that discontinuity. The plot from the "as built" filter matches the computer predicted plot in terms of general response shape for the four pole filter. That is significant in that had I made an error with the capacitor values or the matching network or in wiring the filter or in the test methodology the proof would have been in the plot I made. The plots matched!
So what conclusion can you draw? Is there a problem with the software since my evaluation is that the filter was actually too wide and the opposite sideband was slipping through? The capacitors were extremely close to those specified -- so what happened? I used great care in building the filter and the capacitors were premium quality high accuracy and less than 0.5 PF different than the specified values. I am left with a conclusion that there may be an error in the software that gives values for the caps that are wrong for the bandwidth specified.
I received some sharp criticism that I failed as an engineer to root out why my filter while looking like a Dishal was not a good filter. Frankly you cannot repeat the same exercise with nothing changed and expect to get a different result. Stay away from the Dishal software!
There are other filter design programs including the one from AADE and one in the EMRFD DVD. Frankly I have had better results using four highly matched crystals measured in a loaded state, using five 68 PF NPO couplings caps (for SSB) and in/out matching of 150 Ohms. The opposite sideband suppression is far better than the Dishal and the signal reports indicate it does the job on transmit.

The Art of a Non-Fake Deal: Start Small


I am amazed at how many homebrewer's tackle a very complex project as their first foray into the world of "rolling your own".  The fact that there was success may be attributed to pure luck but as often happens you end up with a non-working pile of parts. In a prior post I recommended starting with a project like the Michigan Mighty Mite which has only 7 parts --8 if you count the "one capacitor key click filter." What is amazing is that many new homebrewer's were not even successful with that project. If you can't do it with 7 parts what makes you think your odds would improve with 500 parts?
But with only 7 parts you have a chance to develop some troubleshooting skills and things to look for when the completed circuit simply does not work. I think back to my earlier discussion about supplying the schematic with voltage measurements for the audio amplifier circuit. As a homebrewer the 1st thing you do even before applying power is to double check the wiring. Then if "no workie" take some voltage readings -- very likely neither was done by the person who emailed me.
You have this large pile of parts (that you paid $125 for) and nothing works --where do you start? This is where having a disciplined troubleshooting process can pay big dividends. But such a process is best developed in building your confidence by starting small -- and learning the how to do it.
Hope this was helpful -- keep in mind: separate the nostalgia from working circuits that are most easily implemented.
Pete N6QW

Thursday, August 3, 2017

A New Line of Transceivers ~ DifX

So What Do You Want To Build?

8/7/2017 Mikele, 9A3XZ has provided this video of two rigs he built and enables you to listen to a  Bitx40 and a DifX side by side.
Pete N6QW


8/5/2017 I mentioned a new project in the text of this posting. Shown below  is a before and after photo of V.1 of the 20 Meter Shirt Pocket Transceiver.
Updates include:
  1. Full 20 Meter Band Coverage versus only 100 kHz in the original VXO version
  2. USB/LSB front panel selectable
  3. Si5351 and Arduino Nano versus the Crystal Switched Heterodyne VXO
  4. 1/2 inch high OLED versus crude Analog circular dial
  5. Tune Function using a 988 Hz Pulsed Tone (you Bitx guys need one of these)
  6. MOX switch
  7. Homebrew 4.9152 MHs four pole filter --this really beats the Dishal or QER filters. Nice flat response with superb opposite sideband rejection
  8. Smaller than the rig shown on the masthead.
  9. Pout about 2 watts.
Pete N6QW
2017 Updated Version of the 20 Meter Shirt Pocket Transceiver
Original Version I of the Shirt Pocket Transceiver


Compare the size of the new rig to the D-104 Microphone
Typically a week doesn't go by where  I receive emails inquiring about projects that I built several (or many) years ago. Much of this traffic is a  result of having posted about 125 You Tube videos and the subject matter of these video will spark a bit of interest which is followed by an email.
There was even a video I made on how to keep your feet dry while driving your Jeep Wrangler which suffers from a water intrusion problem. That one video had many hits (like 31,000) and even a few inquiries if I would manufacture the device that fixed the problem.
But fundamentally I think there is a desire today to return to our early ham radio roots where if you wanted to get on the air --you were faced with two problems: 1) there were no credit cards (or paypal) and 2) ICOM/Yaseu/Kenwood were yet to be in business. Thus to get on the air --you had to entirely build your own station -- everything was "homebrew". There is just something about the buzz you get when you can say " the rig on this end is homebrew". Which is often followed by "gee it sounds good for a homebrew radio!" Arrgh##$$$$!!!
Just yesterday (8/2)  I received an inquiry about the Let's Build Something (LBS) project that appeared in QRP Quarterly several years ago. Basically I was asked to create a special sub-channel on You Tube that collected all of the videos involving the LBS (to make easier for this person to look at all of the videos) and to provide a detailed parts list of all the parts need to build the project. I will do neither of these. If you want to build this project then you have to do some of the work! In fact the one major design criteria for the project was the use of all common parts that were readily available in most junk boxes. One fundamental axiom -- if you want to homebrew then you need a junk box!
The best piece of advice I can give if you want to homebrew a rig is to spend $59 and buy the Bitx40 kit from VU2ESE. The $59 gets you a nearly complete rig including being tested and the shipping to your QTH. Your homebrew contribution involves slapping it in a case and making any mods that are available on the Bitx reflector. For $59, you get all the parts, you get a tested board and you even get a digital VFO with LCD readout. Search on for a link to his site. If you spend another $10 for expedited shipping you can have the kit in less than 3 days. My website has a tutorial on how to actually implement the Bitx40 project.
There is a bit of sound advice in this approach as taking on one of my projects for most hams new to home brewing, you simply do not have enough solder burns on your hands. You'll get there in time but there are many fundamental pitfalls to homebrewing a rig such as appears on the masthead. I didn't learn how to do that last week. My very first solid state transceiver was built in the early 1970's.
There was a lot of luck in getting that 1970's rig to work. Now today there is less luck involved as evidenced by the second version built just a few months ago. I also spend a lot of time learning about the techniques, processes and technology. That in itself is not a guarantee that everything will work the first time; but at least it gives a leg up on having no grounding in what it takes to build a rig. The very first thing I don't do is heat up the iron and start tack soldering  parts together. I also have learned to be skeptical about claimed results. My recent experience with the Dishal Filter bears testimony to that view.
I have a new transceiver on the work bench and this one is not so much new technology but more in the line of repackaging circuits to make a small rig even smaller. For those who have followed some of those 125 You Tube Videos, several years ago I built a 20 Meter Shirt Pocket SSB Transceiver. This new project takes aim at that former project; but with added capabilities in about the same size --yes the 1/2 size OLED display and the Si5351 are part of the mix.
Going back to yesterday's LBS email -- that project featured a building block approach where you started first with a direct conversion receiver (DCR) where almost all that was built initially was reused in the upgraded SSB transceiver. The simple idea --get the direct conversion receiver working and then advance to the more difficult sideband transceiver. By the way if you go to there is a link to the LBS project with all of the detail on how to build one.
Many new homebrewer's who undertook the LBS were unsuccessful, I might add, as they simply jumped right into the transceiver part thus bypassing the DCR. When things didn't work they emailed me with "your project sucks" or your design doesn't work. Ultimately those persons finally discovered that they had a wiring error, bad solder joints or failed to apply power to all parts of the circuit. These are rookie mistakes that are only overcome with practice, experience and starting small! It was important to build the DCR first!
Bill, N2CQR and I in our monthly podcasts (Bill, especially) have advocated building the Michigan Mighty Mite CW transmitter --about 8 parts total -- as a good place to start the home brew Odyssey. Yes even with 8 parts there were many learning failures as evidenced by emails that Bill and I received. But that project is a much better place to start than having limited experience or knowledge and undertaking a 500 part project and then trying to figure out the problem/issue why the board is inoperative.
But it all comes around to having a solid grounding in the basics before you undertake the complex. At the risk of being politically incorrect, those at the seat of our government should well understand this concept and tenet as applied to governance --and that is not Fake News! This comes back to the Bitx40 -- start there and learn all you can. Then and only then should you undertake an entirely scratch built rig.
Pete N6QW