Sunday, April 14, 2024

April 14, 2024. It is all about FOAM.

In our posting about The Four Body Problem, I cited the four major elements of a SSB Transceiver. These were identified in alphabetical order. Thinking about it and as a memory aid I rearranged the letters to read FOAM. Now that may be easier to remember like FOAM on a Beer!

But today's post is connected to FOAM and raises the question as to why most of the radio circuits we see use NPN transistors to the almost exclusion of any PNP devices. A notable exception was the SBE-33 and 34 SSB transceivers where most of the circuitry was not only PNP but also Germanium devices.

I have no good answer as to why we are not using PNP devices? If you look at the PNP datasheets for devices like the 2N2905 or the 2N2907 they certainly pass the smell test as to Ft (cutoff frequency) and Device Dissipation. 

So, with that in mind we turn firstly to simulation software (LT Spice) to see if there is something not seen. The circuit simulated below with a 1-volt input has 19dB of gain at 20M. I have even designed a circuit board for this Driver stage and that is shown too.





 Does the answer get us to something like well it seemed strange to ground the Collector or it always has been that way. I suspect both are part of the answer. But we have proven the circuit in simulation and now is the time to build the hardware. 

A pictorial of the parts layout is provided for your dining and dancing pleasure. Gulp, you could even do this with Manhattan squares and the awful glue down process.

A good test of the use of PNP transistors in a homebrew SSB design would be to convert the PSSST-20 (N6QW website) design to PNP devices. In place of the 2N2222A, I would use the 2N2907 (TO-18). Here is a simulation of the IF Module converted to the 2N2907 PNP device. The 2N2907 has an Ft = 200MHz, a device dissipation of 400 milliwatts and can do a Collector current of 600 milliamps. Thusly, for the 2N2222A we use the 2N2907 and for the 2N2219A we use the 2N2905. What I have not explored is a P Type MOSFET to replace the IRF510. 
 


 
 
With an IF of 9MHz, we have 42 dB of gain. Considering a loss of about 6dB in the filter this module is good for 36dB as a gain block. I am super encouraged.
 
I have seen some devices on eBay that a bag of 10 Motorola units is going for $1 each. You only need five units for the conversion. 

If I get some time, I will tackle the conversion of the Mic/Audio Amplifier stages using the 2N2907.
 
 
 
 
This is the value of LT Spice -- design conversion. No Nano VNA's were used in this process and my soldering iron is off.

73's
Pete N6QW
(The home of innovation)

Friday, April 12, 2024

April 12, 2024. The Four Body Problem.

Currently on Netflix is a movie series called "The Three Body Problem". It is Sci-Fi at its best and scary as hell. I have only watched a couple of episodes but can share that I have had at least one nightmare. 

While a Sci Fi movie, it is a foretelling of the future. Yes, they laughed at DaVinci's helicopter design!

So that you don't have nightmares and are down to earth as you contemplate building a SSB transceiver, I want to introduce the Four Body Problem.

Friend N2CQR shared that his close friend and colleague VU2ESE, Farhan had identified four building blocks to any project. [AFMO]
  1. Amplifiers
  2. Filters
  3. Mixers
  4. Oscillators
Thus, The Four Body Problem is to identify and flesh out the sub-elements of AFMO.

Quite honestly, I had never thought about these broad classifications of what I call modules. But more aptly AFMO can be thought of as the LEGO building blocks of our hobby.



Once again, we see the sheer brilliance of VU2ESE! With his AFMO concept, Farhan, has given us a way to formulate an ordered approach to fabricating a homebrew SSB Transceiver. The LEGO analogy is very apt as in effect we snap together various circuit pieces to form a much larger circuit structure.

Our job now is to create a catalog of circuits within the AFMO scheme and then we have fill for the block diagram which is the initial step in creating a new design.

Amplifiers
For amplifiers we have two broad categories such as Audio and RF. Within these are subsets such as Microphone, Audio Pre-Amp and Audio Output. For the RF side we have Small Signal RF and RF Power Amplifiers.

Filters
Filters are represented by elements such as Band Pass, IF Crystal Band Pass Filters (or Digital Filters), Low Pass (or High Pass) Filters and lastly Power Rail filters. 

Mixers
Mixers essentially take two frequencies and create a series of frequency combinations. A Receiver or Transmit typically converts one frequency into the sum or difference of a supplied frequency. In a single conversion stage, we have an incoming signal "mixed down" to an IF frequency. A Balanced Modulator impresses an Audio signal on to a Carrier frequency. A Product Detector reverses that process to create an Audio Output. 

Oscillators
We have two major sub-sets to the Oscillator category to include Analog and Digital Generation. On the Analog side we have Beat Frequency Oscillators that can be either Crystal Derived of using an LC Network. For the Variable Frequency Oscillator, we have the LC or Varactor Tuned and using certain Crystal properties a Variable Crystal Oscillator or VXO. For the Digital generation of frequencies, the Si-570 or Si5351 driven by an Arduino or Raspberry Pi can do the BFO VFO task. In passing the Digital generation is a simple handshake with SDR homebrew transceivers.

A critical step is to learn the use of LT Spice where each of the aforementioned modules are documented and tweaked to perfection. If they work in the Spice simulation and assuming a circuit is built per the design--it should work. The value of LT Spice is the experimentation and evaluation capability without even turning on the soldering iron. 

Now the trap. Often you simply pull a proven module and install it in a rig. Likely most of the rig is done this way. So, the end result is just another warmed over Bitx with a different front panel. 

Thus, another part of The Four Body Problem is to constantly add to the AFMO catalog. One of my new adds will be the use of the 2N2905 (PNP version of the 2N2219A) as a Driver transistor. A bag arrived yesterday.

73's
Pete N6QW

Thursday, April 11, 2024

April 11, 2024. Your computer has done you in!

That terrible band noise you are hearing in your homebrew rig is not a CME (corona mass ejection) or EMP (electromagnetic pulse) from North Korea. No, it is your computer power supply.

I was hearing something in my SDR radios that also was heard in radios like my KWM-2. It was a constant noise that drove me nuts. Thoughts for the source drifted back to when the nice-looking divorced woman down the block was growing weed in her garage and the noise from the grow lights was S9.  But she had moved. 

My non-scientific research on the Mean Well power supply caused me to make this video and share that with you.

I have 3 computers in close proximity at one of the operating positions which are used for SDR work. Two are Windows 10 machines and one is a Linux Mint 20. Two of the machines are noise generators but the third has no RF Hash from its switching power supply.  This non-hash computer is a small form factor refurbished HP that was bought at Amazon for $75. 
 
The noise is loud and fills the spectrum with blips. So, if you are hearing noise go beyond just powering down the computer as it may not be off, and you will hear the noise. Either flip the power switch on the power supply to "OFF" or unplug it. 
 
Many computers have a Wake on LAN capability and even though the computer is powered down and appears off -- its power supply may be active awaiting a Wake on LAN signal.



A replacement power supply that is "clean" starts at about $100 which is a trade off at buying a refurbished computer at $75 that doesn't have the noisy switcher. 

73's
Pete N6QW


Wednesday, April 10, 2024

April 10.2024. T S Eliot once wrote April is the cruelest month. It might be about paying income taxes.

I received two comments about power inverter switching noise and my choice for the Mean Well DC to DC. [Five Days to April 15th.]

Energizing the Mean Well you do not hear any consistent hash noise in the SDR. But I did note after a short while you would hear a random noise for a short while and then it would move on. During one of those noise events, I unplugged the Mean Well power supply, and the noise went away. Later I had a chance to repeat that event and the same outcome. I can best describe it like a random oscillation. The switching frequency of the Mean Well is 150kHz.

The following You Tube documents the unscientific test.


 

So, likely the Mean Well would be objectionable for use "as is" without addressing the observed condition. What we have is a SOB (Sneaky Oscillation on Board). This may be a tough nut to crack! 

If it were a consistent and constant Hash Noise, then perhaps large choke values and large value capacitors could filter those from the DC output. The Mean Well is encapsulated and no way to access the innards to abate what appears to be a random oscillation that manifests itself as a noise burst. 

I have a bench supply good to 30VDC @ 5 amps. I ran a quick test and observed even more output from the IRF510 enough so to drive the SB 200 to 200 watts out. It was late and I did not have time to evaluate any Random Noise from this supply. 

But certainly, from my less than rigorous tests there is an argument to run the IRF510 with something larger than 12VDC. That said there is a significant technical argument that unless you had sufficient drive, increasing the voltage to 24 or 28VDC bought you nothing. 

My "unwashed tests" indicate there is sufficient drive from the 2N2219A present to give your signal a big boost. The RF Chain is the same as in my non-conventional PSSST-20 SSB Transceiver found at https://www.n6qw.com

There are no contests or operating tips involved --just some quickie homebrew testing.

Now for some TKT (Tribal Knowledge Tips). Realizing heat would be involved, the standard TO-220 small heat sink was not a good choice for testing the IRF510 at higher power. Thus, a homebrew fabrication using a U-Shaped thick web aluminum channel. That jewel is about 1 inch wide and 3 inches long. 

After running a series of tests at 30vdc I tested the efficacy of the homebrew heat sink by placing my thumb on the sink. While plenty hot I did not burn my thumb nor smoke the IRF510. Test Successful! Incidentally, I used a lot of ferrite chokes on the open HB SDR that snap on and that Ferrite core is on the bias line to the IRF510. 



Cross Section U Channel

Yes, you likely did two things: (1) you checked Amazon for the U Channel and (2) then crapped your pants at the cost. Try eBay.


73's
Pete N6QW

Tuesday, April 9, 2024

April 9, 2024. An Unscientific Study.

Fools rush in and at times get the best seats -- but not always. So, I tried something fool hardy and now I need to prove the results with the math stuff and electronic theory.

I have seen some IRF510 designs (ON6MU) where a single IRF510 can produce 20 watts out at HF. 

Consistent with my lack of time and as Larry the cable guy would say Get er Done, I did a quick and dirty test to see if I could squeeze a bit more power out of an IRF510 at 7 MHz.


I have a bunch of stock boards in the CNC files, and one is the IRF510. About 10 minutes start to finish I had a board. The two changes I made were to use 24VDC on the Drain and to provide a 10:1 step down match to the input of the IRF510. The 24VDC I know the why, but not sure why I thought using a 10:1 would be a smart move. 

Prior to installing this on the 40M SDR I used the signal generator as the source and delivered 14V PTP to the input side and with a variable DC supply set to 24VDC was the source voltage. 

The IRF510 bias was set wide open at about 4.2 VDC. I got 64Volts PTP across 50 OHMs. I did note that at 13 Volts DC that the pattern was not a pure sine wave but at 24VDC it was pure as the driven snow. 

For the installation in the SDR I used a Mean Well DC to DC inverter (Black Square) good for 600ma at 24VDC. The supply to the inverter can be anywhere from 9 to 18VDC.


The 10 watts is >2 times what I got straight into the Gate and using 12VDC for source. 

So, now we need to know what the real input impedance to the IRF510 is and to match that input and to run some rigorous, repeatable tests. This was just a place to start.

I am able to drive the SB-200 directly with this amp using the IRF510 to hit 150 watts on voice peaks. So, a choice of either 10 watts or 150 watts would be a good outcome.

BTW 500 milliwatts to 10000 milliwatts is a 13dB gain.

73's
Pete N6QW

Monday, April 8, 2024

April 8, 2024. More Pencil Tube Radios

When I got my 1st CK722 transistor in about 1954, finding applications was mostly limited to print publications like Popular Electronics or CQ Magazine and even QST. 

Today with the Internet, at 5AM while munching on my oatmeal it is amazing what can be found. Following the Ted Henry Little Argonaut cited yesterday, I searched for Pencil Tube Receivers, and this is what showed up.


 https://www.pa3esy.nl/zelfbouw/penciltube-rx/html/pencilM_set.html


A pencil tube radio from PA3ESY as found on the internet.



A Rig ready for SOTA or POTA ops.

This is not unlike finding some circuits for the CK722 and the itch to build something with pencil tubes.

73's
Pete N6QW

Sunday, April 7, 2024

April 7, 2024. HDR versus SDR

Ham radio always boils down to some sort of competition. Whether it is on a large scale such as Software Defined Radios or Hardware Defined Radios or on a smaller subset such as appliances boxes versus homebrew. Of late it is the stupid contest weekends ongoing for the last 4 weekends. So, OK for contests but not 4 weeks in a row!







Pencil Type Vacuum Tube

I only need to look back on my past experience when the competition was Tubes versus Solid State or the huge late 1950's competition AM versus SSB. The above pencil type vacuum tube was developed for hearing aids and proximity switches in munitions. A true sub-miniature tube. 

Several of these pencil type tubes found their way into a custom built 15M SSB Transmitter project sponsored by Ted Henry (W6UOU of Henry Radio). This transmitter, called the Little Argonaut, was shipped out to various Pacific islands and used to provide DX Expeditions to places sought after for DXCC. See the 1st Photo, CQ Magazine cover May 1960.

A Type 5902 good for 1 watt on 20M

These competitions never get fully or definitively resolved -- as a group we tend to cope and/or move on to the next controversy. Now it is which SDR Software is the best. Or FT-8 is not real ham radio! This also extends to the competition between Analog versus Digital VFO's.

I was looking at a video the other day of a radio I built in 1998/99 based on the W7ZOI famous 20M QRP SSB transceiver (Dec89, Jan90 QST). It had an analog VFO --not just any VFO but one well-constructed in a shielded metal box with 1/4-inch walls, with an Air Core coil isolated for vibration and of course the multiple caps for current sharing. Yep, it drifted and were it not for the EI9GQ VFO stabilizer it would have been a complete bust. Today a $20 bill would buy the parts for a digital VFO that could have dual VFO's with memory and no drift.




But one thing for certain, things will change, and they do change. Truth is that some of the former leading-edge approaches can be revisited and thence a new track and new competition.


In yesterday's email inbox was a note from a ham who had previously started work on a phasing receiver and now was headed back to working on it. Of interest he is using "Sputnik Rod Tubes" for the detectors just like the two pencil tubes shown earlier. Now there is a challenge as the detectors are not unlike a pair of Direct Conversion Receivers 90 degrees out of phase and prone to noise pickup -- that will be a special challenge to the build using rod tubes.

My most recent experience with SDR versus HDR has given me an eye-opening experience. The first piece of evidence is that either mode will net a logbook full of contacts given the same power and same antenna. So, factoid one, either will provide QSO's. But the second factoid for me, I prefer to listen to the SDR as the "presence factor" heavily weights the preference to the SDR and I have a lot more HDR radios. If I want listening pleasure, then SDR is the choice for me.




73's
Pete N6QW

BTW at one time I had a bag of those pencil tubes and built several modules of a SSB transceiver using those tubes.

In case you are keeping score...

 Interesting results... Iowa : Joni Ernst's Republican seat was won by Democrat Josh Turek. [senatordb.com] , [ballotpedia.org] Kansas :...