I hate homebrew crystal filters! They are difficult to build and unless you have:
A decent process to build one and
Decent equipment to measure what you have built you will be doomed to failure.
A typical result failing 1 & 2, does and will sound like crap.
Given the recent article in the GQRP SPRAT by ZL2BMI for a tramping SSB transceiver I have given pause to building another small transceiver --- very small. It is a sort of challenge from ZL2BMI as in a recent round of emails between Eric & myself resulted in his stating: "Mine is SMALLER than Yours".
I shared that my shirt pocket SSB transceiver is only 16 cubic inches to which he shared that his rig is less than that and that it has an OLED display. Mine used a crystal switched VXO covering only about 60 kHz of 20M. His is a tri-bander for 160-80 -40 Meters.
So you now know where I am headed.
Firstly I did build a four pole filter using 4.9152 MHz crystals. You can see the plot below! Man that is text book and using the Antuino I now can find the Filter Center Frequency. It will also help me make a 1st run at the LSB and USB BFO frequencies. I used 47PF coupling caps so the BW is wider than I normally use --but it will have presence and not be yellowy! I also only had 47PF in SMD.
JFF (Just For Fun--that must be what the emperor (small e) did when he made his Syria decision.) I installed the filter in a working 80M SSB XCVR and also built an outboard Digi VFO. It works!!!!!!!
Stay Tuned as I work on the ZL2BMI SIZE challenge.
As I sit riveted to cable news, one thought keeps surfacing, that great movie based on Agatha Christie's novel. Yes, I am talking about Murder on the Orient Express!
The ending of the movie shows clear evidence that they all did it! Wow that was a shocker! Now we have those foreign country scandals and more players are "being rounded up in the list of usual suspects". (Like in the movie Casablanca.)
Is it possible that the drag net becomes so large and snares so many that the final outcome is we have the first female President? Wow that is sure to give rise to a lot of sleepless nights for the emperor (small e) and his retinue. Look what the dragnet turned up. Some in the Italian community are now calling Rudy IINO (Italian In Name Only).
My soldering iron is cold as I try to keep a scorecard of who is who on the list of alleged suspects.
Is there that one possibility --they all did it? We also think of that famous line from Shakespeare. "Me thinks he doth protest too much." Stay tuned to cable news-- the best parts are yet to come. There was a trial balloon premise on one of the internet news sources... President Pence … wow some one is sure moving the needle fast forward. Some in Washington today might be thinking of potential future attire in lieu of those $3000 suits they now wear.
BTW the Orange Jump Suit comes in Size 3X --easily will fit a large man or woman -- about 250 pounds is max. So OK -- some ham stuff …
Mechanical Engineering in our Radios, Rigs and RADIGS
Being hams, we often think about the electronic hardware in our rigs. Does it have an IRF510 in the Output or a Real RF Device like a RD06HHF1? Is it being "Run" with a digital VFO or one of those drifty, shifty Analog boxes.
BUT we often forget that some of these beloved radios rely heavily on our brother Mechanical Engineers to make it all play. You only need to look at James Millen who did all of the ME on the HRO receivers. Or pop the hood on that Collins R390A and savor the mechanical tuning mechanisms. Perhaps you overlooked the same in your KWM-2 or Drake TR-4. The ganged tuning of various networks is a marvel. Hey don't forget the SBE-33/34 and the "Geneva" Turret tuning arrangement.
But there is a special place in my shack for those really "cool guys" at Ten Tec. If you perhaps have some earlier Ten Tec radios like the 505, 509 or 540 that string driven dial system was pure Mechanical Engineering as was their PTO's. Can you imagine -- you could rebuild your PTO right at your kitchen table -- pretty clever.
But at times some of the Mechanical Engineering was hidden behind the panel and we simply used it without giving it a second thought. The Model 540 has a band switch located in the lower left hand corner of the radio. Actually it is THREE band switches with two visible on the front panel. When you put the main band switch on 10 meters you have an additional 4 position switch that is engaged to give you four band segments. But there is a third hidden band switch that is engaged as you change bands. Here is that Mechanical Engineering stuff at work
Before you gulp -- this is a Model 540 that was upgraded by me to a Model 544 and was the subject of a QRP Quarterly article in 2013.
When we think about the Model 540 and the principal band switch you must think about what that switch is really doing. It has many jobs/functions.
The way the frequency scheme of the Model 540 works is to have a singular fixed VFO range operating at 5 to 5.5 Megahertz --yes you guessed it a 9 MHz IF. But to work the various bands this basic VFO signal is mixed with various crystal frequencies to produce the LO Signals that results in the 9 MHz IF --it is a single conversion radio.
Now we have all seen the 9 MHz IF and 5 MHz VFO that gives you two bands -- add the IF and VFO and it is 20 Meters. Subtract the VFO from the IF and you have 80 Meters. But you do have to pay attention to any possibility of sideband inversion.
But Ten Tec used for all bands except 20 Meters a VFO signal that is crystal mixed to give an injection frequency above the incoming signal on 80/40 Meters and then for 20 Meters and above the LO injection is below the incoming with no crystal mixing on 20 Meters.
So this mysterious hidden band switch is THE mechanism that "on's" the crystal mixing for the various bands.
Another Mechanical Engineering problem -- packaging. The Ten Tec Model 540 and Model 544 were very "desk top" friendly having a very small compact footprint. Thus a very long shaft could engage many band switch segments to engage multi-function circuits. But you are faced with the trade off of long band switch size and compact footprint.
Another function of that band switch was to switch in various LOW PASS Filter Networks following the final amplifier "brick". Along the way other circuits are switched into play with changing of the band switch.
Did I mention yet another Mechanical Engineering problem -- circuit shielding and isolation. That crystal oscillator and mixer circuit were shielded--yes in a shielded box there is also a MC1496 DBM that mixes the Crystal Frequency with the VFO to produce the proper injection frequency. Another wafer on that third band switch adds some filtering to the output of the driver stage ahead of the final amplifier brick. Believe it or not -- that wafer is mounted on the back side of the shielded box, right adjacent to the driver board.
Oh more ME stuff -- all of the major circuit boards have through board pin connections that on the bottom end plug into chassis mounted sockets and actually protrude through the top side -- this give you access to literally hundreds of test points. Smart Mechanical Engineering!
Back to our third hidden band switch. So if this switch is not on the same shaft as the main band switch how is it engaged. Enter a piece of brilliant Mechanical Engineering which I call "a slider crank mechanism". Two mechanical linkages affixed to the main tuning shaft essentially clock the third band switch so that when you place the main band switch on 40 Meters the Crystal Oscillator puts in line a 11 MHz crystal so the output when mixed with the 5 MHz VFO output a 16 MHz injection signal results. When that injection signal is mixed with the incoming signal at 7 MHz the result is 9 MHz at the IF.
BTW if you look at the Model 540 BFO, a single BFO frequency will result in the "Normal" sidebands for all of the ham bands and with a bit of capacitor padding will shift the BFO frequency to the "Reverse" sideband. Ten Tec does not call them USB and LSB. I keep digressing
Here is a photo of that slider crank mechanism. Essentially what you have is two pieces of metal coat hanger with a loop formed at each end. These loops fit around a drum like spool assembly so that as you clock the band switch the coat hanger material is free to move around the spool; but the push pull action moves the third band switch in unison with the main band switch. If you don't know that the spool assembly comes apart and you somehow get the metal, bars out of the spool -- they simply cannot be fed back in place without removing the cap
The reason I mention this is that I have a project on the bench where the "slider crank mechanism" was in shambles, the dial cord was chewed up and it was obvious some non-engineer maybe one of those BTE's tried to "fix" something without knowing what they were doing. Pretty obvious who they voted for in the last election.
The metal bars were pretty bent up and I did a bit of straightening. The interesting engineering aspects of the nylon assemblies that fit on the two shafts they are molded so that the fit only one way on the half moon shafts of the two band switches and the bonus --there are really small set screws that further lock the nylon pieces to the two shafts.
Good thing I have acquired many small Allen wrenches. BTW the same Allen wrench that works with the main tuning knob will work with these set screws on the nylon assemblies. This is a long term project and you will be amazed at the final product.
I must admit that a I own a few Boat Anchors --like maybe more than I should; but they include names like Collins, Drake, Ten Tec, SBE, National, heathkit, WRL and even a hallicrafters FPM 300 (reincarnated). When I first started in ham radio, several of those names dominated the ham market. They are no more save for Ten Tec that is still operating; but with a cloudy future.
These radios graced the pages of QST, CQ, 73, Ham Radio and even Popular Electronics. They were the mainstay of many OT's and on the want list of those of us just beginning the sojourn in the 1950's.
Some of the boat anchor's today can be had for pennies. Two eBay auctions this week for National NCX-3's ended with a price of $26 and $36.50 respectively. A year ago I bought an NCX3 with a power supply for $69. With the exception of the one heathkit I have, they all work and are on the air. But you know what --- they are worth the amount for which they are being sold. I have recorded two snippets of the Left Coast Vintage SSB Net -- the most recent using the other National radio, a NCX200 which I acquired two years ago. The second recording was made using the hallicrafters FPM300 which has been "souped up" and is all solid state.
I now have a standard of comparison and that is with my RADIG SDR. The old boat anchors were FB in their day; but technology has made light years of advancement and when you listen to the OBA's you try to find the menu selection that will improve the received signals.
One of the huge differences is "being and staying on" the right frequency. The recordings make that point. Yet another is signal handling capability. The AGC circuits in these radios have but one aim -- make the S Meter jump up and down. I have had to rebuild the AGC circuit on the NCX3 using components that had not drifted out of tolerance -- there was no way to even zero the S Meter until I added in spec resistors and capacitors. Hey get a grip many of those components are 60 years old --so you can see why some of original short comings have been magnified (greatly) with age).
For those Collins enthusiasts -- the early KWM-2's had a terrible AGC circuit and that resulted in Service Bulletin #8 which truly fixed the problem --but even with the Collins engineering clout --AGC on the initial production runs (and before the KWM-2A) basically sucked. If you check my website www.jessystems.com you can see a better implementation of SB#8.
My RADIG is not a FLEX or an Apache ANON league; but it runs rings around any or all of the OBA's I have. That also now sheds light on even modern hardware radios such as the many transceivers I have built in recent times. They too suffer from some issues such as those ones with homebrew crystal filters. Do these work --yes; but are they as good as the RADIG -- NO.
Thus I think about the $26 (plus shipping) I would have spent had I won the NCX3 recent auction and for that amount I could build a main RADIG Board and the USB Controller. Add in a $15 sound card and a $35 raspberry Pi3 and I have a very modern radio with lots of bells and whistles. That RADIG will fit in a brief case and not take up the full open space of my work bench.
Some bonus features -- with Quisk and WSJTX and my RADIG --it will do FT-4 and FT-8 without any additional hardware. So now the digital modes are built in to some very simple hardware running with some very sophisticated and FREE software. Try running FT-8 with a NCX-3!
Thus I am making a business case -- if you are considering buying a boat anchor or undertaking a homebrew HDR QRP radio ---- Don't! You can have a far superior and home brew SDR RADIG for the same amount of pocket change. No matter how much I use the beer goggle glasses at the end of the day the OBA's are still a pig with lipstick.
On September 18th, I participated in a Skype Presentation with the Peel Amateur Radio Club (PARC) Homebrew Group. What a blast to visit with a group of fellow homebrewer's and discuss some of my homebrew activities with a specific focus on my work with the RADIG. Mention is made about this change from the IRF510 final amp. This is over an hour long so it is a snoozing event; but the real import is that there are in fact "homebrew groups" that are building their own rigs. Several years ago the PARC made a group build of "The LBS" a joint project from N6QW and AI6YR. Two of their built units made the trip to FDIM where they won first place. Thank You PARC for inviting me to speak to your group.
[Once you have done this upgrade you should consider putting your homebrew SSB Transceiver on FT-8 and here is a great digital adapter kit from:
The price point is about $15 and there is an assembled version for an additional fee of $5. The digital adapter I built cost more than $15 --so I think this is a bargain.However I did not look at the specific interface to a computer aspect, so be sure you check to see if you can trigger this from a USB port on a computer. Most new computers including the RPi do not have a plug in serial port. (You might be able to jury rig a serial port on the Pi with the GPIO pins. I just don't know.)
My adapter took a small kit board that required a serial interface --I had to purchase another board (Adafruit) so you could have a USB to serial hand shake. I think the assembled version should be appealing to many wannabe homebrewers who still ponder which is the hot end of the soldering iron. For those who don't have a homebrew SSB Transceiver it most likely WILL work with that uBitx or an Icom 7300.]
Many of us homebrewers have one or several IRF510 Linear Power Amps as a final in our rigs. Typically below 15 MHz you get a lot of bang for less than a buck ($1). But at times I like to foray out above 20 meters and see somewhat of a drop off in RF output.
I do not believe I am misquoting; but even the uBitx documentation notes that on 10 meters the Pout is somewhat reduced.
So what does it take to take to change out the IRF510 so that there is a better gain distribution at the higher frequencies. Secondly what is the device that would replace the IRF510. The chosen device is the Mitsubishi RD06HHF1 which is good for 6 watts out across the HF Bands.
BTW before those with itchy trigger fingers start emailing me I am not disparaging the IRF510 other than to note is that it was never intended to be an RF device --we just used it that way.
Now what I am about to share has worked for me -- YMMV and there were no Tayloe Detectors or elaborate simulation programs used for my evaluation. Just some simple changes.
First the bias can be higher and since I always use three terminal devices (78L05) to supply the bias a small modification will let you run a bias level close to 6 VDC. The middle pin on the 78L05 is ground. Simply lift the middle pin from ground and install a RED LED with the Anode to the middle pin and the Cathode to ground. Thus when voltage is applied to the bias circuit during transmit --the LED is ON. A bonus you will definitely know that you have bias as the LED shines a bright RED when the bias trigger voltage is applied
The RF FET has a very high idling current with bias applied so you will need something more than a TO-220 clip on heat sink! You will definitely need a robust heat sink!
The Pin Out for the IRF510 (face up) left to right is Gate, Drain and Source. KB1GMX (Allison) has recommended cutting off the center pin on the IRF510 and make all Drain connection to the Tab. In the case of the Mitsubishi RF FET the Pin Out is Gate, Source and Drain. There is good Karma here as the tab is the Source so you can directly screw that puppy down on the heat sink face. But do use the thermal grease.
All other circuit elements remain the same and you should carefully advance the bias so it does not exceed the recommended level. Yes you will be forced to look up the data sheet to discover that level.
Here is my reworked Linear RF Amp Board with the Mitsubishi RD06HHF1 installed. You can see the LED near the 78L05 and the RD06HHF1 is screwed down to the heatsink directly and the center tab soldered to the circuit board. The heatsink is two by four inches and multi-ribbed on the bottom side
At times we might forget the import of ham radio but to a group of Italian polar explorers in 1928, our hobby was a true lifesaver. Do an internet search on Giuseppe Biagi the explorer (not the painter).
You have heard of single board computers --well here is a single board rig! Custom LCD Characters and S Meters See https://maxpromer.github.io/LCD-Character-Creator/ If you are not wiring direct but using I2C select I2C and binary
