Building a 450VDC Power Supply.

How to design a 450VDC at 0.1-amp Plate Supply for the 12GE5? Copilot jumps in and gives an opinion.

Today

Full wave Bridge power supply: 520 VAC, Choke Input at 10Hy, Filter Cap 100 Microfarad at 500 VDC. Bleeder resistor 10K. Is this OK?

You're working with a pretty serious high-voltage power supply setup, so let's break it down and assess whether it's safe and functional:

⚡ System Overview

• Input Voltage: 520 VAC (RMS)

• Rectifier Type: Full-wave bridge

• Filter Type: Choke input

• Choke: 10 Henries

• Capacitor: 100 µF rated at 500 VDC

• Bleeder Resistor: 10 kΩ

✅ What Looks Good

• Choke Input Design: This is a classic approach for high-voltage supplies. A 10H choke provides strong filtering and helps maintain continuous current flow, which reduces peak charging currents and improves regulation.

• Bleeder Resistor: A 10kΩ resistor is reasonable for safety. It will slowly discharge the capacitor when the supply is turned off, helping prevent electric shock.

• Capacitor Value: 100 µF is a decent size for smoothing ripple in a choke-input filter. It won't overburden the rectifier and will help stabilize the output.

⚠️ Potential Concerns

• Output Voltage Estimate: With a choke input, the DC output is roughly 0.9 × RMS input. So:

  • $0.9\times 520\text{ VAC}\approx 468\text{ VDC}0.9\; \backslash times\; 520\; \backslash text\{\; VAC\}\; \backslash approx\; 468\; \backslash text\{\; VDC\}$

  • Your capacitor is rated at 500 VDC, which gives you only ~7% headroom. That’s tight. If your line voltage spikes or the choke saturates, you could exceed the cap’s rating. Consider upgrading to a 600 VDC-rated capacitor for safety.

• Bleeder Power Dissipation:

  • At ~470 VDC, the 10kΩ resistor will dissipate:

    • $P=\frac{V^2}{R}=\frac{{470}^2}{10000}\approx 22\text{ W}P\; =\; \backslash frac\{V^2\}\{R\}\; =\; \backslash frac\{470^2\}\{10000\}\; \backslash approx\; 22\; \backslash text\{\; W\}$

  • That’s a lot of heat. You’ll need a high-wattage resistor, ideally 25W or more, possibly mounted with a heatsink.

I will change the Bleeder to 50K and upsize the Voltage Rating on the Filter Cap to 100UFd at 600VDC and this should do the deed. The Panel Meter reads 0 to 500 VDC.

Integrated Power Supply Front Panel

Started work on building the supply and this is what it will look like.

Power Supply for the Hybrid Wireless

In the photo above, along the left side is the 520VCT power transformer and 10 HY choke. The blank PC board will have the Bridge rectifier installed and pad connections for the choke, plus installation of the Filter cap and he Bleeder. That I built yesterday.

Along the right edge is a 12VDC 5.6-amp supply and at the back our 280VDC 0.3-amp supply. 

The Plate Transformer has a 6.3VAC winding which will be used to feed the 12.6 VAC side of a filament transformer rated at 1 amp and pretty small (not installed as yet). This will provide about 60 Volts AC at the terminals normally used for the Primary.

My plan is to add a back panel and mount small filament transformer on that vertical panel and that becomes the Bias Supply at -60 VDC. The Bias Set Pot will be on the Hybrid Wireless Chassis.

The back panel will also contain an 8-pin socket where all of the voltages required by the Hybrid Wireless will be connected.

1. Power Switch

2. Power Switch

3. 450 VDC

4. 280 VDC

5. -60 VDC

6. 12 VDC Filament

7. Ground

8. 12 VDC Relay Supply 

Close up of Bridge Rectifier PC Board

(Parts to be added when they arrive)

A Final Act will be to put some sort of enclosures around any exposed high voltage portion of the circuitry. I have a couple of ideas in mind that will eliminate any potential shock hazards.

Them that know can make things go.

Old what is his name was MIA for a couple of days and has reappeared so to all watchers just a blip on the radar. It is an interesting cultural phenomenon; you are the news even when not making news.

73's

Pete N6QW

The Front Panel ~ Hybrid Wireless

Not unlike the Final RF Cage, a lot of effort will be put into the layout of the front panel. It all started with the 12X12 Inch 5052 Aluminum panel that I cut into two pieces 6X12 inches. I took care in the cutting using the Home Depot newly acquired hacksaw blades to cut a straight line and then went to the CNC mill to true up the edges and to make matching pieces. Along the way I ran the X Axis past the stop and had to do some CNC Mill maintenance work to get that back working properly.  That consumed most of yesterday.

Suddenly it became apparent that I needed to rethink the front panel layout as 6X12 is 72 square inches seems like a lot. However, we do have a large tuning knob, a very large display, five knobs, three toggle switches and two audio jacks (Mic and Speaker) and finally a panel meter. 

Key to the layout is ergonomics and not getting cables in the way of the critical controls like main tuning, audio output, transmit controls like mic gain and Final Tune. Time for the cardboard cutouts.

Still another factor is that the main assembly plate will sit two inches from the bottom and four inches from the top. I may even have to cut a notch in that assembly plate to accommodate the optical encoder. If you look at your KWM-1 or KWM-2 or some of the Heathkit and Swan gear you will see that notch at the front of the chassis to accommodate the large analog dial. My notch may only be 1inch wide and 3/4 inch deep. But the carboard cutouts will provide the answer.

 Front Panel Ouija Board

The four controls as the upper left are for Volume, RF Gain, Carrier Balance and Mic Gain. The one control along the right edge is the Plate Tuning Cap.

This simple "put the knobs on the front panel exercise" indicates I likely will not need the notch. The open space to the left of the knob would be for the Mic Jack, USB/LSB Select and MOX/Tune. To the right would be Main Power and Audio Output. The 3.2-inch Color TFT would shift to the right to be more centered. 

Mind you a 1/2-inch wide by 12 inches long area at the top and bottom edges is unavailable as that accommodates the aluminum angle stock for the bottom plate and stiffening structure to make the box rigid. 

More work installing boards today which involves more drilling on the main assembly plate.

Oh, have been rethinking the power supply and I found in my stash a 520 VCT 100ma transformer and using a bridge rectifier with a choke input would give me 450 VDC at 100 milliamps and that would be for the plate supply just for the 12GE5. The supposed 300VDC 0.3-amp supply will be for the low-level stages and the screen supply for the 12GE5. 

Them that know can make things go. 

73's

Pete N6QW

 

More Metal Bashing and Precision Locations.

It was hot in the garage yesterday and I took advantage of the heat to sweat off a few pounds. I just sweated but didn't lose many pounds!

Yesterday's work scope was to install the shielded RF Final enclosure, install the 12GE5 tube socket, the Pi Network and line up the area allocation. This singular installation will drive further installations both on the top and underside of the mounting plate. 

I near got a hernia using the chassis punch to cut and 1- and 1/4-inch diameter hole for the 12GE5 in the 1/8-inch aluminum plate. This is where you have to eat your Wheaties.

There was extensive measuring, more measuring and still more measuring before cutting any metal. As a starting place I am using a Pi Network from a Junker SBE-33. The loading cap is a high value padder cap. So, the plan is to set that padder value once and then the front panel Tune Control will bring the tank into resonance. This is like the Pi Network in the Hallicrafters SR-160. (No panel mounted Loading Capacitor.)

Another factor is the panel meter which will be mounted on the control panel in front of the RF Enclosure and above the Plate Tune control. It is all about clearances and space allocation.

Topside Space Allocation

Them that know can make things go! Yesterday was a very productive day!

Surprisingly there were lots of views of the Home Depot Biometric Surveillance issue. A nagging question, we all have, is the Biometric data collected by Home Depot and Others used only for Theft Deterrence and/or are ICE and DHS also getting a feed as a part of the alien roundup? There is no doubt you are being watched and if your immigration status is in question... Caveat Emptor!

73's

Pete N6QW

The New Power Supply Test = Not as Expected!

In a nutshell, most likely overhyped and definitely underperforming. We as Americans, all of us, so much want to get "such a deal" and the Chinese play into that American weakness: Almost a deal but not quite the one we wanted

The supply never produced 300VDC but did hold at 280VDC all the way up to 70 MA of current draw. That 20-volt difference at 0.3 amps is 6 watts of power you won't see going into the 12GE5.

Shame on me as two of the comments on the product did spell out that deficiency, yet the supply has a high rating. Certainly not from me! How could I be so blind! Dire Straits said it best: Something for Nothing (But the Chicks are Free... BUT No Freebies Here). This supply might be a $1.25 Hooker disguised to look $2.50.

I did test this power supply on a CW transmitter to see how it holds up. It did OK with a 6L6 in the socket, where we saw 60 MA of Cathode current, and the Drake Wattmeter said 7 watts in 40M. I attribute the puny output to the lower Plate Voltage at 280VDC but with say 350 VDC or more on the plate I could conceivably see about 20 Watts output. (Do the math as 70X0.3 = 21 watts more input at a higher plate voltage.) The signal sounded clean and that was both the case on 7030 and 7060 kHz.

While my earlier tests (in the video) gave pause for the use of the supply on the Hybrid wireless, the CW tests suggest that it may possibly function on SSB in the Hybrid Wireless. That will have to be tested, but it is a disappointment that you can't get 300VDC out of the supply as advertised.

Them that know sometimes get fooled by the bright shiny object. I included some shots in the video of how I built the test fixture. The Manual Mill is sure a handy tool.

73's

Pete N6QW

 

Random Musings Today.

The watched pot never boils as the new DC power supply didn't arrive until about 7:30 PM last night so no testing until today. The size is deceiving as it is really small. 

At the first look, the Amazon bag looked too small. But sure, enough it was the supply. This is good news for making new compact equipment and/or retrofitting existing hardware. I hope it works as this small size aspect shows real promise. Even without a test, the terminals for the 6.3 volts say Ground and 6.3 Volts, thus most likely DC.

My post the other day netted a response from a blog reader with a link to a news article regarding Home Depot citing a lawsuit alleging video profiling customers during checkout. The basis of the suit is that this profiling is being done without notice to or permission from the customers. If this is a case my admonition about being watched is more than just a possibility and Home Depot may not be the Lone Ranger.

In our legal system (or it used to be) is the concept of innocent until proven guilty and the suit has applicability in the state where the suit was filed because of an existing law regarding collection of biometric data. That may not be the case in other states where there is no such law on the books. Bottom line "AI Watching Technology" is at a store near you.

The basis of facial Biometric collection is to thwart theft. Old school: you look like a thief so you must be one. Yet, just yesterday in Ventura County, where I live, 14 people were Busted for $10M in thefts from Home Depot.  Obviously, this will be a critical piece of defense for the Biometric data collection... The $10M theft is likely not a singular event but only a sample.

Suddenly the pieces are coming together about why the Home Depot salesclerk opened up the machine to validate I put $6 in the machine. True I hadn't shaved on Saturday, was wearing a KN-95 mask and had a ball cap on and must have looked "theify". BTW I still wear a mask at Covid breeding grounds like Home Depot, Costco and the Grocery Store. 

Yesterday while awaiting the power supply, I did fabricate the RF Cage that will be installed around the 12GE5 final RF amp circuitry. More real metal bashing.

Unabashedly I say you are being watched and profiled.

Them that know can make things go.

73's

Pete N6QW

Initial Testing of the new DC Power Supply.

Later today my new $30 power supply is scheduled to land at my front door. The supply, a switcher, is rated at a wide voltage range input, and supposedly good for 6.3 Volts at 5 amps and 300 VDC at 300 milliamps. Targeted at vacuum tube audio circuits, it is touted as low noise.

Before I place all eggs in one basket, I thought it best to do some lab testing and this led me to thinking about test criteria and the actual conduct of the tests. I also generated a concerns list.

The first is the price and to put it in crass terms what would you expect from a $2.50 hooker? The price of $30 for the supply raises that same level of concerns.

Basically, we heard some good things about this switcher, but I do wonder if the glowing reports may belie that there have been known component failures in the supply. Gasp were those glowing reports AI generated?

If it fails at plug in, then we know the awful truth thus there should not be a mushroom cloud of smoke at the 1st application of power. That is followed by a 15 Minute test where the supply is idling (no load connected) where nothing gets really hot, or no burning odors are observed.

Now the big issue is how does this supply hold up under load. The current vision is that most of the load will be seen in the high voltage section and so we should look at some math. 

There is also a power duty cycle factor to consider in that in receive we likely would see a constant load (9 tubes total) but it is in receive where only six are drawing current of less than 70ma.  The max current draw is the PEP of the 12GE5 final on voice peaks and there we could see 200 milliamps or better total current draw on peaks. 

Essentially 300 VDC at 0.3 amps is 90VA. (90 watts). If the load is a 1000 Ohm resistor, then the current is 0.3 amps. That resistor better be 100-watt rated! If the specs hold, then we should see no sag in the voltage and the current supplied should hold at up to 0.3 amps. I don't have such a resistor. 

Now if we would theorize that even at 0.06 amp the supply would hold that would give us a really "good indicator". If it flunks the 0.06-amp test with a voltage sag or failure to supply that current level that is a sure sign it will flunk the max power test. 

Two 10K resistors (20W) in parallel is 5K yielding 60ma with 300VDC applied and only needing 18 watts of power rating. The 60-milliamp current is close to the load that will be seen on receive.  The 2nd part is holding that voltage and current over a 10 Minute period without sagging or smoking itself. This is a critical test as if fails this part ... it fails totally. The duty cycle with SSB voice is easily within the 0.3-amp range and transitory. It is the constant load on receive that is the most demanding.

Other parts of the testing would be to look at the ripple content and noise generation. A DSO will measure the ripple across a series of dropping resistors as you don't want to smoke the scope probes. For the noise I can use my Hermes Lite 2 SDR to look at the noise and evidence of switching transients across the HF spectrum

Now a concern that came up after the unit was purchased. The filament voltage is stated at 6.3 Volts at 5 amps. As a victim of habit, I just assumed that was 6.3 Volts AC, but that could be DC. I could not find that info from the product advertisement. 

Perhaps a hidden bonus is to change over to all 6.3 VDC tubes (the 12Ax7 and 12AU7 can be wired for 6.3VDC) and this would mean I would power the filaments from this single supply and a smaller 12VDC supply could be used for the solid-state circuits.

So, DC would not work with my idea of back feeding a filament transformer for the Bias supply. I might have to add to the mix a small 115VAC to 115VAC transformer for a bias supply. eBay has them!

Things are certainly in a churn, and it is interesting of how other countries view the US. If you have any friends outside the USA, you might ask them. A majority of you voted for this option and was this your expectation.

Reduce your digital footprint as you are being watched.

Them that know can make things go.

73's

Pete N6QW

Sundays when I was a kid back in Pennsylvania.

This past Sunday morning Pasta Pete was at work in the kitchen. I even bought a name tag so when I go to the Grief Group Sessions people know who I am. Should mention that so far, 80 to 90% of the GG attendees are older YL's and the rest are guys like me.  

(I had thought about a 2nd Line: I can cook!)

Back to Sunday mornings in western Pennsylvania. It was a ritual starting with the family going to 8 AM mass at St. Margaret Mary's Catholic Church followed by breakfast at home and then the magic. Everyone out of the kitchen as my mom whipped up Sunday Dinner served promptly at noon. 

There was the homebrew spaghetti sauce with meatballs, breaded eggplant, with a side of zucchini, a tossed green salad with tomatoes and cucumbers and for dessert something out of this world. Supper was usually just a snack as the Sunday Dinner was enough food for a week with the calories to match.

It was not unusual to have midafternoon family visitors where mom would hit the freezer for a previously made dessert to serve to family.

I liked the summer Sunday feasts as the meal was usually served outdoors, and the dessert often was cannoli or an ice cream sandwich made from waffle cookies and a slab of vanilla ice cream. If you check pastapete.com the recipe for the waffle cookies can be found along with a how to do it you tube video. Also on the website is the magic spaghetti sauce recipe.

Those were funny times as the Sunday ritual was watched by others. If you were not at 8AM mass with the whole family, it was like the Spanish inquisition. One Sunday I was on a trip, and my mom got a call from a "Nosey Nellie" wanting to know if I had stopped going to church.

As my sauce came together, and the rich aroma of the simmering onions and garlic permeated the house it was a stark reminder of some really great past times and the impact of a simple family meal. 

This past Sunday it was just me eating spaghetti and vegetables and the cannoli has given way to 70 calorie Non-Fat, Low-Fat, No-Taste, Small Portion and Strange Appearing Danon Yogurt. [Sunday Mass from our local Catholic Church is live streamed on YouTube. Times do change!]

Potent: Olive Oil, Onion and Garlic, Hot Pepper

On Sunday I did write a CNC program for the panel meter cutout, and I bought a DC power supply for the Hybrid Wireless. This supply will provide 6.3 Volts at 5 amps and 300 VDC at 300ma. It should be enough juice to provide 20 or so Watts of RF output from the 12GE5 Final. The price is less than $30.

This supply will be used along with a 12 VDC 5.6 Amp power supply board that will source the SS circuits and handle the filaments. The 6.3 Volts will be used to reverse feed a 110VAC to 12VAC filament transformer which then will produce about a -60VDC for the bias. This will make for a much smaller footprint and perfect for this rig.

Hide your digital footprint.

Them that know can make things go

73's

Pete N6QW

Pox on Home Depot!

As I continue the work on the Hybrid Wireless metal bashing, I will need to cut some aluminum plate using a hacksaw. One piece of stock I have as it stands is too large for my CNC but once cut, I can true it up using the CNC Mill. The problem is the 1st cut and the need now for some fine-toothed hacksaw blades. 

Being Saturday, I needed to run some errands and one place I had to visit was very close to the local Home Depot. So, I get to Home Depot and find the Tool Section and the hacksaw blades... $6+ dollars for two blades made in China and likely good for one cut in soft metal before going dull. (I did get a lot of steps on my Smart Watch so that was good.)

Finally, you get to the checkout area where there is an array of self-service terminals and no manned cashier stations. There are some terminals that only take cash and some that only take credit cards. I opt for the cash as there is my internal discipline to leave no digital footprint, so I wait for a cash terminal to be available.

I hate those machines as they always seemed to get hosed up. This morning was no exception! I scanned the code, and it read OK and the price with tax was $7.48. I opted for cash and was instructed to introduce the cash into the slot. 

I had at the ready $5 and three $1 bills. I put in the $5 and then a $1 and then the machine shuts down. I use the touch screen to call for assistance and waited about two minutes for someone to show up. 

The attendant showed up and did all sorts of things with various keypads which did nothing and then unlocked another cabinet to secure a key to open the machine and then unlocked a box that was the cash repository where she affirmed the last two bills put in the machine was a $5 and a $1.

She then says we have to go to a special terminal where she scans the hacksaw blades and puts that transaction on hold, while she goes over to a special manager's terminal and gets into a cash drawer where she removes $6 and hands it to me. Then she says put your $8 into the slot and I told her NO. You do it! She did and I got 52 cents in change, two hacksaw blades and a receipt.

So, then I asked her what happened, and she says I don't know, but it happens all of the time. The total elapsed checkout time was 15 minutes. Unless it is a dire emergency that is the last time I will go to Home Depot. 

I hate those machines and their lack of efficient and reliable operation. The same for my grocery store as it is hard to checkout items like fruit and vegetables. If everything is bar coded, then maybe not as big of a problem.

This will only get worse, as more AI is introduced under the guise of speedy checkout with attendant layoffs of people who would actually check you out. 

In the end I had to have a real person finish the transaction for $7.48 and took 15 minutes of my time. There is not one blog readers who is reading this that did not have a bad experience with these AI checkout machines. Shudder to think if I had used my credit card!!!!!!!

Did install some additional angle stock on the base plate but it was 90F in the garage, so I stopped.

Them that know can make things go.

73's

Pete N6QW

Travelog: A visit to the Academy Museum 0f Motion Pictures

A non-radio day yesterday, the local family, son Tim, his XYL Amy and daughter Gina had a fun-filled outing in LA. The first stop was lunch at the Farmers Market (3rd and Fairfax) followed by a tour of the Academy Museum of Motion Pictures located close by.

N6QW behind the Godfather's Desk

Museum highlights included sets from various movies, including models used in filming of movies like 2001, The Space Odyssey, Star Trek, Barbie and of course a desk from The Godfather.

Many costumes used in the movies were on display and that was a treat to see a photo of a movie scene right next to the costume.

A large tabletop computer generated display was a map of the LA area showing the various studios that literally ringed Los Angeles. A factoid was that in 1902 when the movie business essentially moved from New York to LA the reining industry was oil. In 1929 LA had grown 12X and the driver was the movie business which replaced oil as the industry. 

At its heyday LA had over a dozen major studios. Some studios were owned by major actors. Something I never knew was that actor Sessue Hayakawa who was the Japanese Prisoner of War camp Commander in Bridge on the River Kwai actually had been making movies in Hollywood since 1914 and owned Haworth Motion Picture Studio. BTW although born in Japan he was a graduate of the University of Chicago.

A technology old dog new trick was Pixar Animation which was a series of small character figurines arranged on a circular disc which was spun up at high speed. A blur indeed. Then they turned on the strobe lights and it was an animated show. Can anyone remember the strobe light kits used for car repair? Stop motion technology used for movie making.

A fun day being with family which started with my Friday visit to the cemetery to bring flowers to my XYL. My smart watch gleefully announced by about 5PM, I hit 6000 steps.

Many may be questioning my continual admonition about all of us being monitored, observed and watched. You need only to call up the Microsoft copilot app and type in your name and call sign and see what results. In short order you will see anyone can do this. Reduce your digital footprint. One thing for certain never say negative things about old what is his name. If you do, your name goes on a list.

Them that know can make things go.

73's

Pete N6QW 

Bashing Metal

In today's world, homebrew frequently uses pre-made circuit boards or that awful ugly unsightly approach with superglue and small copper squares.

When I started homebrewing in the 1950's the soup du jour was an aluminum chassis and the use of Greenlee punches for cutting the holes for the Tube sockets. This was frequently called "metal bashing".

If you want to do a bit of bash and mash start with the type 3003 or 5052 aluminum as these are the variants that are easy to mill, drill, hammer and beat. The type 6061 is not too good for working with simple hand tools.

I had a meeting with myself yesterday and decided to use a base plate of 12X8 inches as I found I had two pieces that were 1/8 inch thick. One piece will serve as the main chassis plate and the second piece will be the bottom plate. I also have a 12X12X1/16-inch piece of aluminum stock that I will cut in half so that the front and back panels will be two pieces 6X12 by 1/16 inch thick. Indeed, metal bashing.

 

 

 

 Top Plate Layout

 

Underside Layout

In an earlier post I mentioned the use of cardboard cutouts, and this is where they come in handy to allocate real estate on the mounting plate.

The bash part went pretty quickly but the measure 300 times and cut once took longer.

The use of 1/2 X 1/2-inch angle stock ties the front and back panels to the base mounting plate as well as the bottom plate. Angle stock at the very top of the front and back panels tied together with cross pieces rounds out a box like structure. Very rigid and solid!

Basically, we positioned the nine-tube chassis and located 4 holes where I installed the mounting pillars, so the tube chassis is elevated above the mounting plate. I also located, a 1.125 inch in diameter pass through hole so cabling can pass to the underside where the power distribution board and the audio amp board will be installed. Also on the underside is the Driver, the IRF510, and Low Pass Filter.

Other controls such as Power On/Off, MOX, mic input and mic gain, carrier balance and the audio output will be on the front panel at about .75 inches above the bottom plate reference. 

The 12GE5 will be located along the right side of the chassis as will the Plate Tune and a RF Output Meter. 

Reduce your digital footprint. Things are not looking good on the horizon. It seems like an important Republican operative was raided by the FBI. A coincidence? The signs are clear.

 

Them that know make things go.

 

73's

Pete N6QW 

Hardware Fabrication

C'mon you belong to all of the obscure social media sites, you have the latest iPhone, and Smart Watch, two FT710's and matching Mercury Amps, an 84-inch wall mounted TV with 500 channels and gasp you even voted for what is his name... yet you don't have a CNC Milling Machine.

Yesterday I spent a bit of quality time fabricating boards for the new Hybrid Wireless Set made all the easier because I have a CNC Milling Machine. No messy super glue and cut copper pads, just chuck in a blank board and push the start button. 

The additional advantage of the CNC is the board layout is standard and pays attention to preventing unwanted feedback loops, provisions for heatsinks and standard mounting arrangements. I would rather spend my bench time soldering parts to a board versus trying to unstick my fingers or remove superglue from my fingers so I can use biometrics with my smart phone. 

Two BPF's and an IRF510 RF Amp

According to a guy who should know, he stated that smart phones are dead! Gee, you just downloaded an app that tells when your Girl Friend might get pregnant. You don't need that app if you are homebrewing radios using your CNC machine.

Them that know can make things go. In a prior blog post the schematic for the BPF's was shared. The IRF510 is a standard design and can be found throughout the n6qw.com website.

73's

Pete N6QW

Fabricating two Band Pass Filters for the 20M Hybrid Wireless

In the development of the original 40M Wireless, I purposefully used two separate Band Pass Filters. The original Swan 120, which was the template for the wireless it had tuned coils. I converted that part of the design to broad band with the BPF ahead of the 12BA6 RF amplifier stage

Another innovation was the use of a homebrew 2.5 mHy ferrite core choke in the plate circuit of the 12BE6 transmit mixer where a couple of turns of wire on the FT-82-43 core provided a 50 Ohm tap point. From this tap point the transmitted signal passes through a second Band Pass Filter and on up the RF transmit chain.

A simple mod to the 40M Hybrid Wireless gave proof of life that receive on 20M was quite good. Two more changes are needed to prove 20M transmit and these include a 20M transmit BPF and the 20M LPF following the genuine IRF510's (from Jameco... no BoJack's here.)

This design uses 9-50pF Trimmer caps available from Jameco Electronics. If you parallel a 150pF cap with C1 and C2 and replace C4 with a 5pF Trimmer cap this can be tuned to 40M. Note R1 is for the simulation and not installed as hardware.

20M BPF

Our LT Spice simulation reveals a nice flat response across the SSB Phone band and so this design should do well in the 20M Hybrid.

Using an existing stored CNC program about 5 minutes of work created two PC boards for the new BPF's. That is on the agenda for today's work.

Hide your digital footprint as we are all being watched.

Them that know can make things go.

73's

Pete N6QW

A slight change (for the better) in direction.

About 3 AM this morning, my brain said: Hey Pete, how about doing this. The suggestion was to take the 40M Hybrid Wireless and make that into the 20M version. The brain also suggested "Plug In" networks so that I could operate on either 40 or 20 Meters without resorting to a band switch. 

Plan View of the 40M Wireless

The above photo is the 40M Wireless sans any Final Output stage. Yes, a real mess that would have to be totally recreated for a new singular 20M version. For starters, this version works, and we have a good sense of the physical space required and a layout that could fit in our 12X12 space. But could this work on 20M is the question.

It was an easy test to ascertain the feasibility of such a project change and the short path was to program one of the spare ESP32 boards for 20M and dig into the junk box for a 20M Band Pass Filter Board. Boom... it worked as I copied stations on 20M and so that is a good idea worth pursuing.

This approach gets me a finished radio in a shorter time span, and we can focus on keeping much of the original 40M build, while also cleaning up this mess. In the process of doing this test I also learned a hidden (to me) feature to the JF3HZB code. 

Most of the Arduino code I have previously used when it comes to the Si5351 has a line in the code to set the correction factor for the Si5351. I looked in the Si53515 correction code set for the ESP32 and somehow could never find the lines of Si5351 code. This was an annoying nit as I was not 40Hz low but 2 kHz low, for which the SDR police would have my neck.

When I set up the 20M test code I noted that the BFO was off and now we were 5 kHz low. That caused me to look at the lines of code and that is when I realized there are two lines of code that need a numerical value for offset correction and the second was the carrier frequency. In the original GitHub download these were the same frequency.

Then I said let's fix the BFO frequency and a few trial runs fixed that issue. But we were not on the correct frequency on the dial. A few more runs of adjusting the frequency correction factor and that was now fixed. I even tested the code across the band, and it was correct. 

I am guessing here but suspect that the JF3HZB Si5351 correction is being done in another manner by using a specific modified BFO frequency and a separate numeric dial correction factor. All roads lead to Rome!

The 20M Code

I am glad my brain awoke me as now the task is not a mountain to climb but a large hill. A follow-on test is to build the Transmit 20M BPF and add in a 20M LPF. These two small changes would give me a 20M transceiver and that will seal the deal for this approach.

Gavin Newsome sent me an email to sign a letter he has drafted regarding some sort of political action in California. I would not sign such a letter, and it has nothing to do with whether it has my support. 

But in today's political environment that signature will become a part of a list of names to be investigated by the IRS and DHS as a part of Make America (you know). Any opposition to what is his name agenda brands you as anti-American. Shades of Joe McCarthy. We are indeed here at that point in the USA and freedom of speech is gone! Reduce your digital footprint.

Them that know can make things go.

73's

Pete N6QW

A Really Nifty $40 Tool from Amazon

Every Friday I go to the cemetery for my weekly visit with the XYL. While the Santa Clara Catholic Cemetery in Oxnard does a pretty good job of mowing the lawn, that is it. So, I always have to do some grass trimming around the headstone and the ground mounted flower holders.

In some ways the additional time it takes to trim the grass and tidy up lets me spend more time on the visit. Up until about three weeks ago I was using a newly acquired pair of hand operated shears. Then I realized my constantly sore right hand was a result of carpal tunnel shearing. That led me to Amazon and this jewel.

Fully rechargeable (even from a laptop) it is good for about 20 to 30 minutes of cutting. Fitted with two attachments it can trim the grass as well as trim bush branches up to about 1/4 inch in diameter. Charging takes about 1/2 hour. It is well made and worth the $40.

There is always joy in seeing the two young Hispanic sisters who sell flowers at the cemetery entrance as they are still here. Anyone who looks Hispanic or speaks Spanish are at grave risk in Oxnard, a special target area of you know who.  

Them that know can make things go.

73's

Pete N6QW

Sideband Select Now Works on the ESP32.

The culprit was a flag that had to be set to 1 in software and not to 0. With a heap of thanks to KK4DAS, Dean, who alerted me to the issue which resulted in the code being operational. So now I will go back and fix the 40M Hybrid Wireless so it can do FT-8 on USB. (Done!)

But my fumbling was an excellent learning experience, as I saw the opportunity for other improvements which I made. The bonus is the learning of a skill much akin to field stripping my .45 Automatic blindfolded when I was in the Seabees (or Mary Jo in the backseat of the 57 VW Beetle). When you get it, you get it!

So, I now can move forward with laying out the 12X12 inch mounting plate. to include the eight tubes in the lower-level stages, the solid-state stages (Audio Amp and RF Driver) and the final RF Amp, the 12GE5. 

The Finalized Digital LO/BFO

The above module took almost a month to build and troubleshoot. But there was real significant learning in how to work with the JF3HZB code. One shortcoming (sideband select) has been conquered. If you look in the lower right-hand corner to the right of the 5-pin cable connector (optical encoder) you can spot two LEDs. I added code so as you shift sidebands one of the LEDs is lit to show there is a change and the action carried out. That was a big plus when I was trouble shooting the assembly.

For the next layout process I will use cardboard cut outs which represent the actual sizes of the modules. It is an action not unlike using a Ouija Board where the variations in circuit placement enables an optimized final layout. 

One target is a 7X6 inch area for the eight tubes. The Audio Amp and Driver Boards will be on small PC Boards mounted on the underside of the chassis. Some shielding on the underside may be required. 

This was a giant step forward in having the Digital LO / BFO built and operational.

Them that know can make things go.

Reduce your digital footprint... it is a vital move. Anyone wondering what really happened in Alaska? Usually someone gets sold out. 

73's

Pete N6QW