Wednesday, December 13, 2017

Brake Delay for Ham-M and Ham-II Control Boxes

A $30 hamfest find - CDE Ham-II rotator control box.
Proudly installed in the operating position.
When I put up my tower in 2001, I'd had read enough on the TowerTalk mailing list over the years to know about the importance of brake delay. This allows the rotating parts on the tower to coast to a complete stop before the brake wedge on the CDE rotator design engages. Without it, the rotator comes to a sudden and abrupt stop, causing a lot of torsional stress and making the antenna and tower elements wave around unnecessarily. That repeated stress could end up causing failures over time.

The original Ham-M design has no delay. Pushing the lever a little bit illuminates the meter and shows the position indication on the meter. A little more turns the rotator. The moment you stop pushing far enough to turn, the brake solenoid releases and engages on the rotator shell -- causing the sudden stop.

The original design schematic, sketched out in 2001.
Back in 2001, I came up with a very simple circuit. I had some Allied Controls 12 volt relays with a 1000 ohm coil. I figured I could rectify some of the AC from the motor terminals with a big enough capacitor to power the relay for a few seconds. I ended up drawing the circuit to the left.

50 volt PIV rectifiers might not be quite enough, so I wrote 100 v PIV on the design. I ended up using 1N4003 rectifiers which are 200 volt PIV.

I estimated 10 mA of current is necessary to keep the relay closed. Since the left and right motor connections are about 30 volts AC, rectified this would be about 42 volts. Putting a 2.2 K ohm resistor in series drops the voltage down to something less than 15 volts across the relay. All that remained was choosing the right sized capacitor. I had some 1000 uF low-profile capacitors in my junk box, so that was perfect, right? The RC value was around 3, so I expected a delay around  three to four seconds.

Imagine my shock when I actually built the thing and found it took something closer to fifteen seconds for the relay to open! Clearly, a lot less than the full 10 mA was necessary to hold the closed relay closed. I ended up modifying the circuit by placing a 1 K ohm resistor in parallel with the relay. This resulted in a delay closer to six seconds, which was just fine. The relay contacts are wired across the switch contacts for the solenoid, so the solenoid remains activated until the relay opens.
Relay board shoehorned in control box.

I built the circuit on a tiny piece of perfboard just a little bit bigger than the relay. Biggest problem was shoehorning it into the very crowded Ham-M control box.  I figured I could mount it on a right-angle bracket just behind the meter.

Wiring it into the circuit was harder. Assembled, the switch contacts of the Ham-M are inaccessible. I had to remove two sheet metal screws holding the meter panel to the base. That gave me access to the switch. After soldering up the leads, the meter panel goes back in place.

This circuit has been in use since 2001, and has worked great.

At a recent hamfest, I found a deal: a relatively pristine Ham-II control box for only $30. I bought it, brought it home and immediately determined it worked great. It would take the place of the Ham-M control box -- but it needed a brake delay.

The Ham-II design has three buttons. The center button supplies power to the brake solenoid and must be pressed for any power to reach the motor windings. Proper technique would have you press the center button and hold it, then engage the left or right button, holding until the antenna had moved to where you want it, releasing the left or right button, while continuing to hold the brake in with the center button.

Ham-II control box with brake delay installed.
However, if you just pop off the buttons, the brake wedge would slam into the rotator shell.... Not what we want.

It wasn't hard to find the same parts in the junk box. Another scrap of perfboard, a few minutes of soldering, and the same circuit was again created. Mounting in the Ham-II box is a much easier exercise, since there's plenty of empty space. Because four of the connections are to the switches, I elected to mount the board on the bottom of the box, just behind the switches. A threaded standoff was used on an existing screw to hold the board slightly above the switch contacts. The standoff also supplies the ground connection.

This circuit works great. In the Ham-M, the meter and lights would go off once the relay disengages. In the Ham-II, the meter and lights are on any time the box is turned on. It might be nice to have an indicator that the brake is engaged. Perhaps this will be a future addition.

Friday, November 24, 2017

Forty Years of Personal Computing - First Steps

As my present for Christmas of 1974, my older brother gave me a one year subscription to 73 Magazine, starting with the January 1975 issue. This was pure genius on his part. He knew I was showing an interest in amateur radio, plus he would have the opportunity to read my copies, so he didn't need to continue his own subscription. I enjoyed the magazine so much, I continued my subscription until the magazine's demise in 2003.

1975 an exciting time in the world of the electronic hobbyist. MITS introduced the Altair 8800 in November of 1974, and other vendors soon jumped into the market. Microprocessors and related components were coming available, and 73 Magazine was writing all about it.

With the February 1976 edition, 73 Magazine added a special section called I/O, to talk about the developments with these Microprocessors. To my teenage mind, I just had to have one of these machines. They were expensive, but I slowly saved my money from my paper route.

By the spring of 1977, I fell prey to a surplus equipment advertisement from Verada 214. The ad promised a complete system with a keyboard, display, microprocessor and dual cassette decks -- the Viatron 2111 -- only $699! It seemed like a perfect way to enter into the budding world of microprocessors without spending a fortune. Although the equipment arrived as advertised, there was no microprocessor to be found -- nor any way to program the device. After studying it thoroughly for a week, I decided the purchase was a mistake. My father managed to convince the company to refund my purchase price for the cost of shipping it back.

My father could see the seriousness of my interest, so he signed me up for a summer course at the local college where he worked. This was a six week course, and the first three weeks were spent on learning about VCR and other television technology. (It really is too bad that Quasar Alpha-wrap video cassettes didn't catch on -- but they were HUGE) But the last three weeks included plenty of hands-on time with a KIM-1, the 6502-powered single-board computer. Forty years later, I can't remember what programs I wrote, but I thoroughly enjoyed that first taste of programming.

About this time, a computer shop had opened up in Morgantown, WV, about 20 miles away from my home. My parents took me there a few times. I was fascinated to see the MITS and IMSAI equipment I had read so much about over the last couple of years. But the prices were beyond my reach. These systems cost upwards of $2000, much more than I had to spend. I'd have to wait for something more affordable.



Saturday, October 7, 2017

WARC Dipole Repair Complete

Yours truly mounting the antenna near the
top of the tower.
Well, if that didn't seem like it took forever. Back in early September, I took down the WARC Dipole. Although I quickly determined that one of the 17m traps was irreparably damaged, I didn't think it was going to take more than a month before I got the antenna fixed and back up in the air.

I made two new traps using identical lengths (35.5") of RG-58 coax. I didn't bother to try to trim these to any precise frequency. I just made the two traps as identical as I could, and that seemed sufficient.

With the traps back in place I didn't want to re-mount the antenna until I was sure all three bands were squarely in a useful part of the SWR curve. To accomplish this, I used the rope and pulley I had attached near the top of the tower. I used a carabiner through the U-bolt to lift the assembly near the top of the tower.

This actually worked pretty well, as I was able to get the apex of the antenna within a foot of it's mounted height. A 55 foot piece of coax allowed attachment to my MFJ-259 antenna analyzer.

First measurements showed the antenna to be off on all three bands. That wouldn't do. Since I had originally hacked in the 12m traps to a 17/30m dipole, I hadn't done a good job of making the wire lengths symmetrical. This was contributing to bad SWR at resonance. So my first trimming was to make the two legs equal, and subsequent trims were to both legs.

The process goes something like this. Attach the antenna to the carabiner, haul to top of tower. Take rope from one leg and attach it to end of old playground. Pull the other rope through a pulley on the other side of the house until it is above the roof line. Attach the coax to the antenna analyzer, take measurements of resonance and the 2:1 SWR points. Loosen the rope on the side of the house and get the antenna past the roof line. Untie the other rope from the playground. Lower the antenna to the ground. Untangle it all and drag it into the basement so it can be trimmed. Repeat this process.

Honestly, I didn't think I would have to do this eleven times! Both 12m and 17m were resonating slightly low, so I carefully trimmed the 12m portion until resonance occurred in the band. Then repeated the process with 17m. Once each band was in the right place, I soldered the connections between the wire and the trap. Trimming past the trap does not appear to affect the inner doublet, which is what we expect.

30m was a problem. It was resonating slightly high. I ended up splicing in about 14" of wire, split between the two legs. This was too much. In the end, I only needed 3 1/2" added, and that only to one unbalanced leg.

Managed to work ZD7BG just a few minutes after getting this antenna back up on the tower. It's really nice to have a working antenna again. Now the next repair is to figure out what's wrong with the A3S.

Friday, September 22, 2017

Life after 5BDXCC

For the last several years, I've started off the late summer thinking about my plans to complete 5BDXCC. This has meant figuring how who to make more DX contacts on 80m, whether that meant improving antennas or just being on the air more.

But, this June, I managed to wrangle enough confirmations to qualify. I'll likely submit these in November.

As that time of the season rolls around again, I have to think - what's next? What comes after 5BDXCC? It's a good question.

I still have a few other bands to add. On 17m, I have 87/87 confirmations, so that's likely next. It will be easier now that I've identified the problem with my WARC dipole. ON 12m, I have 73/73 confirmations is not that far behind, although old Sol is not likely to be active enough to sustain much propagation on that band for a few years.

160m seems less likely, where I have 42/41 confirmations. Given how hard 80m was, it may take quite a while to add many more entities there.

The DXCC challenge is still unclaimed. I have 835 current credits, and my unapplied LoTW confirmations will bring me up to 1125, if I applied them all. Plus at least five more QSLs I'll submit this year could get me all the way to 1130. I've decided a couple of years ago not to push for this too early. The 80m DXCC submission this year, plus all the band endorsements should get me over 950 credits. I'll make it over the 1000 credit requirement as soon as I complete another band.

And, of course, there's the DXCC Honor Roll. At the moment, I have 259 current confirmations, which puts be 80 away from the 339 total. And 71 away from the 330 required to make Honor Roll.  That may take quiet a while. I have two cards that I'll submit this year, and there's a dozen more entities I've worked that I need to obtain confirmations from.

Naturally, some of these remaining entities are increasingly rare. The good news is there are several exciting DXpeditions this year. I'll have to make plans to work them when I can.

So, yeah, still a lot of challenges left in DXCC.

Monday, September 4, 2017

Back in the Harness Again

Yours truly on the tower.
When I moved back from Floyd County a couple of years ago, I did a tower inspection. When I got to the top section joint, about 10 feet above the roofline, I lost my nerve. I checked the bolts, then came slowly down the tower. I haven't climbed since. Until today.

I really had to. All the antennas on the top of the tower show signs of being messed up. Mission today was the WARC dipole. While still resonant on 12m, 17m showed no match, and 30m was way out of band.

I took the ends of the dipole loose, you can see two of the traps around my foot level on the tower. Then the climb to the 44 foot level of the tower. The center of the dipole is a NEMA box U-bolted to the tower leg.

Being up there is a little unnerving at first. You try to stay as still as possible, because each move causes the tower to sway.

Of course, as soon as I got up there, I find I don't have a screwdriver. This is why you always use a ground crew. My daughter Lauren did an excellent job in this department.

With the lid to the NEMA box off, took off the U-bolt, and lowered the dipole to the bottom of the tower.

Once back on the ground, a quick inspection showed what was wrong. One of the 17m traps had a clear open, and it wasn't something that was easily fixed. After 12  years on the tower, the shield braid from the trap coax had disintegrated. The trap needs to be rebuilt. It's companion doesn't look much better. In the interest of keeping them symmetrical, I decided to rebuild both the 17m traps. The 12m traps are a couple of years younger, and look to be in good shape.

After I get the traps rebuilt, it will be back up the tower to re-install. And then I need to lower the A3S down so I can check out the driven element -- figure out why I've lost resonance on 15m and 40m is so messed up.

Monday, July 24, 2017

Forty Years of Personal Computing - Prelude

1971 Logix 0-600 "Electronic" Computer kit
When you are young, you rarely realize how smart your parents were. Mine bought me all kinds of educational toys. One birthday, I got an  optics experiment lab that culminated with building a working 35mm camera. I did every step in that lab, and eventually built the camera. OK, it was inexpensive plastic, yes. And while I never did manage to buy film for that camera, I learned all about concave, convex, chromatic aberration and other principles of optics.

One of my favorite educational toys was a subscription to "Things of Science." Every month, I'd receive a little envelope or a blue box with science stuff. There would be a small booklet, often with plastic or cardboard parts in the box. The people who came up with this were geniuses, as they managed to package a lot of science into a tiny, inexpensive kit every month.

It was hit or miss each month. Sometimes, the article wouldn't interest me, or I'd just not be patient enough to do the experiments. Sometimes, it would be really cool -- like they'd send enough chemicals to grow a small plant hydroponically. Or send some plastic lenses in cardboard holders to do optics experiments. I played with those lenses for months and months.

And while not every kit was a hit, it at least got me to study some element of science for a few hours each month.

The actual "computer", sans wiring.
A couple of years ago, I found something from my childhood in my father's basement. It was a 1971 Logix 0-600 "Electronic" Computer kit. I use the quotes only because this kit contained no electronic components. It's just a battery holder and ten lamps, a SPST pushbutton, and ten ganged ten-way SPST switches. The switches provide data input, the lamps show the output, and the switch connections are wired by sticking pieces of wire into holes that make contact with metal clips -- very much like a modern solder-less breadboard.

The first part of the manual contains assembly instructions. These are easy, provided you can turn a screwdriver and strip a wire. I must have been 11 or 12 when I received this kit, and I did the assembly entirely myself up in my room.

The manual is the brilliant
part of the whole kit.
This kit was really brilliant. Despite the lack of any electronics, the manual shows how to build 39 different "programs." The switches were used to enter the input, and tissue-paper overlays are slipped into the transparent amber lamp cover to label the output. Each program had a clear wiring diagram.

All you had to do is read the program description, and follow the diagram.  So long as you wired it up correctly, the program would work as advertised. The ganged switches were sometimes a little flaky, but not enough to detract from the experiment.

Through the series of programs, the kit teaches some of the basic logic and computer principles -- AND gates, OR gates, NOT gates, Addition, Counting, binary numbers, Venn diagrams and more.

I do remember working through all the programs, and even trying to develop a few myself. At one point, I figured out how to wire up a program, but realized it required a 12-way ganged switch, instead of 10-way. It was beyond the capabilities of the 0-600.

After I ran through the programs in this kit, I don't remember using this machine much more. But it was a real learning experience. The experiments taught me a number of important logic and computer concepts that I later used in my education and in my career. Having it now brings back fond memories.


Friday, June 16, 2017

5BDXCC - YES!

For the last four low-band seasons, I've been hoping I'd capture enough confirmations for 5BDXCC. I had already confirmed 40, 20, 15 and 10m DXCC, I just needed to get to the magic 100 entities on 80m. As I wrote in the fall, I just needed a few more to put me over the top.

At that time, I had 91/90 confirmed on LoTW, plus a couple of cards.

Today, I received a confirmation on 80m from CE2VMF, bringing my totals to 98/97 on LoTW. Plus, I have THREE cards with additional confirmations, which brings the overall total to 101/100.

I remember talking to Jim Streible K4DLI (who now, unfortunately, is a silent key) about his 5BDXCC plaque. Jim said he had spent 35 years earning that award. And Jim earned it back in the day before electronic confirmations, too.

So, when I apply for DXCC awards later this year, I'll be petitioning for 6BDXCC (got 30m, too!).

I'll have to seriously consider ordering the plaque.

Tuesday, May 23, 2017

PIC-based Push-button K9AY Controller

Push-button K9AY controller works!
I can't believe I have been working on this project for three years. I wrote about it earlier, describing the project and how it didn't work, because of the bouncy mechanical switches.

Yes, I bought the PIC16F1503 chips. It also took a while to figure out how to program them.

Conceptually, the software is simple. I program the Timer1 timer to fire every millisecond. Then the cpu loops and waits for an interrupt.

When the interrupt fires, I gather the state of all four of the direction buttons from PortA, and shift each into a byte in memory. This gives me the state of each button for the last eight milliseconds.

When a button is pressed, even if it bounces a bit, it will eventually transition to an on state. If it holds that state for seven milliseconds, it will result in a distinctive pattern: 01111111.

That's what I look for. If a button has this pattern on any interrupt, I consider it to be on. Otherwise, it is off. Take the value of all four buttons and jam them together side by side, and you have a number between 0 and 15. Sixteen values in a lookup table. Look up the current value. If it is zero, do nothing. If non-zero, I write the value out to PortC. The PortC bits control the LEDs on the buttons as well as the control lines for the two relays which select the control voltage for the K9AY loops.

All the work is in the lookup table. It has all zeros except for the four entries where one button has been pressed -- one each for the four directions.

It's all less than 100 instructions.

Most importantly, it works!

My prototype has wiring that's a little flaky. The board flexes a little bit when the buttons are pressed and sometimes the wiring moves a bit and shorts out. I'm thinking that now that I have something that works, I should make a printed circuit board.

Friday, March 31, 2017

Radio 2017 - Progress March

I don't want to admit it, but no, not a lot of progress on this. No long vacation excuses this time, just busy with life. What can I say?

I did get to review the WWDC videos regarding storyboards. I (re-)discovered a few interesting things:
  • Set the main storyboard in the info.plist - I had done this.
  • You can build up window content by using content segues to multiple view controllers. I didn't have a need for this yet, but I'm sure it's coming.
  • Control-Drag directly from a menu or button to a window content view. This will produce a segue to the view controller. Easy and useful. The host view controller receives a method call prepareForSegue:sender: before the window appears for the segue. I'm not using this because there doesn't seem to be a way to associate the document window for this action when it appears on the application menu. 
So, my main question became -- how does all the other document stuff work? When you select File / Save, how does that get to the NSPersistentDocument class? Turns out, it uses a basic target/action to the First Responder. 

I just had to figure out is how to add my action method for my modes window to the First Responder. That turns out to be easy. Select the First Responder icon of the main menu. Then, use the Attributes inspector, and you'll see a list of User-Defined actions. The weird thing is, though, once you've defined them, they disappear from this list.
No matter, just add an action, then edit it to be the method name you want. Then wire it up to the menu with control-drag as before. Implement the method anywhere in the responder chain. I implemented this in my document class. And my mode window comes up correctly, and I have a unique instance of it for each document. Cool.

Next is the hard part -- trying to figure out how to associate the managed object context of the document with the modes view controller. This is were it gets weird. The window controller has a property document which points to the document for that window. How you get to that value from the view controller is something I have not figured out yet. 

Tuesday, March 28, 2017

Cycle 25

Smoothed SSN of the last six sunspot cycles.
Cycle 24 sucked.

I mean, really. Look at this graph. It's pretty clear that Cycle 24 was barely half as tall as the three cycles before it. It wasn't even as good as Cycle 20.

Smoothed SSN in Cycle 24 never made it over 82. The previous four cycles had several years over 100.

And that was after a sunspot low that lasted a couple of extra years before Cycle 24 began an upswing.

Now we're in that downward slide again. I know from experience that we're not seeing the end of the lull between cycles until the 2800 MHz solar flux gets down to 66 or 65. That's probably three years away. And then it will take a couple of years to build up. I'm thinking it will be about 2022 before Cycle 25 shows it's colors.

In the meantime, propagation on 160m and 80m should be pretty good in the winter months for the next several years.

I've told my wife, I want to be ready for Cycle 25. I want to have my antennas and station ready to rock when the sunspots come back.


Tuesday, February 28, 2017

Radio 2017 - Progress February

Well, I can't say that I've made much of any progress since the last posting. I took a 10-day trip to Israel to visit Bethlehem, Masada, Galilee, Jericho, Jerusalem -- all the holy sites. The trip was very intensive, and after I returned home it took me about a week to recover and then catch up on all my missed work.

I did spend a little time going over my test programs trying to figure out the right way to write Cocoa code for MacOS X document-based apps with Storyboards. What I'm doing seems too complicated.

So, I've decided I need to go back and find the WWDC videos regarding using storyboards with MacOS X, and it also might be useful to view the the ones for iOS docment-based apps as well, to see if that gives me any insight.

With any luck, I'll make much more progress in March.

The Station Notebook

When I originally set up my station as a novice, I didn't make a lot of notes. I wrote a few things on the back-side of my logbook pages: radios I used, antennas I put up. But, for the most part, I freely made changes to my setup without a lot of documentation.

Some years later, I had a small steno-style notebook lying around, and I used it to record certain information. Like the exact color-code of the RS-232 cables I wired up for my computer. That sort of thing was handy to have written down, because it saved a lot of reverse-engineering later I went to make another cable. But that was about it.

Perhaps a dozen years ago, I read a couple of articles about James Lawson, W2PV, and how he kept copious notebooks about his station. He even solved a sticky problem during a contest, because he kept notes about his antenna installation.

After that, I started writing more things down in my little Steno notebook. Most of it was changes I made to the station -- radios or antenna modifications. Or if something big happened -- like I blew up the amplifier. I recorded a lot more details, and I put dates on everything.

But when I began to spend most of my time out in Floyd County it put my little steno notebook out of reach. I had a new station and I still wanted to keep notes. My solution was dead simple -- I created a text document using TextEdit. I called it Floyd County Station Log.

Looking at it today, it's more like a diary. Each month, I put in a new heading: March 2017. And below that I record anything interesting that I do. I record the things that work, and I also record the things that don't work! If I don't do anything that month, I'll just change the heading to the next month.

The bottom of this document has a list of projects to be worked on. Silly now, since I'm no longer at that QTH: 6m antenna, K9AY controller, 20m and 30m traps for the dipole, tribander-Moxon  times 3 (now that's an interesting project I'll need to write about some day!), remote antenna switch, etc.

This method worked so well, in May 2015, I adopted it for my main station log. Not only did I make new entries, but I also entered all the data from the steno notebook. Plus, I went back through all of my contest comment submissions and recorded all the changes I had made to the station from November 1994 to date.

Today the Gwinnett County Station Log is a reasonably complete record of all the things I've done with the station for the last 22+ years. It has six pages of project notes at the end. Lots of stuff to work on! And, yes, I also have a smaller document for the Walton County Station Log.

I guess my point is that having a station notebook is extremely valuable. It doesn't matter if you use a small steno notebook, a spiral-bound notebook, a composition book with graph paper, a loose-leaf notebook, or a computer document. The important thing is to find a style that works for you to keep notes. After a while, there's just too much to try to remember. So keep good notes.



Tuesday, January 31, 2017

Radio 2017 - Progress January

Quick progress report on the Radio 2017 project. I've managed to get started on the casual logging program. I've created the basic framework and a local Git repository. An early cut of the Core Data model is in place. You can't do much more than run the program and create, save and close documents so far.

I have discovered that Storyboards on MacOS don't quite seem ready for prime time. First, there's a bug that windows don't cascade when creating a new document. Second, the split of the original NSDocument class into NSViewController and NSPersistentDocument subclasses creates something of a conundrum. The NSPersistentDocument owns the Managed Object Context, but the NSViewController subclass is the place that owns all of the UI actions.

To make things worse, the main menus actions only appear to be able to effectively target the application delegate. If you want to target something in the NSViewController or NSDocument, you have to write a handler in the application delegate to find the current document view controller and call the right method.

I've figured out the proper way to do the menu actions, but I'm still puzzling out how to handle the managed object context from the view controller actions.

None of this is rocket science, but it's certainly slowed things down, as I have had to create separate projects in order to figure out the most appropriate way to do things.

I'm still learning Swift. The nuances of optionals and when to use the various operators is something that will come with practice. The compiler is keeping me on the straight and narrow.

Monday, January 23, 2017

5BDXCC? - Well, Not Really.

Back in December, I made my annual submission to the DXCC desk. In addition to adding a few endorsement entities to my DXCC Mixed, CW, Phone, Digital, 40m, 20m, 15m and 10m, I also submitted enough credits for DXCC 30m as well.

Let's see, thats 40m, 30m, 20m, 15m, and 10m. That's five bands! 5BDXCC, right?

No, not really. 5BDXCC requires DXCC on 80m, which I haven't accomplished yet. Getting there.