Thursday, May 19, 2011

Project Support Parts

I just received new SMT parts from Mouser, these will be used as Support Parts in many of my small projects. These parts are not expensive, especially when they are purchased in bulk (i.e., 10 or 25 at a time).

As previously stated, my Ham Radio Goals, include; I continue to build projects with the smallest parts that I can find.

Project Support Parts

The Coin Cell Battery Holder will be used for low Voltage projects, currently I have an idea for a Key and Keyer that incorporates many published ideas.

The very small push button switch in the center will be used in place of much larger switches that I have previously used in my projects. This switch is about 1/4 of the size of previous used switches, which were more like the lower right switch. This new little switch has a nice snappy feel, it is very small but it provides an easy to press 1/8th inch button (which can not be see very well in the photo). This will become my switch of choice.

I purchased the two push button switches (lower-right) because they were similar to the salvaged switches that I had previously been using, each switch provides dual contacts for the single action. I may not use these much now that I have seen and like the new center switch. The lower-center button is typically seen on small CPU evaluation boards and used as a program set/reset. This switch is physically the smallest switch of the group in volume, but the center switch still has the smallest foot print.

The 1/8 inch stereo connector, photo left, will be used instead of salvaged connectors from old sound cards. This new connector is a bit longer than previously used, but it incorporates a switched element. It will be used for all audio input/output, and/or a CW key input.

The DPDT Switch on the center-right will be used as a mode switch on my beacon projects. It is small enough to easily incorporated into my micro beacons.

The 4 Gang SPST will be used for power ON/OFF and mode configurations for small projects.

Data Sheets and Price for these parts can be found on my Mouser Bulk Order 5/12 Project Cart.


Wednesday, May 18, 2011

DorkbotPDX PCBs

On a previous post, I reported the results of a test of PCB Manufacturing located in China, which is a very good value. My trial of ten 1x2 inch boards was $16.00 and a 22 days wait. That is, the boards cost me $1.60 each.

Recently I found local (US) company "DorkbotPDX" that combines many projects into a full manufacturing PCB panel, and then they send the panel to a Lab for production. The web page can be seen at:

Update: This manufacture has change their name to OSH Park

First DorkbotPDX Test

For a test, I used the next Revision of the same board used for the previous Production test. I am very impressed with the results. But, in my hast to make the lay up deadline, this Revision of my board did not have the right holes sizes for some of the parts (the results of a multitude of issues; with a new Software Revision and Part Foot Print problems). But, what the heck, . . . this was just a test.

The board is very nicely Gold plated, and as they state on their web page, "You can have any color of Solder Mask that you like, as long as it is Purple".

The price of $5.00 per square inch, which includes shipping and three copies. Once sent to the Lab, the production and shipping time is about 14 days. A collection of projects are combined onto a panel about one each week. Their lay up web site, and Phone App, reports panel lay up progress and shipping status.

For this test, these boards cost me $3.25 each, which is a little more expensive than before, but eight days faster.

The price and turn around time is great for small quantity of prototype boards.


Sunday, May 8, 2011

A CPU State Machine

This is not exactly Ham Radio, but it does help satisfy my need to do some programming.

Late last year I saw a YouTube Video showing the efforts of a MineCraft Gamer building an Arithmetic Login Unit (ALU) using gaming blocks. It was very cool. Later I read Mark's (of description of a simple ALU.

Recently I have been playing with a program called DSPExplorer by Ward - AE6TY (see previous post). The program is normally used to interactively create and explore DSP circuits and functions. The program is a very small Java app and it runs on most systems; MS Windows, Mac, and Linux.

While working with the program, it occurred to me that I should be able to build some simple logic circuits, and then connect them together and watch the logic flow. DSPExplorer is intended to provide easy manipulation of arrays of double precession numbers and has support for imaginary numbers used with FFT and DSP computation. It is a very comprehensive software package.

But, DSPExplorer does NOT support boolean or bit operations, it does NOT have a proper "int" data type.

For my first simple logic experiments, double precession integers work fairly well, for constructing logic blocks that could be interactive connected together to form complex logic operations. I built the usual FlipFlops, AND, OR, and other gates. It was fun watching the DSPExplorer Simulation work it's magic with the defined modules.

Then I remembered the MineCraft Videos, if they can build a simple but complete ALU using "rocks and fire", I should be able to do the same with DSPExplorer programmable interactive math units.

I decided to build an ALU. After many frustration attempts, to do bit operations with double precession number without bit operators, I finally got it to work. I now have a DSPExplorer Logic block made using double precession math (it was much harder than I had thought). If I remember correctly, Ward mentioned that "double" is loosely defined to be as large as the user requires (I think). That means my ALU could potentially be very wide.

ALU (center right) and State Test Modules (upper left) are Shown

The current defined ALU's width is user specified and it implements, the following:
  • X input
  • Y input
  • Flag input (which a 6-bit word that controls the action)
  • O output (the Accumulator)
Most ALU functions are provided, the same as the MineCraft implementation.

Later, decided to expand my initial effort.

Now I have a ROM block to hold Assemble programs, and a State Machine (or CPU) for program flow control. Currently the CPU implements simple: NOP, HALT, JMP, GOTO, PAUSE and ALU OPs, with the output going to the ALU's Flag input word. But, I have not figured out how to return the ALU Accumulator value, but that will be later.

The CPU reads the ROM, and executes the programmed instructions. A partial ROM example is shown on the right, the lower part is a section of Assembly program to be executed by the implemented CPU.

Soon, I should have all of the modules; ROM, RAM, StateMachine, ALU, Buffers,  all needed to implement a more complex system.

With it I may be able to implement a simple compiler, if so, I could write an implementation of java, and therefore, run Ward's DSPExplorer from within this simulation, . . .  Nah, . .  just kidding!

So, why did I do this? Because it was fun, and I am learning a lot about the low level CPU internals.

Note: This is a very Perverse use of the Ward's wonderful DSPExplorer Software, Sorry about that Ward, I will return your program to it's normal use, later.


SMite via SimSmith

A pQRP friend and fellow builder, Ilkka - KF7LDX, has been Homebrewing a Knightlite SMiTe Build, see his writeup at:

The original circuit and documentation can be found at:

As seen by my previous posts, I have been playing with the SimSmith program. Ilkka's build provided an opportunity for me to model another output tanks circuit. I have been collecting data, and modeling the circuits that I find. I layout the components, insert their values, estimate the driver impedance, and let the SimSmith Program do the work.

The Circuit and Impedance Match

As per SimSmith, it appears that the published SMiTe circuit provides one of the best matches that I have found, assuming the PA 2N2222 drive impedance of about 100 ohms.

The computed SWR and Second Harmonic Suppression

As can been seen, the output circuit provides a good match at the Oscillator Frequency (solid line) and good suppression at the Second Harmonic (dotted line).

I would not mind building this circuit.

LOAD R:50.0 X:0.0 load:
SHUNT_CAP Fd:820p Q:1.0K @MHz:0.0
SERIES_PARALLEL_TRAP Fd:220p Hn:2.2u Q:1.0K @MHz:0.0
SHUNT_CAP Fd:820p Q:1.0K @MHz:0.0
SERIES_CAP Fd:.01u Q:1.0K @MHz:0.0
SHUNT_IND Hn:22u Q:1.0K @MHz:0.0
GENERATOR MHz:3.6864 R:100 X:0.0
CONTROLLER lower__Scan__Freq:1.0 upper__Scan__Freq:9.0 #__of__Segments:500.0 SWR__R:100 SWR__X:0.0 SWR:1.2


Saturday, May 7, 2011

Standwood Hamfest 2011

The Standwood -W7PIG- HamFest has come and gone.

Tess (my dog) and I was there early and sitting in the HamFest parking lot before 9:00AM. While my Van's motor is not running I enjoyed operating the FT-817 on the bands below 2 meters.

About a year ago, I had some repairs done to my Van's engine. The repair shop felt it was necessary to replace the ignition wires. I did not think to check HF radio operation, as I did not have my FT-817 at the time, Although, afterwards I did notice a little new ignition noise on the AM radio.

So, for now (with maybe solid, noisy ignition wires), I only operated on HF when the Van is stopped. Which may be OK, as I do not like driving down the road with the tip of my 40 meter Hustler Antenna up at 14 plus feet (it is mount on the top rear).

I was surprised, 20 meter was jumping, I made several contacts. There was some sort of contest going on, people looking for Washington counties. My 5 watt FT-817 did me proud, maybe my tuning-up of the antenna a few days earlier paid off.

I may ask the Repair Shop if they have access to resistive "noise suppression" ignition wire, I am due for another tune-up.

Back to the HamFest.

Standwood HamFest

A lot of enthusiastic Hams showed up,  we were some of the first few in the door. I did a quick turn around the floor and found a medium size box full of Altoid Tins - For FREE. The fellow said he had several other boxes about the same size if more were needed. I added six Tins to my collection. Later, I noticed the whole box was gone.

The Full Box of Altoids Tins (-6)
At a good price!

As usual, I found the Air Variable Cap, there were some nice ones, but I again resisted the need.

Air Variable Caps

After several trips around the floor, I found my next purchase, it was setting on the table just above the Altoids Tins, I guess I'd missed it before.  I am now the owner of a dozen 12x12 inch sheets of single-sided Copper Clad - all for only $6.00. I think it must be 1/2 oz copper, but I have not checked for sure. What ever it is, it will be useful.

Standwood Washington is only about 30 minutes from home, so the trip was not expensive.

Actually, that is NOT true!  While driving up the freeway to Standwood, I noticed a slight vibration on the steering wheel. I had noticed it before, but thought it was only very cold stiff winter tires, that had not been rolling much.

After, the HamFest and near home, I went by Les Schwab to have things checked. One of the front tires had some of it's steel belting sticking out, near the tread edge. I check my tires regularly, and at the last check, the tread was showing some ware, but nothing to be concerned about.

After two new front tires were installed, and while still in the Les Schwab parking lot, the local High School ROTC Kids were holding a car wash, the Van needed a wash. I let Tess out while they washed and we were waiting. Tess loves kids, of all sizes, and they had water, and her attention! I am not sure who enjoyed it the most, Tess or the Kids? They washed cars for donations, so it was another $10.00, but it was worth it. Here I have called them, Kids, but in reality, they were Young Adults.

So, this trip did, in fact, cost me more than just my $6.00 for PCB material. That is, an additional $10.00 plus, $415.70 for tires. But now, the Van is clean, and rolls very smooooooth, even at 75mph.


Tuesday, May 3, 2011

Matching Network Problem/Solution

From my previous post you can see I have been working/playing with Smith Charts and having fun learning about matching networks. Several posts contain additional information.

For most people, reading about Smith Charts Problems will make your head hurt, including mine.

Wikipedia has a very intense page on Smith Charts, with several examples and the supporting math.

After reading and agonizing over Wikipedia's solution to their 100mHz "lumped element circuit" example, I decided to try to solve the same problem with SimSmith.

The task became trivial.

The SimSmith Solution:
---Cut, Save, Load into SimSmith----------------
LOAD R:50.0 X:0.0 load:
SERIES_CAP Fd:40p Q:1.0K @MHz:0.0
SHUNT_IND Hn:53n Q:1.0K @MHz:0.0
SERIES_CAP Fd:138p Q:1.0K @MHz:0.0
SHUNT_CAP Fd:36p Q:1.0K @MHz:0.0
GENERATOR MHz:100.0 R:50.0 X:0.0
CONTROLLER lower__Scan__Freq:50 upper__Scan__Freq:250 #__of__Segments:500.0 SWR__R:50.0 SWR__X:0.0 SWR:1.0

For us Hams, the above solution would not be optimal for a Transmitter Output Tank Circuit, it lacks Second Harmonic Suppression. With the addition of a Low Q Trap, the following solution provides additional 16db of second harmonic rejection.

As shown here, the dotted cursor set at 200mHz, read the results below the graph.

But then, of course 100Mhz is not a Ham Frequency :-)

---Cut, Save, Load into SimSmith----------------
OAD R:50.0 X:0.0 load:
SERIES_CAP Fd:44p Q:1.0K @MHz:0.0
SHUNT_IND Hn:68n Q:1.0K @MHz:0.0
SERIES_CAP Fd:194p Q:1.0K @MHz:0.0
SHUNT_CAP Fd:36p Q:1.0K @MHz:0.0
SERIES_PARALLEL_TRAP Fd:20p Hn:31n Q:10 @MHz:0.0
GENERATOR MHz:100.0 R:50.0 X:0.0
CONTROLLER lower__Scan__Freq:50 upper__Scan__Freq:250 #__of__Segments:500.0 SWR__R:50.0 SWR__X:0.0 SWR:1.0


Sunday, May 1, 2011

SimSmith Design for Hi Z Tube Tank CIrcuit

Peak the GRID   -   Increase the LOAD   -   Dip the PLATE

  1. How many thousands of times have you done that on older Tube Transmitters? - Many.
  2. Did you know what you were really doing? - No, just following the documented procedure.
  3. Is the adjustment order important? - Yes, maybe, but did you know why.
  4. When building Transmitters, Did you know why a particular part value was used in a Tank Circuit? - No, just followed other previous designs.

The last few days I have been working with the new SimSmith program, another program by Ward - AE6TY, see previous blog.

With SimSmith I can experiment with the a Transmitters Output Tank Circuit, to see and understand the importance of each component and the effects of each.

I have always had questions about the magic of a Tank Circuit.

For example;
  • How important is the "Low Q Harmonic Suppression Inductor" on the resulting Harmonic Suppression. 
  • What if I had (or could) NOT use the exact value as suggested in a build instructions?
  • How important is the Tank Values on Harmonic Suppression?
  • Does the value of the DC blocking Capacitor effect the suggested component values for the rest of the Tank Circuit?
  • How can a match be achieved, for all bands with only the Plate Cap, the Loading Cap, and a few Standard Values for the Band Select Inductor?
  • What range of antenna miss-match can the Tank Circuit correct?
  • What is the loss when there is a miss match?
  • How important is the Antenna to coax match?
  • Does the length of the coax really matter?

With SimSmith, I now know most of the answers, and can experiment to learn, and ask more.

The following table of numbers are for a "Tube Tank Circuit" and were obtained by using SimSmith within about 30 minute. I have not checked these results against standard published circuits, but I think they are close, or at least a starting point for further experiments.

The circuit was modeled after a standard Tube Style Output Circuit, 10K ohm Impedance at the plate  matched to a 50 ohm load.

The Tank Circuit Schematic
The Elements in order left-to-right are:
Load - Coax - Ant Load Cap - Ind - Plate Tuning Cap - Bulk Ind - DC Blocking Cap - Harmonic Sup - Src

Note: on SimSmith the layout of components are with power flowing right-to-left ( << ), and impedance is defined left-to-right ( >>, it is a standard Smith Chart and Engineering thing).

The table include the older standard Bands and 30 meters (my favorite band). The other newer bands could have been easily included.

Low 10Q
Var Cap
Var Cap
1.9MHz 100uH.01uF2.8mH43pF180uH450pF50+j019.6db
3.75MHz 100uH.01uF450uH45.2pF43uH523pF50+j025.0db
7.25MHz 100uH.01uF450uH12.1pF20uH142.5pF50+j025.7db
10.14MHz 100uH.01uF120uH13.1pF20uH43.7pF50+j030.5db
14.25MHz 100uH.01uF120uH7.45pF18uH62.7pF50+j028.3db
21.4MHz 100uH.01uF0uH10.6pF5.6uH212.1pF50+j038.8db
28.25MHz 100uH.01uF0uH10.9pF3.0uH359.1pF50+j046.7db
Do not design a project with these numbers, check the values for yourself

Note: With additional effort, a slightly better matches can be achieved, for this experimental purpose I was looking for anything better than a 1.2 to 1 ratio.

Working with SimSmith allows for quick and easy adjustments of component values for  "what if" games to be played. The following SWR Plot shows the results, similar to that expected or obtained while tuning a rig with meters.

30 Meter, 10.140MHz SWR Plot
SWR Value, and Power Ratio (In/Out, Red Dotted line) on the Right
With Black Dotted Cursor placed at the Second Harmonic Freq

The following Smith Chart Plot shows how each component contributes to the match and it circular impedance significants. Without a Smith Chart Plot component interaction is difficult to understand.

Smith Chart, Showing Match at Center 50 ohms
Hi Resistance Values are on the left, Low Values are on the Right
With 50 ohms in the Center
Colors are the same as corresponding schematic elements above

The Transmitting Tube High Impedance (source) is represented by the "dot" on the left, the 50 ohm load is represented by the "X" in the center of the Smith Chart Plot. It is easy to see the significants of the Plate Tuning Cap (the long light blue arc). All reactive components plot as an arced line parallel to the arc lines of Smith Chart Grid. Knowing that component reactance values change along an arc, makes adjustments easy to understand and predict.

I plan to repeat this experiment with a modern Transistor or FET output circuit, where source impedances are much lower (typically less the 20 ohms).

In the future, SimSmith with be used when I build any transmitting or impedance matching circuits.

It is easy, no longer do you have to let a complex Multi Frequency Smith Chart scare you off (as seen below). In this view, the point of interest is that of the last schematic element (line colored blue, same as schematic) which runs through the center for a 50 ohm match.

This Multi Frequency Chart is Interesting
But not necessary to understand normal usage of the Software

SimSmith is very interactive. Give it a try, see if you agree, you will learn a lot about your Tanks.

Thanks, Ward for providing the Software.

If you would like to try the 10.14mHz configuration, cut and save between the lines, and then load the file into SimSmith.

LOAD R:50.0 X:0.0 load:
SERIES_TLINE deg:554.56 @MHz:10.0 ft:100 vf:0.66 Zo:50.0 /100f:0.5 @frq:10.0
SHUNT_CAP Fd:43p Q:1.0K @MHz:0.0
SERIES_IND Hn:20u Q:1.0K @MHz:0.0
SHUNT_CAP Fd:13.167p Q:1.0K @MHz:0.0
SERIES_IND Hn:120u Q:1.0K @MHz:0.0
SERIES_CAP Fd:.01u Q:1.0K @MHz:0.0
SERIES_IND Hn:100u Q:10 @MHz:0.0
GENERATOR MHz:10.14 R:10000 X:100.0
CONTROLLER lower__Scan__Freq:5 upper__Scan__Freq:25 #__of__Segments:500.0 SWR__R:10000 SWR__X:0.0 SWR:1.2