Friday, November 27, 2009

Frustrations with Audio

Frustration with Linux Audio System is getting the best of me.

I have been working with Large UNIX Systems for more than 30 years - I have specified, configured, installed and managed many multi-multi-million dollar UNIX systems and nothing is as frustrating as "audio set up and configuration" on a simple Linux system - currently I am using Ubuntu 8.10 with many (maybe conflicting) audio subsystems. I started out looking at  and using Linux for a personal workstation when it was still called Minx.

My problem is there does not appear to be a simple solution to this problem. I do not like MS Windows, but at least the sound system on most Windows systems - just WORKS.

Several years ago, with an older Linux (Red Hat) system - sound was not easy to configure, but it was not difficult either. Now, everything about a Liunx sound system is just plain HARD.

Currently, I would grade the XP Windows Sound System with an "B+", and would grade the Linux Sound System (Ubuntu 8.10) with an "F".

I have OSS, ALSA and PulseAudio all installed which maybe the problem, but nothing suggests that another configuration is more correct.

One minute the Linux sound system appears to work, then the next nothing works, a reboot is necessary to correct the problem, restarting the daemons does not seem to fix the problems. Simple configuration change do not take place when you expect, they may take effect sometime later, when you are not expecting changes. It is difficult to keep up with what is happening to the sound system.

The Linux Sound system is all magic, there is nothing that I have found that diagrams the sound system and attempts to provide help when things do not work.  The forums describe the Linux sound system as delta's on previous information - it is hard to collect enough information to configure the sound system - when it does not work as published.

I would like a simple app that diagrams the sound system from input jacks, to kernel, to application and onto output jacks, with diagnostics showing where sound is coming from and going to.

I am sure for new Linux users, this is as frustrating as anything.  It will cause users to return to MS Window. From what I have seen, the Linux Sound System is the most effective externally generated advertisement for Microsoft that exists to date.

Sorry for the Rant, .  .  .  .  maybe I will figure it out someday.


WSPR 2.0 Working at Remote Location


I now have WSPR 2.0 working on my remote Data Centre Direct Conversion Receiver.

The DC receiver is used as both a QRSS ARGO Grabber and now a WSPR spot database reporter, both at the same time.

The audio is decode by the two programs, both running on Ubuntu 8.10, WSPR runs native on the console window, while ARGO runs under "wine" in a virtual (xvnc4server) window.  The virtual window is necessary as ARGO uses screen scrapes (or captures) to obtain images, that are used for the Web Server Grabber presentation. A non-virtual window application would often be covered by pop-up's or other normal window elements that obliterates the resulting Grabber image.

WSPR 2.0 supports CALLs with a suffix, for my remote receiver I use WA0UWH/R as it's ID, as shown near Seattle on the WSPR map.

My goal is to also report PropNet spots via the same systems.

While I was setting up WSPR (as described above), I did an experiment. I inserted a amall transformer into the computer sound card input line - the "hum bars" seen on the ARGO and WSPR console windows went away! I think a ground loop exists between the computer and the antenna. The coax is grounded at the base of the antenna.

Now I need to redesign the receiver output circuit to include proper ground loop isolation (a fix to be done another day).


Wednesday, November 25, 2009

PCB Software Review

I just did a quick review of AutoTRAX for my potential use to make Hobby Toner Transfer PCB's.

It was looking good until I looked at the copper pour implementation.

See previous reviews.


Sunday, November 22, 2009

November 2009 - eBuild-A-Thon and BBQ

WOW, the November 2009 - eBuild-A-Thon and BBQ has come and gone, we had several projects being worked in different parts of the Shop at the same time -  I could not keep up with all of the very interesting projects, activities and chats. Sorry, photos were not taken.

The QRSS "Beac-in" (not talk-in) beacon was not found, no one brought DF equipment - therefore the prize will be saved for the next event in December.

By my count, there were 22 people that came by to work on projects or just chat throughout the day. Unfortunately, I did not get a complete list of names and calls, forgive me if I don't mention your name or project. Next time I will try to have a official "meet and greeter", to help with names, notes and project photos.

Roger K7RXV brought his network analyser, but I think he worked mostly on a PCB case for an Intermod Meter.

John KC7NVE brought his Transceiver, Portable Antenna and Laptop, to demonstrate station operations in; SSB, CW, and WSPR modes.

Doug W7RDP brought his Lineman Crimping tool for demonstration and use, Doug brought  a section of very small (1/8 inch) hard-line. It will be used for future project intra case connections.

Lyle KK7P worked on a PCB case for one of his  projects, by the count of the accumulated  punched circles at the base of the punch he was a very busy person. I am looking forward to see the resulting case and finished project.

The Next Day Photo
I love the evidence of intent work

Thanks to Lyle, we will now have many large "manhattan style" PCB construction components.

Alan K6ZY worked on repair of a watt meter via meter replacement and face plate adaptor.

For a while, there was competition for the punch access with so many PCB projects needing holes punched. Die changes were coordinated for best overall effective use.

Thanks to my sister, Carolyn, Lunch was served adhoc and projects were continued though out the day.

Tess (my dog) thought it was great - she had so many people that were willing played ball with her - she was exhausted by the end of the day, but she loved every minute.

Thanks to all for a enjoyable event.

I had a great time at Eldon's eBuild-a-Thon today.

Eldon, your shop facility is, like, awesome, dude. Totally.

And the hospitality was superb. Thank you, than a big thanks to Carolyn
as well.

I'm hooked!


Lyle KK7P


Just wanted to mention Richard L's Thames & Komos Radio Ace medium and shortwave dual-triode regenerative receiver he took to a past Build-A-Thon does work. It seemed to not work at the past Build-A-Thon because it has much poorer sensitivity than we expect from regenerative receivers. But today Richard turned it on in the low lands of Snohomiosh county and received multiple BCB (AM broadcast band) stations. Selectivity seems good for a simple receiver.

I'm hypothesizing using B+ = 12 volts on a 12AU7 handicaps the tube. (Or is it a 12AT7 or 12AX7?) Unlike most one or two active device receivers I've played with, an adjustable antenna loading coil did not improve volume on the AM BCB.

Annie, Richard, and I very much regret missing the Build-A-Thon. Richard is in pain from extensive (3-hours) dental work yesterday. I've not fully recovered my strength from mild food poisoning on Thursday. After describing a past Build-A-Thon to some e-pen-pals, some (hams in Kentucky, and Japan, and a technical university professor in Lima, Peru) have emailed me with very positive comments.

Very best regards and wishes to all,

Steven, KD7YTE


Thursday, November 19, 2009

Small DPDT Switches

I have several projects needing very small configuration switches (611-AYZ0202AGRL) , which arrived today.  The switches are much smaller than I had thought, even though I had the specs. These switches will be fun to use and will be seen in future projects.

It is interesting to note that the new small DPDT SMT switch on the right has almost the same terminal spacing as the old large SPDT switch on the left, yet the switch body is maybe 1/10th the overall size.


WSPR 2.0 is Now Available

The new version of WSPR 2.0 is now available.

With the new version, proper support for station CALL suffixes are supported. I my case this means that I can run one (receive only) version at my Web Server Site 24x7 as WA0UWH/R and another at my home QTH (receive/transmit) adhoc as WA0UWH. I will use the "/R" as an indicator for Remote - until someone has a better idea.

Also, this version is available native on Linux, which was a cludge via wine before.  My Web Server is Ubuntu therefore I will beable to use the Linux version on that system.

WSPR is available at:

Tuesday, November 17, 2009

Printed Inductors - Continued

To learn more information about Printed Inductors (see previous post) I decided to check the self-resonate frequency - by disconnecting the capacitors. This experimental coil is self resonate at 22.2mHz.

Scale = 1 to 41mHz
 Resonate at 22.2mHz

Above 60mHz all kinds of strange minor resonate modes occur.

Note: Flat-topping on the graph is over range.


A Balloon QRSS Idea - Continued

To support a QRSS Beacon Balloon effort (see previous post), I checked with the local Welding Gas Supplier, a large standing tank of Helium is about $110 plus tank rental.  The tank can be rented for about a week (used at parties) for a few dollars, or by the month for commercial accounts at $20 / month.  To purchase a tank it is about $450 with refills at the same $110 rate.

I think the rental rates includes a regulator.  But, to buy a tank would require an additional $45 for a regulator.

I need to do some more Beacon Balloon planning before investing in Helium.


Monday, November 16, 2009

Printed Inductors

Today I started experimenting with Printed Inductors, which is something that I have always wanted to try. The idea is to print and etch inductors directly on the PCB, many commercial products and some kits use this technique.

For my first experiment I created a 25.5 Turn Inductor with a single 5mil spiral trace, 12mil centre-to-centre spacing, within a 1 inch diameter. The tuning capacitors are soldered to the upper two pads.

 A Tuned Tank Circuit

Note: it is difficult to see the spiral with the naked eye, it is just a blur of shiny copper.

This is the Centre of the Spiral Conductor

My goal was to resonate at my favourite QRSS frequency of 10.140mHz. With the help of a 47pF cap and a parallel 4-50pF trimmer, resonates was obtained on the desired frequency.

I thought the distributed capacitance of the spiral and of the 32mil double sided PCB would kill the Q of the circuit, but it seems to do well as indicated on an AIM-4170B Analyser. But, this will need more investigation.

Scale = 8 to 12mHz

I may have to plan a project using this technique for the tuned inductors.

As I have time, more experiments and information on Printed Inductors will follow!


November 2009 - eBuild-A-Thon and BBQ


This Saturday, Nov 21, 2009, my Shop is available by Invitation for Building Electronic Projects as before, see link:

Unlike most similar events, there will NOT be the usual "talk-in" frequency,  but in keeping with my interest of Beacons, there is instead a QRSS "Beac-in" frequency to help guide your way at 10.140050mHz (+- a few).  For a real map see the link below.  Also, once you arrive, there is a prize for the first person to find the actual Beacon, hidden on the 40 acres. If found, don't reveal it's location, while others try. Bring your DF equipment and join in the fun.

Due to the possibility of cold unpredictable weather, the outside "BBQ" will be replaced with indoor "Hot Dogs" (or other snacks), bring additions to share if your like. For brave soles, the wood outdoor fire pit/BBQ is available if you tend the fire.

If wanted, I will re-demo PCB generation, Etch and SMD soldering. If you would like to try your hand at this, junk boards will be available.

Please bring "Show and Tell", Build Projects, QRP Rigs, Small Tools, Magnifiers and Laptops (WIFI, Internet and Laser Printer are available).

See this link for All the Details, Map, and RSVP:


Sunday, November 15, 2009


I have started a companion Power Amplifier (PA) for my 43uW QRSS Transmitters, the circuit, as bread boarded, produces about 27mW via Push-Pull output into a 75 ohm load, from a 9Volt Battery, and up to 250mW at 18Volts (two 9Volt Batteries).

In the Experimental Methods spirit of EMRFD, this is a experiment with the following goals:
  • Small as possible (because I like to work on very small things). 
  • Small enough to fit into the centre insulator of my 30m Dipole.
  • Low part count.
  • Low output Harmonics.
  • High Efficiency.
  • Minimum Input Matching Requirements.
  • Minimum Antenna Matching Requirements.
  • Low Idle current between QRSS FSK or CW transmitted messages.
  • Flexible DC Power Requirements, 9 to 18 Volts.
Note: I do not have means to measure or evaluate success of all of the above requirements.

It is of my own design and uses a JFET driver and two 2N7002 FETs for PA's (the black specks seen between the two Toroids). For more power two 9Volt Batteries can be used in series.

Unlike the Oscillator's which uses small 5mil traces, the PA will use a more standard 12mil trace.

I still may do more experiments with different Toroid Turns Ratios for better performance.

The Current Bread Board Configuration.
The 75 ohm Load is standing air cooled :-) on the right.

The PA is designed to support both FSK and CW QRSS from the Oscillator. It is currently a jumble of wires and SMT components, but I have on the computer the PCB layout for a more civilised configuration on a 1x2.2 inches PCB using all SMT parts, including SMT POTs.


Above is a sample output from my local Grabber with CW mode selected on the Bread Boarded QRSS Transmitter.

Currently, the largest component will be a switch, which is used to change modes. It is the smallest switch that I have in the junk box. It is a SPDT, which works OK, but what I really need is a DPDT switch for proper operation.

The Junk Box SPDT Switch

I checked with Mouser, and plan to order a much smaller SMT 611-AYZ0202AGRL DPDT switch.

When finished, I plan to join the Oscillator and PA on a single PCB as a complete Fun and Flexible QRSS Transmitter.


Finished - Small Trace 5mils 30m QRSS Transmitter

I received in the mail a KID2 Keyer to finish the Small Trace 5mils 30m QRSS Transmitter (see previous post).

This circuit produces 43uW of QRSS FSK power into a 75 ohm load on 10.140080mHz from 9Volts at 4.5ma.

The goals of this experimental project was to:
  • build the smallest PCB that I could produce using my current Home Brew techniques
  • use all SMT parts (except Crystal, Battery and Keyer)
  • use smaller than industry standard trace sizes (used 5mils)
  • use no more parts than necessary
  • build a real and useful working circuit

This transmitter will be installed in the centre insulator of a 30m Dipole (see previous test)

Note: the frequency dose not seem as "hand proximity" stable as other similar circuits that I have produced. I think the lack of stability is due to the very small (5mil) traces used, each small trace contributing a small amount of Inductance to each component and therefore hand proximity capacitance is more determinately effective. More tests are needed?


Monday, November 9, 2009

Soldering Iron Casualty - Work Lamp

I worked on a SMT QRP RF Power Amp Circuit until very late last night, near the end, my cordless Soldering Iron was just keeping up, if I let it set in it's charger every few minutes. It was very late (early AM) when I quit. To avoid a low battery for the next day (today), the last thing that I did was make sure that the Soldering Iron was on Charge mode.

This morning, I grabbed the Soldering Iron - fully charge and started work - on the very first try the work lamp blew out, with a bright flash. Too much charge or an old lamp on my new Soldering Iron. Dang!

The replacement lamp is about $5.00 and it would take several days to arrive.

The lamp probably draws about 250ma, which is a waste of my cordless Soldering Iron Charge.

I think, I am going to look into wiring up an LED - I normally do not need the light to see the work.  For me, the light brightness is an indicator of heating status.

So, now I have yet another project!


Wednesday, November 4, 2009

Soldering Iron Fix

While working on a SMT project, my finger was getting tired of holding the "heat" button on my new Soldering Iron. See previous post.

Sometimes the Iron would NOT heat up as fast as normal, and the work light would not be as bright.  Wiggling the switch would correct the problem, but a lot of pressure was necessary to hold it ON.

I had a similar problem before, and noticed then that the plastic button seemed to be melting on the inside at the point where it pushes on the copper spring switch.  When pressed, the copper spring is pushed down onto a screw head to make electrical contact. To help avoid the button from getting hot and melting, I place a small piece of heat tape over that area of the spring. I thought that was going to solve the problem - but NO. The  heat from the copper spring was being generated by the "bad contact" between it and the screw. And the added pressure needed was still making my finger tired and the strain would cause my hand to shake (a little) - which is not good for SMT work.

This had to be fixed!

I though, I may have to replace the switch with something with a more positive contact. Why did "they" make such a dumb switch for this thing?

I decided to take it apart and fix it, again. I could see where the contact seemed to be marred from heat, but what to do about it? What I need is something that would help make a better positive connection.

I found it - I placed a small drop of NOALOX - an Anti-Oxidant Joint Compound, on the head of the phillips screw, see the black dot at centre of photo.

Now the Soldering Iron works like a champ, I can still recommend the Iron, but suggest, fixing the switch contacts before use.

The back side point of my plastic switch button has been melted a little, but still acceptable, I may need to find a replacement.

Also, I think a small dot of NOALOX on the Tip wire holes will be beneficial, or at least that what I am trying.

OK, the implementation of switch is not dumb, it just had bad contacts.


A Balloon QRSS Idea

My micro watt QRSS Transmitter is so small, light and inexpensive, I should "Fly it on a Balloon"?

See previous post.
  • At 25K feet, I wonder what range could be expected of 43uW QRSS transmitter?
  • I wonder what the Rules/Regs are regarding an experiment like this?
  • I wonder if I could get special permission from the FCC to have an MEPT transmitter out of my reach/control - until the battery runs out?
  • Just wondering, does the FCC have jurisdiction over 43uW Transmitters?
  • A centre feed 1.5 wave 30m Antenna could be the trailing support.
  • I need to do some more checking.
  • Would anyone like to join me on this adventure?

My Small Trace 5mils 30m QRSS Transmitter

My small trace (5mils) 30m QRSS Transmitter board has been loaded with most of the parts (see previous post). Still needs a Crystal, Keyer, Battery, and Antenna connections.

So far, no screw-ups, and no shorts!

Now, awaiting the KID2 to arrive in the mail.

QRSS Tx Almost Finished


Tuesday, November 3, 2009

Industry Standard Small Traces

Today I read that "Industry Standard Small PCB Traces" are normally as small as 6mils.

Because I like small things, I wanted to try to build a circuit with even smaller traces.  The circuit I choose is a new iteration on my Micro Power QRSS Transmitter - the same circuit that was used on the QRSS test that I ran last week.

I have re-designed the board for ALL 1206 SMD parts, using 5mil traces and 8mil clearance. I would have liked to use smaller parts, but 1206 size parts are the smallest that I have available. The traces do not need to be this small, I am just doing it to see if I can? - It is kind of a QRP thing for the Home Brew PCB Fabrication process!

The Toner Transfer and Etch process went well - The board DC Resistance checks OK, without loss of connections or shorts.

The size of the board is 1000mil x 850mil,
shown here at about 4X actual size.

The signal traces between pads are 5mils wide, 8mil clearance.
The GND Grid is 5mil Traces on 20mil Centres.
The Font is 3pt.

The major components of the circuit are; a 5 Volt Regulator, Crystal, 2N3904 Oscillator, and a KID2 QRSS ID'er.

Note: SMD 1206 parts are: 12mils long and 6mils wide.

The next step is to mount the parts - which could be a real challenge, . . . . for tomorrow.


Tin Plating with Solder

Tin Plating with Solder on Home Brew PCB Projects.

On a previous post, a viewer ask via a comment; how did I do a "Solder Wipe"? as shown on one of the example HB PCB Projects.

At the time, I used a Soldering Iron with a little bit of solder and moving back-and-forth over the traces coating each with solder and then while the board is still hot, "Wipe" the solder with a cloth. This produced satisfactory results - but, you have to be quick!

Because I like the looks of bare copper for my projects, "Solder Wipe" is not used very often.

Now, many of my projects are showing their age as the copper tarnishes. I need to start using something more protective on all of my projects.

While removing parts and cleaning-up several old junk board (see below), that are going to be re-used for a SMD solder technique demonstration, I noticed that Solder Wick produced a nice clean dry finish, if it was rubbed over the traces with the heat of the soldering iron.

It looked great! I tried it on the rest of the board - with fantastic results.

Without and With Solder Plating

I recommend using a short section of used "Solder Wick" to tin plate (solder plate) new HB PCB projects before parts are loaded.

Used-Wick stays together and provides the small amount of needed solder, New-Wick tends to fall apart (see top left) before it collects enough solder to stay together. Do this task quickly, as you do not want to heat the board more than necessary - it will get HOT.

Be Careful, it is HOT

A little bit of liquid "Rosin Flux" helps the plating process (I use MG 835-100ml).

After plating, and Alcohol cleaning,  a quick buff with Scotch Bright removes the shiny surface and avoids the reflections that are so distractive while loading parts.

Note; more Alcohol cleaning is necessary after using Scotch Bright as the LEAD from the solder mars the PCB between the traces.

CAUTION: Remember,
this is "LEAD" that you are touching!

With previous HB PCB (without plating) I always liked to rub a little liquid "Rosin Flux" over the entire board before loading - it gets sticky (rosin is tree sap) and must be dried first, but it provided excellent soldering conditions for SMD parts - I still may continue this practice, even with Solder Plated PCBs. Note: I think rosin will also help seal the edge of the fibreglass PCB.

Now, . . . . if I could only find a way to do a decent Home Brew PCB Solder Resist Mask to seal the board?


Monday, November 2, 2009

A new QRSS Contact

This morning while working in my shop getting frustrated with progress on trying to get a new DC Receiver working, I abandoned work for a while.  I plugged the computer sound card back into the R-2000 general coverage receiver and turned the dial. I noticed a several signals as reported by ARGO on the computer screen, that is normal noise within the sound card and the ground loops from the receiver, I really need to get a better interface. Most of the displayed signal lines on the screen are steady and associated with 60Hz.  Nothing unusual here.

A minute or so later, I check the screen - there was a single perfectly formed QRSS3 "C" on one of the hum lines!!! What, did I touch something that caused the receiver to jump frequency in time with a perfect QRSS3 "C" - I don't think so.  Where did it come from???  I continued to watch, nothing suggested it was anything but a 60Hz hum line.

Several minutes later, on the same line was a perfect QRSS3 "O"!!!

Now this has my interest!

Who would be sending QRSS3 letters at such a slow rate??, And what was the rate??

I started a timer and put the Grabber into capture mode, and waited, and waited - Eleven minutes later - a perfect QRSS3 "N" - WOW - I have to watch this! Someone is sending some very slow QRSS - I am excited - I have to figure this out!

I can't touch anything - not even the dial - I'll just have to wait!!

At the next Eleventh minute, right on schedule, I could see the leading edge of a letter, BUT at that very moment my "WSPR" transmitter obliterated the receive signal. I did not get that letter.

Who ever is doing this must have an synchronised clock and is transmitting single QRSS3 letter at the begining of minute interval. This has really got me interested! WSPR transmits precisely on even two minutes intervals and then receive-only for (in my case) 20 minutes.  To avoid having WSPR obliterate another letter, I turned WSPR off. I checked my (other) standard local Grabber, I could see the same signal there - but a little weaker - a different antenna is used and maybe my long wire on the R-2000 had an advantage at the right direction.

I computed the time for the next letter, and waited. Right on schedule, A perfect QRSS3 "W" was recorded. I now have "CON_W" received over a 50 minutes period. I wonder what the message would say and to/from whom?

Ten minutes later, a perfect QRSS "A" was received - the first part of a CALL?? - maybe the first few letters, was the last part - I checked the QRZ data base for "WA something CON" - the QRZ database does not have option letter search (that I know of) - no joy.

Ten minutes later, a perfect QRSS "0" (zero) was received, although the frequency was shifting down - NOW this is part of a CALL! - Checked the QRZ database again with "WA0CON", checked Google - still no joy.  I will just have to wait for more of the message.

Now, I started wondering, who would transmit QRSS3 at such a slow rate?  Who ever it was must be on a battery because I could see the tell-tale frequency shift of a heavy loaded transmitter - transmitting a long "zero" for that last received letter.

What could have been the letter that I missed, while my WSPR was transmitting?? an "R", "K", or a comma?? - What kind of message could this be??

I started thinking about my old beacon messages, which consists of something like the following:




BUT, I don't have anything that transmits single characters at 10 minute intervals - or, do I?

There, On the shelf, was the previous weeks experiment, the dead 9Volt battery QRSS Transmitter, with the battery still attached, just as I retrieved it from the cold, wet, centre-insulator of the dipole several days go. See previous experiment.

To make sure this was NOT transmitting, I pulled the battery clip off - the signal went away! Dang!

The extremely low battery voltage must have put the KID2 Keyer (PIC) into some kind of very slow mode, as the characters were being generated one per 10 minute intervals. This is not a normal KID2 Keyer or QRSS mode! Transmit power must have been extremely low, the warmth of the shop must have re-activated the battery (a little), when I turned on the shop heat this morning.

For about 90 minutes, this was a very exciting chase, with a little bit of a let down, but that is what ham radio is all about, chasing signals, even when they are your own!

It was a real rush, even though it was from my own system.

But, I will never ever trust dead batteries again!