Monday, December 27, 2010

Beacon keyer IK0WRB

If you ever build the beacon, or if you have the idea of building one, there is a part that every beacon needs, the keyer. Various beacon keyer designs are available on the internet but IK0WRB solution was the one that I prefer to use. Up to now I built a several keyers and all of them a working perfectly. As a matter of fact, the 144 MHz 9A0BVS beacon is running more that 3 years with this keyer where small reed relay !! is switching the power for the buffer stage. At the beginning nobody believed that the relay will last so long. It is still working.

As a spare solution, there is a keyer based on the transistor switch, where no mechanical parts are included. The schematic is standard IK0WRB keyer with a small modification where two transistors are used in a switch configuration. The first NPN transistor is used to key the input to the ground, and the second PNP transistor is used to switch the positive voltage. The jumper is used to chose beteween this two options, depending on the type of the beacon you are planing to use.

One of the things that make this project unique is the possibility to change the text and the format of the message on a very convenient way, just by tweaking the options in the windows based application written  by the author. The same application is generating at the end the HEX file to be programmed in the PIC 16F84. Of course, the message speed and the duration of the beacon carrier can be selected. Another nice feature is the possibility to use any convenient crystal frequency for this project. As I have a bunch of standard 3.579 MHz crystals, this is my favorite.

All the project can be done using the SMD technology where complete PCB can be really small. Between using the tiny SMD components and drilling the holes on the PCB for the standard components I prefer to use the SMD :-) . As this can be a beginner's project I made a PCB using the standard components with the socket for the micro-controller where the same can be easily reprogrammed with the new messages for a different beacons.

The keyer from the photo is a part of the 23cm beacon where positive voltage is keyed for the 648>1296 frequency doubler. The shape of the signal is looking good without any clicks or chirp. Simple keyer for all beacon builders....

Thursday, December 23, 2010

W1GHZ rover transverter for 1296 MHz

The idea of building this project was how cheap and simple you can build the 23cm transverter that actually works, not just on the bench but also having you on the microwave bands making some contacts. So this project was one of the simplest to build not requiring much understanding and fancy lab equipment to tune. In this case simple means no tune, just solder assemble and run. There is really nothing to tune in this project if you follow the "For dummies" style instructions written by the author W1GHZ . AD6IW likes to say for this kind of project "Plug and pray" :-)  So this was simple.

What about cheap? How cheap you can go for the project like 23cm transverter? The most critical parts you can buy from the author, the rest you can buy on the e-bay or on the flea market, or even use some old microwave stock equipment. Wise guys will find almost all parts from the old SAT TV LNB converters. Let's check our shopping basket:

Local oscillator board + one Transverter board (1296 MHz)  $20
Relay board for SMD relay  $3
Mixer ADE-5 for 1296MHz $5

Relay G6Z-1F-A-DC12  $5
MMIC (8pcs)  $30
SMA female (5pcs) 5$
SMA male (5pcs)  5$
Rest of the parts (SMD components, Oscillator 80MHz) $7

At the end the total price will be around $80. Is this cheap or what? Just to compare, the good quality crystal for the microwave local oscillator will cost you in the Europe between EUR 20-25 ($30) !! Building complete transverter for $80 is really fair price, best buy price. So this was cheap.

Ordering the PCBs and mixers from the gentleman like Paul was an easy task, custom officers (Croatian) was another story. The rest of the parts I found in my "microwave scrap stock". Let's start.

This type of transverter is the earliest one using the Local Oscillator frequency of 720MHz. The latest model (Right side up) is using Local Oscillator working on the frequency of 576MHz. Assembling and soldering the Local oscillator PCB and the SMD components should not be difficult. Just follow the instructions and pay the attention to the MMIC bias resistor. Depending on the type of MMIC you are planing to use, calculate the value of the bias resistor. If you are not familiar with schematic and where to install the components check the following page where all details are clearly seen and well documented.

Canned oscillator is mounted on the ground side of the PCB together with the output connector and few other components. 80Mhz canned oscillator should not be expensive, most probably the shipping will be much  higher then the value of oscillator. I found my on the old DX-80 PC motherboard, used as a clock. The 64Mhz canned oscillator can be found in some ADSL wireless routers. I prefer to use good quality SMA connectors instead of cheap F connectors.

First test was successful and 4 mW of output power was easily measured. No additional MMIC in the amplifier chain was required in my case. First downside of this project is the Local oscillator frequency. We can not expect high frequency accuracy from the cheap canned oscillator. After the multiplier chain I get the 10 kHz offset on the 720 MHz. After the final multiplying to 1440 MHz the offset is doubled. 20 Khz is the final offset from the required frequency. I can live with that, considering the price of  the complete oscillator. It is important just to know the offset, later you can calculate your band start and band stop IF frequency. Changing and looking for the other canned oscillator can help sometimes, but I was not so lucky. It is not worth spending time on this problem until you know what is your working frequency.

With a good working oscillator the next step was assembling the transverter board. I used all the parts that are listed in the author's article. The only change I made was the MMIC bias resistor to meet the maximum allowed current and the best MMIC performances. Find the MMIC datas heet and check for the Vd and Ibias where using the simple formula:
              Vcc - Vd
Rbias =  -----------
bias resistor should be calculated. Use the next higher value then the one you get from the mentioned formula. Assembling was straightforward and the only thing to take care was soldering the components to the PCB ground where more heat should be applied to ensure good connection. The good practice is checking all the mounted parts before applying the power to the transverter. Once you have the white smoke out from the component it is not easy to bring the smoke back :-) , so be careful. The worst thing that can happen is loosing some MMICs. Apply the RX power to the terminal and measure the current where the value should not exceed the MMIC INA10386 bias current. After that connect the power to the 1440 MHz multiplier chain on the same PCB and measure the current using the same method. The total current consumption should be calculated according to all three MMIC bias current.

Now it is time to connect the local oscillator, 23cm antenna and the IF rig. Power up your IF radio first and then after applying the power to the transverter RX part you should notice the slightly higher noise on your IF radio which leads to conclusion that your RX part might work properly. Remember to use LSB on your 144 MHz radio where 143.900 MHz correspond to the 1296.100 MHz. If you have some local beacon in your area check for the signal, or just ask some local HAM to help you with some transmission test. There is nothing to tune except to check and measure the MMIC bias current.

The TX side of the transverter should be checked the same way. The maximum output power was 8 mW in my case. Changing the MMICs did not result with higher power despite the author article where he stated that 15 mw should be the achieved without any problem. I heard from some hams that they manage to get 15 + mW but with the Right side up version. I did not measure the filter pass band, but I suspect on the filter  loos. FR4 laminate have high loos on the microwave frequencies and the Er may vary between the various types of used laminates. Any how this is enough power to make some local qso or to drive some small amplifier. Do not forget to use the attenuator on the IF TX side. 1mW of the drive IF power is enough for this type of mixer. I prefer to use home-brew "sequencer" where attenuator and RX-TX voltage switching is controlled by "RF-cox" or PTT external signal.

Instead of the author PC board with relay i prefer to use a coaxial relay with SMA connectors. The one I use in this project was found on the flea market for a really good price. The truth is that the maximum usable frequency according to the data sheet is 2 GHz what is good enough for our needs. Moreover the coil power was 12V, so just a coil protection diode was required.

As this is intended to be a cheap project, the housing should be the same style. Old CD-ROM player was cannibalized for this purpose. A sheet of the FR-4 laminate was cut to fit inside the CD-ROM. All RF transverter PC boards are mounted on the top side while on the bottom side there is place for the sequencer board with the attenuator. For the moment no LNA or power amplifier is attached, but there is enough room  left on the bottom side to accommodate this as well. Front side was modified from the PC cover leftover. Two LEDs and a switch gives a finish touch to the project.

At the end you should have a nice rover transverter that can give you some extra points in the local microwave contest or you can give a chance to somebody to work some new squares. Both way you should have a lot of fun using this simple transverter. If you are lucky enough to live in the area where the hams are active in the microwave bands you will be able to make a lot of qsos. And remember: rover does not mean EME. This is a simple transverter and this kind of the design suffer from the facts that you should be aware of.

I already mention the canned oscillator frequency accuracy.
Sometimes tuning reverse side (LSB) can be confusing but sometimes this can help you, avoiding the big gun signals on the 144 MHz IF frequency working in the contest. Some of the older 2m radios can not tune below the band, some new radios can do that with some modifications but still they suffer from the lower sensitivity on this part of the band.
Roughly calculating, the conversion gain is low or none. Loos in the splitter 6dB, filter 7dB, mixer 6db, connectors 1 dB, coaxial switch 1 dB give us a total of 21dB. The only active amplifier on the receiving side is INA10386 delivering same or maybe a few dB higher gain on this frequency.

Knowing all this, I decide to build this rover transverter just to see how easy is to start with the microwaves not spending the big bucks.

Thursday, December 9, 2010

13cm Pipe-cap filter

I needed a filter for my 13cm transverter project. What was the best buy option? Thinking and looking for the simple filter but with the characteristics that suits my needs, the answer was Pipe-cap. Simple enough, cheap, and easy to tune. After all, so many articles written about it (WA5VJB, W1GHZ, etc.) showing that this can actually work not losing so many time on mechanical design.

Check the local plumbing store for the pipe-caps used in the central heating plant. It should look like the ones on the photo. The one we need for the 13cm filter is 1" pipe-cap. Remember that the inner dia. is approx. 28-29 mm and the inner height is approx. 23-24 mm.

After choosing the right pipe-cap, a M5 brass screw with two brass nuts should be purchased as well. Beside this we need a two SMA connectors and a piece of double sided FR-4 laminate, 0.8mm will be fine.

Drill the 5mm hole in the center of the pipe-cap and solder the brass nut on top of it. Prepare the peace of FR-4 laminate where you etch the 50 ohms micro-strip lines where SMA connectors and probes should be soldered. The length of the probe is 10mm (length inside the pipe-cap) and the probe spacing is 18mm. For the probe use the wire (silvered center conductor rigid coaxial cable should work) or any coper 0.6-0.8mm dia. wire. Don't forget to solder the SMA ground to the PCB. Soldering 1" cooper pipe-cap might be tricky if using just the low power soldering tool. To avoid bad connections use the clothes iron to heat the pipe cap and then even with the low power solder you should not have a problem to solder properly cooper or brass. When finished, the pipe-cap filter should look like this.

Tuning the filter to the required frequency can be done with simple RF signal generator and a power meter. For this purpose I used the 13cm Comtech ATV transmitter controlled with the I2C protocol and a simple diode power meter made from SAT TV LNB mixer diode. Tune the generator to 2320 MHz. Ensure that the power meter sensitivity is set to range 10-30 mW center scale. The screw should be all the way down touching the bottom of the filter. Back the tuning screw out and after 4 to 5 turns you will notice the signal on the power meter. Gently tune to the maximum and that's it. This is the best what you can do just with this type of "tuning set". Use  the second nut to stop the tuning screw. After all, a drop of the paint or nail varnish can help to fix the screw and the nuts. If you can read the power, measure the power with and without the filter, the difference will give you a pretty good idea about the I.L. Changing the frequency and measuring the power with the help of the Excel spreadsheet and plotting the chart can give you an idea about the filter pass-band.

If you have the access to the network analyzer more accurate tuning and readings can be done. The one I made showing the filter characteristics based on the tune to the maximum method. The filter response is as expected with I.L. down to -2.85 dB and the stop band rejection down to -32 dB for 144 Mhz from the center frequency, not bad at all for such a simple filter. Playing with the size of the probes can give us more or less sharp skirts, higher or lower I.L., all affecting the stop band rejection.

Final notes:
* Tuning the filter with the screw is quite sensitive. Trying to stop the screw with the second brass nut may become tricky and not easy for everybody. Try to replace stop brass nut with the nylon one what will produce some tension to the tuning screw and prevent detuning due vibrations.

* Instead of using the micro-strip PCB lines, semi rigid coaxial cable can be used as well, extending the center conductor inside the pipe-cap and soldering the coaxial shield to the filter grounding.

* The same filter has been used in the home brew 2320/432 MHz transverter with very good results.