---

OVERVIEW:

This project is an RS232 controlled Scalar network analyzer for personal use. The design uses a Harris 45102 DDS source (10 Hz-16 MHz), and two receivers made from Philips NE604 IF strips. Using the IF strip RSSI output I was able to achieve about 50-60 dB of dynamic range. My usual control method was used, i.e. A PIC in the box to control the RS232 interface and decode commands for the hardware. The control program is a full fledged Visual Basic program that provides a network analyzer interface to the user. Click HERE for an example screen shot of the analyzer measuring a 455 KHz ceramic IF filter, or click HERE to see a screen shot of the analyzer measuring a 10.7 MHz IF filter, this shot shows the on-screen cursors being used to measure the 3 dB bandwidth of the filter. One really notable feature of this design is that using chips developed for the wireless market, it took only 15 IC's to make this thing.

Who say's you can't build anything anymore as a hobbyist!

Some of the uses are:

• Testing filters
• Testing audio amp circuits
• Testing / tuning IF circuits
• Use of the source as a precise (1 Hz resolution, 100 ppm accuracy), adjustable output amplitude bench source
• Learning

The goal of all this stuff is to make the job of designing low frequency circuits a science, instead of guesswork. If you can measure what you build quickly, you will get better results than if just build and see if it smokes!!

SOFTWARE:

A complete software package is available free from the ARRL files site as 'hageman.zip'. The program runs on Win 3.1, 3.11, 95, 98 and was tested on one NT4.0 PC. I can't support NT if it does not work however, sorry. A good test is to get the program files from the ARRL and run the setport utility. It won't set up the Analyzer, but it should show if the program will run.

If you can program your own PIC16C71 download the object code for the PIC here (Right click and save as...). Be sure to set the programmer to WDT OFF, Power Up Timer ON, and XT oscillator (Don't code protect a EPROM device!!!!! It won't program again). This code is in Intel HEX8 format and works with the Microchip PicStart and other programmers. You must manually configure the fuses! Fuse info is not in the hex file.

PCB's:

Fred at FAR Circuits has PCB's available for this project (a 4 board set: 2 receivers, 1 source/uP and a power supply board).

Please e-mail with your questions and if you have found parts anywhere. That's how this all works, you share with me, I'll post it so everyone knows -- and I'll give you credit too!

Thanks for every ones kind words so far! It really helps to inspire me to do more of this.

1) The software package (see above) contains a HEX file for programming your own PIC (if you have a programmer). Otherwise you can obtain a pre-programmed PIC from me for $15.00 USD (Check/Money order only, no credit cards).

Overseas: Please see what it costs to ship 2 ounces (60 grams) to your destination from California USA, and add that amount also. If you tell me how you want it shipped with your order and include enough to cover the method, I'll do it that way. Thanks!

I usually ship the same day I receive your order.

2) The schematic printed shows U5 as being a 'HSP14502', the part number really is a 'HSP45102'.

3) The Receiver schematic shows R3 as being a 5k, 10 turn pot and the parts list says '1k'. 5k is the correct value.

4) Be sure to check the FAR PCB's for shorted traces before building. These are pretty easy to spot. The look like a blob on a trace or at a trace corner that should not be there. If you find one, us an XACTO knife to remove the short.

5) As of 1/98 the NE604AN IC seems to be hard to find. I suggest that you look for an SA604AN (Same part, industrial temp. range). Another possibility is to buy a SMT device and 'air' wire it to the PCB using small wire wrap wire. Or you can buy an SMT adapter from Mouser get a Philips SA604AD, SMT part and put that on the FAR circuit board. The part numbers are,

Mouser lists that they have a 16 pin SOIC to DIP adaptor P/N 535-16-350000-10

Philips web site has links to many distributors on line (follow the links to find parts). Allied generally has SMT versions of the part.

Bob reports that TechAmerica has SA604AN's (DIP Parts)!!!!! Search for SA604. Thanks Bob :)

6) One thing I did to keep 60 Hz noise out of the box was to roll the receiver bandwidth off below 100 Hz. This lowers the dynamic range at 10 Hz. It can make for some less than optimum wide band plots (i.e. 10 Hz to 100 KHz for example). Usually at 100 Hz the source may switch attenuators to re-optimize the SNR. This can create a 'bump' in the display that does not look really pretty. There are two way's to handle this: 1) Set the sources to a fixed value during a sweep. 2) Sweep 10 Hz to 100 Hz then 100 Hz to whatever and look at the plots individually. If this is not acceptable, you may want to lower the receiver response rolloff point to 10 Hz (just be careful of 60 Hz noise!). To lower the receiver response all the way to 10 Hz, change the following parts on BOTH receiver boards,

C12, 13 from 1 uF to 10 uF Tantalum, 25V

C1, 2, 3, 4 from 22 uF to 47 uF Tantalum, 25V

7) Allied electronics has AD847's and most all of the other semiconductors.

8) Digi Key has the CA3338's (if Allied does not).

9) Mouser has a large selection of box's for putting the PNA in (as does Allied) and SMT to DIP adapters (Thanks to AB6KS for the lead!).

10) Harris semiconductor has a web link that lists all their distributors: www.semi.harris.com/contact/index.htm I last found parts at Allied, Rochester and Gerber. It is possible to use an Aries SMT adaptor for SMT versions of the 45102 also. See #5 above.

11) A few questions have come up about the shielding that was talked about in the article. If you use the FAR circuit boards, no shielding is required on the PCB itself. It is a good idea to put each receiver in its own mini-box though. Then tie the receivers boxes to the chassis. I would recommend that a metal box be used for the chassis. The receiver bandwidth is >25 MHz so they are susceptible to RF pickup. Use good RF grounding for the receiver portions.

12) The Pinout for the BS250 shown on page 40 of the QST article is correct. It turns out that the silkscreen on the FAR circuit boards is backwards (due to my error in the information I sent FAR). The lettering on the FAR board is correct, it's just the outline that is backwards. The VN2222 has the correct pinout and part silkscreen on the FAR boards.

13) Make sure that you buy a LM2931AT-5.0 part for the +5 volt regulator not just a LM2931CT. The LM2931CT is an adjustable type, not fixed at 5 volts, it also has 5 pins (thanks VE7CA for the heads up).

14) A sharp eyed reader noticed that figures 9 and 13 (Feb issue) are quite susceptible to loading errors caused by long lead lengths back to the PNA. He is right of course, and the text leads one to believe that the fabricated probes that I showed are being used for these measurements. The probe construction as I presented them may have 20-50 pF of capacitance at the probe tip. I actually used small fixtures that plugged in right at the input spigots of the PNA for these measurements (with very short lead lengths). Please keep in mind probe capacitance and possible associated loading errors when using the PNA. When in doubt keep those leads short! The probes are still very useful for low impedance work (as shown in the other examples) and they keep noise out of the analyzer + they are easy to use. If you wish to make precision impedance measurements the best bet is to use a return loss bridge as I mentioned in the article (see the ARRL handbook).

15) Jay reports that C15 is shown on the schematic as 0.1 uF (Figure 2), but the FAR silkscreen is shown to be 0.01 uF. Either value in the circuit will be OK. - Thanks Jay for the info.

16) Jay also confirms that everyone (I've run into this before) is having problems with C33 in figure 4. The ARRL says that 'Decimal' values of capacitance are uF all others are pF. This means that 0.1 is 0.1 uF and 4.7 is 4.7 pF. Tricky -- eh? First time I looked at it I thought they meant 4.7 uF also, but not so the ARRL informed me. Thanks Jay I'm sure this will clear things up for others also. BTW - C33 is 4.7pF!

17) Jay is busy, he has also found that the silkscreen on C12 and C8 on the FAR receiver boards do not match the schematic in QST. C12 in the QST schematic is actually backwards from my schematic that I sent them. So the QST schematic is incorrect and the FAR PCB is correct. C8 must be as per the QST schematic for the linearity to be correct if the capacitor leaks at all. Thanks, Jay.

18) Q: When I click on the cursor menu all I get is a box with nothing in it (also no cursors). A: This is normal, you need to move one of the cursors before the box will fill with text or the cursor(s) will display. To move the red cursor press the left or right arrow keys. To move the green cursor press and hold the shift key while pressing the right or left arrow keys. After either cursor is moved the box should display the frequency and dB value for the red (R), green (G) cursors and the delta (D) reading between the two cursors.

19) Q: When I go to a sweep for the first time the Hz, kHz and MHz list box does not display properly next to the start and stop frequency input boxes. When I sweep a second time they (may) appear. I am running Win3.1. What is going on? A: I cannot recreate this problem on any of my computers. These types of screen writing problems are usually caused by the display driver in use by windows. One thing I can suggest right now is to switch to Windows generic VGA driver in 640x480 mode, or I have a modified version of the program available that removes the offending controls and allows input of start and stop frequencies in kHz directly. If you have this problem and would like a copy, e-mail me and I will e-mail you a copy (I only need to send a 60k exe file).

20) Q: The control program does not seem to run on really old hardware (i.e. before 486DX2 platforms) A: If you run into this, edit the delay parameter in the analyzer.ini file (instructions are in the file). This adds a slight delay in the RS232 readback functions so that the hardware and software can catch up to each other. This can also happen if your UART type is not a high speed 16650 type (most brand name 486DX and above computers have this type of UART).

21) A free program called PartMiner is available from partminer.com. This nifty program will search a dozen US sites for parts that you may need. Try it, it's how I keep up to date on what is available where. Sorry they went bust!

22) The only 'generic' problem that has shown up several times now is that some people have tried to build the PNA by substituting Aluminim Electrolytic capacitors instead of the Tantalum ones called out for in the schematic. This is not a good idea, especially around the NE/SA604. The Tantalums were called out because of their leakage characteristics. If you use Aluminum Electrolytics instead the end result is really poor linearity and attenuation accuracy.