2-250 MHz Synthesized VHF Source Project

By: Steve Hageman

The VHF Source Project: Inside and Out!


As seen on the cover of the January, 2000 issue of
QEX magazine.

For information on article reprints, click here.


 

How about putting a computer controlled, 2-250 MHz ,1 Hz resolution synthesizer that you built on your own workbench! The Hagtronics VHF source project published in QEX is just that!

 

The basic synthesizer uses two PLL's mixed together to get a 2-250 MHz output. The PLL's use Mini-Circuits VCO modules and Motorola MC145191 PLL chips to produce either a fixed 750 MHz or a variable 500-748 MHz output. These two PLL's operate with a reference frequency of 100 kHz for excellent reference spur rejection and fast lockup times. The two PLL outputs are difference mixed and filtered to produce a 2-250 MHz output. The 1 Hz frequency resolution is achieved by driving the variable PLL's reference input with the output of a Intersil HSP45102 NCO (Numerically Controlled Oscillator). The Intersil NCO chip has 32 bits of frequency resolution, so achieving an overall 1 Hz synthesizer resolution is a snap. The output of the mixer is amplified to +15 dBm (+17 dBm Typical) with Agilent Technologies MSA-1105 MMIC Amplifiers. By using low noise and low intermodulation design techniques, the sources Phase Noise and Spurious Responses are kept at very respectable levels. To see a plot of the sources Phase Noise click here. To see the spectral purity and reference spur suppression, click here.

The last output stage uses a directional coupler on the output to provide a true leveled output. The output level is sensed with an Analog Devices RF Log Amplifier (AD8307). This sensed level is compared to a DAC value that is programmed to the desired output in dBm, the error value from this comparison is used to servo a PIN diode attenuator and Viola, we have a true leveled output. The output leveling circuitry frequency response has been carefully shaped to give very low two tone intermodulation when two sources are combined.

The source is controlled by a PC via an RS232 link running at 19.2k BAUD. Up to four sources may share a single RS232 connection via a novel, low cost RS232 link port expansion technique. The source uses a Microchip Technology PIC16C63 microprocessor to translate the RS232 commands from the PC to the proper setups required by the PLL's, NCO and other hardware. The PIC also monitors the status of the PLL's and leveling loop and can report this back to the PC for display.

The software 'Virtual front Panel' is a true 32 bit Windows 95/98/NT program. The program is a easy to use as it will search all RS232 ports in your system automatically looking for any connected source. The control program also allows "Set and Forget" operation, that is the software can set the desired frequency and amplitude then be exited from and the source will remain at the last setting until power is removed (Even if the RS232 cable is disconnected). The software employs many advanced capabilities such as spur avoidance. With any wide range source LO to RF crossing spurs occur, to prevent any interference to the desired signal the software automatically changes the fixed and variable PLL's so as to keep the crossing spurs at least 4 MHz away from the desired carrier.

Support for other programming languages, such as Agilent Technologies VEE is supplied by a 32 bit ActiveX control. The ActiveX control contains all of the required properties and methods needed to control the source from any modern programming language. Using the ActiveX control and VEE I have achieved programming speeds of 13 mSec per frequency/amplitude point.

As with all good homebrew projects, this project started out as a breadboard. Click here to see a picture of the first prototype. The prototype is still in use today, it just does not have any shielding so the spurious performance is not as good as the final units.

The project FAQ follows...


VHF SOURCE TYPICAL PERFORMANCE

Frequency Range....................................................2-250 MHz

Frequency Resolution........................................... 1 Hz

Amplitude Range (Calibrated)............................. -15 to 15 dBm

(Using internal PIN Diode attenuator)

Amplitude Resolution............................................ 0.1 dB

Maximum output.................................................... +17 dBm

Amplitude Accuracy............................................. +/-0.5 dB over full frequency and power range

Frequency Accuracy Error................................... Calibratable to 0 Hz

Frequency Drift per hour (After warmup).......... 0.0002 %

Frequency Stability............................................... 0.001 %

Over temperature of 15-35 deg C ambient

Phase Noise:

-64 dBc/Hz @ 1 kHz Offset

-90 dBc/Hz @ 10 kHz

-110 dBc/Hz @ 100 kHz

Harmonics: 2 MHz to 20 MHz............................. < -30 dBc

20 MHz to 250 MHz.............................................. < -45 dBc

Non Harmonics: 2 to 220 Mhz............................ < -65 dBc

220 to 250 MHz..................................................... < -35 dBc

LO and RF feedthrough...................................... < -55 dBm

Clock feedthrough (fund and harmonics)........ < -85 dBm

 

Output Match (VSWR) 2 MHz to 7 MHz............ < 2:1 @ +15 dBm output

7 MHz to 250 Mhz............ < 1.3:1 @ +15 dBm output

 

Note: Output at 100 MHz, 0 dBm output power and temperature of 25 degrees C, unless otherwise specified.


VHF Source Project FAQ's

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!

 

Parts FAQ:

As with the Personal Network Analyzer Project and the 2 Meter receiver, I will be keeping folks abreast of where to get parts for the project here:

1) The complete parts lists for all the assemblies are available here as a zip file (4k file). The format is comma separated values or CSV. This format is easily read into any spreadsheet, or you can read the files with notepad.

2) A PCB is available for the project from FAR Circuits. Check with FAR directly please, as they supply the PCB's, not me.

3) A preprogrammed 16C63 PIC is available from me for $30 US Dollars. Shipping to the US of A (and Canada, Eh?) is FREE. Shipping to Europe is $5.00 (Via US Post Office Global Priority Mail). Shipping to South America is $10 US Dollars (Via anyway I think it will get there!). Sorry South America, the postal system down there just looses too much of what I send.

4) Note: Motorola is now ON Semiconductor! Just as HP is now Agilent Technologies! And Harris is now Intersil! At least for the parts we want to use here.

5) DC Kits may have the MC145191F PLL IC's, give them a try at www.dckits.com

6) Future (a big Canadian distributor, eh?) is a big distributor of Motorola (ON) Parts. The minimum order may be quite high however, I'm not exactly sure.

7) Newark is an old standby for Motorola (ON) parts. They have a $25 minimum order, the good news is that they have many other parts such as IC's (i.e. AD8307), resistors, capacitors and a nice selection of suitable chassis in which to house the project.

8) C13 and C15 on the PLL boards are rather expensive Polyester types. To save cost, equivalent Tantalum capacitors can be used (i.e. 1 uF and 3.3 uF @ 25 or 35 volts). The Tantalums will give about 10 dB worse 100 kHz reference feedthrough rejection and the low frequency Phase Noise will increase somewhat. Unless you are watching the source with a spectrum analyzer, you probably will never notice the increased noise.

9) Bob M. brought this to my attention: I managed to get the part number of the PCB mount, SMB jacks messed up in the parts lists (They were listed as J648-ND). Here is a clarification (I hope):

You have three options (as I see it anyway),

1) The expensive way -- The way I built my modules is with nice right angle bulkhead mounts. These have a screw collar that can hold themselves tight to the box wall. This also serves to hold the PCB in place. The Digi Key part number for this is: J522-ND

2) A little less expensive is to use a right angle jack without the bulkhead mount. Just drill the case hole big enough for the jack and the cable end plug. This won't hold the PCB in place, but it will work. It will even be shielded very well if you wrap the completed jack with conductive copper tape when you are done. The Digi-Key part for this is: J472-ND

3) Hybrid -- The hybrid approach uses two bulkhead jacks at opposite ends of the PCB with the less expensive right angle jacks in between. here you can easily hold the PCB in place with the bulkhead jacks and still connect to the center jacks.

My opinion is to use #3, If I had to do it again, this is what I would have used.

10) The description for D1 on the PLL board was inadvertently left off of the parts list in figure 3. D1 is a 1N4148, 1N4448, 1N914 or really any general purpose Silicon diode. D1 is a clamp to prevent the VCO tune line from going too far negative if OPAMP U2 rails negative for any reason, normally D1 is back biased off.

NEW - 11) John P. reports that the MC145193 PLL chip is pin for pin compatable and works with the source project. You must change the charge pump resistor on pin 8 from 17.8k to 3.92k (R7 on both PLL boards).

 

Hardware FAQ's:

Here are the FAQ's as they relate to the hardware portion of the project:

1) A few folks always seem to ask for plots of the PCB for debugging purposes. Well since I laid out this PCB I will succumb this time and post them here. The plots are in PCX format. Many programs can read this format including the Microsoft Paintbrush program. Click here to download the zipped PCX files for The Assembly, Component and Solder sides of the PCB's (113 k). These plots are not suitable for making your own PCB's because they are not scaled. I recommend that you get your PCB's from FAR Circuits because they are reasonably priced and known to work. If you do not have a program that reads PCX files, go to www.zdnet.com and search for the program Irfan View, several versions are available.

2) By using a Mini Circuits frequency doubler (Model FD-1) on the output of the source: 2-500 and 2-1000 MHz may be achieved with very little degradation in phase noise (6 dB per doubling). The power drops about 13 dBm per doubling however, but this is an easy way to reach the upper VHF bands with plenty of power left to test a receiver.

 

Software FAQ's:

Here are the FAQ's as they relate to the Software (Firmware or PC Program) portion of the project:

1) The associated application software is available free as 0100HAGE.ZIP at the ARRL/QEX files site. The program runs on Windows 95, 98 and NT and requires a standard RS232 port driven by a 16550 UART for proper operation (all name brand computers of 486DX and better should have this type of UART). The download from the ARRL includes the executable files only, for other files see below.

2) Only use Winzip or some other 32 bit (i.e. long file name aware) unzipping program, otherwise all the long filenames will be truncated to the old DOS 8.3 format and the program will refuse to install!!!

3) Overwriting of outdated files during install the should be safe as these files are all genuine, Microsoft files. If you don't feel safe about this, note the name of the file name(s) the install program want's to update, respond NO to the prompt(s). Then find these file(s), copy them to a floppy and reinstall the program. This way if anything goes wrong, you can recover from the problem.

4) You can uninstall the program by going to -> Control Panel -> Add/Remove programs.

5) The Visual Basic 5 Source code for the project is available by clicking here. The software is Freeware when used with the source project as described in QEX. If you want to use the software for any commercial purpose, then all I ask is that you contact me for a license. Fair enough?

5) The ActiveX install is available via e-mail. Please e-mail me (shageman@sonic.net) with your request and I'll send it along. You must first install the PC control program (see above) or have Microsoft Visual Basic 5 installed on your system, as I will only send the ActiveX control, not all the required components. You should have a firm grasp on how to register an ActiveX control on a 32 bit Windows system, which should not be a problem if you are developing code yourself (i.e. you need to know how to run Regit.exe or a similar program). NOTE: The source code for the ActiveX control is not currently available.

6) I write all my firmware for the PIC's with the CCS PCM C compiler (Hey, life is just way too short for all that assembler stuff ;). This is really a fine little compiler, and the best $99 dollar compiler since QuickC ! Heartily recommended (the Hagtronics 'Two Thumbs Up"). The PIC firmware can be downloaded by clicking here.

NEW - 7) John P told me about some odd behavior and when I looked at the problem I discovered a problem in the Visual Basic source code. The problem prevents the PC control program from stepping in true 1 Hz increments. I have a new EXE file that fixes the problem. Download the VHF Source Version 1.0.1 EXE file here and unzip it over the one that you installed previously. Version 1.0.1 allows the source to increment in true 1 Hz steps.

 8) Copyright: The software is Freeware when used with the source project as described in QEX. If you want to use the software for any commercial purpose, then all I ask is that you contact me for a license. Fair enough?

 

Other Information about RF Sources on the WEB:

Agilent Technologies (formerly Hewlett-Packard), the world leader in test and measurement has many application notes available on RF source technology, please take time to visit their web site at www.agilent.com and browse the selected applications below,

You might be able to find the application notes by following this link... www.agilent.com/find/apps


|
Comments?| |To Steve's Workshop Home|


Updated - 7Jan01