| EIP 545/548 Power Meter Calibration |
| GBPPR Radiation Laboratory |
The instructions for manual calibration of the Power Meter option in the Service Manual are very confusing. Here's an attempt to decipher them...
Note that there are a total of 150 "power vs. power" correction factors to compensate for variations from square law in the detector diode and power meter circuits. They are divided into three tables. The first tables corrects variations below 10 GHz. The second corrects variations between 10 and 20 GHz. The third corrects variations above 20 GHz.
There are also 540 "power vs. frequency" correction factors to compensate for variations in the diode detector output at different frequencies. The "power vs. frequency" corrections cover 950 MHz to 27 GHz, ever 50 MHz.
0.) Remove, read (and save) the contents from U20 EPROM (TMS2516 or M2716) on board A107. It should look like this: 6400010-05A
1.) Set to 0x0000-0x02EC (or 0x0000-0x02B2) to 0x00. The address range changes slightly with different firmware versions.
2.) Set to 0x02F6 to 0xFF. I don't think this step is needed with older firmware versions?
3.) Get a new TMS2516 or M2716 EPROM and UV erase it for around 30 minutes or so. Might as well just get a new EPROM and keep the old one as a back up.
4.) Program the new zeroized BIN file into the new TMS2516 or M2716 EPROM.
5.) Tune a RF SigGen to 2 GHz at -35.0 dBm. Connect it to the Band 3 input on your EIP Counter and enable the Power Meter option. Adjust the RF output of the RF SigGen until the EIP Counter reads -35.0 dBm.
6.) Connect the output of the RF SigGen to an accurate RF Power Meter. Subtract the EIP Counter's reading from the RF Power Meter reading. Round the result (R1) to 0.1 dBm.
7. Using the following formulas, justify the correction to a number between 10 and 20 (decimal). (What if it's out of range?)
Perl Example:
RF Power Meter: -30.2 dBm EIP Counter: -35 dBm N: 3 CORRdec: 18 (Should be between 10 and 20) CORRhex: 0x12This is where it starts to get hard to follow in the Service Manual...
8.) Program the correction in these locations in the EPROM's memory:
9.) Connect the RF SigGen to the EIP Counter. Increase the RF SigGen's RF output power until the EIP Counter reads 1.0 dBm higher.
10) Connect the RF SigGen to the RF Power Meter. Subtract the EIP Counter reading from the RF Power Meter reading. Round the result (R2) to 0.1 dBm.
11.) Using the following formulas, calculate the correction:
12.) Program the correction in the three EPROM addresses found by the following formulas:
40 + P = Address110 Convert Address110 (dec) to Address116 (hex) 40 + P + 50 = Address210 Convert Address210 (dec) to Address216 (hex) 40 + P + 100 = Address310 Convert Address310 (dec) to Address316 (hex) Where P is the power the EIP Counter was set at in Step 9.13.) Repeat Step 9 through Step 12 until OVERLOAD is reached on the EIP Counter.
14.) Program and install the partially corrected EPROM in the EIP Counter.
Now for the fun part...
15.) Set the RF SigGen to 950 MHz at about -15 dBm.
16.) Measure the RF power on the EIP Counter and the RF Power Meter. Subtract the EIP Counter reading from the RF Power Meter reading. Round the results to 0.1 dBm. Multiply the results by 10 and convert to hexadecimal.
17.) Program the correction in the EPROM address found by the following formula:
18.) Increase the frequency by 50 MHz. Repeat Step 16 and Step 17 (540 times!). Do this until you reach the upper-frequency limit (18 or 27 GHz) of the EIP Counter is reached.
19.) Perform a checksum calculation on the corrected EPROM BIN data.
20.) Calculate the Program Number by the following formula:
21.) Program EPROM address 0x2F6 with the Program Number found in Step 20.
Below is an example from my EIP 545A. Note that the three address ranges (0x0000-0x0005, 0x0032-0x0037, and 0x0064-0x0069) have different values (0x11, 0x11, 0x11, 0x10, 0x10, 0x12) whereas the manual makes it seem like they should be the same value.
I think you are suppose to start at -40.0 dBm (not -35 dBm like the manual says) and increase the power 1.0 dBm during each address increment... Note how it kinda tracks if you start at -40 dBm and go to around +9.0 dBm. This is just a guess!
Below is a chart of the 150 "power vs. power" calibration factors (left side of chart) from my EIP 545A counter. Note they are the same for all the frequency ranges. It doesn't say in the manual to change frequency during this calibration process, but it makes sense you should. For example, use 5.5 GHz for Row 1, 15 GHz for Row 2, and 23.5 GHz for Row 3.
Knowledge is Power