Author Topic: HP 3588A Spectrum Analyzer Gamble  (Read 198 times)

0 Members and 1 Guest are viewing this topic.

Offline Smoky

  • Frequent Contributor
  • **
  • Posts: 509
  • Country: us
HP 3588A Spectrum Analyzer Gamble
« on: July 17, 2021, 05:34:36 am »
I went searching for a spectrum analyzer that would span low frequencies to just beyond the FM radio range. I wanted it to be an American-made boat-anchor job like the HP 3585B, but the price for those in good condition are breaking a thousand dollars.

I found this HP 3588A that weighs-in at a svelte 61 pounds and it seems to fit the bill. The frequency range of the analyzer is 10Hz to 150MHz and so I decided to take a gamble. The seller on Ebay didn't know how to test it so it was sold as (expensive) parts.

The CRT display looks just like my HP 35660A DSA but unlike the appearance of the HP 3585B CRT. I haven't compared the specs or what sets the two analyzers apart but I first wanted to see if this thing worked. It was made in 1990:



The back of the beast has a BNC marked "oven reference out" and option 001. It also has a separate "Ref Out 10 MHz" BNC too? I used an RG400 cable from the "oven reference out" to bring the signal to the front BNC input:



I narrowed the span to 2MHz and sent the marker to the peak:



The input seems to be working.

Anywho, I lifted the cover to measure the battery mounted on the CPU board. It read .03 VDC. So a new one is needed.

One thing is for sure, you can't help but notice the quality built into these analyzers!



What I did here is feed the spectrum analyzer an external waveform from the Siglent SDG1032X. I set the waveform generator to 10 MHz at 4.25 dBm:



It appears to fall very close:



I may not have let the waveform generator warm up before I took the picture. I just looked at the spectrum analyzer after 5 minutes and it now reads 4.36 dBm and hasn't budged. The spectrum analyzer had been on for about 30 minutes beforehand.

The Siglent requires the "load" to be set to 50 Ohms for it to allow the amplitude to be in dBm. The spectrum analyzer's input is set to "couple to input." The spectrum analyzer still reads 4.36 dBm after 10 or more minutes now so I'd say the waveform generator needed to stabilize for a little while. I'm using a 36" RG400 patch cable for this.

These specs are from the sales flyer:



So far, so good :-+

I skimmed through the service manual and came upon a chapter relating to calibration correction curves. I followed along with the steps to see where this machine lands. There are no errors in the fault log so it seems to be in calibration but I thought I'd go through the process anyway.





Here is my unit's 50 Ohm input path result. I apologize for the crappy picture:



Here's the 1 Meg Ohm typical path from the service manual:



And the 1 Meg Ohm path from my machine:



If I understand this correctly, as long as any point on the curve remains +/- 10dB from the center of the graph, the analyzer can take it from there and compensate for the error.

One other topic related to the results of the "quick confidence test" and the numbers at the end of the test results:



And here are the results of this HP 3588A:



Not so bad :-+

I'd like the oven reference signal to be calibrated but I don't have a 10 MHz frequency standard. The internal frequency counter in this analyzer reads the 10 MHz oven reference signal as 10,000,021.7 Hz. Plus, the analyzer must be warmed up for 48 hours!

« Last Edit: July 22, 2021, 02:38:14 am by Smoky »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf