Author Topic: HP8970A with Agilent 346B noise source 80mV 5µs negative peak on powerup.  (Read 560 times)

0 Members and 1 Guest are viewing this topic.

Offline _Wim_

  • Frequent Contributor
  • **
  • Posts: 751
  • Country: be
Hi,

I recently acquired an Agilent 346B noise source and an older HP 8970A noise figure meter. Both were sold as "untested/parts only".

The 8970A needed only 1 replacement cap in the power supply (and 2 new X2 capacitors because the RIFA's released the magic smoke), and seems to be working ok. As far as I can tell, the noise source was ok upon arrival.  :-+

When using this combination to test some LNA's, I get the same results for gain that I get using my CMU200 (with FreRes software), the measured noise figure matches quite ok with the datasheet of the DUTs I tested so far.

However some LNA's are clipping during the test. When I connect the noise source to the scope (with 50 ohm input adaptor), I get a negative spike of 80mV every time the noise source is powered on (yellow trace). The 28Vdc drive signal for the noise source looks ok (bleu trace).

I this normal behavior, or is there something wrong with the 346B or with the rise time of the 28Vdc excitation signal from the HP 8970A?

Schematic for the 346B can be found here:
http://www.simonsdialogs.com/2015/06/hp-agilent-keysight-346b-noise-source-finally-a-calibrated-enr-standard-and-a-temperature-compensated-current-source/

edit: forgot to add attachement
 

Offline VK5RC

  • Supporter
  • ****
  • Posts: 2295
  • Country: au
Hi,
I have an old hp 8970a, but (I think) a good Keysight 346B noise source, i repeated your set up - and got similar results at high horizontal CRO speed.
At slower speeds (the other picture) if you 'squint' I think you can see the difference between the 'on' and 'off' periods.
I think your gear is ok.

I got my set up to test some very low noise factor rf amplifiers for a pre-amplifier in amateur radio earth - moon-earth experiments. I had modest success being able to measure the thermal rf noise of sun, tree and cold sky. Set up was for 1296MHz. Using the low level noise source 346a, I am reasonably confident the amps I built are about 35dB gain, NF about 0.6, they are a G4DDK design.
Regards Rob
Whoah! Watch where that landed we might need it later.
 
The following users thanked this post: _Wim_

Offline _Wim_

  • Frequent Contributor
  • **
  • Posts: 751
  • Country: be
Hi,
I have an old hp 8970a, but (I think) a good Keysight 346B noise source, i repeated your set up - and got similar results at high horizontal CRO speed.
At slower speeds (the other picture) if you 'squint' I think you can see the difference between the 'on' and 'off' periods.
I think your gear is ok.
 

Many thanks for repeating this test! Much appreciated!
I looks almost identical, so that confirms this is normal. Still quite surprised though that this happens. I was testing a 40db gain LNA with clipping indicator. 80mV x 100 caused the amp to clip, so that how I noticed.
When I power the 346B from my Rigol DP832  (much slower rise time for the 28Vdc)
I do not get this behaviour  at all.

I got my set up to test some very low noise factor rf amplifiers for a pre-amplifier in amateur radio earth - moon-earth experiments. I had modest success being able to measure the thermal rf noise of sun, tree and cold sky. Set up was for 1296MHz. Using the low level noise source 346a, I am reasonably confident the amps I built are about 35dB gain, NF about 0.6, they are a G4DDK design.
Regards Rob

Nice! My use is more for testing low noise  pre-amps  used for low level industrial sensor signals, but doing earth-moon-earth experiments does sound like a fun and difficult challenge.  Is it this one you have built?
http://g4ddk.com/G4DDKVLNA2B3.pdf
 

Offline VK5RC

  • Supporter
  • ****
  • Posts: 2295
  • Country: au
Yes, that is the one, I built 3 - but keep blowing up the final transistor amp when I put it all together, and try the actual bounce, I think it is a sequencing problem and that either some dc or high power rf is getting up the back end of the pre-amp. The transmit section is 5W out of the tranceiver, a low amp to get to approx 15W then another amp to get to approx 350W. The big problem is that the tranceiver uses the same rf connector for both receive and transmit - the small sequencer circuit (that controls relays and power into the amps) is not configured correctly - after blowing up 3 of these amps I have decided to take a break from the project. HiHi.
Regards Rob
Whoah! Watch where that landed we might need it later.
 

Offline _Wim_

  • Frequent Contributor
  • **
  • Posts: 751
  • Country: be
Yikes, 350 Watt of RF power, that can do some serious damage :o

As I do not have a HAM licence, this "might" not be not the ideal project for me…  :-DD


 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf