Author Topic: Noise source for noise figure measurements  (Read 5503 times)

0 Members and 1 Guest are viewing this topic.

Offline TimeokTopic starter

  • Contributor
  • Posts: 13
  • Country: it
Noise source for noise figure measurements
« on: July 27, 2023, 12:31:50 pm »
I made a series of noise sources using a BFR90A. The basic features are:
Power supply: 28V
Freq. 1MHz - 2GHz
ENR about 15dB
Flatness +/- 0.6dB
And most importantly, I am able to calibrate them individually.
I have several prototypes and a final version.
 

Offline mr ed

  • Regular Contributor
  • *
  • Posts: 98
  • Country: ca
Re: Noise source for noise figure measurements
« Reply #1 on: July 28, 2023, 11:28:31 pm »
How did  you get so flat an output?
 

Offline mr ed

  • Regular Contributor
  • *
  • Posts: 98
  • Country: ca
Re: Noise source for noise figure measurements
« Reply #2 on: July 29, 2023, 09:34:15 pm »
It looks like a lot of gain (2 stages to get to 25dBm 18GHz) and a filter may be used. 

Web ref: "Design of a 0.1~18GHz High-Power Broadband Noise Source"
 

Offline rhb

  • Super Contributor
  • ***
  • Posts: 3527
  • Country: us
Re: Noise source for noise figure measurements
« Reply #3 on: July 30, 2023, 02:11:53 am »
How did you calibrate them?  There are quite a few thread going on power measurement and accuracy.

Reg
 

Offline TimeokTopic starter

  • Contributor
  • Posts: 13
  • Country: it
Re: Noise source for noise figure measurements
« Reply #4 on: July 30, 2023, 08:15:16 am »


These noise sources are characterized by wide frequency range, 1MHz to 3GHz and by a flat response.
 Its main features are:
 Nominal ENR 15dB +/-2dB
 Working frequency from 1MHz to 3GHz
 Frequency response +/- 0.6dB or better
 Bias +28V
 Output connector sma male
 Bias connector BNC female
 Individually calibrated with professional instrumentation
 Compatible with NFA HP, Eaton etc.

Using a selected bjt and reverse biasing the be junction generates wide band noise. The selection of the component is very important because they behave differently both in frequency response and in noise level. Bias is also essential because it affects both the level and the slope of the noise. In particular, the variation of the bias has a higher value in the 0-1GHz frequency range than at higher frequencies. Its stability is therefore very important. In the design phase it was decided to use a constant current generator to guarantee a stable polarization. To ensure acorrect level and swr the source uses a 10dB 18GHz attenuator. The pcb has been manufactured using microwave material to ensure minimum losses in the higher frequency range. The calibration of each single source was done using an Agilent NFA N8974A and a HP436C source as reference.

 

Offline TimeokTopic starter

  • Contributor
  • Posts: 13
  • Country: it
Re: Noise source for noise figure measurements
« Reply #5 on: July 30, 2023, 08:17:56 am »
this source is designed for noise figure analyzer, not for "trackig source".
 

Offline mr ed

  • Regular Contributor
  • *
  • Posts: 98
  • Country: ca
Re: Noise source for noise figure measurements
« Reply #6 on: July 30, 2023, 09:13:37 am »
What is special about the BFR90A?   How many transistors does it take to select out a good performing noise gen? Some designs  use diodes, some zener etc.
 

Offline rhb

  • Super Contributor
  • ***
  • Posts: 3527
  • Country: us
Re: Noise source for noise figure measurements
« Reply #7 on: July 30, 2023, 12:30:09 pm »
Please.  Will you  chronicle your design and development?  I realize that's a lot to ask, so I'd suggest an outline of the process you used and the significant issues you had to resolve.

There is not a lot of literature.  The best I've seen was on the ones made for the Florence EME conference attendees.

I have one of the metal box sources from China.  Unlike the BG7TBL it's quite flat out to ~1 GHz.  It lacks proper voltage regulation to the MMICs though. I've got several other hobby grade sources   I have  a 346B and 8970B, but have not learned how to operate it.  Too much TEA ;-)

Have Fun!
Reg
 

Offline TimeokTopic starter

  • Contributor
  • Posts: 13
  • Country: it
Re: Noise source for noise figure measurements
« Reply #8 on: July 30, 2023, 02:41:57 pm »
Hi,
I'll try to answer the questions. My first approach was to build a series of noise sources of high-level characteristics, calibrated, with frequency 1-500MHz and costing less than 200 Euros.
Looking at the solutions on the market I saw that, apart from the HP sources which are similar if not the same as the Noise com, it consisted of a capacitor, a diode and a bias resistor in addition to the input decoupling. consequently I built a test set by mounting a spider web to test various diodes. I was lucky that a friend lent me a Noise Source analyzer and an HP source to do the first tests. The first diode was a 1N4148 which works very well with a high noise level which however was not flat, the ENR variation at low frequencies was more than 10dB higher than at 500MHz which is also its maximum operating frequency.

So I thought of using a classic BFR93 reverse biasing the BE junction which works at higher frequencies but it too had a diode-like linearity at low frequencies, so to linearize I decreased the value of the output capacitor from 1kpF to 300pF. A solution that works but I don't like it because in this way part of the stability of the level, below 100MHz, is shifted to the temperature stability of the capacitor.
So I bought a dozen of BFR90, BFR90A and BFR91. These BJTs have very good frequency linearity up to 2 or 3 GHz and beyond. At this point I understood how the diode bias current works and I mounted a constant current generator which must also have the characteristic of having a fast response to follow the 28V which are switched on and off at a frequency of about 1kHz. In the end I opted for a different circuit from that of the heads of the EME conference in Florence, much simpler but functional. This is composed of a FET and a resistor which determine the BIAS current which is in my case less than 2mA.

I would like to add that at the output of all the sources I have mounted a 50Ohm 10dB attenuator to have a low SWR, as the diode and the generator components do not have an impedance close to 50Ohm, especially when the diode is in the off state. It is a configuration that all manufacturers use. In my case the worst return loss is 24dB like the HP version I have.
But back to the bias current, I've noticed that it has a different effect depending on the frequency. The rule is that for frequencies lower than 1-2GHz the variation of the current causes the noise level to vary greatly and much less at higher frequencies. So I took the noise level at 1.5GHz as the base noise level of the single diode and I varied the bias so that the response curve aligns with that value even below that frequency.
In addition to this I have verified that even using the same transistors, but from different lots, the generated noise level varies. Don't ask me for scientific explanations.
For example, two BFR90s from different batches generate, one 13 and the other 16db (after 10dB of attenuator). Therefore, the rule is that these devices are all quite good but they must be calibrated individually, each one has its own specific current to have excellent linearity.

Now let's talk about pcb, I decided to use a material for microwaves, the FR4 loses about 0.2dB at 2GHz, not much, but I preferred to start from an excellent quality support also because I can use it in the future for higher frequency sources.
I did the calibration with the Agilent NFA which gives me the response with a graph on a color LCD display. Having a sample source, first I normalized the trace, then I connected each of my sources and the reading difference with the markers will be subtracted or added to the ENR table of the reference in order to have its own personal table.
It is very difficult to do bias adjustment with Noise Figure Meter without graphic display. I hope I have answered your questions.

additinal note: testing several diodes I found one is +/- 0.3dB from 10M to 3GHz.

Luciano
« Last Edit: July 30, 2023, 02:52:03 pm by Timeok »
 
The following users thanked this post: rhb, KE5FX, chuckb, ch_scr, eplpwr

Offline TimeokTopic starter

  • Contributor
  • Posts: 13
  • Country: it
Re: Noise source for noise figure measurements
« Reply #9 on: July 30, 2023, 02:54:07 pm »
pic1
 

Online G0HZU

  • Super Contributor
  • ***
  • Posts: 3239
  • Country: gb
Re: Noise source for noise figure measurements
« Reply #10 on: July 30, 2023, 04:30:40 pm »
The main issue I've found when making a homebrew noise source is achieving good performance over temperature (and time).
The classic HP 346A noise source really does seem to excel in these areas. I think the diode used in the HP 346A achieves about 0.01dB/degC thermal drift. The Noisecom diodes are fairly similar although nothing I've found competes with the 346A here.

Have you tested your source over temperature? Noise diodes are designed a certain way to provide consistent avalanche performance over time and temperature.

I've always tried to get really low VSWR in both hot and cold states. The homebrew noise source I made using a Noisecom diode has an ENR of about 13.5dB and it works up to about 2GHz. See the plot below. The VSWR is better than 1.03:1 up to 2GHz and better than 1.02:1 up to 1GHz in both hot and cold states. In my case I opened up a NC502 noise source module from Noisecom and took out the glass diode from it. I then used it in my low cost noise source module. By removing it from the Noisecom module the frequency response was improved.

I took the diode out of one of these metal can sources. It was very easy to pop the lid off the metal package and remove the diode:

https://noisecom.com/Portals/0/Datasheets/NC500REV3_datasheet_WEB.pdf

I've also made other noise sources that use cheaper diodes (covering up to about 7GHz) but these can't compete with the 346A or the Noisecom diode based source when it comes to thermal stability. One of my cheapo sources manages about 0.1dB/degC and the other manages about 0.03dB/degC. These use diodes that cost about £0.50. The Noisecom diode costs about £70 just for the diode but it offers 0.01dB/degC.

Note that the plot below is titled NC502 but this isn't the VSWR of the NC502 module, it's the VSWR of my homebrew noise source that uses a diode taken from an NC502 noise source.
« Last Edit: July 30, 2023, 04:40:18 pm by G0HZU »
 

Offline mr ed

  • Regular Contributor
  • *
  • Posts: 98
  • Country: ca
Re: Noise source for noise figure measurements
« Reply #11 on: July 31, 2023, 02:47:36 am »
The HP 346 manual shows only an internal power supply for a constant current source and the majic avalanche diode followed by  a matching network. The C version generates 15dbm ENR (over -174dbm? Is that correct?) to 26GHz which is impressive to me.

15db ENR won't be seen on a spectrum analizer with a typical noise floor of around -125dbm to -140dbm so it's only for checking amps etc. - unless amplified.

I have tried making a flat noise gen using a tiny uwave diode and avalanching it with 20v and about 2ma. It generates lots of noise but output drops something like many db per GHz to 5Ghz. But it's output is strong to see it on my SA. Thomas Lee's Planar Microwave book says not only are the diodes hand picked but the flatness is adjusted for each diode too.
 

Offline TimeokTopic starter

  • Contributor
  • Posts: 13
  • Country: it
Re: Noise source for noise figure measurements
« Reply #12 on: July 31, 2023, 08:18:17 am »
Hi G0HZU,

*The main issue I've found when making a homebrew noise source is achieving good performance over temperature (and time).
*The classic HP 346A noise source really does seem to excel in these areas. I think the diode used in the HP 346A achieves about 0.01dB/degC thermal drift. The Noisecom diodes are fairly similar although *nothing I've found competes with the 346A here.
*Have you tested your source over temperature? Noise diodes are designed a certain way to provide consistent avalanche performance over time and temperature.

R: I have not yet tested the variation with time and temperature. At the moment I have six calibrated sources, five with temperature compensated current generator and one with the FET generator + resistor. Soon I'll feed them all with 28V for 2000 hours to "age" them a bit, then I'll do a test campaign at different temperatures. It is clear that the stability will be the sum of the two factors, the diode threshold and the bis stability.

*I've always tried to get really low VSWR in both hot and cold states. The homebrew noise source I made using a Noisecom diode has an ENR of about 13.5dB and it works up to about 2GHz. See the plot below. *The VSWR is better than 1.03:1 up to 2GHz and better than 1.02:1 up to 1GHz in both hot and cold states. In my case I opened up a NC502 noise source module from Noisecom and took out the glass diode *from it. I then used it in my low cost noise source module. By removing it from the Noisecom module the frequency response was improved.
*I took the diode out of one of these metal can sources. It was very easy to pop the lid off the metal package and remove the diode:
*https://noisecom.com/Portals/0/Datasheets/NC500REV3_datasheet_WEB.pdf
*I've also made other noise sources that use cheaper diodes (covering up to about 7GHz) but these can't compete with the 346A or the Noisecom diode based source when it comes to thermal stability. One of *my cheapo sources manages about 0.1dB/degC and the other manages about 0.03dB/degC. These use diodes that cost about £0.50. The Noisecom diode costs about £70 just for the diode but it offers *0.01dB/degC.

R:My goal, as already said, was to create a low cost source and in any case I wouldn't know where to find Noisecom diodes at low prices. Maybe I could get two 12GHz Nc diodes in the surplus market, but I'll know about it only after mid-September. We will see. In the meantime I want to continue studying and optimizing the solution of a bias current compensation which, however, from the tests done, works in a non-linear way in frequency.
Another thing to study would be the stability of the BE junction as a function of the temperature, I'm sure that different devices have different responses due to the different chip structure. These tests can be done initially in DC.

Luciano
« Last Edit: July 31, 2023, 08:41:52 am by Timeok »
 

Offline TimeokTopic starter

  • Contributor
  • Posts: 13
  • Country: it
Re: Noise source for noise figure measurements
« Reply #13 on: July 31, 2023, 08:41:04 am »
Hi MR ED,

*The HP 346 manual shows only an internal power supply for a constant current source and the majic avalanche diode followed by  a matching network. The C version generates 15dbm ENR (over -174dbm? Is *that correct?) to 26GHz which is impressive to me.

R: The 436C they lent me has an ENR ranging from 13.1 to 14.9 depending on the frequency and I've seen some with even greater fluctuations. I confirm that the level in dBm is very low, also because it is used to measure the noise levels of amplifiers and other devices and therefore it would be counterproductive to have sources with very high noise.

*I have tried making a flat noise gen using a tiny uwave diode and avalanching it with 20v and about 2ma. It generates lots of noise but output drops something like many db per GHz to 5Ghz. But it's output is *strong to see it on my SA. Thomas Lee's Planar Microwave book says not only are the diodes hand picked but the flatness is adjusted for each diode too.

R: I also confirm that the circuit is very simple but the variables are many and that each diode needs a specific bias to obtain good linearity. To make some measurements with the spectrum analyzer you need to add a preamplifier with at least 30dB of gain and a noise lower than 3dB. If your intention is to use it as a source to measure filters with a spectrum analyzer without Tracking generator, you need to amplify the signal of a good diode already calibrated by at least 60-80dB to have a decent measurement. Honestly, this way seems impractical and very unreliable to me, I advise you to buy a tiny Network analyzer for just over 100 Euros, more reliable and with many more functions.

Luciano
 

Offline TimeokTopic starter

  • Contributor
  • Posts: 13
  • Country: it
Re: Noise source for noise figure measurements
« Reply #14 on: July 31, 2023, 09:00:38 am »
start 2000 Hours
 

Offline mr ed

  • Regular Contributor
  • *
  • Posts: 98
  • Country: ca
Re: Noise source for noise figure measurements
« Reply #15 on: July 31, 2023, 10:07:12 am »
I have a tiny vna and it's great.  I am curious as  to how HP does it. Over the years they have made better and better outputs as designated by the product suffix letters A, B and C.  Seems to me that the diode is the trick, selection and biasing. Good uwave technique on low loss rogers boards would  also be important. It looks so easy and tempting  but in reality it's not.
 

Offline rhb

  • Super Contributor
  • ***
  • Posts: 3527
  • Country: us
Re: Noise source for noise figure measurements
« Reply #16 on: July 31, 2023, 12:13:02 pm »
I have a tiny vna and it's great.  I am curious as  to how HP does it. Over the years they have made better and better outputs as designated by the product suffix letters A, B and C.  Seems to me that the diode is the trick, selection and biasing. Good uwave technique on low loss rogers boards would  also be important. It looks so easy and tempting  but in reality it's not.

HPAK commonly uses the last letter to designate different upper frequency limits, e.g. 8648A/B/C/D = 1/2/3/4GHz maximum frequency in similar fashion as seen with the 346A/B/C.

Reg
 

Online G0HZU

  • Super Contributor
  • ***
  • Posts: 3239
  • Country: gb
Re: Noise source for noise figure measurements
« Reply #17 on: July 31, 2023, 12:33:49 pm »
I don't think you should consider the A, B and C versions of the 346 noise source as representing improved performance across A to C. They are just different devices optimised for different tasks.

For typical amateur radio and commercial radio stuff up at VHF and UHF, the HP 346A is by far and away the best one to have as it typically has an ENR of about 5.7dB and it has the lowest source VSWR by some margin. This makes it best suited for LNA design where the noise figure of the LNA might be 1.5dB and the input VSWR of the LNA might be 3:1 for example. The B and C version would be a poor choice to use here because they have a relatively poor source match.

You can always add an external attenuator to a 346B to improve the source match and lower the ENR, but you would have to then recalculate the new ENR for the noise source.


 

Online G0HZU

  • Super Contributor
  • ***
  • Posts: 3239
  • Country: gb
Re: Noise source for noise figure measurements
« Reply #18 on: July 31, 2023, 01:33:56 pm »
To show what I mean. here's a hot and cold VSWR plot of my Agilent 346A noise source from 300kHz to 6GHz. I'm not sure how accurate my VSWR measurement is up at 6GHz but this was taken with an Agilent E5071B VNA and it was calibrated with a N4431B 13GHz Ecal module. So it should be fairly close to reality.

You can see that the 346A noise source has an extremely low source VSWR across much of this frequency range.  The 346B and 346C versions will not show a VSWR plot as low as this.
 

Offline Uky

  • Regular Contributor
  • *
  • Posts: 125
  • Country: se
Re: Noise source for noise figure measurements
« Reply #19 on: July 31, 2023, 01:39:26 pm »
For amateur radio hobby purposes, I have used home built sources where I purchased NoiceCom diodes (expenceive and minimum order qty shared with friends) for years until I stumbled on professional sources which I grabbed faster than the speed of light. It seems as it is easier to find  an HP8970 or Eaton 2075 than a working calibrated noise source.The (lack of) temperature stability can cause problems. Comparing an Eaton 7626-1 with different HP 346x went in favor of the HP's. The Eaton source is housed in a much smaller aluminium case which clearly affects the thermal drift. The HP units were much more stable.

Thru-hole diodes manufactured by NoiseCom are useable well above their stated frequency range. They come at a price though. I have not tested bipolar transistors but if given the opportunity, testing an old NEC 21935 or 64535 would have been interesting. (Ceramic Micro-X package) When designing noise sources, I have used a number of very small (0402)  resistors connected in series feeding the noise source, selecting them for a high resistance that would hopefully prevent any "backward" frequency dependence and also by not using any ground plane around them as they connect to the noise generating component.

One such home made source made at my work employed a chip type diode (beam lead NoiseCom BL303) and a semi-ceramic board material. It was verified to be relatively flat all the way up to 8 GHz as specified. Here, a constant current regulator IC was used.

NoiseCom has a series of small noise sources intended for BITE. ENR around 30dB's or so (totally unattenuated). At my work, one of them was found to be broken. Not wanting to spend $1000 for a repair, I decided to try to fix it myself. The supply circuitry was built around a simple 78L three terminal regulator configured as a constant current source. The diode was OK. The regulator was broken. It took less than an hour to fix it. The only hurdle was reassembling the components properly after replacing the IC.

 :)

 

Offline TimeokTopic starter

  • Contributor
  • Posts: 13
  • Country: it
Re: Noise source for noise figure measurements
« Reply #20 on: August 01, 2023, 08:31:35 am »
Just a curiosity, I don't think the HP436x sources are produced by HP, but they are produced by Noisecom which, apart from the prefix, have the same number and the same case.

Luciano
 

Offline Uky

  • Regular Contributor
  • *
  • Posts: 125
  • Country: se
Re: Noise source for noise figure measurements
« Reply #21 on: August 01, 2023, 10:32:38 am »
I have noticed the same. It is a "strange" coincidence.

Somewhere amongst my files I have a downloaded hand-written schematic someone made after opening up a 346.
Judging from it, the design of the constant current circuitry is not that straight forward at all.

Some years ago, I had an Eaton 7626-1 (15dB ENR) calibrated with and without a precision 10dB attenuator added. Although that specific attenuator should have been good up to 26.5 GHz, the anticipated 10dB reduced ENR followed the accuracy of the attenuator somewhat but not exactly. Below 1GHz, the difference was marginal +/- 0.1 dB but at 1GHz and above the difference was +/- 0.2...0.3 dB. On spot frequencies even worse than that. This will of course add to the general uncertainty when measuring very low NF's.

I don not know if the telecom business does measure NF anymore. Last time I had anything to do with it, measuring BER and Vector Errors was more important.

 :)
 

Offline pdenisowski

  • Frequent Contributor
  • **
  • Posts: 941
  • Country: us
  • Product Management Engineer, Rohde & Schwarz
    • Test and Measurement Fundamentals Playlist on the R&S YouTube channel
Re: Noise source for noise figure measurements
« Reply #22 on: August 01, 2023, 01:39:01 pm »
The (lack of) temperature stability can cause problems.

Indeed.  One of the motivators in introducing so-called "smart" noise sources is that they can measure and automatically adjust for temperature changes.

https://scdn.rohde-schwarz.com/ur/pws/dl_downloads/pdm/cl_brochures_and_datasheets/product_brochure/5216_2718_12/FS-SNS_bro_en_5216_2718_12_v0400.pdf
Test and Measurement Fundamentals video series on the Rohde & Schwarz YouTube channel:  https://www.youtube.com/playlist?list=PLKxVoO5jUTlvsVtDcqrVn0ybqBVlLj2z8
 

Offline TimeokTopic starter

  • Contributor
  • Posts: 13
  • Country: it
Re: Noise source for noise figure measurements
« Reply #23 on: August 14, 2023, 08:05:26 am »
Hi all,

I did some tests about the ENR stability over  temperature.
To make a temperature step of about ten degrees I built a temperature controlled heater and inside it I mounted four NS.
The result is a shift of -0.05dB per degree Celsius. HP gives its sources for 0.01 dB per degree.

Finally I have  compensate this variation with an NTC which will vary the current in the diode to obtain 0.01dB. 


Luciano
timeok@timeok.it
« Last Edit: August 14, 2023, 12:04:58 pm by Timeok »
 
The following users thanked this post: pdenisowski

Offline zmetzing

  • Contributor
  • Posts: 41
  • Country: us
Re: Noise source for noise figure measurements
« Reply #24 on: December 13, 2023, 11:34:35 pm »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf