NF measurement on 24 GHz preamps & mixers

Introduction

After building a preamp and adjusting it to minimum Noise Figure (NF) it is nice to know if the claimed NF by the designer can be reproduced. There are several ways to determine the NF of a preamp or mixer. But NF measurement isn't easy, there are many pitfalls.

Mixers & preamps to measure

Three 24 GHz preamps and three subharmonic mixers were available for measuring.
  • Preamp1 PA3CEG for reference 
  • Preamp2 PA3CEG new design 
  • Preamp3 PE1AEO new design 
  • Subh mixer1 PE1AEO 
  • Subh mixer2 PA3CEG (PA6C) 
  • Subh mixer3 PA3CEG 
Preamps at the left PE1AEO to the right PA3CEG.

The NF of one of the three preamps was measured on the Dutch Microwave rally in Heelweg (1999). This amplifier is for reference used. The two others are a newer design of the one for reference used.

The 24 GHz noise source was calibrated at a microwave meeting in Belgium (1999). i.e. the ENR (7.3 dB) was determined.

Measurement setup

For connecting the noise source to the preamp a Narda coax to waveguide adapter was used. Al preamps have a waveguide input and coaxial (sma) output. One of the subharmonic mixers was used to mix down to 144 MHz. Between the preamp and mixer there was an original OE3PMJ filter. For measuring the NF the well known design from DJ9BV in DUBUS was used. This piece of measuring equipment is a must for everyone who is building on microwave.

Results

After finishing construction of the new preamps they were aligned to minimum NF.

The first measurement was preamp1. The expected result should be close to the value measured at the Heelweg 1999 meting. The result was disappointing.

Preamp2 showed a higher NF with respect to preamp3. Investigation of the (HP) coax-wg transition showed a to short inner conductor. The coax-wg transition was replaced by a better one from Narda.
Preamp name
Reference NF
Preamp+db6nt MKII mixer
Preamp1
2.4
3.6
Preamp2
-
3.7
Preamp3
-
3.4
Preamp2 (Newly Aligned)
-
3.4
Table 1 

Curious why the NF of a system was still very high I started an investigation. First the gain of the preamps was measured. Al were in the range 25-28 dB. The insertion loss of the filter was measured. With 4-5 dB a little too high but should not be a big problem with almost 30 dB preamp gain.

With an other mixer the measured NF's were about the same. This left two possibilities:

  1. The ENR from the noise source was not determined correctly (3.8 in stead of 5.0 dB)
  2. The NF of the Subh mixer MKII was much higher as the claimed 8dB.
At this point it was time try to measure the NF of the MKII mixers. The LO for the mixers was checked 35mW @ 12 GHz. This should be enough refering to the author. A filter with low insertion loss was used between mixer and noise source. The first measurement didn't give any indication on the NF meter. The PANFI scale stops at 10 dB. Fortunately the DUBUS PANFI can also measuring NF's with the Y-method. With this method the results in Table 2 were measured.
Mixer name
NF (dB)
Subh mixer1
23
Subh mixer1
18
Subh mixer1
21
PE1JBK mixer (*1)
21
Table 2 

 *1 This NF was measured by someone else. Just mentioned for reference.

When measuring high NF's with a low ENR the Y-value isn't large. This will lead to some dB's of uncertainty in the NF. Typical xx dB Y-factor was measured.
 
 
 
 
CALL preamp 
Reference NF 
Preamp+db6nt MKII mixer 
 NF Corrected for mixer
PA3CEG Ref preamp 
2.4 
3.6 
 2.4
PA3CEG New 
- 
3.7 
 2.6
PE1AEO 
- 
3.4 
 2.2
PA3CEG New Aligned 
- 
3.4 
 2.2
Table 3 

Finally in 2001 preamp 1 and 2 were measured again on a professional site. An NF of 2.3 and 2.0 dB was measured.

Talking with Ron PA3BPC/DL3BPC made clear that there are some important factors on subharmonic mixers. He also had high NF's with the original MKII mixer. But replacing the suggested diodes by better ones and tuning the mixer resulted in a lower NF. Also important is the distance between filter and mixer (phase).

Final remarks

Noise figure claims from certain designs must be taken with some sepsis. HB9MIN writes in a Weinheim lecture: "NF from subharmonic mixers one should expect around 16 dB with good diodes".

The claim for the MKII mixer from db6nt is 10 dB and with tuning one can achieve <8dB is somewhat optimistic.

References

  1. Novel Approach To Automatic Noise Figure Measurement, Rainer Bertelsmeier, DJ9BV, Dubus 2/1990
  2. Noise Figure Meter MKII, Rainer Bertelsmeier, DJ9BV, Dubus 3/1994
  3. NOISE: Measurement and Generation, Paul Wade N1BWT (c) 1996
  4. Fundamental of RF and Microwave Noise Figure Measurement, Agilent Technologies, Application Note 57-2, Oct 2000
  5. Noise Figure Measurement Accuracy - The Y-Factor Method, Agilent Technologies, Application Note 57-2, 16 Feb 2001
  6. 10 Hints for Making Successful Noise Figure Measurements, Agilent Technologies, Application Note 57-3, 21 Nov 2000
  7. Practical Noise-Figure Measurement and Analysis for Low-Noise Amplifier Designs, Agilent Technologies, Application Note 1354, Sep 2000
  8. Applications and Operation of the 8970A Noise Figure Meter, HP Product Note 8970A-1.
  9. Harmonic Mixing with the HSCH-5500 Series Dual Diode, HP Application note 991, June 1984.
  10. K-Band Quadrature Mixers with Plastic-Packaged Diodes, Matjaz Vidmar, Microwave Journal, January 2000 pp. 22-40.

Back to PE1CQQ Page.
(c) PE1CQQ Last update 3-3-2007