![]() |
![]() |
![]() |
![]() |
Signalhound 12 GHz Spectrum analyzer and tracking generator |
![]() |
Maj : 02/04/13
This small page is a quick evaluation of small and low cost spectrum analyzer and tracking, to replace my a classical equipment, big, heavy, obsolete and expensive, the 40 GHz HP8564E + HP 85645A Before to read this page you must have an attentive look at the constructor's site, and don't forget in chapter "Demo" the Youtube intro which is a good presentation of the products, but in the reality of work, the sweep speed is not so idyllic... |
Cette petite page est une évaluation rapide du petit analyseur de spectre et tracking à faible coût, pour remplacer mon équipement classique, gros, lourd, obsolète et coûteux, le 40 GHz HP8564E + HP 85645A Avant de lire cette page, vous devez examiner attentivement le site du constructeur et ne pas oublier dans le chapitre "Demo" l'intro Youtube qui est une bonne présentation des produits, mais dans la réalité du travail, la vitesse de balayage n'est pas aussi idyllique ... |
|
|
|
|
|
|
Click to zoom this picture of the installation. |
Cliquez pour agrandir cette image de l'installation. |
First simple example of one of strange problems using sweep / Premier exemple simple de l'un des problèmes étranges en balayage
|
|
For control, GPIB screen picture of SA HP 8564E, showing spectrum up to 50 MHz |
Pour contrôle, image écran GPIB du HP 8564E, montrant le spectre jusqu'à 50 MHz |
There was a big problem on software versions before 2.18b. Always use the last version! |
Il y avait un gros problème sur les versions logicielles avant la 2.18b.
|
Second example : spurious frequency bug / Deuxième exemple: bug fréquence parasite
|
Image de l'écran du Signalhound SA (marqueur est de -10 à 6 = 16 dBm): Tout semble nominal |
For control, screen GPIB picture of HP 8564E, showing spectrum with span 50 MHz |
Pour contrôle, copie d'écran du HP 8564E, montrant le spectre en span 50 MHz |
With old softwares, major bug when clicking on "stop" sweep, high level parasitic signals remains. |
Avec les anciens logiciels, bug majeur en cliquant sur "stop" balayage, les signaux parasites de haut niveau restent. |
Accuracy of frequency / Exactitude de frequence
But now the good surprise. To verify frequency value, the easiest way is to look an external reference signal, it is a 10 MHz rubidium, very stable and verified with a caesium source. For everyday frequency measurements, I don't need an external reference, I know the small error to correct. |
Mais maintenant, la bonne surprise. Pour vérifier la valeur de fréquence, le plus simple est de regarder un signal de référence externe, il s'agit d'un 10 MHz rubidium, très stable et vérifié avec une source au césium. |
Questions to team development, and answers from Justin Crooks (
english only)
Thank you for your thorough review. I think I can provide a few answers:
Point 1: Did you plan shortly to make a new release to correct these major bugs for tracking output (2.18 software)?
Answer from Justin: The TG, at the end of a sweep, does not turn off. Rather, it continues to output the last frequency of its sweep. To turn it off, you can simply set it to 0 Hz, –30 dBm. We will integrate this into the software next revision so this will be the default action at the end of a sweep. Both of the big spurs seem to be the TG output still on. We will correct this so you will not have to manually set it to 0 Hz, –30 dBm.
If turning the TG “off” after each sweep is your primary requirement, I think I could add this fairly soon (over the next week or so). Let me know if this addresses the core of your concerns, or if I missed something major.
Closed, corrected (02/2013) by release 2.18b
Point 2: Smith chart utility tool works on a single frequency. What is the problem to improve for a frequency sweep?
Answer from Justin: The problem is there is no hardware mechanism for a phase locked sweep. We can frequency lock at any 10 Hz step, and measure the relative phase to null it out, but each frequency change breaks phase lock.
OK, no possible evolution
Point 3: How to read firmware version of SA and TG?
Answer from Justin: Help —> About will give you software and firmware versions.
To be completed: There is not an indication if version is for SA or TG, and no serial number of the boxes, waiting upgrade for correction for a more detailed information
Point 4: A lack of ergonomic
When powering on there is four times the panel "settings changes, RBW adjusted" that needs 4 clicks to start the panel "verify...". This is stupid, there is no other choice, why these traps, and not to start directly on panel "verify..."
Question 4: waiting upgrade for correction
Point 5:
A bug in subtracting traces
Very simple test, SA is connected to TG / Sweep on a band, for instance 10 MHz to 2 GHz / Copy A to B / Shows A-B
The result on the violet trace is about -20dBm (or anything else) and not zero ! The strange problem is that this error is not always present. I have repeated many times exactly the same protocol, sometimes it works well, sometimes I read a false value.
Exemple inserting a line with loss
Looks at this example, for instance at 1 GHz!
A (yellow) is at -21.5 dBm
B (blue) is at -10 dBm
A-B (violet) is at -21 dB
With classical arithmetic -21.5 - (-10) is about -11.5 but not -21...
Answer from Justin: waiting 2013 summer new software for correction
Other: To avoid the Error 5 (common in laptops), set your laptop power management settings, from the control panel, to “high performance.” This tells your laptop to optimize data transfer rather than optimizing battery life.
Things seems to be on the good way!
At the end of march, Justin answered to numerous complains with this next mail, we try to be optimistic and gives a reprieve waiting a sunny summer...
The mail from Justin: < I have your review page in my "Favorites". We will be re-designing the software
this summer, and will make sure to fix the trace math, as well as dozens of
other issues customers have brought to our attention. I know the process is
slow, and I am aware that the trace math seems like an easy fix. Unfortunately,
the trace math was written and tested for the spectrum analyzer, with the "STORE
THRU" function for the TG.
If you can, use the "STORE THRU" with the TG, which automatically subtracts the
"THRU" from all subsequent TG sweeps. Use the "Trace Math" when operating
without the TG. We will sort through all of the features and implement them in a
solid, robust fashion for our next release. >
|
|
The problem of USB / Le problème de l'USB
USB is not the best solution for a scientific instrument, Ethernet is faster, more reliable and at the same price in production. USB versions have a short life and compatibility problems, Ethernet is stable and not dependent or computer versions of software. SH made the choice of USB and claims that it needs two USB plugs, its a big problem on a laptop with only 3 USB plugs, one is used for mouse, no other one remains free.
In reality, this sentence is false, SH can works with some USB hubs on a 1 Ampere supply, but don't works on some other models. |
L'USB n'est pas la meilleure solution pour un instrument scientifique, Ethernet est plus rapide, plus fiable et au même prix à la production. Les versions USB ont une vie courte et des problèmes de compatibilité, Ethernet est stable et ne dépend pas des versions de logiciels ou de l'ordinateur.
En réalité, cette phrase est fausse, SH peut fonctionner avec certains hubs USB sur alimentation 1 Ampère, mais ne fonctionne pas sur certains autres modèles. |
Conclusion
After many years using classic solutions, it is a revolution to works with these small boxes. Software evolutions are expected, and in the future, many competitors will come on this market, but for the moment, despite various annoying bugs, and waiting next summer new software, this Signalhound is a good choice.
|
Après de nombreuses années d'utilisation des solutions classiques, c'est une révolution de travailler avec ces petites boîtes. Les évolutions logicielles sont attendues, et à l'avenir, de nombreux concurrents vont arriver sur ce marché, mais pour le moment, en dépit de divers bugs gênants, et en attendant le nouveau software l'été prochain, ce Signalhound est un bon choix.
|
There is some other competitors for spectrum analyzer measurement. If you works only on VHF, UHF, Rigol has a cheap solution. Rohde & Schwarz with the ZVH 8GHz gives an autonomous product completely finished, with vector analyse and without major bugs, with Ethernet link. It is a bit more expansive than SH, but if you add a very expansive separate vector analyzer, price of the RS single unit is very competitive. The limit of 8 GHz is too low for my operations on 10 GHz band |
Il y a quelques autres concurrents pour la mesure de l'analyseur de spectre. Si vous ne fonctionnez que sur VHF, UHF, Rigol a une solution économique. Rohde & Schwarz avec le 8GHz ZVH propose un produit autonome entièrement terminé, avec un analyseur vectoriel et sans bugs majeurs, avec liaison Ethernet. Il est un peu plus cher que SH, mais si vous ajoutez un très analyseur vectoriel séparé très cher, le prix du RS complet est très concurrentiel. Je n'ai pas fait du RS mon premier choix pour quelques raisons: La limite de 8 GHz est trop faible pour mes opérations sur bande de 10 GHz Quand j'ai commandé SH, je ne pensais pas qu'il était plein de bugs! Le prix du SH était plus bas, mais c'était une mauvaise raison, maintenant je dois multiplier mon budget pour ajouter un analyseur vectoriel 10 GHz très cher. |
To replace all antique material, now good as boat anchors for mooring, with new small ones recent units with USB or Ethernet plugs, some complementary units are needed. |
|
|
|
Liens
matériels de mesure et TV ham
Autres pages sur la mesure et le radioamateurisme :
Achat et vente de matériel électronique :