|Product Performed to Expectations:||7|
|Specifications were sufficient to design with:||8|
|Demo Software was of good quality:||7|
|Product was easy to use:||8|
|Support materials were available:||10|
|The price to performance ratio was good:||8|
|TotalScore:||48 / 60|
Welcome RF enthusiasts. For this road test, the antennas from Molex are surprisingly small, although not the smallest I have observed. It is important to be clear that most designer and makers would be unaware of the full function of RF systems and so this road test is to confirm operation as per specifications in the datasheets and in a way how well they will perform.
For the designer, there are only a few things that you need to know to incorporate antennas in your design.
An antenna at this period in time is more popular for the IoT environment and as many are not RF experts, therefore the datasheet is the most important documentation for determining an RF product is capable for the intended design. So do they though perform as expected? I hope to provide some indication.
For myself this was a disappointing road-test, not from the point of the product, but from changes in University policy made it difficult for me to use the EMC Chamber that I was hoping for. So I will not be able to carry out precise radiation performance, although I may have a chance after the road-test is finished.
Overall the linking of datasheet's are a bit disappointing. Approx 30% of links from the manufacturer's website ended up with 404 errors. Most of the problem is suspected automated publishing errors, or HTML database linking errors - https://www.molex.com/pdm_docs/ps/PS-47948-001-001.pdf.pdf as this link from the MOLEX website has the .pdf doubled. I am though surprised that this has not been picked up*. The issue is the same for almost all the antenna products in this road-test and more. Still, the actual documents are available but require a bit of address bar editing.
The quality of the datasheets is OK, but not really what I would have expected, but nevertheless still has details that are usable. Again, as a designer, I would be concerned trolling through multiple sheets where some of the plots should be present in the actual datasheet. I completed my Thesis for my final year project recently and I know that using the datasheets in this manner would have taken a lot more valuable time looking through to find the relevant information available. I will not compare to other sheets, though there is a need for more plots in the main datasheet that most would get to for looking at the performance of a product. Typically S11, Impedance plots and more should be in the product datasheets instead were found in one of the 404 pages that I edited.
The best of all the products with product details was the 146153-0100 antenna with the MCRF connector. There is a significant amount of information still though a little messed up in my opinion. The link though seems to show an antenna that is not the same. After searching the website the correct product was found - 206995-0150. Even after locating the correct page the link to some of the datasheets was still incorrect with doubled .pdf.
|*Note: Just prior to publishing the issue with the PDF linking had changed and all links were working as they should.|
|MOLEX will need to make a bit more effort in the online documentation for me to feel comfortable using the products. More visualisation is required in the most important document which should the first one to view. (https://www.molex.com/pdm_docs/ps/2069950150-PS.pdf.pdf) oops ( https://www.molex.com/pdm_docs/ps/2069950150-PS.pdf) This is key to getting a designer to browse over many different manufacturers design as which one is more appropriate. For anyone interested in how I know that it is automatically generated from an internal drive is at the bottom of the sheets with the words "This document was generated on".|
For maker and system designers there is a requirement to have a certain amount of information. The plots on Return Loss, Impedance and more after searching found this information in the Application Specification Sheets. With very clear plots and enlarge for easier viewing is a great document. Some of these plots could be well served in a more detailed Part Spec Datasheet. There is an obvious pride in the supply of designer information detailed on each page. These are simply though printout from CAD packages and not specific marketing information. Good for some people, but still not an easy read.
The most interesting read though is Sheet 10 of 25 (page number ?) The near-field plots are at 5500MHz, not the regular 5800MHz that would be used in most ISM band systems around the world. This would light up my eyes when it comes to the actual tests I performed which will be discussed later in this review. The 2450MHz performance though been more aligned to general use and performance was good as well. The 3D plots looked good, but I am not sure about the use of X, Y and Z directions. That may though be very fussy on my behalf and it is possible to look through that.
|These antennas are built for optimisation. I personally am to blame for not looking closer at the heading for the antenna. 5GHz is too simplified and I failed at looking at this before applying for the road test. The performance at the most important frequency is not there as detailed later in this review.|
For the most part, my setup was with an Agilent Fieldfox N9915A, although I attempted tests on a Rohde & Schwartz ZVL, the age of the ZVL made saving images antiquated and cumbersome. The N9915A also had a firmware update and was not performing the way I previously used, and I lost a lot of images taken of the display. Retesting was a grind and still didn't manage to get all the images I lost.
All the good test equipment was out on the job site in the Murchison Radio Observatory. It really wasn't my week for road testing.
The next bit of kit was one of the most important components is VNA calibration. In my case, I was fortunate that it wasn't taken out on the job. The 4 port Agilent E-Cal makes calibration an ease. The step so easy:
I lost most of the images for the calibration steps as well as the ZVL calibration steps. The result after calibration was a nice clean straight line on the S11 display shown to the left.
I struggled with the adaptor for the MCRF Connector so I decided to see how good my soldering skills are on connectors. I placed on an SMA connector on the fly lead ensuring the length was the same from pin to the antenna as with the original setup. It wasn't particularly hard just fiddly. The important aspect as I see it is would it be possible to replace the connector without affecting the performance.
Overall it did not it performed as expected, whether from my connection or from the manufacturer I am not sure.
When I realised that I didn't have an adaptor, it was too late to order one. Although a total cost of $40.00 wasn't in my budget.
It wasn't till the last week of the testing that I saw another connector for about $7.00 plus $15.00 delivery but no time again to order and get it delivered on time.
206995-0150 2.4GHz / 5GHz Wifi Stand Alone Balance Antenna with MHF Connector.
Surprisingly nice looking antenna, although for my applications for my project not really what I was looking for although it was great to test what I could. The antenna requires accurate placement, possibly one of the trade-offs when designing an antenna that does not conform to typical 1/4, 1/2 wavelength specifications.
This equates to a surprising amount of reflected power that will come back to the transmitter, although not as a severe effect on the received signal.
Yet as also seen on the Fieldfox there was a good amount of Return Loss (Lower the better) on other frequencies. Below left, the image is a simple translation of the return loss also know as VSWR (Voltage Standing Wave Ratio).
This is one of the key measurements for field technicians when tuning systems to the best performance. Unfortunately tuning this antenna is not as easy as a standard 1/4 or 1/2 wave whip style antenna.
Performance does change when moving the lead around although not as much to get the performance for 5.8GHz.
I have a lot more plots of the system but the 2 images shown here are the most relevant for evaluation. Impedance plots are not as good as expected.
|This antenna will more than likely perform but most people will not know the performance that can be seen using a VNA. RF Engineering is a lot of compromises. Did the manufacturer lie in the datasheet? For the most part no. There are so many variables to consider when testing, and it is likely my access to an anechoic chamber may have let my testing down a bit in the real environment, I wouldn't be overly impressed. What my most disappointing aspect is not reading the datasheet properly, I assumed the performance at 5.8GHz would be the best. but the tests carried out by MOLEX are at 5.5GHz. I also found during my test this is the antenna best performance as shown in the above is around this frequency. At 2.4GHz though I am not impressed. Could this bee the connector? Maybe. but the need to make changes in a design is crucial when selecting an antenna and was one of the main reason for changing the connector. The supplier does specify changes at various lengths and I had adhered to the same length and the connector choice shouldn't make a large difference.|
206513-0001 - 2.4GHz SMT Ceramic Antenna
This was the type of antenna that I will possibly use in the near future. Chip antenna is a common requirement for IoT applications and this one is a candidate for a possible Master by research or PhD that I may be starting from next year. The requirement for small and efficient antennas is a necessity now as products are becoming more portable. This will play a key role in my chosen project/s. These antennas are extremely fragile but reasonably resilient. I had to re-solder this chip antenna on twice, both for silly accidents. But for every-time the tests always came back the same. Unfortunately, I lost a lot of images for this antenna and the next one.
From the image to the left the S11 plot was bit disappointing, the peaks for return loss was centred just on 5.4GHz not 2.4GHz as it should be. Again the antenna was a bit of a let down for a general test. As mentioned I did have to re-solder the antenna but can confirm the results on this test is the same as the very first test I carried out prior to the unfortunate accidents.
I did have a VSWR plot, but was one of the lost images. It did reflect similarly to the plot shown on the left with great VSWR at 5.5GHz.
|This antenna could be re-configured to improve performance at the required 2.4GHz but not enough time could be devoted to the changes. There was a decreasing return loss at 2.4GHz and substantially broad but it appeared the antenna was more suited for the 5.5GHz than the required 2.4GHz that was specified. What is the main concern is that this antenna even if tunable could not be relied upon for the maker or hobbyist. For the professional time is money adding component to a development board is not value for money.|
47948-0001 - 2.4GHz SMT On-ground MID Chip Antenna
I am hoping to provide good news at least once and I must say I was happy to announce that this antenna performed just nicely. I was so relieved that the performance was great.
To the left is the S11 return loss plot, and as like last time the VSWR plot missing also show excellent result smack on the 2.4GHz frequency of interest. I was starting to consider my test equipment as a problem but relieve with this display.
Spreading the display across the 4GHz bandwidth was just an observation and it seems the antenna also performs a bit below 5.8GHz but is of no concern. Coming in a < -11dB return loss at 2.4GHz, it matches the Molex datasheet for performance.
For the datasheet, though I am pleasantly surprised the data I captured was so close to the application spec sheet.
|Very simply I am impressed. I cannot fault this antenna at all. It performed exactly as specified and would be comfortable as a maker or hobbyist to say it works and "should" perform well for any 2.4GHz IoT applications and any other application I the 2.4GHz ISM band. The only thing I would be careful of is the ground plane. I asked for the antenna to be supplied if possible without being soldered as I was planning to carry out the installation myself. Although the antenna was pre-soldered the risk is how well it will work on a different ground plane arrangement. In some way i am please as I may not have had enough time to add the necessary components to tune the antenna.|
Unfortunately, my final year project took a lot more of my time than I expected and started to run out of time to evaluate. Along with the new University policy I really hoped to use the university's EMC chamber. but as I was not specifically looking at 2.4GHz application (instead my project was looking around 150MHz) I had to look for a quick alternative to testing the radiation pattern.
The process quick and dirty. First, I fed the output of the Agilent N9915A in S21 mode into a Mini-Circuits ZX60-3018G amplifier (20-3000 MHz) then connected the amplifier to the 206995-0150 2.4GHz / 5GHz WiFi Stand Alone Balance Antenna with MHF Connector using this as a radiator.
As it was fixed in location to any radiation was considered fixed to the receiving antenna which I decided to use my favourite 47948-0001 - 2.4GHz SMT On-ground MID Chip Antenna. The result again was impressive. It mimicked the specification sheet al-be-it a bit broader in reception than expected was still nice to see.
As expected from a quick setup and not located in the best environment the result applies confidence in the product. The other results though were as expected but lost image hit me again. With the correct setup (EMC Chamber) the roughness is highly likely reflection around the lab bench and all the metal work surrounding.
I will be re-visiting this product in the future if I am awarded the scholarship for the PhD or Masters research project. I have already decided that 2.4GHz will be my IoT platform as antenna size is critically important. Although I do have other chip antennas to compare with as well.
If I knew how my project was going to end, I would have tested the antenna on 1.5GHz as only to see what potential there was.
My power supply is vertically polarised.
For the most part, I am more disappointed with myself on this road test. I was hoping to provide much better content and better experimental conditions. It is likely something went wrong with the tests but as the last antenna was so good and was performed under the same conditions it will be unlikely that the cause of the poor results. It shows a more real-world test environment test that some locations and physical conditions may upset the real results compared the datasheet, and that antennas are not easy products to deal with.
I personally would not be using the first 2 antennae for any projects in the future especially at 5.8GHz, but for the last antenna I will be hanging onto it and test it along with others, in line to be awarded their place in my project.
I very much hope I will have the time to continue the test soon, but my workload has picked up significantly with a number of components that are obsolete and busy locating replacements and retrofitting, also the holiday period is soon upon us and the University will close down for a couple of weeks. Maybe next year something else will come up.
The biggest concern on whether I will perform a road test on this type of product is having my own test gear. So I will no longer doing road tests like this until I have my own test gear or better access to the right gear. The only notable test equipment I have is two really great multimeters (Agilent and Keysight). To hang on for the use of test gear at university is not the best outcome for my sake.