Antenna for cycling race

[Christian C.]

Hello,

What are some brand/model of good antenna that have been tested with CrossMgr? Also is there any recommended setup to use for MTB, Road and CX racing?

Thank you

C.

[stuart lynne]

We have been using this: http://www.superpassantenna.com/Antenna/902-928MHz/ProductSPPC12.html

It is sold as a vertically polarized antenna when used upright. But works as horizontally polarized when used on it's side.

It has a tighter beam and higher gain than the previously suggested SPLC11H from SuperPass.

We have two boxes like this, one for each side of the finish line. Used for road and cx. 

--
You received this message because you are subscribed to the Google Groups "CrossMgrSoftware" group.
To unsubscribe from this group and stop receiving emails from it, send an email to crossmgrsoftware+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--

__________O___________
_______-\<,____________
_____(_)/_(_)___________
_________________________
Stuart_Lynne____<stuart...@gmail.com>____604-518-1749(m)__604-461-7532(h)

[Jonathan Rosen]

I found a 14dbi 900mhz Yagi antenna on eBay, delivered for $20. I bought two and use them for road and criterium races. They are 4-feet long and mounted on a stand made out of PVC pipe from Home Depot. It works great, they have a very narrow 30-degree beam pattern and are very high quality. 

[Jonathan Rosen]

14dbi Yagi pair with 1-meter LMR400 cables on a PVC stand. 

[Carsten Machholdt]

Hi Jonathan

Could you telle little about read rate and accuracy with
these.

How many riders was participated and how many came in, in a group?

How far away can the chip be read?

Best regards 
Carsten 

[Jonathan Rosen]

Using the Impinj R220 reader, its rated at 600 tags per second. At criterium races with 34 riders going by at 30mph, every tag gets read; I don't consider rider density as a limitation.  With the 14dbi Yagi I can read H47 chips at 10 meters or more. Accuracy is now increased due to the new QR and RANSAC algorithms in CrossMgrImpinj; read the other posts for more info on this advancement. 

https://www.ebay.com/itm/YA90012-DMS-Wireless-900-MHz-Directional-Yagi-Antenna-with-12-0dBi-gain/122616432146?hash=item1c8c824612:g:5RAAAOSw1WJZMpY5:sc:USPSPriority!14221!US!-1

[stuart lynne]

The eBay link is for a 12dBi. Do you have a pointer for the 14dBi version?

How and where are the tags mounted? In what orientation?

Still at about $25 each with shipping (to canada)... I just grabbed two to play with.

On Sun, Jun 24, 2018 at 1:46 PM Jonathan Rosen <jonrose...@gmail.com> wrote:

Using the Impinj R220 reader, its rated at 600 tags per second. At criterium races with 34 riders going by at 30mph, every tag gets read; I don't consider rider density as a limitation.  With the 14dbi Yagi I can read H47 chips at 10 meters or more. Accuracy is now increased due to the new QR and RANSAC algorithms in CrossMgrImpinj; read the other posts for more info on this advancement. 

https://www.ebay.com/itm/YA90012-DMS-Wireless-900-MHz-Directional-Yagi-Antenna-with-12-0dBi-gain/122616432146?hash=item1c8c824612:g:5RAAAOSw1WJZMpY5:sc:USPSPriority!14221!US!-1

--

You received this message because you are subscribed to the Google Groups "CrossMgrSoftware" group.

To unsubscribe from this group and stop receiving emails from it, send an email to crossmgrsoftwa...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

[Jonathan Rosen]

I didn't find the 14dbi for sale on eBay today, but you can contact the supplier of the 12dbi to inquire if they have any 14dbi available. They have a nice tight 30 degree beam width in both directions (horizontal and vertical.) Before CrossMgrImpinj had quadratric regression, I really wanted an antenna with the highest gain possible and the narrowest beam width, since it would use the first detected chip time.  

Now that we have QR, having a wider beam width could provide an advantage, letting the Impinj reader gather more chip reads over a wider area. Looking at pictures in CrossMgrVideo, I can see that 14dbi Yagi would read chips up to 5 meters or more before the finish line, well beyond the 30 degree beam width. I think the 12dbi Superpass antenna you have is fine, I doubt you'll see any difference or improvement between that and the 12dbi Yagi. The Yagi is a nice upgrade from the common circular patch 6dbi antennas. 

[stuart lynne]

One question, for the Yagi style antennas, do you think they would be better or worse for reflected reads.

I have one venue where the finish line is next to a large metal stage, with metal safety fencing on the other side. I get lots of reflected reads. Especially from the antenna on the far side of the finish line (facing the stage.)

On Sun, Jun 24, 2018 at 5:19 PM Jonathan Rosen <jonrose...@gmail.com> wrote:

I didn't find the 14dbi for sale on eBay today, but you can contact the supplier of the 12dbi to inquire if they have any 14dbi available. They have a nice tight 30 degree beam width in both directions (horizontal and vertical.) Before CrossMgrImpinj had quadratric regression, I really wanted an antenna with the highest gain possible and the narrowest beam width, since it would use the first detected chip time.  

Now that we have QR, having a wider beam width could provide an advantage, letting the Impinj reader gather more chip reads over a wider area. Looking at pictures in CrossMgrVideo, I can see that 14dbi Yagi would read chips up to 5 meters or more before the finish line, well beyond the 30 degree beam width. I think the 12dbi Superpass antenna you have is fine, I doubt you'll see any difference or improvement between that and the 12dbi Yagi. The Yagi is a nice upgrade from the common circular patch 6dbi antennas. 

--

You received this message because you are subscribed to the Google Groups "CrossMgrSoftware" group.
To unsubscribe from this group and stop receiving emails from it, send an email to crossmgrsoftwa...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Jonathan Rosen]

I really don't know anything about direct vs. reflected reads. How can you tell the difference? 

[stuart lynne]

With the old scheme (first read counts) by seeing people drastically early in the finish order compared to what the finish line judge or camera saw.

What happens is the beam from the antenna heads out at an angle, bounces of something metal hits the tag, gets the read back in the same fashion. So effective distance has 2-3 meters added to it (plus width of road etc.) But more importantly, you can now read tags a few meters early. Which generally you don't want (or we would have just bought antennas with a wide angle etc.)

It is quite possible with the continuous reads and quadratic fitting, these can be weeded out. But I need to get enough bad examples so we can understand the characteristics better. Generally (I think), when we see reads that are far away (in time) from the center line, then probably they should be ignored. The trick will be trying to recognize that algorithmically. 

On Thu, Jun 28, 2018 at 8:40 PM Jonathan Rosen <jonrose...@gmail.com> wrote:

I really don't know anything about direct vs. reflected reads. How can you tell the difference? 

--

You received this message because you are subscribed to the Google Groups "CrossMgrSoftware" group.
To unsubscribe from this group and stop receiving emails from it, send an email to crossmgrsoftwa...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Jonathan Rosen]

I don't think the early reads are due to reflections. I would expect a reflected signal to be weaker than a direct signal. Looking at the QRplots you provided, I think the outliers are a better example of reflected reads. 

All of the antennas have beam widths from 30 to 60 degrees or more. The 14dbi Yagi I have has the narrowest beam width at 30 degrees. However, the beam width doesn't confine the read zone, it simply defines the boundary for the highest sensitivity. Chips can be detected at around -70 dbi RSSI, so a chip can frequently be read far away from the finish line. For a 30 degree beam (15 degrees on each side of the antenna) that results in a 2:1 ratio between chip distance and finish line/antenna line distance. 

Looking at the pictures in CrossMgrCamera and CrossMgrVideo taken before QR was available, I can see that chip times are being recorded up to a full second before the bike reaches the finish line. At 17mph, that's 25 feet per second. Now with QR, the chip times in the pictures consistently show the bikes on the finish line. It can't be overstated how important the use of QR is in improving the accuracy of identifying the chip time. 

[Jonathan Rosen]

Do you have any data from races where there were two antennas adjacent or stacked? 

Ideally, I'd like to see the QRplot data for two antennas that are stacked and reading the same chip. I would expect they should both have a nearly identical curve, but due to the height difference, one antenna should be able to get some chip reads that the other one misses. It would be useful to see a comparison between two separate plots and the combined plot from both antennas to see if it yields an improvement. 

[Carsten Machholdt]

Hi

 

I'm very interested in the debate in this groupe.  I have not considered placing the antennas over each other. I always place them next to each other. I'm not sure it's the best solution.

 

I only make time for cycling. My setup consists of the following: 

Motorola FX9500 4 Port Reader
HuTax XC1 and XC2
3 linear antennaes 

Max participates around 50 riders

 

http://timeu.dk/resultat/20180620/resultat.html

 

I this race there was around 350 reads and I missed around 8. Raceline was around 5 meters.

Should I try to stack antennas on top of each other or still use side by side.

Please see exampels

[Carsten Machholdt]

The Hutag chip is located on the right ankle of all riders

[Jonathan Rosen]

With the antennas side by side, it should pickup more reads, but using the QR feature will result in three different finish lines. Each antenna will create its own virtual finish line, so while the antennas may pickup more reads, the results will be out of order and require more corrections. Switching to the stacked antennas will get the finish line times to line up in the correct order. 

I think if you put all three linear antennas in a stack separated by around 30-40cm, it might yield better results. Certainly, with stacked antennas you should have fewer corrections to the finish order. With the HuTag on the rider, there is going to be a lot of movement of the chip, and the chip will be low, so having antennas at different heights may help see the tags better. 

It would be useful if CrossMgrImpinj had an option to specify how far apart each antenna was for a side-by-side configuration. 

[Robin Karinen]

Where are the chips attached when using the yagi-antennas?

Thanks

[Jonathan Rosen]

Here's a picture of the ChipCard attached to the seat post with two Velcro straps.  The Impinj H47 chip is on the other side of the CR80 plastic card.

[stuart lynne]

Here is my current version. We print the number on demand during check-in.

So far it appears to be reasonably good at eliminating wobbly frame plates. But we have had a few people find ingenious ways to mount them vertically. 

On Tue, Aug 14, 2018 at 1:26 PM Jonathan Rosen <jonrose...@gmail.com> wrote:

Here's a picture of the ChipCard attached to the seat post with two Velcro straps.  The Impinj H47 chip is on the other side of the CR80 plastic card.

--
You received this message because you are subscribed to the Google Groups "CrossMgrSoftware" group.
To unsubscribe from this group and stop receiving emails from it, send an email to crossmgrsoftwa...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Robin Karinen]

What do you think of this antenna? https://www.aliexpress.com/item/RFID-868mhz-helical-antenna-14dBi-Circular-Polarization-outdoor-directional-lorawan-antenna-Best-quality/32829800440.html

I know most of the usual UHF RFID panel antennas are usually using circular polarization. This is done to better be able to transmit and receive to tags that are not in optimal orientation to the antenna. More info: https://blog.atlasrfidstore.com/circular-polarization-vs-linear-polarization

Maybe even a combination could be used? 4 different types of antennas :) (panel(circular), helical(circular), yagi(linear), helical-yagi(circular))

[stuart lynne]

The chief problem with circularized antennas is the approximate 3dB signal loss.

Compared to linear polarized antennas of the same gain, circular polarized antennas will have a shorter read range because they lose about 3 dB splitting their power across two separate planes.

The tradeoff is to try and accommodate poorly mounted tags or get best reads with properly mounted tags.

Hence the interest in tag mounts that lead to well-mounted tags that don't move once mounted. With a bit of persuasion for people that come up with innovative ways to do it badly.

That said, the Dogbone tags do work when mounted 90 degrees out of phase. But only when they are alone in the FOV. If they are in a big group of riders the read rate goes down. So OK for cyclocross which mostly has singles and doubles finishing together. But not so good for crits with big mass finishes. 

--

You received this message because you are subscribed to the Google Groups "CrossMgrSoftware" group.
To unsubscribe from this group and stop receiving emails from it, send an email to crossmgrsoftwa...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Robin Karinen]

I was now able to purchase a RFID -reader for a reasonable price. I am now trying to figure out the antenna situation. I wonder how much the actual power will be limited using a yagi that is circularly polarized compared to horizontally/vertically polarized antenna.

For example, the antenna you listed has a gain of ">12dBi” and beamwidths of 62 and 34 degrees. The helical antenna that I linked to lists 14dBi gain and 35 & 35 degree beamwidths. So comparing it would be a 11dBi antenna with nearly half the beamwidth of the horizontally polarized antenna in the horizontal plane. How much should this increase the power delivered or received? Also doesn’t the circular polarization add the advantage for when the tag might be even a few degrees of from a horizontal antenna’s plain, it still reads or gets a better signal?

[Robin Karinen]

Also here are some results that were done with a 6dBi circularly polarized antenna.. (Thinkify TAC-60) I’d imagine a helical antenna with 5dBi more should be able to perform similarly.

https://youtu.be/9Ia9ZIMGEOE?t=22m33s

I guess there is only one way to find out, really..

[Jonathan Rosen]

I think the formula for calculating transmit power involves subtracting the antenna gain (6, 9, or 12dbi) from the transit power (29 to 32db gain) and then adding back the loss from the tag (20db to 70db) I think that formula was designed for a static situation, where the tag position and RSSi are very consistent and close. The goal was to avoid saturating the receiver antenna with a signal that was too powerful for a tag that was close to the antenna (probably 1 meter or less.) This is not the situation with bike sports, where the tag is relatively far from the antenna (1 to 10 meters) and always moving. With bike racing, we always get a curve on the tag RSSI because its moving past the antenna. 

The bottom line is that the receiver defaults to max transmit power and we want the most sensitive antenna possible using a tag that provides the lowest (best) RSSI. Basically, setup your antenna and position a tag at a typical distance (3 to 5 meters) and then evaluate the RSSI on the tag in different positions. The latest version of TagReadWrite provides a snapshot value for RSSI, so in a static test, you can easily and quickly evaluate the tags performance. Rotating the tag will reveal if polarization is significant or not. 

Regarding antenna types, I'm a big fan of the Yagi since it provides higher gain, which means the tag gets read at long distances. For bike racing, the tags remain in a consistent position, they aren't moving around or showing up in different positions like a runner with a bib tag, so the circular polarized (CP) antenna doesn't provide an advantage. Also, the circular polarized antennas always have lower gain than a Yagi, so that's not useful. Once again, the proof is in the testing. Its easy to connect a CP and a Yagi to a 2-port reader and then use TagReadWrite to select an antenna and compare the RSSI. 

Regarding beam width, I have used a 14dbi yagi for all of my races, and the relatively narrow 30 degree beam width was still wide enough to generate up to 20 reads from a passing chip at 30kmh, and at the highest speeds (50kmh) I was still able to get 6 reads from a chip, more than enough data to perform quadratic regression with CrossMgrImpinj. Whatever the antenna, if I'm getting 3 or more reads from the chip, I'm getting the best possible result. The final measure of success is looking at the percentage of tag reads that have QR data. Ideally, we want every tag to be read 3 or more times for QR.