Comparison of FX9600-8 and FX9500-8 readers

[Hronometar]

After 2 months of testing my new 8 port reader FX9600, I have noted it doesn't give such a good performances compared with 8 port reader FX9500.

I have started using 2 RR mats this Spring and always have had 100% read rates with two FX9500-8 readers. All chips have been read always on the first line, doesn't mater I use DogBone MR6 wet inlays or the inlays covered with foam. It looks I could time all finish lines without backup system, but I do it with backup.

In July I bought a FX9600-8 reader to have it like spare reader and started testing and using it in August. The reader doesn't provide 100% read rate with RR mats. On some races that I was timing with side antennas, I also didn't have 100% read rate. It has never happened I have missed on finish line since I switched to RR mats and FX9500 readers. Doesn't mater it is a 5/10 k event with 1000 participants or 21 k race with 100 participants. I will make a few comparison and show you the setup I use.

Side antennas (10-11 dB circular antennas used with LMR-400-UF cables):

• FX9500 setup for OCR with ~250 participants with HuTags and 100% read rate on the first line with only 2 side antennas,
• FX9600 setup for OWS with ~75 participants with HuTags and 2 missed (same setup and equipment like for OCR but FX9600 was used).

RR mats on half marathons finish line:

• sample 1 - 2 lines of RR mats where the first line is with FX9600-8 reader (more that 10 missed on the first line and all were read on the second with FX9500-8 reader),
• sample 2 - 2 lines of RR mats where the first line is with FX9600-8 reader (more that 10 missed on the first line and all were read on the second with FX9500-8 reader),
• sample 3 - 2 lines of RR mats where the first line is with FX9600-8 reader (I was counting missed chips and there were 27 missed on the first line and all were read on the second with FX9500-8 reader - it was yesterday and that race had around 175 finishers).

Yesterday I was maximally concentrated what was happening with timing with FX9600 reader, and I found that there were finishers that finishing pretty fast and pretty slow that their chips were not read. It looks like ART takes more time for each antenna to be active with FX9600 compared to FX9500 and some participants are not caught because of that. Brian, please, can you check this in your code?

FX9600 definitely works better with my side antennas than RR mats. I had some races where I was timing turning points on the course with only 2 side antennas with old FX9500 readers and I had usually 100% read rate doesn't mater how many laps that race have. That race I showed you setup had 6 laps and there were around 200 participants. It means there were 1200 splits on that timing point and there were no any missed.

But a few weeks ago, I tried the same with new FX9600 and there were 4 missed in the first lap when it was more crowded than in 2nd lap. Here is the setup. I had one more turning point on that course I was timing with FX9500-4 reader with exactly the same setup, and there were 100% read rate on both laps for all 350 participants. Setup of another turning point.

I will give you more details of equipment I use and my setup. For most of my races I always use DogBone MR6 vertically attached on the back of racing number because of RR mats antennas. Sometimes with foam sometimes without foam. I still have some amount of integrated inlays with foam, and when I spend all of them I will use only wet inlays as I always have 100% read rate with wet inlays when I use RR mats and FX9500 readers on finish line. Often on start line I have 100% read rate with 2 lines of the RR mats. Distance between 2 lines of RR mats is 4-5 m. Never less than 4 m. So, the same distance is between two FX9500 readers or between a FX9500 and a FX9600 reader. As you could see from photos, they are powered with Zebra's power supplies. All readers are with static addresses.

My FX9600 is configured for EU region. All 4 frequencies are enabled and an auto switching between frequencies is activated. I didn't find FX9500 has this setup for EU frequencies.

I didn't find any of you have complained on FX9600 readers. Maybe it is about my reader, but if the reader has a problem, I would have worse read rate on turning points with side antennas.

Anders, I have read in some of your posts that you have started using FX9600 reader. What are your experiences about it? How did you set your frequencies for EU region?

Brian, I think there are 2-3 reasons that FX9600 doesn't provide 100% read rate:

• FX9600 reader is not so strong like old FX9500 reader or a firmware in FX9600 should be better tuned, or
• ART doesn't take max performance that FX9600 provides.

Did any of you test those 8 port readers similar like me?

[blueridgetiming]

I've experienced similar results with the 9600 compared to the 9500 using RR mats. At times, it almost appears as if the reader is off since it seems to miss reads in chucks for me. For example, it'll miss 2 or 3 within the same second, but then read fine at other times. 

[Hronometar]

Thanks! We are gonna see if there are more timers experiencing similar problems.

Didn't have a situation that finishing small group of participants and any of them are missed. Yesterday I had two participants they were racing last 50 m before finish and faster one was read and slower one wasn't. The slower one was maybe 1 m behind the faster one.

After 10th missed I started writing how many participants finished when they were misses:

number of missed	number of finishers in that moment
10	88
15	108
20	129
25	154

[JK Race Timing Systems]

Hi

I know nothing about the FX9600 but one thing you could check  is the multiplex sequence just in case it is different to the 9500. Might be no harm checking what the antenna sequence is and if they all have the same activation times by default. There is a load of customization that can be done on this. You have the ability on these readers to leave them activated longer or shorter on different antennae.

Example If for example someone decided to run 8 antennae on a gantry all in line( Madness but just pretend) and the antennae are numbered 1 to 8 from left to right.

In this example the worst antenna Multiplex sequence would be 1,2,3,4,5,6,7,8

It would mean that sections of the gantry would be "Dead"  over at side 4,5,6,7,8, for unacceptably long periods of time 

Really what you would ideally looking for is a sequence like 1,8,5,2,7,3,4,6 or the like . This sequence will give the maximum coverage on the gantry for a given time period.

That is what I would be checking first as it would have an immediate bad influence.

Check aswell of course that they are ALL set to monostatic mode. The 9500 has the ability to set Any antenna ( or pairs) to be either Monostatic or Bistatic or Both.

Log into both readers and print out the full setup, That may shine a light pretty fast if you have a setting interfering on your setup

Best regards

JK 

[Brian Agee]

Hmm, very interesting!

The FX9600 I purchased to test the interface is at my head timers house. His son is old enough now to be his own timing crew - however he times the smaller/easier races since this is his first year on his own. His son is the one that has the FX9600, he hasn't reported any problems, but again his races are typically less than 150 participants. My head timer and myself are using the FX9500's that we've had since the beginning. I have another FX9600 at my house that is defective - for some reason, in order to establish a connection, a factory reset is required every time the reader is powered on. However I can use this device to compare the default settings for the FX9600 against the FX9500.

Now that you have indicated that there may be a difference between the readers, I have a suspicion on what the issue might be:

The FX9600 uses the same 'guts' as the FX7500. When I first tested the FX9600 I found that it had the same problem as the FX7500 did when I first tested it: See here

That video was created after many weeks (maybe months) of trying to convince Zebra that something was terribly wrong with their new FX7500. After this video they finally acknowledged the problem and send me some code (which was not documented in their interface guide) that appeared to resolve the issue. It appears the code simply sets the stop time between each antenna to 100ms. I will have to look back on our emails to see why it is set to 100ms and not something lower, I would imagine I would have preferred to set it as low as possible. I'll do some digging and see if I can find this setting in the FX9500 and what value is being used by default in it.

The FX9600 had the same problem as shown in the video, but of course I immediately knew what to do and applied the same fix and it appeared to work.

[MarkE]

I too have a 9500 & 9600 and run most larger events with 2 RR mats.  One of my mats is the older all black style and 1 is the newer red/black accordion style.  I typically set up with the 9500 and the newer mat on the front and the 9600 with the older mat ~4-5m later.  Front mat would always read 100% and the rear would read around 98%.  So this past week I had an event with 1425 entries and switched to the 9600 up front with the new mat and the 9500 in back with the old mat.  For me, the low read rates stayed with the mat and not the reader.  No misses with the 9600/New Mat and ~98% with the 9500/Old Mat.

It did seem like there was a bit more lag between a finisher and a time being displayed with the 9600, but I was attributing that more to the busy finish line and not the reader.

[blueridgetiming]

Here is some more data. A few weeks ago at a 750 finisher HS xc event, on a very sloppy course with 90% humidity (but no rain), I missed 6 on the main system which was a 9500 with newer RR mats, and missed 25 on the backup which was a 9600 with newer RR mats positioned 15 feet past the finish. The 9600 was POE (I wanted to try it), so that could be another factor.

All chips were vertical on back of bib.

[Anders]

What I have learnt is that 9600 inside is the same as 7500. That means that it has a really low performance compared with 9500.
We have five 7400/7500, one 9500 and one Impinje reader. The 74/75 is only for spare since they have to low performance. When we have 95-100% read with 9500 and Inpinj, the 74/75 has only 60-80%.

[blueridgetiming]

I'll give the 9600 way more credit than my 7500 paperweight. I've gotten 100% reads in 600+ events with the 9600. It's just when I noticed misses with the 9600 they came in chunks

[Muddy Dog Sports]

I've got a question, when I use both of my RR mats on the same finish line I have both readers connected to a single laptop, which of course means that if the laptop goes down the entire system goes down. Am I correct in understanding that you guys are using two laptops which in effect gives you two completely stand alone systems? If yes, I'm wondering how you go about checking for missed reads on the primary system and combining the results? 

I know my question isn't related to the 9500/9600 dilemma, but appreciate any advice on how others are setting up for these larger races - thanks!

[MarkE]

I run 2 mats on 2 computers so, as you say, two stand alone systems.  I do this for the very reason you mention, I wanted a true backup.

The way I use it, I watch the finish line and just count finishers.  As frequently as I can, I glance down and verify the number of finishers I have matches the number of finishers in ART.  If there is a mis-match, I write down the finish range for later reference.  After the race, I double check and compare the two computers for the areas I noted as concerns.  There is likely a better way, but this is how I've been doing it.  Hope it helps.

[Peak Performance Timing]

I do it the same way MarKE does with the addition after the race I make an ART export from each computer and I have written an Excel Macro that allows me to compare the two databases more quickly.

[BWRTiming]

Hmmm. It seems like the Import Splits option should do the trick. So long as you import reads from a backup mat that is after the primary mat, missing numbers should show up and anything that's already there should be ignored since it should already have a smaller time than the Import.

But I've never used The Import Splits option and in my test tonight, it failed. Got a debug error. Was I just doing something wrong or is that option some how out of sync with all the recent enhancements?

[Tim Irvine]

BWRT were you setup for splits (multiple laps)? If not, that may explain the error

[Peak Performance Timing]

BWRTiming,

I just tried your method of bringing in missed splits for the finish on a single lap race......It does indeed draw anything in that was not in the database ( in a single lap race anyway...) This is GREAT to know. No matter whether the backup time is faster or slower than the primary time, if a time is in the primary it leaves it alone.....if it is missing, then it brings it in. I would like to have Brian chime in on this and tell us if this is the way it is supposed to work, if it is just coincidental, or a planned feature I was unaware of until now.

[Brian Agee]

I have been eager to dig into this after the Fall timing season ended,  I wanted to update everyone on what I've found so far and also ask for you help testing a possible solution.

In an earlier post on this thread I wrote about how, when the FX7500 was first released, I had to work with Zebra to figure out why the default switching rate between the antennas was TERRIBLY slow. They showed me some code that I could call when connecting to the reader that would drop it down to 100ms. I can't remember the conversation we had back then, but I assume that I would have told them that I need it to be the same as whatever is used by the FX7400 and FX9500. However it's possible that I did not ask this or that they provided a value that was not correct (or as good as the FX9500's).

I have looked through the web console on both the FX9500 and FX9600 and I cannot find the switching rate in there, so I hooked up my FX9500 ran a function to query the reader for the switching rate. Oddly it returned zero. However I'm not sure if zero is returned if the property that I tried to access doesn't exist (the function I called may only work on the FX9600/FX7500) or if zero is the true value...

So here is where I need your help!

I updated the software so that the switching rate for the FX7500 and FX9600 is set to zero. However I cannot test it because I sent my FX9600 to SenThys so that they can benchmark their new rugged 8-port reader (which is not available yet) against the FX9600 and FX9500. When they asked me to send them the FX9600 and FX9500 the only thing I asked for in return was to find the switching rate of the FX9500 so that I can set it to be the same for the FX9600.  SenThys will share the results on how each reader compared when they are finished and I'm sure we'll all be interested to see those results. So the point is, I cannot test this out until I get my reader back, which may be a couple more weeks and I'd rather get this fix as soon as possible.

1. Let's first test the FX9600 BEFORE getting the update so that we can know if this update improves the performance.

2. Place 4 antennas (two on each side) so that they are pointing to a single tag. Using the FX9500, open the Settings screen and click on the RFID Testing Center screen. Read for exactly 10 seconds (I'll use 10 seconds as an example, but it doesn't have to be 10) and see how many reads show up in the software. The problem is that the software clears the listview every 1000 reads - otherwise the list can get so long that it freezes up the software. So you may want to record the results or get someone to count how many times it reached 1000 and started back at 1. Ultimately you should be able to tell exactly how many reads were sent to the reader (if it cleared 3 times and you now see 214 reads in the list then obviously there were 3,214 reads).

Oh, and I just remembered that the RFID Testing Center screen does not have a "stop listening" option, so you may have to find a way to block the tag from the antennas at exactly 10 seconds so that you can see how many reads occurred before closing the screen. :(  Hopefully that's not too hard though.

Repeat this process a handful of time until you feel that you've got a good grasp on how many reads occur over a 10 second period. Now do this same test using the FX9600.

3. Do this same test listed above, but you only need to use the FX9600, and this update to the software. Hopefully you'll find that this update does improve the FX9600's performance.

In case you're curious, this update also has the following fixes/enhancements:

• You can now pull an unlimited number of races from RunSignUp. I was unaware that it was pulling only the newest 50 races.
• Fixed AthleteType bug in OverallResults report.
• Fixed place field in AthleteType reports.
• Added option to export a USATF Result File.

If you want to go back to the update that everyone else has, you can manually uninstall the software (via Control Panel / Add or Remove Programs), and then reinstall it using the Free Trial button on my website.

I am looking forward to hearing what you guys find on this. Thanks!

[Brian Agee]

I have no idea why I didn't think of this until now. But I have one (or more) FX7500's at my house that I can test it with - since the guts are the same as the FX9600 I can test this myself instead of asking you guys to do it for me. If anyone does test it then of course I appreciate it, but I'll try to test it tonight and share the results.

[Oscar's Race Results]

I know my 4 port 9500 whines when you hold a transistor radio antenna near it and an FX 7400 rattles at about 5 Hz.  And if you thought your microwave was completely clean, hold the radio antenna next to it.  Interesting sound changes with number of antennas hooked up and whether or not you have them turned down to nothing in Session One.

Be interesting to hear what the 9600 and 7500 sound like.  

Oscar 

[Brian Agee]

Very interesting results in tonight's test:

I altered the RFID Testing Center screen so that it would count how many reads were passed to the software in EXACTLY a 10 second period. It would start the 'timer' when the first read came in and then stop the counter at exactly 10 seconds.

Hardware setup:   I had 4 antennas setup where all antennas were touching each other (see attached image), facing straight down on my couch, with an RFID tag laying on the ground under the antennas. This way all antennas could see the tag. My goal was to see if there is a difference in read rates based on how fast the reader was switching between the antennas.

I started off with the FX7500, then switched to the FX7400, and finally the FX9500. I was careful not to move the antennas or tag while switching out each reader. I also make sure to plug the same cable into the same port on each reader. For each reader I did 10 sessions of reading for 10 seconds and recorded the results. Here is what I found:

Here are the results:

FX7500 Reads:

1. 887
2. 924
3. 944
4. 907
5. 926
6. 943
7. 914
8. 928
9. 902
10. 924

FX7400 Reads:

1. 216
2. 217
3. 218
4. 203
5. 217
6. 214
7. 213
8. 214
9. 213
10. 219

FX9500 Reads:

1. 1319
2. 1475
3. 1564
4. 1593
5. 1606
6. 1643
7. 1621
8. 1635
9. 1649
10. 1599

I tried dropping the switching rate of the FX7500/FX9600 down lower but it looks like 100ms is as low as it will go. The software was  using 100ms already. Anything lower than that and it will not read any tags at all. I played around by making the switching rate higher (200, 500, and then 1000ms) and the number of reads didn't change much at all since each antenna could see the tag at all times.

I was expecting the FX7400 to have a slightly higher number of reads than the FX7500 so I was totally shocked by those numbers.

I haven't really had time to think about these results. I did this test kind of quickly and now I have to rush the kids off to bed, but I wanted to share what I have so far. I'll probably do some more tests in the morning. I'll see if I have another FX7400 to test with to make sure the one I used isn't defective.

[JK Race Timing Systems]

Hi Brian
Try the same test again with just 1 antenna to see the max count possible with just 1 antenna switching between rx and tx.
It would be interesting to see absolute max read count minus the antenna multiplexing time just in case there is delay in the antenna switching from tag powering mode to receive mode
Regards
Jk

[Brian Agee]

Single antenna connected to Port 1:

FX7500

1. 817
2. 782
3. 831
4. 764
5. 810
6. 817
7. 847
8. 817
9. 828
10. 849

FX7400:

1. 216
2. 217
3. 215
4. 210
5. 214
6. 217
7. 214
8. 225
9. 216
10. 211

FX9500

1. 2120
2. 2143
3. 2171
4. 2149
5. 2171
6. 2193
7. 2148
8. 2166
9. 2179
10. 2149

Thinkify TR200:

1. 767
2. 768
3. 768
4. 767
5. 767

I plugged up two other Thinkify reader and it got less than 100 reads every 10 seconds (range was from 60 - 80). This is concerning because I have no idea why one would produce hundreds and two other produce less than 100.

If anyone else wants to try this themselves I put an installer out here: Download Link

If you installed the software the other day from the link I provided then you may have to manually uninstall the software and install it with the new link. I'm trying not to burn through version numbers.

With this update, when you open the RFID Testing Center, it will not show any reads in the listview. It will simply connect to the reader and after 10 seconds it will show you how many reads were passed to the computer.

This update also includes the following features in case you're curious:

1. At this past weekends race we had quiet a few males register as females or enter the wrong DOB. So I wanted to add a quick way to know if there are possible problems. You can now right click on the Athlete list when you open a race and select "Verify Age and Gender" and it will highlight possible problems. This works by highlighting anyone with an age less than 4 or more than 100, and it compares the most recent occurrence of the Gender for the last person in the system with that same first name. If the gender doesn't match the most recent occurrence then it will highlight that person as a possible mistake.
2. When switching someone from one race to another, the software will now wipe out their RunSignUp ID (assuming this person was pulled from or pushed to RSU). Otherwise it will cause an error if you push results up for a person and RSU thinks they are in a different race.

[JK Race Timing Systems]

Those results are very strange for all except the FX9500 which is exactly what would be expected to happen when you reduce the amount of antennae, you expect the read rate to increase as the power down time has been reduced by a divisor of 4( or 3 is it).  As for the other readers they seem to have remained unchanged which is very strange as they too should have exhibited the same increase in read rate as the 9500 when reduced  to 1 antenna.

wierd for sure

[Peak Performance Timing]

Some of the differences between the Fx7500 and the FX9500 can be explained by the max power output and max receive sensitivity differences.

FX7500

Max power output = 31.5 dBm
Max receive sensitivity = -82 dBm (monostatic)

FX9500

Max power output = 33 dBm
Max receive sensitivity = -84.5 dBm (monostatic)

The specs on the FX9600 are:
MPO = 33 dBm
MRS = -86 dBm

[JK Race Timing Systems]

That wouldn't explain why the read count remained the same between 4 antennae and 1 antenna.
It should go up the less antennae you have fitted (just like the 9500 did) on a monostatic setup. Something else going on there

[Peak Performance Timing]

Not necessarily true.....maybe the return signal was right on the edge of the FX7500's limit for the antenna used....so it couldn't return a signal for a read every time, but the FX9500 could because it has better sensitivity and had a better return rate.....and as you indicated the number of reads increased because there was no antenna multiplexing.

[JK Race Timing Systems]

Ok
I don't buy it but let's run with that.
Re run the test but turn the antenna right way up and throw the tag on it.
I suspect you will see same results but let's rule it in or out

[Tim Irvine]

Shot in the dark idea here: Is it possible that on the 7400 and the 7500, the ports are still cycling through no matter if antennas are connected or not, where as the 9500 is not? I know the software shows them off, but what if they still power on and off and change regardless of a connection? Thoughts?

[JK Race Timing Systems]

Hi
That is exactly what I would be thinking. Easily tested with a scope or spectrum analyser.

[Hronometar]

Brian,

Thanks much for those tests and for providing a beta update that we can make our tests in RFID Testing Center. I was hoping I will do my tests last week, but unfortunately I was knocked down to the bed by a virus for almost a week.

Here are my tests I made today and on last Sat during a 10 k trail race.

Today I was using single side antenna placed on couch facing down, connected to reader(s) with a 5 m LMR-195-ULTRAFLEX cable. A HuTag was placed below each antenna when it was used. As it is shown in a table below, I did tests with linear and circular antennas. Here are the results:

#	R420		FX9500		FX9600
	Linear	Circular	Linear	Circular	Linear	Circular
1	230	231	2406	2480	3761	3730
2	232	227	2466	2454	3679	3811
3	233	226	2428	2449	3650	3828
4	228	232	2473	2501	3915	3818
5	234	228	2529	2520	3701	3847
6	232	233	2480	2451	3710	3718
7	234	233	2459	2505	3709	3866
8	235	232	2000	2527	3677	3753
9	234	232	2596	2617	3770	3706
10	235	232	2460	2503	3746	3840
Average	232.7	230.6	2429.7	2500.7	3731.8	3791.7

On last Saturday, I did a test like I was earlier explained FX9600 problem in the first email of this thread. There were around 300 participants and I was using 2 lines of RR mats with single DogBone MR6 wet inlay. For timing start line, I used 2 x FX9500-8 readers and had only one missed tag. For timing finish line, I used a FX9600-8 and a FX9500-8 readers. Exactly the same setup I use for most of my races and the same setup I used with testing the FX9600 last Fall. The FX9600 was connected to the first RR mat in direction how participants were finishing. It missed around 20+ tags on the first line, and on the second line with the FX9500 there were no missed tags. I have counted 16, but there were some situations people finishing in groups of 3-5 and I wasn't sure if their tags were read on the first line. Distance between 2 lines were around 5 m. Here is the setup.

After the tests I did today, I really do not know what to think about FX9600 reader. I looks like it has better performance than FX9500, but when I use it on races, it is not even close to a FX9500 reader. I was thinking a firmware update could help as my FX9600 is made in Jan 2018 and it probably uses a firmware from the beginning that it was produced in end of 2017. Here is the firmware version that my FX9600-8 reader uses: 2.4.30.

Brian, please, could you contact your Zebra dealer and ask for a latest firmware for the FX9600 readers? Also, it would be perfect if you could talk with them to provide a new firmware that allow you set a switching rate below 100 ms for FX7500/FX9600 readers. Also, can you borrow the FX9600 reader that your head timer's son uses and make same tests you did with other readers?

Thank you!

[blueridgetiming]

Hronometar,

I feel the same way with the 9600. Last year at an XC meet I used a 9500 with RR mats as the main and then a 9600 with RR as a backup 10' past the line to test it. Single dogbone on the back of the bibs. Main missed 2 out of 1200, where as the 9600 missed over 100. 

I know there has been much discussion on how the 9600 uses many of the same components as a 7500, but I would like to see the 9500 vs the 9600. 

Doug

[Peak Performance Timing]

Hronometar,

Why don't you try putting the FX 9500 for the first line at the finish line and put the 9600 second and see if you have the same misses on each line... Maybe it has something to do with people covering their bib with their hand trying to get their time at the first line?

[Peak Performance Timing]

Doug,

[Peak Performance Timing]

Doug,

I think part of your problem may have been that the mats were too close together?

[blueridgetiming]

Peak,

I’ve never had a problem with mats being too close and they certainly weren’t even remotely close at that event (see attached)and that would nullify the stopping watch conjecture

[Peak Performance Timing]

Blueridge,

I thought you meant they were 10' apart......was just tossing out ideas.

[Tim Irvine]

Just a shot in the dark here. Are you hooking the antennas up in the same sequence on both readers? Thinking you might have a signal hole if the antennas aren't staggered correctly.

[Rivalry Events]

I have a 9600 and 9500 and haven't been able to tell the difference, even if there is a noticeable difference to you guys. Occasionally I'll get a few missed reads on either one of the readers, there doesn't seem to be a rhyme or reason.

However, I have 2 RR mats, AND I don't use every single antenna in either of them since there's 8. In fact I've taken some of the antennas out along with the associated cables. Currently in one of my RR mats I have 1, 3, 5 and 7 taken out and on the other RR mat I have I've taken out antennas 2,4,6, and 8. I think that's what I did anyway. Then i used the available channels to run a couple 30 ft cables to the other side for side panel antennas on tripods. And I use one or two side panel antennas on tripods for the near side as well. The benefit of staggering some circularly polarized side panels a little after the RR mat antennas, while being on one 8 port reader is to cover more area of course, have proper firing sequence, and keep reader costs down, and provide more read angles. 

[Hronometar]

Peak, thanks but it is not case that a hand that stopping wrist watch cause tag reads. Not for a FX9500 reader in my experiences. Please, read my first email in this thread to see and better understand comparison.

Tim, yep, they are exactly on the same sequence. I usually connect my RR mat connectors to match the readers connectors (1-1, 2-2, 3-3, etc.). I was testing different combinations and there is no any difference with reads. And the sequence I wrote give me possibility to know in any time which antenna is connected to which port without looking to cables and reader.

Rivalry, you do not need to place cables through 'free' channels and always open and close red cover. On bottom black surface of the met there are 2 channels that you can use for 240 series cables, extension cord and network cables. If you are curious about difference in tag reads of those two readers, read my first email in this conversation and make same tests. I know that side antenna can help read rate in combination with the mats, but the point was of using only the mats.

Brian, have you had some time to test your FX9600 reader that your head timers' son uses? Do you have any info from SenThys about differences between FX9500 and FX9600 readers?

Can anyone with FX9600 do the same tests I did with beta update that counts total tag reads in 'RFID Testing Center' for 10 seconds. Candidates can take a look in my email I wrote on Feb 25, 2019. I have saved beta update and it can be downloaded here.

[Rivalry Events]

Yea those under belly cord channels of those RR mats are nice! 

I've used the interior channels because I had already taken the red panel covers off to take out a few of the antennae and use them as split points and such. So the channels were already accessible. And I kinda like the semi-permanency of the long cables I put in for attaching circularly polarized panel antennas on tripods. The bottom channels are more for temporary usage. 

The goal of the RR mat for me is to use its power and RFID orientation to cut upwards through dense amounts of people, like chip starting a large race. And also for the convenience of folding it out fast as setting up overhead panel antennas is cumbersome. BUT it's linear polarization, and without a wide cone shaped RFID field, so ultimately I think when used in conjunction with circularly polarized antennas it's sort of a "cover your bases" mentality where you get more angles and area coverage, so why not?

But going back to the Zebra readers, maybe the circular antennas is why i don't notice a difference between the 9500 and 9600. It's hard for a person to cross the entire RFID field of a circularly polarized antenna in a fraction of a second. I think Brian has YouTube videos of people crossing a single circularly polarized antenna on a motorcycle going 40+ mph. I wonder if an RR mat speed test could replicate that consistency, although that sounds a bit dangerous.

[Race Pros Sports Timing]

Hronometar - I see you have an Impinj R420 in your tests, and I was surprised it performed so poorly. I own both an Impinj R420 and a Zebra FX7500.

Why do you think the Impinj perform so poorly with the ART tests? Other sources like the Atlas RFID store and Webscorer software say the Impinj is the fastest 4 port reader available for race timing. It works great with Webscorer. FYI, I use both Webscorer and ART RFID versions. I like Webscorer’s live results, but I like ART for all other aspects of the race (Annoucer station, encoding bibs, connecting to Race Roster, etc).

Does ART use a different way of connecting with the reader or something that leads to poor performance with the Impinj R420? Webscorer uses LLRP.

[JK Race Timing Systems]

Hi
I use an impinj r420 on occasion with art and the results are perfect.
In the testing mode it is definitely faster than the 9500( it has more reads per second)
Hope this helps
Jk

[Hronometar]

I have discovered it in Summer 2016 before I become ART user that I do not have such a max performances using a R420 with ART. I still have a timing box Ultra8 from RFID Race Timing Systems from Australia that I was using before ART. It has 2 Impinj R420 inside the box. Read distance and number of collected reads were much different when I was using the box and directly connecting to the readers with ART. After some period of using ART, I got a new R420 and I have noticed that all 4 port lights blinks when 1-3 antennas are connected to the reader (could not see that in other 2 readers as they are hidden in the box). I have noticed Brian about that but he told me that the software for sure tries to connect only to ports that have an antenna connected. Also, I have tested it with some Impinj's free software with using only the port that has an antenna connected, and other lights were not blinking. Biggest differences a felt when I got my first FX9500 reader. From that period missed reads are pretty rare, but same setup. I was also testing 2 R420 connected to RR mats, and also there were situations I had missed with that setup in comparing with FX9500 reader.

However, I didn't have bad read rate in that period, but there were always missed reads with using only R420.

I suggest you to make same test and to check it by your own. Impinj software I was using is MultiReader. A software that I was using to connect to the Ultra8 unit says that 'Comms protocol' is LLRP. I can see that Brian still has the same Beta version in this folder that you can test total number of reads for each reader in period of 10 seconds that starts with the first caught read in RFID Testing Center.

Here is some of Brian code for the Impinj reader that he has published.

About the test I made a few months ago, it looks that ART tries to connect to other 3 free ports that do not have antennas connected and for those 10 seconds the system doesn't take max number of reads it could. For summary, R420 doesn't perform such a strong in comparing with FX9500 reader when it is used with ART - my own opinion covered and proof with many tests and race experiences.

John, please, could you make the same test I did when you get some free moment? I would really like to see if I did something wrong or my readers have some issues. But chance that 3 R420 readers have a problem is pretty low.

[Race Pros Sports Timing]

Thanks for the replies JK and Hronometar. I should do my own tests using Brian's beta software from the link provided. It's good to know that ART will work well with both the 4-port Impinj R420 and 4-port Zebra FX7500. I'd like to get a third reader for 24-32' start/finish areas with panel antennae, or to use with Race Result mats. So the question is buy a second R420 to run alongside my first one, or to go with a Motorola 9600. I use the FX7500 is for announcer stations and backups. I have some time to figure his out, since I probably won't buy the mats and another reader until the fall. 

[JK Race Timing Systems]

Hi Yes I can do

It has been a while since I have used it and I couldnt be sure how many antennae I had connected when testing but I definitely spotted the lights being activated, much like your observation

All I remember is the read rate was extremely high compared to the 9500

Job for the weekend I think

Best regards

JK

[Tim Irvine]

I had an Impinj R420 when I first started timing. When I started using ART, it worked fine, then I added an FX9500 to the system. When I did that (years ago now) I could not get the Impinj to connect when the FX9500 was in use also, so I ended up selling the Impinj and got another FX9500 and they both work great together. I am sure whatever issues I was having then have been addressed by now, but to speak on the performance of the Impinj reader, I found it to be pretty good. Antenna placement with that reader like with any reader makes a difference. Tim

[Brian Agee]

Today I sent a very detailed email to Zebra with a lot of the evidence I could find that makes it clear that the FX9600 is not performing quiet as well as the FX9500. So thank you to everyone that took the time to document what they found and post it here or sent it to me by email. One of the most interesting video that show what is happening is this one: https://youtu.be/TroUwxReQeU

As I (and others in this thread) have suspected, the cycle (antenna switching) rate still appears to be too slow. In an earlier post on this thread I shared another YouTube video that shows what the default cycle rate for the FX7500 is FAR too slow. I found that this was true for the FX9600 too (no surprise there since they use the same boards). Zebra sent me a code snippet that I plugged into my software that improved it dramatically, however even at the fastest allowed switching rate, the video above appears to show that it's still not fast enough.

Hopefully they'll jump on the email and send a response back quickly with instructions on how to fix it.

[Oscar's Race Results]

Tim,

  Do you run in bistatic mode?   My understanding is that the main value of 8 ports is you have 4 channels with 100% up time vs 50% due multiplexing.

Thanks,

Oscar 

[Jason Swan]

Is there anything I can do on my end to optimize my fx9600 reader?  I will be doing a large meet near the end of the month and my experience thus far has been spotty and right in line with some of the experiences related here.  I am waiting for macsha mats to be delivered and in the interim have done several smaller xc meets with the 8pt fx9600 and two Alien ALR9611 antennas, spaced several feet apart on the same side of the line.  In these meets I had many instances of a single runner striding through the finish line and not being picked up at all while other similar instances or even bunches of runners were picked up just fine.

As I will have only one mat for the finish line, my plan is to double chip and to add 2 panel antennas to the setup.  So I will have the 4m mat,and then one antenna on each side several feet behind the mat as my second chance to pickup runners. 

Obviously any information on what I can do to optimize the fx9600 settings will be appreciated.  I also have two fx7400, but they are two port readers and intended for split points or check stations.

[Tim Irvine]

Oscar I have never messed with changing Monostatic or Bistatic. So whatever the default is, that is how I run it. If I get a chance to look I will.

[JK Race Timing Systems]

Hi Jason

What type of system have you exactly. Can you draw out your system with measurements.

regrards

JK

[Brian Agee]

This is how I spent my Sunday: https://youtu.be/SLEhI0DtkNs

Summary of video:

1. (History)

Back before the FX7500 was officially launched I received a beta unit and fairly quickly discovered that unlike the FX7400 and FX9500 the FX7500 does not come from the factory setup and ready for maximum performance. In fact, the switching rate was so slow (roughly around a half second) that it was basically unusable for race timing with the factory default settings. I informed Zebra of the problem and sent a comparison video (linked above in an earlier post in this thread) and I explained to them that we need it to at least match the switching rate of the FX7400 or FX9500. They sent me back some extra code I needed to plug in after connecting to the reader. That extra code dramatically improved the performance and everything appeared to be working great, but I did not have a RFID meter at that time. My crews and I (and of course many of you) have used the FX7500 and FX9600 for over a year now and since most races are < 500 read rates have been comparable with the FX7400 and FX9500. However for larger events (or XC events) where the finish line is typically busier or where you are capturing the chip start times, I've had a growing number of users tell me that they noticed a slightly higher number of misses with the FX9600.

2. (Performance Testing)

The video shows me testing all FX series models to see what the switching rate is by using a UHF meter.

3. (Results)

The FX7500 and FX9600 had a noticably slower switching rate.

What I thought was the "minimum" switching rate actually was not.

4. (Resolution)

I will try to have an update out before the day is over with the setting that lets you try the new settings applied for the FX9500 and FX7500.

[Peak Performance Timing]

Brian You're AWESOME!

[Motor City Timing (Nick)]

Can't wait to get the update and run some tests! Thanks, Brian!

[Brian Agee]

I ended up not getting to prepare and push the update out last night. I also decided just to push out the fix for the antenna switching rate without putting a toggle in where you can revert back to the old switching rate. So the 1.5.3.11 update by default will increase the switching rate by 80%.

[Bruce B]

What effect does the update have on the 9500? As I am perfectly happy with my current read rates with that model.

[Brian Brian.S.Agee]

None. The software only adds those extra commands if the model name contains "FX7500" or "FX9600". For all other FX series readers it's business as usual.

[Joe Fertsch]

Thank you Brian  - That's Great, So as long as it is connected as a FX7500 it will speed up those  rates,  Is it important that it connects under those readers names FX7500 or FX9600 names.  I have so many readers on the list it generally connects to the first one that is enabled.

[Peak Performance Timing]

I have a question about this whole FX7500/FX9600 sparked by Joe's comment. I usually just put the IP address of my reader in ART settings. So will does ART automatically know if the reader is and FX7500/FX9500 when it connects? Brian?

[Brian Brian.S.Agee]

No sir, when it connects I then grab the official hostname directly from the reader. For example, imagine if you only used the IP address to connect. So basically those extra settings will apply to any FX7500 or FX9600 no matter what you enter into ART.

[Brian Brian.S.Agee]

Oops, peak responded right before I did. That "no sir" was for Joe's post.

[blueridgetiming]

Here are some of my findings with the updated FX9600 code.

• Version 1.5.3.11
• Main system -- FX9500 with Race Result mats
• Backup system located 10' past the main -- FX9600 with Race Result mats
• 1350 finishers
• Main system read 1349
• Backup read 1334 (5 were within the top 250...)

While it's much improved, I still feel as if the 9600 is lacking in comparison to the 9500. 

[Peak Performance Timing]

Blueridge,

I think you backup mat is too close to the main mat.Just my suspicion.

[Brian Agee]

I suspect that Blueridge is right. When I was playing with the switching rate of the antennas with the FX9600 I could get it close to the FX9500, but not quiet as fast. In the beginning the difference between the switching rate was obvious when measured with a meter, and after the adjusting the switching rate the difference is barely noticeable, but still slightly slower.

I can set the switching rate to be faster, but the problem is that the FX9600 can't seem to handle it. When I drop the switching rate down to where it appears to match the rate of the FX9500, the FX9600 doesn't read hardly any tags. It's like it's shooting out RFID waves but doesn't take the time to listen to what tags are reporting back before it switches to the next antenna. Somehow the FX9500 can switch faster and still process tags in view of the antenna.

Look at the 11:42 mark of the video I posted and you'll see the activity indicator barely flashing at all even though a tag is directly in front of the antenna. Right after recording that video I slowed the switching rate down one millisecond at a time until I found what I thought was the right balance of adequate read time and fast switching rate.

Of course I played with all of the other settings I could find to see if they would make a difference, but nothing seemed to help. I'll keep digging to see what I can do, but I wanted to share what I learned while messing with the reader.

[blueridgetiming]

I wonder if hooking up less antennas would help out read rates. That'll be part of my next testing phase.

I must say read rates were drastically improved, but still not still quite 100% trustworthy. Does the power output have anything to do with read rates? I know if you hook up the 9600 via POE it doesn't produce as much power as if it is hooked up to a power source.

[Windsor Running]

As a newbie, I find a lot of this fascinating. I have used the 9600 in a race (two missed reads out of 199 runners) and just bought a 9500 on eBay that I haven't had a chance to test yet so my only points of comparison are an FX7500 which I use as a backup and an Impinj R1000 (which is still for sale by the way for $150! Clearly I priced it too high initially!). 

My 9600 was used prior to the update and was powered over POE with 4 mat antennas (Feibot) and one side antenna (5 total ports) so I'll be interested to see how things change with my next race. 

A couple of questions:

1. I think Race|Result uses a 9600 in their timing box and they seem to get exceptional reads. Is that due to a software difference between theirs and Brian's? It's interesting that with the same reader and the RR results mat setup that BlueRidge is using that he's still missing runners. Again, my background is on the running side, not on the hardware/software side so there's no judgement in this question. 

2. Brian, has Zebra gotten back to you with any answers? I'm curious to hear what they say. 

Jesse

[Steve Hudson]

Hi Brian, following on from Windsor Running...any more news, please ? I've just needed another reader and can only get a FX9600 and I'm concerned about missing reads - all be it minimal.

Can the POE really function on an 8-port race result mat ? if the Cat 5/6 cable runs directly between the reader and Timing PC , will the PC battery take a hit ?

[Ohio Race Day]

Steve Hudson,

POE is not powered by the PC battery.  It has its own power supply.  So the PC battery shouldn't "take a hit."

Has anybody tried POE+ or POE++ for the FX9600?  Here is an article about the difference in the three choices that are now available.

https://community.fs.com/blog/poe-switch-types.html

Just curious,

Mike

Ohio Race Day LLC

[Patrick W]

Mike it's a Poe device so it doesn't matter which poe type of switch you use, it ids and gets the same power it would on any of them. That said the switch having higher power support will run more devices without browning out which is nice.

[Rainier Timing]

Steve, 

"can the PoE really function on an 8-port race result mat?" 

Regardless of how your reader gets its power (AC or PoE), the number of antennas is not a concern because switching rate between ports is the same whether you have 2 or 8 antennas. The power output of the reader is the same as well. Antennas don't "draw" power in the typical sense - they just take INPUT power from the reader and convert it to radio waves, and receive radio waves and send back to reader. When people do talk about power loss it usually refers to "loss of dBi / gain" due to cable size, length, and insertion loss (connectors and adapters), but that has nothing to do with what your reader is outputting from each port.

[Steve Hudson]

Thanks et al for the feedback on POE. Sounds very technical to me. As long as it works, I'll soon find out when the FX9600 arrives. Just hope I don't get too many non-reads. Thanks again.

[Steve Hudson]

Am I correct in thinking that POE isn't just as simple as connecting the Cat 6 cable into the FX9600 POE port and Timing PC ? And that a "POE injector" is required ? which itself requires to be powered up ?

I've googled POE injectors and sellers just show a photo of the unit itself. Do they get sold with a power lead ?

[Patrick W]

Requires an injector or a Poe switch. Get a Poe switch, you're going to want to network things anyway and for a basic setup it is two devices in one (switch+injector) and will handle more than a single Poe device (my track finishline is up to 6 Poe devices with wireless and cameras and whatever).

Pay a little more and get a Poe+/++ one. Standard Poe switch (~15w) will do 2 readers happily, but will brown out at 3, but only when you start reading (ask me how I know lol). Having more power budget makes it less likely you will accidentally brown out.

[Steve Hudson]

Ordered something - will see how it goes. Many thanks

[Steve Hudson]

NETGEAR POE switch arrived today and works a treat with the FX9600. FX9500 naturally worked fine. Thanks Patrick et al

[-Kurt]

I've never owned a 9500 so I don't have a reference, but besides the occasional unexplained missed read (maybe 1 per 100 at most on finishes and a bit more on super-dense chip starts), I feel that I get pretty good performance from my 9600's, but I'm always trying to improve my setups.

My question: I set the power to each port via the control console with my 9600 readers to compensate for losses due to cable length and diameter.  I remember Brian mentioning somewhere that ART overrides the power settings coming from the readers. Does anyone have any solid evidence or information on the usefulness of adjusting the power via the console or does ART still override everything?

[Oscar's Race Results]

Reader	Power Source	Antenna 1	Antenna 2	Idle-Read Watts	Reads per Second
FX7400 - 2	POE	242025	None	6-12
FX7400 - 2	POE	242025	262006	6-12
FX7400 - 2	Brick	242025	None	5-13	120
FX7400 - 2	Brick	242025	262006	6-13	120
FX7500 - 2	Brick	242025	None	8-17	580
FX7500 – 2	Brick	242025	262006	6-17	440
FX9500-4	Brick	242025	None	11-20	275
FX9500-4	Brick	242025	262006	11-20	230

Been playing with alternative reader power sources using Session One program.  Pointed one or two antennas at a jar of chips and recorder power consumption and read rates after both stabilized.  My 9500 is firmware locked.  $300 used, what did I  expect?  

1. No surprise that 1 antenna with no mux is faster than multiple antennas.

2. If 7400 and 9500 were first generation and 7500 and 9600 second generation, is it possible that the 7500 is inherently fastest of any of them?  OR if I had an "untampered-with" 9500, would it be a lot faster than the 7500?

Oscar

[Oscar's Race Results]

Should have put questions at top.

Been playing with alternative reader power sources. Pointed one or two antennas at a jar of chips and recorded power consumption and read rates.   My 9500 is firmware locked.  $300 used, what did I  expect?  

1. No surprise that 1 antenna with no mux is faster than multiple antennas.

2. If 7400 and 9500 were first generation and 7500 and 9600 second generation, is it possible that the 7500 is inherently fastest of any of them?  OR if I had an "untampered-with" 9500, would it be a lot faster than the 7500?

Oscar

Reader

[Hronometar]

It has passed a year since last improvement of FX9600 reader. I was looking on Zebra support page and found there are a few firmware updates from 2020. Brian, could you please give it one more try for FX9600 to improve it more to be as much closer as possible to FX9500 reader? Maybe some of those updates from 2020 were about reducing switching rate.

https://www.zebra.com/us/en/support-downloads/software/operating-system/fx9600-series-operating-system.html

[Matthew Thomas]

In the process of getting grants and building our club's system, should I try to find an FX9500 rather than the FX9600? Seems that way reading this thread.

Thanks in advance.

[Brian Agee]

Months ago when I reached out to Zebra to get their feedback on why the FX9600 didn't appear to perform as well (at least in regards to how quickly/smoothly it switches between antennas), I provided them with a few videos that show evidence of a slower switching rate and the feedback that I've been getting from users that indicates a very slight increase if missed chips when tons of people are flowing through a starting line. They asked me what firmware my reader was on, which turned out to be a very old firmware version. The FX9600 that I had was a pre-release model I received when they first created the FX9600. So they recommended that I update the firmware and give it another test. The problem is that this reader simply will not accept the newest firmware no matter what I try. So a few weeks ago I ordered a brand new FX9600 for myself and ensured that it's running the newest firmware available.

The new FX9600 performs a LOT better than my FX9600 and seems to be stronger than my FX9500's in many aspects, however I'm still not 100% convinced that it's better than the FX9500 in picking up tags in a dense chip start situation because the switching rate of the antennas still seems to be slightly slower.

I wanted to post a video of how I tested it and the results I received. This way anyone else can setup a similar test and compare the results of their reader against a brand new reader running the latest firmware update. If your results are not as good as what you see in the video below then please write down the firmware version that your reader is currently on, try updating to the newest firmware, and if the firmware update improved the results then please post the what firmware version that the reader was on so that we can know which version are good and which could benefit from an update.

FX9600 vs FX9500 Test: https://youtu.be/nCBI6Q_Af2A

[onjw...@hotmail.com]

Brian,

  If you placed one chip on each of the 4 antennas of a Feibot mat, read for a few seconds and compared  "Last Read"s in microseconds, could you then calculate switching rate?

Oscar

[Steve Hudson]

FX9600 8 port , firmware 3.0.35, with 4 antennas connected and 1 tag on an antenna , I got average readings of 2400. The 3 other antennas were face down on the floor as per your video. BUT if I faced the other 3 antennas into free airspace I got average of 3150.

[Jonathan Newton]

Hi, was there ever a solution to this? I have an FX9600-8 and it seems to be giving me all this issues that I've been reading in this thread. I've been missing a LOT of reads with my mascha mat...

When I spoke with the seller he said to disable "auto-detect" on the reader, which I did, but still had the same results (missed reads).

Thanks in advance.