2
Edit: I have added a new card, the Intel Ultimate-N 5100 to the list and performed the same battery of tests. Retested the others at the same time to account for any effects due to the time-of-day.
Situation
Different ping latency averages across different card models, with the newer cards performing more poorly.
Cards Tested
The 6230 is the newest unit, while the 1030 is an entry-level card. The 5100 is also a higher-end unit, but is a couple years older than either of the others.
Network Setup
All tests are done on a 802.11g single-band network, so the dual-band features of the 6230 and 5100 are not being used, to even the playing field.
Testing Process
- Install the card.
- Factory-reset the laptop.
- Run 25 speed tests at several pre-measured points/distances from the wireless router.
- Eliminate outliers and calculate averages.
- Remove card from the laptop, insert into the next one, and repeat from step 2.
Laptops Used
A Dell Inspiron 1545 and two Dell XPS 15 L502X machines.
Question
Is there any explicable reason for the ping differences? I have tried moving both computers, and even swapping the network cards between them. Tried factory reset too.
Update
Added the new results for the 5100 card to the list. It has the best ping yet, with throughput comparable to the 6230. The 6230 still averages the highest upload speeds, while the 5100 consistently returns the lowest ping latency.
Wireless-N 1030 (2nd best download, 2nd best upload, 2nd best ping)
- Bluetooth support: Yes
- Download: 12.1 Mbps
- Upload: 0.6 Mbps
- Ping: 18ms
Advanced-N 6230 (tied for best download, best upload, worst ping)
- Bluetooth support: Yes
- Download: 12.6 Mbps
- Upload: 0.7 Mbps
- Ping: 21ms
Ultimate-N 5100 (tied for best download, 2nd best upload, best ping)
- Bluetooth support: No
- Download: 12.6 Mbps
- Upload: 0.6 Mbps
- Ping: 15ms
Summary
Since Bluetooth capability is not as important to me at the moment, I am using the Ultimate-N 5100, as the ping time is more important to me than throughput. However, I would still like to figure out the reason for the consistent differences.
Do you shutdown other wifi devices during your ping tests to ensure there isn't contention? – Paul – 2012-01-19T04:27:20.530
The actual thruput numbers and ping times would be useful for comparison. – Garrett – 2012-01-19T05:48:32.137
It seems to me that you are measuring those values using an on-line speed test. Unfortunately it is not the proper way to do it. Since you have two machines, you can connect one of them to your access point using Ethernet and perform a bandwidth test between the two machines without using your internet connection, thus measuring the true throughput of the wireless link. For this, you will need Iperf for which you can get a windows binary here installed on both machines. You can also Jperf which is a GUI written in java.
– phil – 2012-01-19T20:42:57.167