Intel vs Marvell WIFI Performance Difference.

Question

One of the problems that I have observed with my Surface Pro 7 has to do with the performance of the Intel modem compared to the Marvell modems used in older products such as my Surface Laptop 1.

Specifically...I recently noticed that the WIFI signal strength meter on my SP7 would frequently lose one bar when connected to the 5GHz band from the repeater located on the corner of my desk.

The signal strength meter on my Surface Laptop 1...which sits right next to the SP7...consistently has all bars illuminated.

Finally I decided to do some research and used the WIFI checker app installed on both devices.

Lo and behold...the app showed a glaring difference between the two units. The Surface Laptop 1 registered a -24 dbm. The Surface Pro 7 registered a -30 dbm. The SP7 reading wavered so to the point where I could see that the signal strength meter dropped one bar when the signal strength fell to -34 dbm.

So I ask....what is the reason for such a discrepancy between the two devices?

If the SP7 sees a weakened signal when the WIFI spot is only a few feet away...imagine what the performance is like when the device is 20 or 30 feet away.

Out of curiosity,  I repeated the test with both devices utilizing the 2.4GHz frequency. In that test, the SP7 registered a stronger reading than the SL1.

I have cropped pictures of the screens showing the test results from the 5GHz test. I can post them if needed.

Before being asked...toggling WIFI off & on has been tried. It really doesn't make a difference.

Is this looking like a hardware issue? A driver issue? 🤔

Answer 1

Is this looking like a hardware issue? A driver issue? ?

what brand and model router/extender are you using?
do you have the same issue if you use the devices near the router itself and
unplug the extender/repeater as a test?

Answer 2

Hi Barb!

The router is supplied by my ISP...Verizon. It is a model G3100. The extender is also from Verizon....an E3200. Now prior to the E3200, I had a Linksys WRT1900AC on my desk. The SP7 experienced the same symptoms then as well, so, I can't say that the extender is the problem.

Here are my two Surface devices. The SP7 is on the left and my SL1 is on the right. The extender is just off camera to the right.

Here is a picture of what the SP7 sees for a signal from the extender...

Now here is what the SL1 sees from the extender...

9db difference connecting to the same device located two feet away...

The Verizon router is in the basement of the house. The SP7 and SL1 are here in my bedroom three floors up. I have not taken the computers down to the basement to connect to the 5GHz signal from the router itself.

Answer 3

Now this is interesting.....

I powered down the extension and took the SP7 to the basement. The SP7 connected to the 5GHz frequency. The signal strength meter was full. (the computer was placed directly underneath the router...which is wall mounted).

The WIFI checker showed the following....

The router in the basement was utilizing a different channel. The SP7 was "happy" with it.

I brought the SP7 back up to the bedroom and powered up the extender. This is what I see....

The extender has switched to a different channel. The SP7 seems happy with this one as well. The reading is shifting around, but, has not fallen below -31 dBm. The signal strength meter is fully illuminated.

So the question now becomes...Does the Intel radio have weaknesses with some of the channels associated with the 5 GHz frequency? Remember that the SL1 had saw a robust signal on channel 44 while the SP7 was indicating a more marginal signal with that same channel.

Answer 4

The whole channel thing is a slippery slope. I don't think it is necessarily

chip or driver dependent as there are other conditions that can influence

connectivity. Sometimes in a router/extender environment there can be a "race"

condition where the connected computer is actually bouncing between the router

and the extender. And if you have neighbors with WiFi in range using the same or

overlapping channels, add that to the mix.

//    In my own environment, I find using the higher number 5GHz channels gives the

best results   //