Radio 3.35 Speed Poll

[Geckotek]

What are your experiences with the new Radio?

For consistency, please use the following test method.

While tethering your phone to your PC, use the speedtest located here: www.speedtest.net

Please post a graphic of your results as Maccaberry has below.

Please only post if you are in a Rev A area.

[supernatedogg18]

Just to make it clear, you cannot test the speed of the broadband connection accurately using the phone and Pocket IE. It will not return the correct results!

The reason for this is that the processor needs to perform double-duty in order to load a page in Pocket IE; this means it processes the incoming data and also has to render the page. While the page is rendering, the data has to wait, and vice-versa.

The correct way to accurately test the speed of the data connection is by tethering the phone to your computer and then perform a speed test, through someone like speedtest.net or something like that. This is because the phone only collects the incoming data and then ports it to the USB for your computer to render in the browser.

My results confirm this:

Download Speed using the new ROM:
Pocket IE: ~950Kbps
Tethered: 1550Kbps

The difference is due to the processor not being able to stream all of the data at the maximum speed AND render the page all at the same time.

Hope this helps!

[maccaberry]

Geckotek,

I agree that the only way to accurately test is via tethered pc. PIE has never given anything like a real number for me.

www.speedtest.net seems to be the best and the one that most posters use as their reference.

Now if you could amend your instructions to use this method we would be onto something.

EDIT: with old radio and rom I get 1400 k DOWN and 120K UP as a point of reference. EV-DO Rev0

[madman34]

sitting at the console, from 3.27 to 3.55.04 RNC says session is REV-A for both, BER is better for new resulting in more reliable session, speed is a little better due to the BER (a timer was tweaked in the leaked radio) S/N ratio a tad better with older ROM but if you are in a good signal area then youe EVDO stays up better because the noise ref figure for fallback is set higher, and EVDO speed drops off faster with lower signal, new radio is wayyyyyyyyy better trust me as I see it's impact on the system.

[hoothrewpoo]

I get about 440Kbps d/l and 104Kbps u/l tehered to my laptop. Seems strange that I can't get better numbers than that with 5 bars sitting in the center of Houston, TX. Certainly we have Rev. A here...

[Geckotek]

Very good point about the processor guys. I'll make the changes.

[drgreenberg]

I like www.speakeasy.net/speedtest. For my posted results (which are > 1Mbps download, >300kbps upload), I used WMWiFiRouter to tether to my laptop and ran the test out of Firefox 3.0b4 running on said laptop. I believe this should be valid.

[whatsit]

Before I changed the ##778# setting for Rev. A to enabled:
http://www.speedtest.net/result/245487845.png (http://www.speedtest.net)

After I changed Rev. A to enabled:
http://www.speedtest.net/result/245491918.png (http://www.speedtest.net)

Do you see something wrong with this picture? I don't think that setting does what some think it does.

[Geckotek]

Before I changed the ##778# setting for Rev. A to enabled:
http://www.speedtest.net/result/245487845.png (http://www.speedtest.net)

After I changed Rev. A to enabled:
http://www.speedtest.net/result/245491918.png (http://www.speedtest.net)

Do you see something wrong with this picture? I don't think that setting does what some think it does.

Yeah, that's been discussed. The entry is to "Disable" Rev A. So "Enabling" the entry would in theory "Disable" Rev A. So the general concensus is to leave it alone.

Dealing with MS registry entries you find this type of logic a lot.

[whatsit]

I'm sure most people know about this, but I'll put it here for posterity's sake.

EVDO Maps (http://evdomaps.com/)

[j-squared]

OK - slightly off topic - sorry - but when I go to speedtest.net it syas I need at least flash player 7.0 or higher. I installed the flash player bundle and thought that would do it. no dice - anyone have an idea what I'm missing?

[Geckotek]

Flashplayer on your laptop.

[maccaberry]

OK - slightly off topic - sorry - but when I go to speedtest.net it syas I need at least flash player 7.0 or higher. I installed the flash player bundle and thought that would do it. no dice - anyone have an idea what I'm missing?

You are not off topic. In my reference to www.speedtest.net I specifically referred to tethering to the laptop because of the issues with running speed tests on the ppc itself.

Yet another reason that they should change the name from PIE to POS. ;)

[simoncinis]

While using wmwifirouter

used two different sites

Last Result:
Download Speed: 2235 kbps (279.4 KB/sec transfer rate)
Upload Speed: 618 kbps (77.3 KB/sec transfer rate)


http://www.speedtest.net/result/245545895.png

[illuminarok]

http://www.speedtest.net/result/245568566.png (http://www.speedtest.net)

[AmazingLarry]

There is no point in doing this at all. Would you do a speed test on a WiFi access point when you have no idea what other people using the same access point are doing? That's exactly what you're doing with EVDO. You could be sharing bandwidth with 50 other people.

[s10onn2o]

VZW here in houston area

http://www.speedtest.net/result/245681974.png (http://www.speedtest.net)

[dalvarez7]

Here is my test I did this morning...

http://www.speedtest.net/result/245809144.png (http://www.speedtest.net)


Speakeasy.net/speedtest

2324kbps Down
316kbps Up

[illuminarok]

http://www.speedtest.net/result/245904651.png (http://www.speedtest.net)

[hoothrewpoo]

Sprint PCS Houston

Rev. A enabled ("Disable" shown in epst)
http://www.speedtest.net/result/245940335.png (http://www.speedtest.net)

Rev. A disabled ("Enable" shown in epst)
http://www.speedtest.net/result/245937645.png (http://www.speedtest.net)

I have a minimum of 4 bars where I am, jumping to 6 at periods. Any clue why I can't break the 500kbps barrier?

[hoothrewpoo]

test done with my T1 on my laptop.
http://www.speedtest.net/result/245942103.png (http://www.speedtest.net)

[rafaelc]

I was thinking on go back to radio 3.27, but with the info published in this post I made the correct test and my speed is:

http://www.speedtest.net/result/245944160.png (http://www.speedtest.net)

[slvrap1]

heres mine via wifirouter
http://www.speedtest.net/result/245582649.png (http://www.speedtest.net)

[dalvarez7]

Sprint PCS Houston

Rev. A enabled ("Disable" shown in epst)
http://www.speedtest.net/result/245940335.png (http://www.speedtest.net)

Rev. A disabled ("Enable" shown in epst)
http://www.speedtest.net/result/245937645.png (http://www.speedtest.net)

I have a minimum of 4 bars where I am, jumping to 6 at periods. Any clue why I can't break the 500kbps barrier?

You might try testing again. Remember, the speed depends on how much bandwith is available. If you are tesing during a peak time speeds may be down.

[coffey_joshua]

http://www.speedtest.net/result/246096233.png (http://www.speedtest.net)


i think i should test again....

[texdoctor]

during normal business hours like 9 to 6, I get 1500-2500 up. 450 down

after 8pm until about 11 pm i am lucky to get 140 up and 100 down.

Maybe they turn the darn thing off lol Oh in the south texas area

[pidsw]

These numbers are all over the place. Unfortunately I did not record any speed test results on my prior radio but with 3.35 on Telus my last result was:
Download: 1572
Upload 249
Ping 162

With Rev A I would have expected a closer ratio (up vs down). All in all I have't noticed too much, but voip is a lot better andf the latency (ping) of everything online is might better.

[hoothrewpoo]

You might try testing again. Remember, the speed depends on how much bandwith is available. If you are tesing during a peak time speeds may be down.

I've been testing for 4 days brother. I've tested at tdifferent times of day/evening, and 660kbps d/l is the best I've managed (had 5 bars). It's strange to me that my phone is faster with Rev. A disabled ("enabled" shown in epst).

[mj41374]

I just ran the test with RevA and got this result (looks like most of yours).