• After 15+ years, we've made a big change: Android Forums is now Early Bird Club. Learn more here.

Root [ALL] FIX your 3G Speed NOW! [PLEASE READ]

I'm on the Sharp Froyo Rom and don't have that in my build.prop, but thought it would be interesting to do a "reverse" test.

So first I did a speed test 3 times without the lines (phone as is)

Test 1: 728 down, 334 up, 410 ping
Test 2: 696 down, 213 up, 126 ping
Test 3: 800 down, 188 up, 100 ping

Then I put the lines in build.prop and rebooted:

Test 1: 349 down, 535 up, 715 ping
Test 2: 276 down, 392 up, 510 ping
Test 3: 176 down, 430 up, 498 ping

Pulled the lines out, rebooted and speeds went more or less back to what the were at the first.

Not sure if any of the Froyo roms actually have it in there, but if they do, I would pull it out.
 
I'm guessing this isn't a fix for anything. It's most likely due to who's using the towers at that time and based on other completely random anomalies.
 
apexcell said:
I'm on CM9. Of those two lines of code, the only one I had was a partial of the second one:

ro.telephony.ril.v3=datacall,signalstrength

Commented out, rebooted, and network connection error, so couldn't test speeds. Rebooted again, and same issue.

Edited back and connected to network fine again.
Click to expand...

I am wondering if this is a needed config for CM9 and it was then moved into CM7? I didn't spend the time going through all of IssacJ's commits as there is so many. However it sounds like this may only work on CM7 ROM's or a ROM that has both lines in it.
 
That's why users are doing multiple tests, the average will smooth out the randomness.

emmaa said:
I'm guessing this isn't a fix for anything. It's most likely due to who's using the towers at that time and based on other completely random anomalies.
Click to expand...
 
+144% download
-19% upload

That is a loss? Maybe if you were using the phone as a file/web server! Most people will download many times more than they will upload.


emmaa said:
You lost upload, that is not a success in my books. It's more like someone stopped downloading something off the tower you're on. Try again in an hour and I bet you get different results.
Click to expand...
 
Didn't really do much for me I get the same speeds I always do.

screenshot-1342048050101.png
 
Another thing to consider is that VM's network is all over the place. I would not make any definitive conclusions or declarations until a bit of use in different areas. I work from an office in the Shareton accross the street from Sesame Place and just about every network has a tower on our Roof including Sprint and I could run 5 back to back speed tests and see something like,

1400 kbps down
83 kbps down
400 kbps down
900 kbps down
450 kbbs down

Another thing would be to consider waiting 15 - 20 minutes after rebooting before running any tests as there is allot going on the first couple of minutes after a reboot that could potentially effect any testing. Especially if your Triumph had been up for a bit when the before tests had been performed.
 
Chairshot215 said:
Another thing would be to consider waiting 15 - 20 minutes after rebooting before running any tests as there is allot going on the first couple of minutes after a reboot that could potentially effect any testing. Especially if your Triumph had been up for a bit when the before tests had been performed.
Click to expand...

Well for what it's worth when I tried this I had to go outside because I get no coverage in my building. Came back in fired up wireless downloaded the fixed ROM, wiped reloaded, setup my account downloaded the speed test and went back out side. Also when I went to lunch I was getting 3-4 bars which I never get. Maybe it's a fluke maybe it's not. All I know is I can't figure out why IssacJ changed this from TickerGuy's code. Need to spend more time looking over at github.
 
g60madman said:
Well for what it's worth when I tried this I had to go outside because I get no coverage in my building. Came back in fired up wireless downloaded the fixed ROM, wiped reloaded, setup my account downloaded the speed test and went back out side. Also when I went to lunch I was getting 3-4 bars which I never get. Maybe it's a fluke maybe it's not. All I know is I can't figure out why IssacJ changed this from TickerGuy's code. Need to spend more time looking over at github.
Click to expand...

I just envisioned users making the changes and then running a speed test the second their Triumph was back up and running and at the same time all their apps are trying to resync. May not be the case but just thought it would be a good idea to do a few more tests in different scenarios before declaring that their is no effect. Seems like the post has not been up all that long. I went from a 12k kbps down average to 93 kbps. I don't get a very good signal from my apartment.
 
There is no explanation as to why this should increase speeds. This isn't magic. These are computers. What is being fixed that is broken in the 3G system?

You'll never get anywhere with these speed tests. The speeds change from minute to minute. To smooth out the randomness of the network traffic, you'll need thousands of data points. Even then, you'd expect different results for 3am vs 3PM. So you would have to get your data points over a small period of time. Maybe if you just look at the maximum speeds possible with and without the mod, you'd have something. Even then, you'll need a lot more than a few data points to make sure you've captured the maximum.
 
Hello G60! I am usually a lurker around here on AF, but I just HAD to post on this! I can assure you these are definitely not placebo! I am using the Starship 7/7/12 ROM. I also did not use this, "persist.sys.use_dithering=0".Here is a before and after picture, the after being the higher one, of course. Thank you so much! I wish you luck on rebuilding CM7 aswell, and CM9.

mTSHN.jpg
 
To confirm, if it's necessary, that the aforementioned code is not present in MIUI's build.prop file. Maybe there are a few lines in MIUI that do more or less the same thing but that's far beyond my own knowledge.
 
bobboberts said:
There is no explanation as to why this should increase speeds. This isn't magic. These are computers. What is being fixed that is broken in the 3G system?

You'll never get anywhere with these speed tests. The speeds change from minute to minute. To smooth out the randomness of the network traffic, you'll need thousands of data points. Even then, you'd expect different results for 3am vs 3PM. So you would have to get your data points over a small period of time. Maybe if you just look at the maximum speeds possible with and without the mod, you'd have something. Even then, you'll need a lot more than a few data points to make sure you've captured the maximum.
Click to expand...

If you've ever taken a statistics course before, you should know that the maximum (and minimum) outliers are irrelevant, and also that, while having tons of data points would be nice, it only gives you more confidence as to the true average of your measurements, but that the average should not continue to change much at all once you've gone beyond 30 measurements or so. Thus, the average of 5 measurements instead of a single data point is still a bigger boost to the reliability of the data than 50 measurements over 5. Yes, I agree that time of day is an important factor however, and when I do further tests with more PRLs, I will be waiting to take them around 2-4pm, the same times I took my earlier measurements.
 
Awesome. Haven't had speeds this fast since I switched from OV. Thought VM just wanted to throttle down all the time!
 
You must log in or register to reply here.
Back
Top Bottom