38 replies to this topic

[mudmanc4]

Afternoon tests brought another wild-and-crazy download result with 112.3 mbps. The blue bar blurred across the screen. Follow-on test was 16.98 mbps. Uploads were 115 and 111 kbps. Changing TOS disable to enable and setting TOS to 136 seems to have made a small change in upload speed. There seem to be fewer excursions down to lower numbers. I have a 2Gb 82566DC network connection to my HN9000 modem. It runs at 100Mbps. What do other people have?

That is odd, once ca3le gets back in I'm sure he'll be interested in looking into that. If he isn't already.

[mawgdrum]

9 AM Eastern results from the Central US site: 13.6 Mb down, which is now my average download (on Hughes!) followed by 143.8 Mb. It would be nice if the blue speed bar had sound effects, maybe like an old hotrod Ford flathead. Upload is in its current usual neighborhood of 118 kb. Sound effects for this kind of speed would be as above, but with all the plug wires loose.

[mudmanc4]

... It would be nice if the blue speed bar had sound effects, maybe like an old hotrod Ford flathead. ...

not an entirely bad idea , in fun. Although the transfer might skew the test results, unless done before the tests start.

[zeddlar]

I saw out on the chart that someone on HN got a result of 139 Mb/s, if only. hehehe

[marsh_0x]

Hi for an update, TMN speed tests remains the same with a consistent, never varies much, now at post time 10.9, 8.6, 8.7, 8.4. 8.3, 8.5 Mbps using the 12MB load.1pm cst.
Wild Blue remains accurate at 1.116Mbps and but later will take a serious drop in speed.
-------------------------------
My base standard has to be Hughes speed test site always gives me accurate tests, confirmed many times with a 118MB with FileHippo "dedicated" servers.
Hughes 3.418Mbps
WB 1.116Mbps
----------------------------------------------
My conclusion the inaccurate tests begain when Hughes doubled our speeds, before here was getting 2.9Mbps almost always.
Happy birthday to (me at 64) just another day, my neighbor is cooking a choice cut beef tenderloin, I don't even like party's.

Best to you all...

Marsh

[mawgdrum]

Last Speed Test of 2011:
HN 1.47mbps down and 124kbps up
TMN 34.97 mbps down and 113 kbps up

I suspect that some of my tweaking has found the Achilles' heel of the TMN speed test.

[zeddlar]

I was getting accurate tests long after my speed doubled, but my speed doubled like 3 or 4 months ago. I think they did something with their compression to counter the speed problems the new FAP system could likely cause and this is what is throwing things off. It was when they did a wide scale rollout of the new FAP rollover that my tests went wild. I am still getting actual download speeds above what the HN test is showing for line speed. I am suppose to talk to HN about this tomorrow and see if we can figure out what the problem is.

[mawgdrum]

Using ping -f -l [MTU VALUE] WWW.HUGHES.COM I arrived at an MTU of 1472, much lower than the generic 1500 I had been using. The new value did not change download values to any marked degree and may have caused a slight increase in my pathetic upload speed. Browser response is (subjectively) a little "springier".

My question for someone techier than I am is, did I do this right? Is the MTU value 1472 correct for Hughes Net?

[mawgdrum]

32, 12, 18, 16, 17 then 128 mbps download. What color is the sky in your world?

[zeddlar]

Well now according to my actual download speeds, the HN test is accurate at reading my speed at an average of 200 to 350 Kb/s. My actual download speeds seem to range most of the day between 40 and 60 KB/s. This happened overnight day before yesterday so I think something has gone haywire at HN. I have had congestion slowdowns before and they have always been gradual and never anything like going from 4 Mb/s overnight to 200Kb/s the next day. I don't think the upgrades would explain this either since everyone at HN was on vacation when this happened to me. Probably need to sit tight for a few days and see if this works out on it's own, because I have a suspision that it will.

And btw, this site is still giving me download speed readings at 5 Mb/s and I should mention that my browsing doesn't seem any slower either which is strange. I mentioned this elsewhere but it seems to me like something is throttling actual file downloads while somehow letting normal browser activity through at full speed.

[mawgdrum]

I'm getting download test results from Hughes that average 2.1 to 2.2 Mbps except during slowdowns due to traffic in the evenings. Uploads average about 122 most of the time. TMS test results are over 17 Mbps download and 103 kbps upload. Recent uploads have been averaging about 117 kbps.

My Hughes service is the $50 a month package running with an HN9000 modem. My computer talks to the modem over an Intel 2GB network that runs at 100 MB.

Following this text is a copy of most of the tweaks I have done to my computer. They were done one or a few at a time. The changes that made the most noticeable improvements were first a disc defrag that perked eveerything up then a big RWIN setting. Setting MTU to 1472 may have increased speed but seems more to have given very steady, repeatable results on the Hughes speed test. I wasn't paying close enough attention to see which change caused an increase in TMN upload speeds, but it was probably resetting the MTU from 1500 to 1472.

I get crazy results sometimes on TMN downloads, up to 140 Mbps, that cannot be even close to reality.

My tweak listing is attached for some brave soul to try and let me know if similar results occur.

Most of my edits were with SpeedGuide's TCP Optimizer. The IRQ tweak was done on sysedit. The IRQ tweak may have done a little for speed but seems to have done a little for stability. I started with a value of 8192 then doubled it and currently have it set at 25600. No differences are apparent.

OS is Windows XP Professional SP3 build 2600 living with 3.23 GB of RAM.

MTU=1472
TCP Receive Window = 1031040
MTU Discovery: Yes
Black Hole Detect: No
Selective Acks: Yes
Max Duplicate Acks: 2
TTL: 64
Checksum Offloading: Enabled [0]
TCP 1323 Options: Window Scaling
MaxConnectionsPerServer: 10
MaxConnectionsPer1_0Server: 10
Host Resolution Priorities:
Local 5
Host 6
DNS 7
Netbt 8
LAN Browsing speedup: Optimized
LAN Request Buffer Size: 16384
Dynamic Port Allocation
MaxUserPort: 65534
TCPTimedWaitDelay: 30
Type/Quality of Service
QoS: NonBestEffortLimit: 0
ToS: DisableUserTOSSetting: 0
ToS: DefaultTOSValue: 136
DNS Error Caching: all cache time DWORDs:
set to zero
In System.ini: IRQ(Network card#)=8192

[zeddlar]

The test seems to be dead on again for me, thanks CA3LE.

[mawgdrum]

Results are still wildly inflated.
Hughes results for the morning on download: 2182 and 2177 kbps. Upload results: 125 and 126 kbps.
TMN results on upload: one test at 118kbps. Download results: 17, 11.91, 75.89 and 17.02 Mbps.
These crazy TMN results are skewing the Direcway averages. The upload reads are always roughly comparable between the two tests. The inflated results on TMN download tests have gotten to be just a curiosity and I still wonder what factor in my setup causes them to occur. The really wacko numbers seem to come when the lag time before the blue speed bar shows run longer than usual.

[zeddlar]

Turn turbo page off and try it. It would seem that something has changed with the turbo page compression that is causing the problem

[marsh_0x]

Turn turbo page off and try it. It would seem that something has changed with the turbo page compression that is causing the problem

Thanks Sir, that certainly corrected MY wrong high tests of 8.4Mbps, you win the prize.
Now back to normal whatever that is to me...

Confirmed using Hughes speed test and will late tonight time a 118MB dedicated server file when the others are sleeping.

[mawgdrum]

What is the turbo page? Where do I access it for changes?

[marsh_0x]

What is the turbo page? Where do I access it for changes?

Please understand the only reason is to make TMN more accurate for tests, none other.
Otherwise you venture into trouble on other web sites not even loading.

http://www.systemcontrolcenter.com/
Click the little guy in the bar going across screen (next to what is) brings up the left side panel.
Turbo page -> expands to Control -> Enable/Reset.

Choice is yours, remember how you got there and don't forget how to get Enable back.

Not worth my effort anymore with TMN as Hughes speed test remains accurate.
Unsubscriping from thread...

[marsh_0x]

Hi a little update still inflated high results.
As I mention early the double speed using HN test of 3.8Mbps I've become a victum of congestion after 21 month of fast service.
Prime time 5pm cst until midnight a sad 330-438 at posting is unacceptable.

My motto ~never do today what can be put off until tomorrow~ went with my neighbor to Cellcom, returned with a demo for her home wireless modem is seeing 1Mbps ping 0.19ms 50GB monthly cap for the lovely price $42 (includes sales tax).

Yep tomorrow will return for mine and sign on the line guaranteed price 2 years.
Testing begins, hiding in plain site I didn't see the light !
I'll be back with stored tests here of my new experience.
WB contract is fullfilled HN has 3 months remaining, so saving $128 monthly is sweet...

Marsh

[sgbotsford]

My understanding is that all satellite providers provide some form of burst mode, in addition to a FAP whack.

It's actually a good policy. The first N megabytes come fairly fast, for me, the first 5 megs about 40-50 KB/s then it tapers off to about 20 some. Then after a longer period it will stabilize around 12 KB/s. This makes web pages reasonably spritely. For Youtube it sucks golf balls through a garden hose.

I'm on xplornet (Anik F2) and for us, FAP cuts in around 25 MB/hour. Once it hits, you're dropped to something like 2xdialup speed for the remainder of that hour and all of the next hour.

The new satellite is more generous. FAP is calculated on the top 10% of users, and is in 15 minute increments.

Anyway, if you want reliable results, I suggest you try running a large number of 10 MB tests or so back to back, and see if the time stabilizes after a few. This will give you an idea of congested speed, or FAP whack speed.

Ultimately you have to decide what speed means for you. This sliding throttle schemes the sat companies come up with make speed as meaningful as speed on your commute to work. It depends. It just depends.

That said, regular speed tests are a good way to determine if you are getting reasonable returns for you money.

Being rural, I don't have a good option beside satellite, and all the companies have pretty much the same plan. When I get really pissed off I remind myself I could be on dialup, which even at the worst the satellite gets is 1/6 the speed at 1/2 the price.

My intent, while we are in the process of converting over to the new satellite, is to record twice daily speed tests. This way if they drop their speed after the one month trial, I can complain to consumer affairs for fraudulent advertising.