speed issues!

[#1]

Quote from: spraxyt

No offence taken. I normally snip posts but I was so excited with the results you had achieved for me  I will do the adjustment you recommend later in the day and report back. Thank you.

[#2]

Try changing your MTU to 1500.  If it makes things worse, then just change it back to what it currently is.  Might need a bit of fiddling, so always worth giving different things a go

[#3]

Hi Spraxyt,
Set the figure to 95200 as suggested by you.  here's the tweak test URL.
http://www.dslreports.com/tweakr/block:4d9ee53?service=dsl&speed=8000&os=winME&via=normal
Here are the latest Thinkbroadband speedtest results: The highest of these was with the setting at your original one of 136000.  Is this significant?

29/08/07 16:37  5096.26 Kbps  377.85 Kbps  87.112.17.244 
29/08/07 16:23 5290.33 Kbps 378.40 Kbps 87.112.17.244
29/08/07 16:17 4176.06 Kbps 378.42 Kbps 87.112.17.244
29/08/07 16:13 3833.49 Kbps 371.73 Kbps 87.112.17.244
29/08/07 16:05 4766.22 Kbps 371.71 Kbps 87.112.17.244
I have left the MTU at 1400 still.
Going back to the Dr.TCP window you originally asked me to enter 1400 in a second box which didn't seem to be there.  There are 2 blank boxes on the window. One is marked 'Max Duplicate ACK' and the other 'TTL'  Nothing is entered in either.
Also there is a box with a drop down menu.  Currently it shows 'Dial Up adapter' but one thing on the drop down menu is 'Voyager USB Remote DDIS device'.  Should this be selected?
thanks for all your work on this problem.  Looks like we're getting there.

[#4]

Yes, you need to select the Voyager as this is the device you are using to connect to the broadband service. If you do select this you should get additional options.
Hope this helps,

[#5]

@artmo
The transmission percentage is slightly better with the smaller RWIN (90% compared with 87%) but that could be random variation rather than a real difference.  Since the larger value gave the highest speed on the TBB (ThinkBroadband) test I agree you should revert to that.
I have Max Duplicate ACKs set to 2 (set by another tweak program).  I believe using that modifies the way that congestion control works, though I haven't delved into that.  All I can really say is that 2 seems fine for me.  It's something you could try (as a single change to try to isolate its effect - if any) if you wish.
I have TTL set to 64, again done by the other tweak program.  This is the Time To Live, the number of hops (intermediate routers) the datagram can pass through before being discarded.  I think the default value (blank in DrTCP) is twice that.  In normal circumstances 20 is a lot of hops.  I suppose using the smaller value might allow "vagrant" datagrams to expire sooner.  I wouldn't expect this parameter to have a major effect in normal environments.
I'm glad you mentioned the Adapter Settings box had two choices on its drop-down menu.  As orbrey said you should select the Voyager line on this.  Doing that may give you the Dial-Up (RAS) MTU box.  You'll probably find the right-hand MTU reverts to blank.  Set that to 1400 and the RAS one too if it appears.
To summarise, set RWIN to 136000, select Voyager on the Adapter Settings drop-down menu and put 1400 in the MTU box(es), then retest on TBB and the tweaktester.

[#6]

Quote from: orbrey

Yes, you need to select the Voyager as this is the device you are using to connect to the broadband service. If you do select this you should get additional options. Hope this helps,

I can select 'Voyager' but when I leave the program and then open it again it has reverted to 'dial-up' Also when 'Voyager' is selected there are no additional options.  Any ideas?

[#7]

Quote from: spraxyt

OK, just re-set RWIN to 136000 and left MTU at 1400. Here's the Tweaktest URL: http://www.dslreports.com/tweakr/block:25316e9?service=dsl&speed=8000&os=winME&via=norma ....and here's the Thinkbroad band speed tests: Date  Downstream  Upstream  IP Address  Connection 30/08/07 13:18 4787.25 Kbps 372.17 Kbps 87.112.17.244 30/08/07 13:09 4835.83 Kbps 372.24 Kbps 87.112.17.244 30/08/07 13:08 4252.81 Kbps 371.99 Kbps 87.112.17.244 I haven't played around with the other figures you mentioned but will do so later and report back. Thank you for the information.

[#8]

When DrTCP starts it always shows the "first" item on the Adapter Settings drop-down list, so the behaviour you describe is normal.  Only the MTU setting is adapter specific and if the Voyager value needed to be changed again it would be necessary to select that item on the list before entering the new value in the MTU box.
Obviously the second (RAS) MTU box doesn't figure in Windows Me settings, so DrTCP doesn't include it (the program does "know" the rules for the supported operating systems and behaves accordingly).  It's helpful for such as myself to be aware of this to reduce the chances of us confusing other users - thank you for mentioning it.
Regarding your tweaktest result, this shows the "Max packet recd (MTU):" value (which actually means those *sent* by your computer) has changed to 1400 (from 1380), so the change you made has been applied.  Transfer efficiency is now 98% (up from 90%); that is good and hopefully is more than random day-to-day variation.  However the RWIN setting is still 95200, so the change to 136000 must have been discarded.  If you can change RWIN to 136000, save the settings and restart your computer, I think we may then have achieved a decent result for you.

[#9]

Quote from: spraxyt

Transfer efficiency is now 98% (up from 90%); that is good and hopefully is more than random day-to-day variation.  However the RWIN setting is still 95200, so the change to 136000 must have been discarded.  If you can change RWIN to 136000, save the settings and restart your computer, I think we may then have achieved a decent result for you.

Sorry, my fault.  I changed the value but didn't reboot.  I've now done that and the results look good.
http://www.dslreports.com/tweakr/block:18a8b6b?service=dsl&speed=8000&os=winME&via=normal
Date  Downstream  Upstream  IP Address  Connection
30/08/07 17:04 5084.85 Kbps 371.83 Kbps 87.114.160.46
I haven't played around yet with the Max Duplicate ACK or TTL.  Shall I still do this?
As you say, a decent result.  Thank you so much for all your efforts on my behalf.

[#10]

Transfer efficiency is 97% this time so it looks like the improvement seen last time was more than random variation.  I think we've now got to the best you'll see at busier times.
As far as trying Max Dup ACKs=2, have a go by all means (not forgetting to reboot); I doubt it will change performance significantly but trying it is the only way to be sure.  If you want to return to the default value, just delete the contents of the box, save and reboot.
Similarly with TTL, though I would expect that to have even less effect (but I could be wrong!).
Putting MTU (for the Voyager) up to 1500 might have a noticeable effect (5% improvement) but it would also be necessary to change the value on the router configuration settings (not doing that is likely to reduce download speed) and there is a risk some sites might not load properly.
My suggestion with MTU is to leave the value alone for now, get a general feeling for how broadband performs doing whatever takes your fancy, a fair slice of your bandwidth must have been used on speed and tweak tests.  After a couple of weeks or so, if you feel inclined to try increasing MTU come back here and (if you need help with this) I'll do my best to guide you through changing the router settings.
If you do try setting Max Dup ACKs and TTL I'd be interested to know the effects.
Thank you for your patience while the changes have been worked through, I think the end results are worth it.

[#11]

Spraxyt,
I've followed your advice and played around with the TTL and Max Duplicate max settings. Here are the speed tests:
With both set to zero  speed is 5.2Mb
With TTL set to 64 speed is 5Mb
With MDA set to 2 speed is 4.9Mb
With both TTL set to 64 and MDA to 2 speed is 5Mb.
Just realised I perhaps needed to do a tweaktest  as well??  Will this help?
Since following your earlier instructions I have constantly been achieving 5Mb+ which is considerably better than getting barely 3Mb before we started. Best speed has been 5.5Mb.
I am very grateful for all your help.  I hope you don't mind but I have nominated you as a 'Bright Spark'
Can I also ask, as I'm going on hols soon should I leave the router on or off?

[#12]

When you say "both set to zero" I assume you mean left blank, which means the default values.  From the test results it looks like that is a reasonable choice to make.  In fact the differences between all these is possibly a small amount of scatter, but it reinforces that leaving the settings blank, i.e. default, is a reasonable choice.  I'm delighted that making the changes has improved Internet performance significantly for you.
I don't think doing tweaktests would significantly add to our knowledge so I wouldn't bother with those now.
When you go away I suggest you turn the router off; this minimises the risk of some external phenomenon (power cut, lightning strike, BT tinkering with a neighbour's line, …) causing line errors and a low sync which can be a struggle to recover from.  When you return the router should re-sync normally again.
I recommend leaving the settings alone for a period now so that you get a feel for what is "normal" - in particular if you visit the Microsoft site (www.microsoft.com) (which tends to have an abundance of fancy graphics) does it load OK?  Similarly if you use eBay or do online banking do those work OK?  With the settings you have (MTU=1400 in particular) I would expect no problems.
Later, if you wanted to try MTU=1500, which would mean changes to router settings as well as those on your PC, if behaviour then became problematical you would know this wasn't "normal" and that reverting to MTU=1400 should fix it.
Thank you for the nomination, I appreciate you taking the trouble to submit it.
Enjoy your holiday! 

[#13]

Yes, you assumed right, I meant left blank.
No problems on the microsoft site and I've been bidding on ebay in tha past couple of days without any difficulties.  The only problem with online banking is my account balance but that's more to do with my trip to the US next week rather than any settings we've made       

[#14]

I also note that different speed testers consistently give better or worse results than other testers. Therefore it is not reliable to compare results from different speed test sites. My actuals downstream results rank them as follows:
PlusNet portal  (>7000 kbps)
Mybroadbandspeed.co.uk (6600-6800 kbps)
Speedtester.bt.com (5600 - 5800 kbps)
Thinkbroadband.com (5600 - 5800 kbps)
As one would expect the speed testers are affected by several characteristics of "the system". To analyse ADSL connections there continues to be a need to have a speed test as close as possible to the PlusNet connection or a means of baselining expectation for speedtesters outside PlusNet
Chris

[#15]

Theoretical maximum is 7.15Mbps and I would generally ignore the PlusNet Portal speedtester.