Broadband speed drop at router - Draytek 2762

[#1]

Please can knowledgeable members help me.

I have an ADSL broadband connection which affords me upto 9064 kbps download speeds and 1088 kbps upstream speeds.

Historically broadband speed tests have always been 1mbps less on the download speeds than shown on the router - this was the case with a Netgear DG834 router i had and remains the case with a Draytek 2762 router i bought 2 months ago.

I have recently started to experience connection drop outs and in response Plusnet sent an Openreach engineer out today.  The engineer reported no line faults and summised plusnets recent hard reset of the line and a resulting lower snr had increased the speed of the line but caused instability issues. The engineer also advised that when he did a hardwired test the router was delivering 1mbps less out of the back than was going into the router. The engineer recommended i contact plusnet to request a plusnet router to see if this resolved the issue.

Please can the plusnet community let me know if there might be a Draytek router setting that i could change to resolve the 1mbps speed loss issue (of note is that the draytek and previous netgear router had the same issue and have been set up with default settings).

Many thanks

[#2]

When you say the download speed out of the router is 1mbps less than the figure shown on the router, are you running some speed test to get that lower figure?

The sync speed of the line reported by the router will always be higher than the achieved payload download speed - each packet of data that is sent has header data that is necessary for the router to know what it is and where to send it (like an envelope for a letter), so that the actual data speed you see from a speed test doesn't take that extra data into account.

A 1Mbps drop on a 10Mbps line is about what I'd expect to see.

[#3]

Many thanks - the lower speed is what i record when i do a broadband speed test - i had thought it might be to do with internet bottle necks and therefore discounted it but the engineers reference to it this morning made me rethink whether there was a router/setting conflict - from what the engineer suggested when he used his equipment plugged intot he master socket he got 1mb higher speeds than he got when he plugged into my router.

[#4]

Just to add - i have an ADSL face plate/master socket so i wouldn't expect any significant speed drop between the face plate and router (1m cable).

[#5]

It really depends on just what the engineer was measuring. If he was looking at the sync rate reported by the router versus the sync rate that his magic box reported then just maybe they were using a different SNR.

[#6]

There needs to be crystal clarity on what is being measured here.  SYNC speed which is what is measured when the engineer plugs a box into the line and DATA speed which is what you measure when you run a speed test on your computer are NOT MEASURING THE SAME METRIC.

SYNC speed is the raw rate of passing bits (ones and noughts) down the phone line to the exchange.

DATA speed is the net rate of passing data from your PC to the remote data server.

The following factors make the metrics different...

1. BT IP profile (the very fastest that bits can be transmitted at) is 88.2% of line sync speed - 9.064 mbps SYNC speed gives a IP profile speed of 8mbps (already 1mbps less than sync speed)
2. 8 bits of data (a byte) needs "wrapping up" with error correction bits
3. A block of data needs wrapping up with protocol headers and flow control information
4. Messages sent need acknowledgements, hand shaking and error recovery

It is generally accepted that to account for 2-4 above (over tcp/ip) we need 2.8% overhead on the data - therefore with an IP profile of 8mbps might reasonably deliver just 7.77mbps NET DATA RATE.

Want a detailed explanation?  See: ::. Kitz - bRAS + IP Profile .:: and TCP Over IP Bandwidth Overhead - Packet Pushers

[#7]

Welcome to our community forums @stephenpk 

I'm sorry to see you're experiencing speed issues. I've tested your line and I've not been able to figure out the issue so I'm attaching the results below in case anybody else has further ideas.

xDSL Status Check
Circuit ID:	[CBUK redacted]	Service ID:	[BBEU redacted]
Telephone NO.:	NA	Test Executed On:	19-05-2021 17:13:53

xDSL Status Test Summary
Sync Status:		Circuit In Sync
General Information
NTE Status:		NTE Power Status:	PowerOn	Bypass Status:

	Upstream DSL Link Information	Downstream DSL Link Information
Loop Loss:	25.5	46.3
SNR Margin:	6.5	7.6
Errored Seconds:	0	0
HEC Errors:	0
Cell Count:	0	0
Speed:	1099	8639

Maximum Stable Rate (KBPS):	576	Fault Threshold Rate (KBPS):	576
Mean Time Between Retrains (Seconds):	86400	Mean Time Between Errors Upstream (Seconds):	2400
Indicative Line Quality:	A	Mean Time Between Errors Downstream (Seconds):	331
Custom Thresholds
MTBR_RED:	MTBE_RED:
MTBR_GREEN:	MTBE_GREEN:

If you'd want to rule out an issue with a setting within your router though I'm happy to send you a Plusnet router.

[#8]

For a 46dB attenuated line, those stats look somewhat impressive!

The slightly elevated SNRM might suggest that there is (or has been) some electrical noise around the place, but without sight of the router's view of the stats and the time when the router last re-sync'd such is a little speculative.

So the question remains, what comparative measures is the user concerned about?

Note it is possible that previously this circuit might have been "flying by the seats of it pants" with a target SNRM of just 3dB which might have added another 1mbps to the raw sync speed ... but at the cost of increased instability, increased retransmissions and a lower data throughput speed.

[#9]

Thank you for all the replies. Townmans response suggesting 9.064 mbps SYNC speed gives a IP profile speed of 8mbps seemed to correlate with my findings BUT i then noticed the comment 'therefore with an IP profile of 8mbps might reasonably deliver just 7.77mbps NET DATA RATE'.

As it stands right now my router reports attainable rate downstream of 8280 kbps - i wont grumble too much BUT with no other devices connected and when i run multiple broadband speed tests using hardwired ethernet connections the highest download speed result i can achieve is just 6.84mbps. The difference appears to be significantly different to the 2.8% overhead of data reported by Townman.  Am i correct to query where the remaining 1196 kbps is?  Thanks

[#10]

Screenshot attached to show Bt wholesale speed test results - now at 6.8mbps. Speedtest.net reports 6.91 mbps results.

[#11]

The ATTAINABLE speed is a theoretical RAW SYNC speed assuming all things optimal, sunny days, with a following wind according to the current line conditions).  Such is not real world needing to deal with normal transient electrical nose which is always present in the environment.

For useable DATA speed, as advise above, a theoretical maximum expectation is...

1. IP profile (effectively maximum useable packet speed) is 88.2% if the sync speed
2. Maximum data transfer rate is then 97.2% if the IP profile

The actual data transfer then needs to factor in what you PC can handle and what is happening in the rest of the connection beyond the exchange to the data speed test service.  Such is why different test services tend to deliver different results.  Speed is not everything, sometimes it is faster to go slower, for in trying to go fast at the margins, the retransmission rate can  increased, thereby having the effect of diminishing throughput capability.

For your line (46dB attenuation) the reported sync speed is way above reasonable expectation (approx 7.5mbps), so you already have a significant  bonus.

Taking your figures - the actual sync speed is over 4% above the reported attainable speed (which is a dynamic figure) and therefore under the current line conditions, you might reasonably expect there to be some level of error correction / retransmission being required.  Such reduces the EFFECTIVE DATA TRANSFER rate.

Wit the figures you have offered, there would appear to be nothing wrong with your data transfer rates.

What is the real world problem you seek to address here?