PPPoE speed issues

Hi everyone,

I am experiencing an unusual throughput issue with a specific PPPoE client and would appreciate some insight.

Environment:

  • AP/Router: MikroTik RB4011
  • CPE: MikroTik LHG5
  • Provisioned Speed: 100Mbps Down / 25Mbps Up (PPPoE)

The Issue: The client is reporting that their download speeds are fluctuating between 50Mbps and 70Mbps instead of reaching the provisioned 100Mbps.

Troubleshooting Performed:

  1. Router to CPE Test: Running a bandwidth test from the RB4011 directly to the LHG5 accounts for the full 100Mbps without issue.
  2. CPE to Router Test: When running the test from the LHG5 back to the RB4011, the throughput drops to 50–70Mbps.
  3. Queue Scaling: If I artificially increase the client’s queue limit to 150Mbps, their actual throughput increases to around 70–80Mbps.
  4. Queue Disabled: If I completely disable the queue for this client, they successfully achieve the full 100Mbps.

It appears that the queue itself is introducing a bottleneck or overhead that prevents the client from pulling the full line speed. Has anyone encountered this specific behavior with this hardware combination, and what is the recommended fix?

Could this be related to queue types, buffer sizes, or a CPU limitation when FastTrack is bypassed? Any advice on optimizing this would be greatly appreciated.

Thank you in advance.

Ps OS version 6.49.21

The queue dropping packets is the most likely cause here. Let’s confirm it first:

In Winbox:

  1. Go to Queues → Simple Queues
  2. Double-click the client’s queue
  3. Open the Traffic tab
  4. Have the client run a speed test while you watch

You’re looking for two things:

  • Does the Dropped counter increment during the test?
  • Do the drops coincide with the speed dipping below 100Mbps?

If yes — the queue is actively discarding packets, which causes TCP to back off and reduces throughput.

Reply back with what you see and we’ll go from there.

It seems i can not upload any screen shots. Yes i can see packets are getting droped.

Is there any souloution to the problem?
the signal is 54/55 on the LHG5 ccq its at around 94%

there’s an issue with this forum, we’ll migrate it soon to a faster and better community forum (currently working on this).

Idea that could help: how about setting it to 110Mbps in the queue (or in Zima that gets propagated to the Mikrotik simple queue).
Burst could also help for example an extra 20Mbps for 10seconds (You could also set this in Zima, it propagates it to mikrotik simple queue).

If other users faced this and solved it, please share your feedback

Good evening, Thant you for that suggestion. I did do the 110Mbps up to 250Mbps, The client i only getting the highest was 80Mbps. It only gets to a 100Mbps, if i disables the simple que from Zima. and i cant leave it like that as my Main router reboots at 3AM every morning and the que will be re-enabled by zima.

Just to clarify, the queue is actually recreated by Zima during the router sync (a few minutes after deleting it), rather than after a reboot. Give it a test and see.

Also, for the 80Mbps calibration, try setting it to 125Mbps. It’s a best-effort hotfix that gets the job done for now, but if anyone else has run into this and fixed it another way, I’d love to hear what worked for you.

Good morning,

Sorry for the delay. I have done the Deletion of the 100Mbps and i have increased the 100Mbps to 250Mbps. The Device was updated but could not get past the 80Mbps. Only getting to the 100Mbps by disabling the que with in the router. But what i did do was i moved the PPPeO to the Dish At my side and Placed the client Dish on a Bridge. now i do get around 90Mbps to 97Mbps at the client.