Fritz 6660 - wrong OFDMA values pulled

[Erroneuz]

Bug Type

Data quality or analysis issue

Affected Area

Monitoring dashboard

DOCSight Version

Latest release

Exact Version or Commit

v2026-07-07.1158

What happened?

I believe docsight is pulling incorrect OFDMA values for Fritz 6660, maybe also other Fritz modems.

Steps to Reproduce

1. Open Docsight
2. Go to channels and select the two OFDMA channel or notice the avg on the dashboard

What did you expect to happen?

OFDMA values that match the Fritz webinterface

What actually happened?

Docsight is reporting OFDMA values of 47.2 and 49.8 dBmV, but when I check the web interface on the modem it self, it reports 41.2 and 43.8, so 6 dB lower. I also have access to the CMTS, the modem connects too (as I run a smaller coalition ISP on a voluntary basis), and the CMTS reports the same values as the Fritz interface.

Deployment Method

Docker Compose

Modem or Router Model

Fritz!Box 6660 Cable - v8.25 firmware

Relevant Logs or Console Errors

Screenshots, Exports, or Extra Context

Docsight Dashboard

Docsight Channel compare

Fritz web interface

Panther - ISP Management software that pulls data from the CMTS

OFDMA values - UsTX is the monitored value in docsight and fritz web

**More detailed OFDMA values - UsTX is the value **

I haven't yet tested with Arris CM3500, Arris tg3442s and Humax HGF310EU, which are the other modems I have access to.

[itsDNNS]

Thanks for the detailed report and for including the CMTS/Panther comparison. The exact 6 dB delta is useful, because it points at DOCSight's intentional FRITZ!Box DOCSIS 3.1 upstream handling rather than a random parser error.

DOCSight currently adds +6 dB to FRITZ!Box DOCSIS 3.1/OFDMA upstream power before analysis. This is based on the long-standing VFKD/FRITZ!Box interpretation note that FRITZ!Box cable devices show DOCSIS 3.1 upstream Power Level about 6 dB lower than the value used for the VFKD interpretation scale. The FRITZ!Box-specific note is documented here:

https://github.com/itsDNNS/docsight/wiki/Modem-FRITZ!Box#docsis-31-upstream-power-6-db-offset

Historically DOCSight first compensated this in the threshold profile, then moved the compensation into the FRITZ!Box driver so that non-FRITZ!Box modems can use the real VFKD thresholds without a global -6 dB shift.

What is especially interesting in your report is that Panther/CMTS-side software appears to show the same lower value as the FRITZ!Box UI. That may mean Panther is exposing the same unadjusted DOCSIS/OFDMA reference value, or it may mean the old FRITZ!Box behavior is firmware/model/CMTS dependent and has changed for your 6660 on FRITZ!OS 8.25. We should not ignore that.

If you are willing to help, the most useful next step would be to identify what Panther's UsTX value actually represents:

• Is it read from a DOCSIS MIB/SNMP field? If yes, which OID or MIB field?
• Is it a modem-reported transmit-power value mirrored by the CMTS/software, or a CMTS-measured receive/ranging value?
• Does Panther distinguish DOCSIS 3.1 OFDMA reference power from total OFDMA channel power?
• Can you compare the same behavior with the Arris CM3500, Arris TG3442S, or Humax HGF310EU devices you mentioned?

For now I would keep this issue open as a data-quality/interpretation check rather than remove the +6 dB handling immediately. The current behavior is intentional and documented, but your CMTS-side comparison is a good chance to verify whether the community-known FRITZ!Box offset still applies universally or needs to become model/firmware/source dependent.

[Erroneuz]

Thanks for the detailed explanation, my apologizes that i haven't seen the wiki with information about the +6dB adjustment.
It might have been a change in their firmware, they have changed quite a bit in the v8 branch, in the start not for the better, but v8.25 has been running stable in our tests.

Of course, I have to dig up some information and get them validated from Panther and Vecima (our CMTS), but will look into it. I'll be back with more information in a couple of days.

[itsDNNS]

Hi @Erroneuz, were you able to confirm what Panther’s UsTX field represents or compare the behavior with any of the other modem models?

The exact source and semantics of the CMTS-side value would be needed before changing DOCSight’s documented FRITZ!Box +6 dB normalization. If no further data is available, we’ll close this for now while preserving the current behavior, and can reopen it when additional evidence becomes available.

[Erroneuz]

Hi, apologize for the lack of response.
Due to work and finally some vacation, I haven't had the time, but will test with the other modems later in this week to get some values, expect to see them posted on Friday.

Regarding some of the details regarding the readouts from Panther, I hope to have some clarification next week. The last 3 weeks has been vacation season, so I decided to wait with my inquiries until week 32.

[Erroneuz]

Hi again, I have some values from other modems, and the OFDMA UsTX values are close to each other.
This is from a test setup from our headend location, where the CMTS are located and all modems are connected to the same splitter.

OFDMA UsTX values

AVM Fritz 6660 v8.25:
Channel 1: 39.9
Channel 2: 40.3

Arris CM3500 v11.05.054.EURO.NCS:
Channel 1: 39.8
Channel 2: 41.0

Arris TG3442S v01.05.042.02.EURO.SIP:
Channel 1: 39.8
Channel 2: 41.0

I'm going to do a test tomorrow (Monday) on my home connection also, to validate if I'm seeing similar results.