xDSL modem from Mikrotik would be great…
No, it would be a can of worms. Every country, every provider uses slightly different setups and wellknown DSL
modem manufacturers make different firmware versions for different situations. You see that happening here
in this thread, where a German distributor has had contact with the manufacturer and has a special version made.
However, what would be great: Support from MikroTik for the protocol used to display the connection information
of these VDSL SFP modules so you can see the training speed, operating mode, SNR etc. That would be very valuable
to the DSL users and MikroTik would not have to deal with the particular issues of DSL.
No. They have their own stickers on the SFPs, but they come up as Proscend 180-T in the SFP information in Mikrotik.
/M
Any known issues with power consumption using these in a CRS125? The supplied power supply is 12v 2.4 amps I believe. I have a CRS125-24G-1S so there’s no WiFi.
The factory supplied power supply is 24V 0.8A. Either way, I don’t think you’ll see any issues.
I’ve ran those on CCR1016, HAP ac and 951s. Currently testing 2 of them in a CRS106-1C-5S. No issues to speak of.
/M
\
Hi folks,
very interesting thread 
In Germany DTAG (Telekom) started to sell “super vectoring” (G.993.2 Annex Q, profile 35b, DTAG requirements on page 62). Are there any experiences out there with modems discussed in this thread? How future proof are these devices? Could there be something like a firmware update to support new profiles or do we need new hardware then?
Regards.
Not. The Allnet ones can handle 20b supervectoring, but not 25b. Tested with AVM Fritz!Box 7582 and 7590. Against the Master sfp of the Client/Master pair.
Haven't had the chance of testing them on a 25b supervectoring DSLAM yet.
/M
For 35b you have to wait for this one:
Interesting that the development goes on. Next would be a model that supports line bonding?
It is used here to get even higher speeds: 2 lines in parallel with ethernet-level load balancing to get 2x the speed.
However, I have abandoned the use of these things as there is apparently no support forthcoming from MikroTik to view the status and it is not acceptable to run connections “in the blind”.
So external modems it will have to be…
There is no point in developing these for line bonding.
First of all, they get hot enough as they are handling 1 phone line. You also have to consider power consumption of the sfp, which is high as it is.
Secondly, there are no problems running multiple in for example a CCR1016 and then either run MLPPP (If your provider allows you to establish multiple lines with the same credentials) or the way we use them is with bonded EoIP tunnels and baby jumbo frames, which works excellent for bonding, of you are in control of both ends.
/M
Sent from my SM-N950F using Tapatalk
\
In the future??? Is DSL the future, I thing PON and SFP is the future!
There are lots of Telcos, that still build on their copper plant and are ignoring the migration to FTTH. Hey .. German Telekom is one of those, that still haven’t got around to fiber thinking.
The biggest joke of that is large areas in eastern Germany, where the copper networks were so bad, that when they replaced it, they put fiber in. But they had not invested in technology to deliver broadband on fiber, so when DSL came around, those on fiber couldn’t get broadband. Don’t think they’ve fixed that yet … 20 years later.
Here in Ireland we got (sort of) lucky, in that we have 3 large FTTH deployments and lots of smaller ones. And one of the larger ones now covers 180k+ rural homes with FTTH. But VDSL is still a big thing, especially in urban areas, and 25b supervectoring is being implemented before the end of the year here. So support for that would be good to have.
/M
Bonding is an option offered by the ISPs here (as part of their competition with cable and fiber which offer a lot higher rates) and it is not using MLPPP.
It is balancing at the ethernet level and a single PPPoE connection is running over the combined link.
For the subscriber it is a single connection with a single address but with double speed.
There are a couple of existing external DSL modem/routers that support bonding and the only visible difference is that they use all four wires of the RJ11 line connector.
I don’t know if there are two single-line DSL chips or a different special dual-line DSL chip in them.
The issue is not the equipment but the cables running to the homes. E.g. where I live there is fiber to streetcabinets that offer DSL to the homes, and I get 100/30 Mbps
over a single Vectored VDSL2 line. And I could have that extended by using 35b profile AND by using bonding, so the speed can be about quadrupled.
Wiring fiber to the individual homes, while being done all over the country, is expensive and not as easy as using the existing copper wiring infrastructure.
Usually it is only done with subsidy from the municipality or a housing council. The Telco/ISP by itself will not do it.
(of course you can get individual fiber but it has to be paid at cost of installation which is usually about 50 euro/meter of digging)
Of course it’s expensive to replace existing copper infrastructure with fiber. No question there. And thats why DSL isn’t dead.
But i’ve seen scenarios where german telecom even has rolled new copper out instead of going fiber. And that’s outright stupid.
/M
Copper perfectly suited to be used in closely build area’s and over cable 1Gbit/s is already used and VDSL is also available in higher speeds.
Fiber is well suited to cover long distances of 20+ km and here in the Netherlands it is more and more used for people living outside the cities. The people run their self the bundeling of the interest people.
A bad thing is when there are enough people interested the big telecom firms try to torpedo the iniative by offering their solution the never will be implemented.
So do you have one you’re willing to sell and ship to UK? Email me at dofrotest123@gmail.com as this board doesn’t seem to support direct messaging?!
I am not going to ship mine as I am not certain it is 100% working. (the past 2 experiments failed to result in link while that worked before)
However, you can easily order it from FMS Internetservice GmbH in Germany.
I bought a ALLNET ALL4781-VDSL2-SFP for my VDSL2 17a G.Vector (ITU G.993.5) line using 1&1 in germany. It’s sold as a 100Mb/s VDSL2 line and syncs at around 109Mb/s on a fritzbox 7412.
The modem synced fine in my RB2011 green light went solid in less than a minute. After some confusion on the PPPOE details (eventually got them with packet captures via the fritz.box diag tools) I was able to get to it “Link established” by having the PPPOE client connect to VLAN7 on the SFP interface. That got me to “link established”. Then messing with the MTU values to set it to 1492 got me to “Authenticated” in the PPPOE client and I could scan and find a service and mac.
However, I didn’t manage to go further (which I assume is the ‘connected’ stage), the PPPOE client just goes in a loop and tries to reconnect after 1 minute. I’m a bit out of ideas but I feel like I’m really close so any help is welcome!
I finally have a FTTC connection provided by UK ISP “SSE” so now I can post an update, which is really a request for help 
With the ISP-supplied router/modem (a Technicolor TG589vac v2) everything works as expected and specifically I can send 1500-byte pings (i.e. 1500 IP packet size) with don’t-fragment bit set successfully to/from my internet server. (I’ll call these “full size pings”). So we know my ISP supports RFC4638 / baby-jumbos.
If I connect my hAP ac running RouterOS v6.43.4 to a TG589 LAN port then full-size pings still work OK.
If I set the TG589 into bridge mode, where the TG589 only handles the VDSL side (including VLAN tagging required for UK FTTC network) and the hAP does PPPoE then it partially works:
/interface pppoe-client add add-default-route=yes disabled=no interface=ether5 keepalive-timeout=disabled max-mru=1500 max-mtu=1500 name=SSE password=PASSWORD profile=pppoe-client-default use-peer-dns=yes user=USERNAME@ISP
13:59:13 pppoe,ppp,info SSE: initializing...
13:59:13 pppoe,ppp,info SSE: connecting...
13:59:13 pppoe,debug,packet ether5: sent PADI to FF:FF:FF:FF:FF:FF
13:59:13 pppoe,debug,packet session-id=0x0000
13:59:13 pppoe,debug,packet host-uniq=0x15b0001
13:59:13 pppoe,debug,packet service-name=
13:59:13 pppoe,debug,packet ppp-max-payload=1500
13:59:13 pppoe,debug,packet ether5: rcvd PADO from 2A:8A:1C:EC:7C:29
13:59:13 pppoe,debug,packet session-id=0x0000
13:59:13 pppoe,debug,packet ac-name=nge001.clc-re0
13:59:13 pppoe,debug,packet host-uniq=0x15b0001
13:59:13 pppoe,debug,packet ppp-max-payload=1500
13:59:13 pppoe,debug,packet service-name=
13:59:13 pppoe,debug,packet ac-cookie=1e 3f 7b 3c df 1e fe 44 ba 13 e9 34 fc 3e 7a 62
13:59:13 pppoe,debug,packet ether5: sent PADR to 2A:8A:1C:EC:7C:29
13:59:13 pppoe,debug,packet session-id=0x0000
13:59:13 pppoe,debug,packet host-uniq=0x15b0002
13:59:13 pppoe,debug,packet service-name=
13:59:13 pppoe,debug,packet ppp-max-payload=1500
13:59:13 pppoe,debug,packet ac-cookie=1e 3f 7b 3c df 1e fe 44 ba 13 e9 34 fc 3e 7a 62
13:59:13 pppoe,debug,packet ether5: rcvd PADS from 2A:8A:1C:EC:7C:29
13:59:13 pppoe,debug,packet session-id=0x015c
13:59:13 pppoe,debug,packet service-name=
13:59:13 pppoe,debug,packet host-uniq=0x15b0002
13:59:13 pppoe,debug,packet ppp-max-payload=1500
13:59:13 pppoe,debug,packet ac-name=nge001.clc-re0
13:59:13 pppoe,debug,packet ac-cookie=1e 3f 7b 3c df 1e fe 44 ba 13 e9 34 fc 3e 7a 62
... then LCP, IPIP, etc.
The hAP doesn’t want to create a PPPoE session with 1500 MTU:
/interface print where type="pppoe-out"
Flags: D - dynamic, X - disabled, R - running, S - slave
# NAME TYPE ACTUAL-MTU L2MTU MAX-L2MTU MAC-ADDRESS
0 R SSE pppoe-out 1480
The hAP ethernet port connected to the TG589 has MTU set to 1512. I’ve tried setting various pppoe-client settings like “max-mtu” and “max-mru” to 1500. Even if I set the “pppoe-out” interface’s MTU to 1500 it reverts back to 1480 after a few seconds.
Trying to offload more work onto the hAP just makes things worse for me.
With VLAN tagging turned off on the TG589 and a vlan interface created on the hAP then PPPoE from the hAP doesn’t work at all:
/interface vlan add interface=ether5 mtu=1512 name=VDSL-vlan vlan-id=101 arp=disabled
/interface pppoe-client set [ find name=SSE ] interface=VDSL-vlan
I’ve no idea why because the “PADI” packets sent have the correct vlan ID but no “PADO” responses arrive:
19:17:12 pppoe,ppp,info SSE: initializing...
19:17:12 pppoe,ppp,info SSE: connecting...
19:17:12 pppoe,debug,packet VDSL-vlan: sent PADI to FF:FF:FF:FF:FF:FF
19:17:12 pppoe,debug,packet session-id=0x0000
19:17:12 pppoe,debug,packet host-uniq=0x13a0001
19:17:12 pppoe,debug,packet service-name=
19:17:12 pppoe,debug,packet ppp-max-payload=1500
19:17:13 pppoe,debug,packet VDSL-vlan: sent PADI to FF:FF:FF:FF:FF:FF
19:17:13 pppoe,debug,packet session-id=0x0000
19:17:13 pppoe,debug,packet host-uniq=0x13a0001
19:17:13 pppoe,debug,packet service-name=
19:17:13 pppoe,debug,packet ppp-max-payload=1500
The ideal final setup is with the Proscend 180-T in the SFP port in the hAP:
/interface ethernet set [ find default-name=sfp1 ] speed=1Gbps arp=disabled auto-negotiation=no mtu=1512
/interface vlan set [ find name=VDSL-vlan ] set interface=sfp1
As other people have said, setting speed=1Gbps and auto-negotiation=no gives the interface “running” state.
The green LED on the Proscend goes solid so I guess DSL synchronization is OK.
Again, no “PADO” responses arrive when I try PPPoE.
I do have a couple of RB260GS with also feature an SFP port so might try some tests using that.
So I guess I’m asking for feedback on:
- Known working/broken RouterOS devices, e.g. hAP ac works, hAP ac2 broken, RB260GS
- Known working/broken RouterOS versions, e.g. 6.43.4 works, 6.40.1 broken
- Confirmation/config examples of PPPoE with vlan tagging, particularly with 1500 MTU allowing full-size pings
Other related threads:
PPPoE MTU problem
PPPoE connection to UK BT ADSL, MTU with RFC 4638 which mentions RouterOS RFC4638 support since 6.33!
PPPoE RFC4638 problem. Not seeing 1500 MTU on pppoe-client?
More investigation into Mikrotik PPPoE negotiated MTUs