Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Bald_Eagle1]

That is very interesting & does indeed suggest that your experimentation does deliver the required outcome. i.e. increasing SNRM & lowering sync speed accordingly.

My query is have you ever been able to lower SNRM below 6dB in order to further increase sync speed?

My connection never drops below 5.5dB SNRM these days & even when it was in need of repair it would still be useable with SNRM occasionally into negative values.

Using my HG612 modem, I have been unable to lower SNRM or increase sync speed at all from whatever the connection negotiates at resync time.
Using the SNRM tweak command, the connection does resync, but SNRM & DS sync speeds don't really alter at all.

I would like to be able to at least test my connection at 3dB SNRM if at all possible.

I notice that US bitswap changes to OFF at the higher DS sync speed/lower SNRM.
I wonder if that actually has any relevance or not?

[Ixel]

That is very interesting & does indeed suggest that your experimentation does deliver the required outcome. i.e. increasing SNRM & lowering sync speed accordingly.

My query is have you ever been able to lower SNRM below 6dB in order to further increase sync speed?

My connection never drops below 5.5dB SNRM these days & even when it was in need of repair it would still be useable with SNRM occasionally into negative values.

Using my HG612 modem, I have been unable to lower SNRM or increase sync speed at all from whatever the connection negotiates at resync time.
Using the SNRM tweak command, the connection does resync, but SNRM & DS sync speeds don't really alter at all.

I would like to be able to at least test my connection at 3dB SNRM if at all possible.

I notice that US bitswap changes to OFF at the higher DS sync speed/lower SNRM.
I wonder if that actually has any relevance or not?

Yes, I have tested (accidentally) at 3dB a long while ago, as I was trying an offset above 100 and it just made it the lowest possible instead. The connection was a little unstable though (loss of sync at one point and more CRC errors). I haven't got a screenshot to prove this sadly, so you'll just have to take my word on that.

My connection usually stays within -0.5 to +0.2 of the original sync'd SNRM, depending on crosstalk (unless an engineer is in the cabinet fiddling around with wires for reasons such as setting someone up on Infinity, in that event I either get a nice increase in SNRM or a horrid drop in SNRM).

I haven't found a working way to do what can be done on the FB7390 on the HG612. On the FB7390 it seems fairly painless and easy. I am going to leave it connected so long as it remains stable (e.g. no crashes, no sudden reboots, no loss of sync) and will eagerly await to see if DLM adjusts my INP from 4 to a lower value, eventually. I expect DLM to intervene and reduce max/min rate on the downstream however, due to my reduced speed as a result of an increased SNRM.

I've found the FB7390 handles bitswap in a way that it's off when you first sync, but it switches on if it encounters a certain level of errors (at least from my observations, perhaps there's another cause I'm not aware of).

[Chrysalis]

Hi, this does interest me also as my snrm barely moves over 24 hours, doesnt even drop at night, so I suspect my line could maybe handle 3db snrm.  I say only maybe as although I think the sync wouldnt drop the increased error rate could push DLM into action.

But like asbokid says (hi by the way nice hg612 site you got). £200 is a lot of cash to spend on a modem.  My current sync is 71mbit so I would be paying £200 to potentially gain 10% speed but risk stability.

Also is there conformation of the syntax on the hg612 asbokid says is right? I would have thought setting 50 means 5.0 not 50%.  On other modems where I used to tweak snr margin it worked with the syntax of target figure with decimal place removed, but those were not broadcom chipsets.  As I understand it tho the device recieving the signal has overide control on target margin which is why upstream cannot be tweaked modem side but downstream can.

[burakkucat]

Also is there conformation of the syntax on the hg612 asbokid says is right? I would have thought setting 50 means 5.0 not 50%.  On other modems where I used to tweak snr margin it worked with the syntax of target figure with decimal place removed, but those were not broadcom chipsets.  As I understand it tho the device recieving the signal has overide control on target margin which is why upstream cannot be tweaked modem side but downstream can.

Please take a look at the graph in the Kitz wiki that shows the change in SNRM (dB) versus the adjustment ('tweak') value.

[Chrysalis]

yes I have read those documents now and understand the broadcom syntax now, I think the explanation is as Ixel said and simply the hg612 has no functionality for vdsl snrm tweaking, any idea if unlocked ECI modems can tweak it?

[Ixel]

Hi, this does interest me also as my snrm barely moves over 24 hours, doesnt even drop at night, so I suspect my line could maybe handle 3db snrm.  I say only maybe as although I think the sync wouldnt drop the increased error rate could push DLM into action.

But like asbokid says (hi by the way nice hg612 site you got). £200 is a lot of cash to spend on a modem.  My current sync is 71mbit so I would be paying £200 to potentially gain 10% speed but risk stability.

Also is there conformation of the syntax on the hg612 asbokid says is right? I would have thought setting 50 means 5.0 not 50%.  On other modems where I used to tweak snr margin it worked with the syntax of target figure with decimal place removed, but those were not broadcom chipsets.  As I understand it tho the device recieving the signal has overide control on target margin which is why upstream cannot be tweaked modem side but downstream can.

On another interesting note, since the last experiment I suddenly found myself one morning with no INP, delay and interleaving depth at 1. However, my speed since nearly 4 weeks now has been capped at 60 megabits, with a minimum rate of 30 megabits on the downstream. I've switched between ECI to HG612 but no difference, even tried to nudge DLM into giving me a penalty such as INP/interleaving again, but it doesn't seem to. It's like I've either confused DLM's algorithm for my connection or something else has happened. I've been on the HG612 (ECI DSLAM, as such makes CRC errors) for over 3 weeks sync time and no changes by DLM whatsoever :s. I'm wondering if DLM removed INP/interleaving based on my past history as having achieved 70-80 megabits (attainable currently 94 megabits~) and saw my sync near the absolute minimum threshold rate (30 megabits), so to prevent the possibility of me going below 30 it switched off all interleaving (that is, interleaving depth 1, no delay, no INP). That's my theory anyway, could be wrong, could be right, who knows.

[burakkucat]

. . . any idea if unlocked ECI modems can tweak it?

The functionality of the ECI B-FOCuS modem is proving to be somewhat difficult to understand. I suspect there is not the means to 'tweak' the DS SNRM. 

[Ixel]

. . . any idea if unlocked ECI modems can tweak it?

The functionality of the ECI B-FOCuS modem is proving to be somewhat difficult to understand. I suspect there is not the means to 'tweak' the DS SNRM. 

I believe there's an option (via telnet of course) to limit the sync speed rates, but I've not tested whether this works. From memory I can't remember what it is, I'll try and find out and post again later if I have.