VCXO Frequency Offset

[Tacitus]

For about 12 months I've been using a v6 2700.  When I first installed it the VCXO Frequency Offset was around -112 and over time and successive reboots this has crept up to the point where it is now -120 and the 2700 reports it as a suspicious hardware frequency mismatch.

Since the same thing happened to my last 2700 - a v5 - I assume -120 is the highest figure the modem reports, but if the current one follows the same pattern it's actually getting higher to the point where it will fail to sync.   I know frequency at the exchange end is tightly locked so I doubt it's happening there, so it must be the 2700 that for some reason is drifting.

Has anyone any idea why this should happen?  Is it down to line conditions, temperature variation, humidity, spooks, whatever.  At current rates I'm getting through 2700s every 12-15 months.  OK so they don't cost a fortune on e-bay, but it's still irritating. 

FWIW I've monitored temperature and it's always been within operating limits, although across a fairly wide range, which made me wonder if that might be the cause. 

Also, I tried the last 2700 that failed on my line over at my sisters  - also iDNET - where it worked perfectly for almost a fortnight, before I reinstalled her modem.  We're both on Max. 

Frankly I'm completely stumped.  It's the sort of thing that probably needs an in depth BT investigation, but the last time I got BT iround I got hit for the full £180 - probably £230 or so by now.

[Rik]

It sounds more BT than you, Tac, tbh. I've never had it happen, and we've not seen other reports of it happening with 2700s as far as I can remember. If the router worked at your sister's place, it also points to the BT end doing the drifting.

[Tacitus]

If the router worked at your sister's place, it also points to the BT end doing the drifting.

It does but if the BT guy comes out he'll plug in his modem which will almost certainly work.  Using another modem seems to reset something somewhere so the whole process starts again.  I suppose I can try using the old V5 and see if that will fool it into thinking it's a 'new' modem which might reset the process.

Not sure what iDNet could do, since BT will just stonewall on the grounds that I have a service, which, since their modem will almost certainly work is arguably correct.  It needs an investigation by someone with serious knowledge and some high end kit neither of which seem in plentiful supply at BT.

I suppose one alternative would be to switch to BT Business and see what happens with their supplied 2-Wire.

[Rik]

Rock and hard place spring to mind, Tac. We can, logically say that it has to be a BT issue, but getting BT to accept it and then do something about it will be next to impossible. What you need is a friendly and knowledgeable engineer, sadly, it's luck that decides what you get.

[Polchraine]

Frequency drift on consumer or office grade electronics is inevitable as the cost of better oscillator will push up the price of the device significantly.     The oscillator at the exchange will be very stable, but may not be dead on frequency and if slightly off then your VCO will need to compensate at day 1 and then any drift over time could pull it in either direction.    Given te nature of DSL services there is no need to lock the DSLAM oscillators together or to a frequency standard.

It would be interesting to find another user on your exchange who uses a 2700 and see what their offset is, and then try your 2700 on their connection to see what figure yours has there.

[Tacitus]

Frequency drift on consumer or office grade electronics is inevitable as the cost of better oscillator will push up the price of the device significantly.     The oscillator at the exchange will be very stable, but may not be dead on frequency and if slightly off then your VCO will need to compensate at day 1 and then any drift over time could pull it in either direction.    Given te nature of DSL services there is no need to lock the DSLAM oscillators together or to a frequency standard.

I can appreciate that, but can't help but wonder why I've had two exhibiting the same failure.  The exchange end is usually stable but I suppose they can drift for whatever reason

It would be interesting to find another user on your exchange who uses a 2700 and see what their offset is, and then try your 2700 on their connection to see what figure yours has there.

Unfortunately the only ones I know who have BT Broadband are using the HomeHub.  Might be worth trying the 2700 on one of those lines - always assuming they'll let me experiment 

[Polchraine]

Going back in my mind, I have come across two others that were similar.   One was mine and it would go across that threshold and then back,  cannot remember much about the other but it was another high negative value.

I'm now on VDSL so cannot see mine down here - and for some reason cannot login to the one in my Scottish home but I'll be there in 10 days and may remember to look.

Do you have your old one?    If so, and you are prepare to fiddle around, open up the case and see if you can locate the VCO/VCXO - a finger across the output to gnd/5v or the control input may bring it back. It is something hard to describe but something electronic design engineers do using "intuition" and experience.  Have a close look, there may be a small adjustment close by - used to set up in production - could be a series of printed links on the PCB which are cut or resoldered.

[Tacitus]

Do you have your old one?    If so, and you are prepare to fiddle around, open up the case and see if you can locate the VCO/VCXO - a finger across the output to gnd/5v or the control input may bring it back. It is something hard to describe but something electronic design engineers do using "intuition" and experience.  Have a close look, there may be a small adjustment close by - used to set up in production - could be a series of printed links on the PCB which are cut or resoldered.

I've still got it somewhere. I kept it as a spare in case Sis ever needed a replacement for her Zyxel, so what you suggest might be worth a try.  The impression I get is that it will report up to -120 but drifts further than that to the point where it won't sync. 

The V6 I currently use has never been below about -110, so I suppose it was only a matter of time.