Hi!
So for everyone who has problems with BS-11 clock accuracy, what would be the theoretical straight-forward way to solve the problem, is to replace the 32.768MHz crystal oscillator on the E1 card with a OCXO of the same frequency.
So what we're looking for is an OCXO with +/- 0.05ppm or higher stability for 3.3V power supply. Unfortunately I was not able to find one at that frequency, even digikey doesn't have one.
If anyone can find a source for a OCXO in the abovementioned parameter range (and small quantities), I think quite a number of people would be interested in buying one and connecting it to their E1 boards. Any help is much appreciated.
Thanks!
Hi,
On Wed, Jul 22, 2009 at 12:28:55PM +0200, Harald Welte wrote:
If anyone can find a source for a OCXO in the abovementioned parameter range (and small quantities), I think quite a number of people would be interested in buying one and connecting it to their E1 boards. Any help is much appreciated.
After having given up on the same task, I decided to try Andreas' isdnsync using my ISDN line. I synchronized the clock of the BS-11 by connecting it to a cheap HFC-S PCI card and connecting that to the ISDN network.
I just connected C4IO and F0IO, added type=0x00800 to the hfcmulti parameters, ran isdnsync and started bsc_hack. I could then watch the oscillator frequency getting slowly adjusted to the ISDN line with bs11_config (the value changed from somewhere around 1840 to 1035). It took about an hour for the work value to stablilize.
Now all available mobiles can see and join the network and are also able to find and join the official networks again.
I've uploaded a (bad) picture of the modified card here: http://www4.cs.fau.de/~gernoth/IMG_8005.JPG The attached cable can be directly plugged in to one of the PCM connectors on the E1 board (PIN 9 and 11).
Regards, Michael
On Thu, Jul 23, 2009 at 10:36:44AM +0200, Michael Gernoth wrote:
I just connected C4IO and F0IO, added type=0x00800 to the hfcmulti parameters, ran isdnsync and started bsc_hack. I could then watch the oscillator frequency getting slowly adjusted to the ISDN line with bs11_config (the value changed from somewhere around 1840 to 1035). It took about an hour for the work value to stablilize.
Ok, thanks. It seems I'll have to do the same thing and pre-calibrate the BS-11 units that we'll be taking to HAR2009, as we certainly won't have any actual ISDN lines there...
Hi,
On Thu, Jul 23, 2009 at 06:09:57PM +0200, Harald Welte wrote:
Ok, thanks. It seems I'll have to do the same thing and pre-calibrate the BS-11 units that we'll be taking to HAR2009, as we certainly won't have any actual ISDN lines there...
One pre-calibrated unit should be enough, as you should be able to use the clock generated by one BS-11 as PCM master clock for another HFC card with a second unit attached.
Regards, Michael
On Fri, Jul 24, 2009 at 01:10:26PM +0200, Michael Gernoth wrote:
Hi,
On Thu, Jul 23, 2009 at 06:09:57PM +0200, Harald Welte wrote:
Ok, thanks. It seems I'll have to do the same thing and pre-calibrate the BS-11 units that we'll be taking to HAR2009, as we certainly won't have any actual ISDN lines there...
One pre-calibrated unit should be enough, as you should be able to use the clock generated by one BS-11 as PCM master clock for another HFC card with a second unit attached.
I'm planning to use only one E1 card. Yes, theoretically that should still work with only one calibrated unit, but I'd rather not take my chances.
-
Harald Welte -
Michael Gernoth