Hardware Mod zur Steigerung der Annex A Qualitäten?

[el_valiente]

@drifting:

It's a pitty you don't have any spectrum before the mod.
Are you willed to replace the capacity by the original one and post the original spectrum?

It seems, the capacity mostly affects the upload stream. And this one looks good in your spectrum.

 

[drifting]

Don't know why but today my DSL line has great values, I think.
At night, is normal that all SNR values goes down a bit, but I never see both upstream and downstrem SNR on this levels.

I will now put back the old capacitor, screenshot all values, put the new capacitor again and check if values still to be good.

 

[el_valiente]

@drifting:

Thank you! I am very interested in your results (and I suppose, others are, too).

 

[drifting]

Hmm... I didn't change the capacitor yet (my sister need internet) but now my line is with poor quality like yesterday (on this image).

Easily we can conclude that with this 1uF capacitor the upstream are good and stable but downstream is much more unstable with bigger variation (best than ever from 7h to 17h and worst than ever, from 18h to 6h). Or this variation is the real variation of my line quality? I live in a village (+/- 30 ADSL users) @ 3.2KM from ADSL repetitor.

Maybe the capacity value of this capacitor isn't the most correct!? What do you think?

 

[el_valiente]

Maybe the capacity value of this capacitor isn't the most correct!?

If you read the first postings in this thread, you will see, that the original annexA boxes have 1uF as capacity.
That's the reason why we believe, that 1uF is a good choice. There are reports using 2uF, too, but this capacity gives no better results.

Any way, any improvement should occur at the lower frequencies in the spectrum.

 

[drifting]

If you read the first postings in this thread, you will see, that the original annexA boxes have 1uF as capacity.
That's the reason why we believe, that 1uF is a good choice. There are reports using 2uF, too, but this capacity gives no better results.

Any way, any improvement should occur at the lower frequencies in the spectrum.

I see... my spectrum has poor qualite with frequencies above 530KHz.

A capacitor with less capacity let pass low frequencies... and low frequencies of my spectrum are good. Maybe a value between 1uF and 33nF... like 500nF.
Tomorrow I will search my electronic stuff to try o find a ~500nF capacitor.

btw, if someone know the polarity of capacitor, please tell us. I have too much electrolytic capacitors with all values.

 

[el_valiente]

btw, if someone know the polarity of capacitor, please tell us.

In the FB print the capacitor changes polarity permanently! It is part of a frequency filter. So better use film capacitors.

 

[Sansan]

I see... my spectrum has poor qualite with frequencies above 530KHz.

A capacitor with less capacity let pass low frequencies... and low frequencies of my spectrum are good. Maybe a value between 1uF and 33nF... like 500nF.
Tomorrow I will search my electronic stuff to try o find a ~500nF capacitor.

btw, if someone know the polarity of capacitor, please tell us. I have too much electrolytic capacitors with all values.

The high pass cutoff frequency increases with smaller capacitor.
Electrolytic capacitors are not suited for high frequencies; in this location only film capacitors should be used.

 

[drifting]

Hi!
Take a look on all my tests.
I test with capacitors:
- 6.8nF 100V
- 33nF 100V
- 100nF 63V
- 0.1?F 63V (for sure it's 0.1uF=100nF)
- 1uF 100V

I made all tests this afternoon. On the end of tests (with 1uF capacitor soldered), I made a call to my ISP and change upload stream to 768kbps.
In the next week, I will try to buy and test capacitor with values like 0.2uF, 0.5uF, 0.8uF

What is your opinion?

 

[el_valiente]

What is your opinion?

You did a good job, thank you!

As far as I can see, all spectra look very similar, beside the one with the 6.8nF. With 6.8nF the upstream is noticeable reduced (from 512 to 288 kbit/s).
All other differences are minor, but anyway the upstream for 100nF and higher capacity looks a little better than for 33nF (the maximum is shifted to lower frequencies).

Conclusion: Only the upstream is affected. 100nF or 1000nF look a little better than 33nF; 6.8nF is definitively worse.
Knowing these results, I don't expect any news from 200nF, 500nF or 800nF.
Furthermore, something between 500nF and 1000nF seems to be a good choice for AnnexA.

 

[jaybee1]

hi,

Easily we can conclude that with this 1uF capacitor the upstream are good and stable but downstream is much more unstable with bigger variation (best than ever from 7h to 17h and worst than ever, from 18h to 6h). Or this variation is the real variation of my line quality? I live in a village (+/- 30 ADSL users) @ 3.2KM from ADSL repetitor.

Maybe the capacity value of this capacitor isn't the most correct!? What do you think?

Your conclusion is wrong...
The cap only affects the upstream range! So 1µF is the sole reasonable value.
And it has to be non polarized, because you want an frequency filter (which is "always" alternate current...).

The variation in downstream is caused by pattern noise (i.e. faulty microwave oven in your neighborhood)...

jay

 

[drifting]

hi,


Your conclusion is wrong...
The cap only affects the upstream range! So 1µF is the sole reasonable value.
And it has to be non polarized, because you want an frequency filter (which is "always" alternate current...).

The variation in downstream is caused by pattern noise (i.e. faulty microwave oven in your neighborhood)...

jay

ok, I understand. Thank you.

It's unfortunate that the fritzbox ADSL modem is not very good. With a minor variation, the line goes down, especially at night.
My ISP distributes Sagem 3102 and Dlink 3170 routers.. I have a friend with Sagem and sometimes at night the SNR hits 2dB and the line never falls.

Maybe I will use one of these routers to provide internet to my fritzbox.

 

[drifting]

The cap only affects the upstream range!

Maybe there will be any component that we can change to improve downstream quality?!

 

[jaybee1]

Hi,

My ISP distributes Sagem 3102 and Dlink 3170 routers.. I have a friend with Sagem and sometimes at night the SNR hits 2dB and the line never falls.

2 dB SNR and no drop? Improbable... That sounds like a faulty report. Normally the dslam drops the line if the snr goes lower than 6db.
What type of dslam shows the first page of the fritzbox stat?

Maybe I will use one of these routers to provide internet to my fritzbox.

Its worth a try Many sole modems are much more stable than integrated ones...

jay

 

[jaybee1]

Maybe there will be any component that we can change to improve downstream quality?!

No.
The cap mod just expands the lower edge of the filter range of the input filter for annex a.
The upper edge is much higher than used by your adsl type 1.

The only way to increase stability at poor lines is to change the datapump and the external filters of the hybrid chip. Both not practicable.

jay

 

[el_valiente]

Maybe I will use one of these routers to provide internet to my fritzbox.

My practical experience confirms that other router/modems could be better in AnnexA line than the FB internal modem.
Before knowing the speed-to-fritz patch I used Zyxel routers in Spain (ADSL Telefónica) in bridged mode (internal FB modem off) - with quite good results.

 

[Blindi0]

Hallo
Habe mich jetzt auch ein bisschen mit den Hardwareunterschieden der 7170 Annex A und B befasst.
Meine 7170 hat Annex B Hardware ("günstig" aus Deutschland importiert)
wie in diesem Forum schon steht hat das Annex A einen 1j63 Kondensator statt dem 33n100 Kondensator und einen 33nJ400 statt dem 223J400.
Hatte vorher kurz nach der 32 ein kleines Loch im Spektrum (siehe Bild) das ist durch das Austauschen der Kondensatoren weg.

Performanceänderungen sind aber leider keine zu verzeichnen.
Das Spektrum ist etwas schöner, weil kein Loch, aber sonst nichts (Kosten aber auch nur 1 Euro und 5Min Aufwand).
Bild unten: 7170 Annex A Platine
Für den Fall das jemand vergleichen will. Habe auch noch ein 9MP Foto falls Interesse besteht



mfg
Blindi

 

[RAMler]

Ich habe hier 2 W701V - einer der beiden synct an meinem Annex B Anschluss höher. Bisher schob ich das auf den Zufall bzw. Glück beim Modem Part, aber nach lesen diesem Themas steckt da ja evt. mehr dahinter.

Der eine Router hat nen 0,033µF Folienkondi (33n J100). Der andere einen Folienkondi mit Aufdruck "33nj63" zu dem ich leider keine Specs finden kann. Kann mir hier jemand weiterhelfen? Ich denke mal das 33n steht für 33nf, also gleicher Kondi, aber für was steht das j100 / j63?

 

[andilao]

Toleranz ±5% und max. Spannung 63 bzw. 100 V.

 

[el_valiente]

... steckt da ja evt. mehr dahinter.

andilao hat Dir ja schon die Erklärung für Kapazitäts-Toleranz und Spannungsfestigkeit genannt. Die Kapazitäten sind also identisch.
Ich bin so gut wie sicher, dass die Unterschiede in der Sync-Geschwindigkeit der beiden W701V nicht an diesen Kondensatoren liegt.