Print

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

[Weaver]

[Apologies if we've talked about this subject before some while back, although from a different angle]

Referring to the later posts in the thread
        http://forum.kitz.co.uk/index.php/topic,17237.msg317161.html#msg317161
I notice that the poster said that he had seen better SNR results reported by his modem / router when using what he considered to be a higher quality DC source than the original wall-wart PSU that was supplied with it.

Q: I wonder if it's possible to find and try out a known very high quality mains-to-<x> V DC alternative unit that is plug compatible with the wall-warts for my DLink DSL-320B-Z1 modems?

1. Would have to be plug-compatible, am not up to doing any physical mods, I'm afraid.
2. What is x? I'd need to look up the specs for the modem. I believe x=9.
3. I can't really test it: Would need firstly (a) a scope to verify the quality of the hardware, and secondly, (b) in order to see the end result, if any, it would be necessary to fire up mydslwebstats or similar and graph the relevant figures from the modem. Since I'm ashamed to say I don't have the resources to do either, I might need help from a volunteer tester?

This unit on amazon
        https://www.amazon.co.uk/dp/B004MZQKZI/ref=wl_it_dp_o_pC_nS_ttl?_encoding=UTF8&colid=1NKM82F39L7B&coliid=IDXP47M9WKZ11

says 9 V. And claims high quality, but of course that cannot reasonably be read as meaning higher quality output, or higher quality in any other sense than the modem manufacturer’s unit. It just perhaps is meant to claim that it's maybe durable or well-built, has good quality components, or forty things. It mentions conversion efficiency somewhere for example. It might be worth a try, but I'm not willing to bet on it.

Could anyone find a spec for the DSL-320B-Z1’s power supply requirements, incl current budget?

High quality output would have to mean all of (a) a stiff supply, more than adequate current delivery, low internal impedance plus no inductance getting in the way of really fast slew rate, high dI/dt. And also (b) low noise and low ripple. And all still true even under real-case, fast-varying load. Have I forgotten anything?

What might definitely qualify as a known, very high-quality, practical, usable 9 V DC source?

* A battery, we've talked about this before in earlier threads. A bit of a usability nightmare and not plug-compatible unless I could get a lot of help / volunteer with a soldering iron etc.
* Other wall-warts that are simply very high-quality designs? Would be ideal.
* Daft idea: What about laptop PSUs? Not plug-compatible, wrong voltages for all I know, and are they actually higher-quality outputs? Certainly high current, presumably stiff too?
* Anything else I've omitted that could give really high quality output plus practicality / ease of use?

Baseline: In any case, I don't know how good or bad the Dlink-supplied wall-warts are, so I don't know what the baseline is. They will be incredibly cheapo, although I don't know how bad or not cheapness might be. Nor do I know how significant the impact of psu quality is.

Measurement: Another point is that it might only be possible to capture representative measurements of modem stats if using my particular ultra-long line, or one of extreme length just like it, because we want a really weak DSL signal, plus (assumption, guesswork) a relatively quiet line noise environment especially considering the length of line. We are then going to be considering the ratio of a low DSL signal to a possible low PSU-contributed noise source.

Thinking again about the poster’s observation, I also ask myself how an improvement that actually reaches the modem might affect performance? Is there a chance of seeing higher sync rates? A reported SNRM holding up better=higher? Some other result visible with detailed graphing? If there is a change in the quality of DC input seen at some level, how significant might that be in terms of affecting DSL modem performance? If a modem has its own smoothing and filtering, then that might simply block any effect, in which case, good for Dlink.

[ Yet another long post from that Weaver. What is that guy on? (See attached A4 sheet  ) ]

[ejs]

I think any differences between power supplies will be due to how much RFI they pump out, rather than the characteristics of the DC output. Chips don't run at 9V or 12V, so the device will be converting it down to lower voltage(s) anyway.

Also, I suspect anything marketed as being a power supply for a specific model of router is going to be overpriced, aimed at being found easily by a search, or for the benefit of people who wouldn't know how to pick out a compatible generic power supply.

I may have got a very small increase in bandwidth from swapping the original Netgear power supply for the DG834Gv3 with a suitable, slightly newer power supply that I happened to have around. It was a PSU that came with an external hard drive. I've actually got quite a few largely interchangeable 12V PSUs. I don't really want to test them all, none of them are for sale, and none of them are particularly new anyway. It's not really possible to determine using an AM radio if the amount of RFI one puts out is more or less than any other, they all put out a fair bit.

[underzone]

You could always add some of these to the PSU cable, just before it enters the modem. Easy to add and remove.

http://www.ebay.co.uk/itm/Clip-On-EMI-RFI-Noise-Ferrite-Beads-Split-Core-Choke-Clip-Filter-5MM-USB-Cable-/281961820637?var=&hash=item41a63bf1dd:m:mEjRDT2y0sKVdTzqNvjGU2A

[Weaver]

@ejs Quite agree with your points about the products advertised on Amazon, they are readvertising the same unit with "Compatible with Xxx" like keyword-spamming. Try googling something like Florists in <Heasta> for example (my village). Indeed you could probably even get a match for "Florists in North Uist" for all I know. Btw there are none.

I hadn't thought about RFI, good point.

@underzone Agreed. Am already using a couple of ferrites on each modem DC cable. (Probably from Maplin.)

[ejs]

One of my 12V PSUs does have a ferrite near the DC plug, but this one has the highest rating of 3A so I'm using it to power an external hard disk. I've also got another which has the ferrite at the other end of the DC cable, close to the PSU.

@Weaver
Have you tried any other modems since the switch to 21CN and ADSL2? Considering the equipment at the other end has changed, what worked best in the past might not be the best now. Evaluating the effects of difference PSUs might be an interesting project in itself, but there may be bigger gains to be found elsewhere.

[PhilipD]

Hi

Only a very bad power-supply (or faulty one) would degrade performance, either directly via the DC output or indirectly by squirting out lots of RF noise.

Generally all correctly functioning power supplies will not give rise to much difference, unless the device is faulty anyway, perhaps suffering from degraded capacitors.

The thing is, the voltage supplied from the power supply (typically 12 volt DC for most routers/modems) goes through several even nosier switch mode DC-DC converters inside the device anyway.

Still if a good supply is required as a replacement, here is a good 12 volt one http://www.ebay.co.uk/itm/401120185273 (I ordered one for a pfSense PC running an Intel CPU and works great and is a genuine one, plus fits 2.1 and 2.5 DC type jacks).  A tip, look out for the efficiency indicator on any replacement power supply and stick with V (tiny V in a circle) as this indicates best efficiency so it runs cooler and shows it is well designed.

Regards

Phil

[Weaver]

> The thing is, the voltage supplied from the power supply (typically 12 volt DC for most routers/modems) goes through several even nosier switch mode DC-DC converters inside the device anyway.

A killer point. Maybe the poster in the other thread referred to earlier was not using a sound test method in some way?

[Weaver]

@ejs -  good point. No, I have not tested other kit since the exchange upgrade to MSAN. I did look into getting a HG612, but no luck there. This was discussed in another thread where Burakkucat gave an enormous amount of generous help, but for various reasons the project was not progressed. ADSL2 gave me an extremely welcome performance boost of something like ~20%, but unfortunately since January a lot of the performance gain has been lost, for reasons not well understood.

I would definitely welcome suggestions for the latest ADSL2 PPPoE modems as I'm short of ideas.

[Weaver]

A thought :- There remains ejs’ point about RFI, is that independent of PhilipD’s objections!

[ejs]

Generally all correctly functioning power supplies will not give rise to much difference

I think the "not much difference" will be exactly the size of the difference people will find by scrutinising their line stats looking for any small differences.

The difference may, of course, depend on various factors such as the proximity of the PSU to the modem and phone line.

About the RFI emitted by any PSU, it's the kind of thing where if you wave an AM radio very close to any electronic device, you'll find it emits some noise. Maybe not lots, but it's still some, and one PSU may emit slightly more or less than another. Perhaps the PSU could be partly enclosed in a metal box or wrapped in metal foil to shield it.

If you don't already have a bunch of different PSUs to try, it's probably not worth buying one or more just to compare them, any difference will be pretty small.

[Weaver]

> Perhaps the PSU could be partly enclosed in a metal box or wrapped in metal foil to shield it.

Good point. I had considered that, but in reverse, thinking about putting the modem in a Faraday cage. But coming back to shielding the psu, would I also need to ground the shielding?

[ejs]

I wouldn't bother. If not grounded, it'll be like a shielded DSL cable being plugged into plastic sockets at each end.

[S.Stephenson]

Would BT let me isolate my line in a faraday cage?   

[Chunkers]

[Apologies if we've talked about this subject before some while back, although from a different angle]

Could anyone find a spec for the DSL-320B-Z1’s power supply requirements, incl current budget?

This is the one you get with it :



Screenshot from
this video on Youtube

I'm not a sparky, but as far as I know wall-wart type PSU's typically use bridge rectifers to convert DC (presumably with smoothing circuitry e.g. capacitors) which inevitably leads to lumpy DC to some extent depending on the quality.  If I was looking for a quality PSU for test purposes I would probably try and find a good bench power supply.  For this kind of stuff I love watching bigclive videos, brilliant.

There are bench power supplies on ebay from China for much cheapness, the trick is finding a good quality one - some of the them are probably OK.

Conscious that there are probably electrically competent people on here, feel free to shoot me down in flames

Cheers big ears

Chunks

[Weaver]

Good point about bench supplies. Perhaps too thirsty for my ups, which is struggling horribly, and needs replacing.

Thanks indeed for the photo. So much for the 9V theory, which just came from the advert mentioned earlier for a supposedly compatible unit. Don't believe ads.