Jump to content

Ryelands

  • Posts

    46
  • Joined

  • Last visited

  • Country

    United Kingdom

2 Followers

Retained

  • Member Title
    Newbie

Recent Profile Visitors

The recent visitors block is disabled and is not being shown to other users.

  1. My hair-shirt server 'design' is so different to that being developed here that I'm not sure how helpful comparisons might be. That said, I've found that the P4 line on the various mobos I've used over the years is almost certainly the most fruitful place to start. I first experimented with a separate linear PSU for it about ten years ago. I'm currently using one of those 10,000uF 'slit foil' caps meant for fancy amps and a DIY'd noise-cancelling cable based on (OK, plagiarised from) a Shunyata leaflet. For the P24 socket, I made up some years back a paddle board with caps for each line but now use a Pico-PSU powered from one of the Seasonic's 12v outputs so the paddle-board is gathering dust. (I did it to save space but the resulting SQ improvement was very welcome.) Whatever, my main point was that the Elfidelity boards are well designed, well made, reasonably priced and IME effective. I'm currently using their PCIe, RAM and USB filters. I also recently devised a Shunyata-style cable for SATA power lines c/w in-line caps but, though it too is strikingly effective, connecting them to HDDs via Elfidelity filters gives a significant SQ pop. HTH Dave
  2. Essentially, it's just that except that it uses three different types of capacitor plus inductors, etc. It's a bit more than 'a bunch of capacitors' and IME (albeit in a slightly different context than here) very effective. Ditto BTW for Elfidelity's other products, all of which are passive devices. Hopefully, other folk will report back in due course. Meanwhile, I've compared 'a bunch of capacitors' (admittedly not of the best quality) with Elfidelity boards on a server using a Seasonic Prime Titanium ATX PSU. The Elfidelity thingies were significantly better. YMMV and all that . . . Dave
  3. With apologies for wandering off-topic, my approach to building a music server is rather different to Nenon's, relying as it does on elderly socket 1155 motherboards with slowish CPUs running a heavily slimmed-down Win-XP build with noise-suppressing cable geometries to power HDDs, the usual fancy PSUs and so on. I am - emphatically - not suggesting that my system is better (or worse . . .) than others, let alone of much interest here but back in January my curiosity was picqued by Nenon's endorsement of Apacer RAM. Several years ago, I and other members of the Audio Asylum PC forum experimented with the effect of different RAM specs on a music server's sound quality. As I recall, we pretty much agreed that, give or take, the lower the latency, the better. As socket 1155 boards use DDR3 RAM, not DDR4, it was a bit of a stretch to assume that Apacer would be better than "stock" in my setup but, as I could e-buy two 2MB sticks for about £12, why not just try them? Today, as last, I did. I should explain that my 'server' connects to a Fit-PC2 'endpoint' via a 20-metre FOC, several modified LAN switches, DIY'd LAN cables that knock spots off such audio-grade bespoke designs as I've tried and other gizmos. Whatever, the effect of switching the server RAM to Apacer on the system's SQ was dramatic - greatly improved clarity, a much more musical tone and so on. I have no idea why but Apacer RAM definitely cuts mustard here: I'm delighted to say thanks for the tip. Dave
  4. I've used the LPS-1 to power EITHER a switch (Zyxel 10/100 ES105a w upgraded Vregs, sonically better than but comparable load-wise with your Netgear) OR an FMC (TP-Link or DeLocker) but using it to power the two together was not succesful. The load as measured was IIRC within the LPS-1 spec but, for whatever reason, the units wouldn't power up. I seem to recall that switching one device on before the other did work but it was a while back. I'm now using a different approach (not appropriate to discuss here) but, though it works well enough and sounds excellent, it's not, as they say, entirely hassle-free. Using two LPS-1s was fine (and also sounded excellent) but, as a permanent solution, too expensive esp as I'm hoping to try the EtherRegen when my funds and UpTone's supplies permit. It might be worth your while experimenting with "sequential switching". That said, even though the LPS-1 has robust overload protection, I'd first use a DVM to check the max load. HTH Dave
  5. I don't dispute that fitting "cleaner" chargers does not affect the output of the LPS-1 - I can't because I've never used anything other than the supplied Meanwells on the five LPS-1s I have in daily use. (I don't think a user can reasonably or even honestly expect a manufacturer to honour a warranty claim if the device is not used as supplied.) But what I can confidently say is that improving the charging configuration of my LPS-1s definitely improved the SQ of my system. I use an isolating device based on Jon Risch two-transformer & capacitance circuit in front of each Meanwell. They're bulky and I'd love to get rid of them but, if I do remove them, SQ definitely falls. Trust me, I've tried several times. I long since connected the Meanwell's negative outputs toi safety earth as instructed to good effect. The only directly mains-powered devices in my system are the power amps - they have a dedicated mains feed, a mains-power regenerator and fancy PSUs. Everything else (CPU, USB gizmo, DAC) is "off-grid" in the sense you describe. The sound is exquisite BUT, if I by-pass the Meanwell's mains isolation, its quality definitely falls. I'm not saying the Risch circuit is mandatory (I had the bits to hand) and suspect that any decent isolating transformer would work just as well. I'm also sure we can agree that, though using a spare JS-2 output is fair game, buying the likes of a JS-2 simply to drive the Meanwells is a waste of money. I wonder of you're not doing actually yourself dowh slightly here. An LPS-1 charged by a directly mains connected Meanwell sounds excellent by any reckoning but the relatively modest effort involved in ensuring that the Meanwells (or, I suspect, the LPS-1.2 equivalents) are thoroughly isolated from the mains definitely gave me a worthwhile improvement. Am I the only user who claims this? Dave
  6. A while back, I made up two short USB cables with a toggle switch in the Gnd lines (PC > Intona > I2S board). The idea was that, once the post-boot handshake was complete, I could break the Gnd lines. SQ was excellent but, as you'd expect, I would forget to restore the connections on next power-up. In the end, I settled for low-value resistors in the wire with little, if any, SQ loss and less hassle. I forget from whom I pinched the toggle-switch idea but seem to recall getting the low-value resistor idea from an old post by JS. (I was going to call the design "The Plagiarise".) However, I've just measured my USPCB and found it has no resistance in the Gnd line. Am I missing something?
  7. If the target device is powered via the USB port but not by the device supplying the data, the switch is handy. If, as I do, you use a DIY'd adapter to connect a LiFePO4 battery (or even an LPS-1 . . .) to the port. it is possible to connect both the host's +5v AND the battery. OK, you need to be pretty dumb to do so but which of us has not done dumb things from time to time? By setting the switch to OFF, you reduce the chance of a potentially interesting error.
  8. Conran:

     

    Hope you don't mind my contacting you but, as it was on your recommendation that I bought one of those "LTE3045 x 4" boards, I'm hoping you can advise me.

     

    Do I simply connect the supplied mains SMPS to the input side and the load to the output side, power up and run, pretty much as other PSUs? The instructions are not clear - they mention "removing batteries" - and I cannot find a pic of the board in use. I've asked the supplier but cannot get a clear answer, probably because my question is badly worded. If I can use it as I hope I can, I plan to buy several of them.

     

    Best wishes

     

    Dave

     

    1. Cornan

      Cornan

      Hi Dave,

       

      No problem! Yes, you just plug the supplied SMPS into the AC wall and the DC plug to the DC socket on the board (input not labled but the output is so it should make sense). The board have a balanced charger and will charge both Li-ions and LiFePO4s. I use the latter with great success. The switch should be to the right (facing DC input) when powering and charging. Switch to the left means that everything is off (use it to save battery life).

       

      Now, the SMPS comes without the AC adapter for your country (atleast mine did). I bought this adapter for Sweden. You’ll probably need to buy one for your country. They are likely easy to find.

       

      1ADC2060-A5EC-411C-A2BD-A028FF800637.thumb.jpeg.abaf58bb4fed7f678b5e8a67d9f91ee2.jpeg

      5C61FE21-A5FD-4BE1-8AC9-8E4FE8E7B37A.thumb.jpeg.9a861cdea1878d17eb71ccca5a8923b6.jpeg

       

      Here is a couple of pictures of my board which I only charge between the music sessions. The light on the SMPS goes from red to orange to green indicating how much the batteries is charged.

      The info from the seller is that the board can be used without batteries as well, meaning that you can for example power it with a LPSU (and without batteries) as long as the input voltage is less than what the LT3045s can handle.

      I hope that helps! ? 

       

      650D8A92-E43D-4873-A4E6-0DA124C9ECB0.thumb.jpeg.5dea18e62e95691c3b43e5055109e12f.jpegBD257530-418C-46E5-BE01-50CD32F93F87.thumb.jpeg.ab0a69485fc7cf8f70dd86929d451426.jpeg6550A274-EC70-4B8B-8D20-89450A74B730.thumb.jpeg.3752271740700f6cddd05c51ade12e94.jpeg

       

      Micael

    2. Ryelands

      Ryelands

      Thank you so much for such a quick and helpful reply! Pictures and everything . . . what's an old man not to like?

       

      (I'd already got round the incompatible mains plug issue on the SMPS by soldering a mains lead to the pins and adding insulation.)

       

      It'll be a day or two before I can try this but I'll let you know how I get on.

       

      Best wishes

       

      Dave

    3. Cornan

      Cornan

      You’re welcome! I’m glad it helped! ?

       

      I’m very interested to hear your impressions. Here that board concured LPS-1.2 with 3A LT3045 at the input at my perticular spot (GI Spdif output of my BluWave USB to Spdif board).

       

      Micael

  9. It wasn't a safety recommendation, it was a slightly snide remark. One-and-a-half's post was much more helpful even if the topic in hand is the cable between a socket and a device, not in-wall wiring. Interestingly, electrically parallel though not physically proximate wiring between a panel and wall sockets has been the rule in the UK since the early 1940s. It is known as a ring mains; stranded cables are not permitted. Thanks but I know why quad-core geometry was developed, when and for what applications. That was why I wrote, "early applications being for long-run signal cables" though telephony cables would have been more accurate. It's been a couple of years since I researched quad-core mains cables - apologies if links are no longer to be found or I'm mis-remembering. (I do seem now dimly to recall that the Supra QC reference was to speaker cables.) I have no interest in doing it again though I did recently see material on VH Audio's site which IIRC discussed a Belden.QC mains cable since superceded by its own product. Can I assume you take my point about DVMs being the wrong tool for assessing power cables?
  10. I'm not your research assistent (Google's your friend here) but, to get you started, recall that the refresh rate of a DVM is typically and for obvious reasons about 2 to 5Hz, that AC mains is 50/60Hz, that the diodes in a typical PSU conduct for a small fraction of each cycle and that, absent appropriate measures, diode switching noise can extend into the KHz and even MHz range and is readily propagated back into the mains supply. Kill-a-Watt meters are excellent gadgets but not in this context. I didn't suggest it was, at least not in the geometry shown here. Quad-core configurations OTOH are a different matter. As to why, Google is again your friend. There's a wealth of material on the topic, much of it on sites hosted by the likes of Supra, Belden and Canare. I was merely pointing out that the suggestion that parallel conductors were not permitted under the Regulations (UK term) was, well, not helpful.
  11. Several reputable manufacturers (Belden, Supra, etc) sell mains cable and assembled cables featuring quad-core geometry, i.e. two pairs (L & N) of parallel cables. The benefits of the geometry have been understood since the 1930s; early applications being for long-run signal cables. In passing, anyone who thinks that a DVM or, so help me, a Kill-a-Watt meter, with their slow-by-design sample rates, is going to help understand (dismiss) the notion of why power cables can affect the performance of amplifiers needs to do a bit of homework.
  12. Though by all accounts Ghent Audio cables are well made and are certainly reasonably priced, I don't off-hand recall John commenting either favourably or unfavourably about them. Correct me if I'm wrong. What I do remember is Alex complaining about the company using the acronym JSSG in its advertising. The point seems reasonable given that the source of the acronym is not acknowledged and that Uptone sells its own cables. Dave
  13. I'd say that decent cables bring the best out of any PS but especially out of the likes of an LPS-1. There's little point in spending big money on an ultra-low output impedance circuit only to connect it to its load with an over-long lead of highish impedance. John's by-now-notorious JSSG quad-core design has repeatedly proved in my system to be superior both to conventional cables and to quad-core unscreened cables. As you say, they're not hard to DIY if you know how to use a soldering iron.( I've yet to try the "360 degree" version.) Today's hot cable tip: if the target device needs a lower voltage than the LPS-1/1.2 can provide, set the LPS-1x to a higher voltage and mount a decent Vreg (LT3045 or equivalent) at the remote end of the cable, as close as possible (closer, even) to the load. I did just that to both my WaveIO USB>I2S board and Intona isolator with excellent results. Ditto for my DAC. (H/T to GStew for making me try it.) Swallow the extra cost, call it a HyperActive Link and show it off to your friends . . .
  14. I don't have a JS2 but I did for several years use John's original C-L-C circuit (published on Audio Asylum in 2009) to power a DIY'd Network Audio Adapter. AFAIK, the 2009 circuit is essentially identical to the JS2 though the latter may have been refined. (It's certainly better built - it even comes in a case . . .) My DIY effort was much better than the conventional linear PSU it replaced and is still in use in an RIAA pre-amp. The same goes for two more builds of the circuit which for some time powered my TDA1541A DAC. But they just don't compare to using LPS-1s in the same role, as I do now. I have no hesitation in recommending you use one for your ultraRendu. The JS2 is a fine circuit that addresses the LPS-1.2's power limitations but, where they provide enough power, the LPS-1/LPS-1.2 are superior devices sonically. HTH Dave
  15. Switch-wise, the best results, at least for me, came from upgrading the 3.3v Vreg in the ES-105. The mod is effective and all but negates the differences between the versions (I've tried versions 1 & 3 but not 2). That said, it's slightly easier to fit the Vreg to the v1 board. I can confirm Greg's finding that powering a Vreg-modified DeLock with an LPS-1 sounds significantly better than using a linear PS. Until recently, I'd had to use the latter. Other LAN-related upgrades include stacking two or even three ES-105s in series though that one is probably no longer worthwhile given the imminent(ish) Uptone EtherRegen. Still well worth trying is removing the redundant pairs from the LAN cable - 100BaseT uses pins 1&2, green & 3&6, orange but not the blue or brown pairs. Though easy to do, it can make a dramatic difference. Decently-made CAT5 cables (e.g. Excel, BlueJean) are IMHO the best choice. Also, where practicable, try running the LAN at 10MB/sec, not 100MB/sec. Dave
×
×
  • Create New...