Jump to content

Gavin1977

Members
  • Content Count

    57
  • Joined

  • Last visited

About Gavin1977

  • Rank
    Freshman Member

Recent Profile Visitors

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

  1. Well just following SGM it’d have to be 6 DIMMs per CPU. Memory bandwidth must therefore be important. Power it all with a Farad super12?
  2. Wow - what’s going on with the memory though, why so many banks?
  3. Farad Super12 should be a really welcome addition at 999 euro ex VAT and is something I'm watching since it's performance is reported to be close to that of an SR7. A single SR7 rail for a NUC is obviously rare, very expensive and the cost of an even higher ampage version deters from these higher end configurations from being used. Having said that I'm still not sure if the difference in sound quality would be worth it. One thing I keep thinking about is if someone had the time to experiment and report back on disabling CPU cores on these processors and running them lean with lower amperage - how low can there power draw be taken? I also think it'd be great if there was some form of current sensing ATX DC-DC converter that switched from linear power to a standard SMPS once say 2-3A had been surpassed - that'd be great and best of both worlds.
  4. I know - really gutted about this, but could be modded I guess 🙂
  5. Thanks Seeteeyou - so what is more important ECC, Wide temperature or Industrial grade?
  6. Appologies - I might have got this wrong, I thought the source of this was earlier in this thread. Anyway - the question still stands.
  7. Given that ECC is unsupported in the PJYH NUC (Pentium J5005 officieally doesn't support ECC) yet is still reported to work, has anyone tried ECC in DNHE NUC? I am wondering which is of most benefit - ECC, Wide Temperature or both?
  8. Both have the same specs and just differ in capacity. CAS latency of 15 is nothing special, 12 or 10 would be better if available. I’m not familiar with EEC memory, but maybe CAS latency is slightly higher with EEC. Have a look and see what you can find.
  9. Hello All! OK I have completed my VRM comparison. The results suprised me somewhat, but perhaps my findings shouldn't have been such a suprise after alll.... read on. So, I was waiting for a Gigabyte Z87 XPower motherboard with 32 Phase VRM to be delivered. Much to my disappointment the eBay vendor sent me the wrong motherboard. So anyway, my testing was now done on a Z87 Mpower motherboard with 16 Phase VRM, vs my older 6 Phase MSI ITX motherboard. Both motherboards use the same capacitors and ferrite chokes. Previously I had ranked some of my earlier system configurations here. My system following the server is Naim DAC V1, Ncore NC500 Monoblocks and IPL Transmission line speakers which remained unchanged during the testing. Playback is direct from the server via USB to the Naim V1 (i.e. not a server/endpoint config). I do have a separate Synology NAS with a load of music on it elsewhere, but to my ears it sounds noticeable poorer than playback via Tidal using Euphony Stylus. I use WifFi, not Cat5. So, subjectively, how would I rank now (ranked according to software used, as I find this to be one of the largest determiners): BEST Euphony (DNHE NUC with Paul Hynes SR7) . . . Euphony (MSI Gaming Z97 ITX 6 Phase VRM, 4790k, Corsair SF450 20-25mV ripple) . . Euphony (MSI Z87 MPower ATX 16 Phase VRM, 4790k, Super Flower Leadex III 550W Gold 5mV ripple) . . Euphony (DNHE NUC with stock power brick) Audiolinux RAMBOOT (DNHE NUC with stock power brick) . . . . Roon Rock or Audiolinux without RAMboot (DNHE NUC with stock power brick) . MacBook Pro, iTunes playback within MacOS WORST This pretty much surmises my journey so far. The audiable differences between each configuration are easy to discern, so this is not a subtle thing. I have stopped using AudioLinux, and favour Euphony, so I haven't and won't be doing any further testing with AudioLinux unless there is some new amazing development. That's why you see it lower down the list. So, what I have found with the VRM experiment is that a very high quality VRM - 16 Phases in this case - when combined with a low ripple power supply sounds really very detailed, I would say it is close to the DNHE / SR7 combination in terms of detail retrieval. The 16 Phase VRM on the MSI Z87 MPower is apparently able to control voltages with an accuracy of 0.001v. So high phase VRM when paired with a low ripple power supply = more detail than a 6 Phase VRM and a higher ripple power supply. This would make sense and is completely logical! However, I have lost tonal colour and dynamics/transient response compared to the Z97 ITX 6 Phase with the Corsair PSU (this is why I rank the Corsair solution higher). I have not tested, but I am pretty sure this is due to the differences in PSU design. Perhaps to achieve the low 5mV ripple the SuperFlower PSU has more voltage regulation applied in its design and this has damped it's performance as a power supply for audio purposes when compared with the Corsair. So what have I learnt. It's all about system design. Paul Hynes SR7 not only sounds good because of it's low ripple, it's also down to Paul knowing how different circuits and components sound (e.g. capacitor brand X vs capacitor brand Y). This is were it becomes a matter of experience and time, and 'perhaps' were commercial companies succeed compared to the DIY route due to the knowledge and resources available. I'm now of the presumption that it not will not be easy to match the performance of a Pink Fawn or Sound Galleries Music Server as you need a specially designed PSU (this was not my aim anyhow!). The Superflower SMPS has the lowest ripple in it's class, so I think we're pretty much maxed out before adopting something like a HDPLEX PSU. Since the combination of a lower power DNHE NUC and a SR7 achieves lashings of detail, dynamics and tonal colour I don't think it would be worth my while investigating different combinations of ATX motherboards and power supplies. I do not have the experience (e.g. capacitor brand X vs capacitor brand Y), time or resources to evaluate lots of different combinations to get the right balance of characteristics. I was lucky to get hold of an SR7 second hand, prior to this I was going to retire the NUC and use the MSI Gaming Z97 with the Corsair PSU. It really does sound great, and provided you don't mind the reduction in detail you find it very cost effective and musically satisfying. I'm going to sell the PC's now, I might try a SOtM tx-USBultra with the proceeds as my next step as I'm lucky to have a spare rail on my SR7. This would be logical as I completely trust the experience of AustinPop and he has done something similar, although I intend to limit my digital frontend to no more than two boxes :-) I hope this helps others.
  10. I’ve not tried x570, but will be trying 32 phase MSI XPower motherboard shortly based on the same logic. Will be interested in your results as well as I’ll be moving from a small 6 phase MSI Z97 Gaming ITX. The XPower has tantalum chokes which are apparently better as blocking high frequency noise then the super ferrite chokes.
  11. Where can I fine more information on this? Is it the Ripple Eater?
  12. What audiable improvement did you experience when moving to i9-9900K? What associated components also changed?
×
×
  • Create New...