Jump to content


  • Content Count

  • Joined

  • Last visited

Community Reputation

1,209 Excellent

About rmpfyf

  • Rank
    1000+ Post Club

Profile Fields

  • Location
  • Country

Recent Profile Visitors

3,509 profile views
  1. rmpfyf

    DSP options

    @greggy the maths involved is broadly the same whichever program you use. Some compute a response function for your room on different considerations, some allow you to play with these but fundamentally the maths isn't broadly different. Denis Sbragion's DRC webpage has an excellent, if extremely technical overview of what's involved. What you chose to implement this much over is up to you and what best suits your rig. Some MiniDSP will work up to certain frequencies, there are solutions that will do 384kHz if you ask nicely and you can use a PC with the right software to do just about anything on a sliding scale of price vs difficulty to implement. I would be wary of any box that involves analogue conversion, do it all digital then let your DAC do it's thing. Be aware that digital room correction shouldn't be used to cure all ills. Some room modes can be particularly severe and whilst you might generate a response function that counters as much, you might similarly be asking a great deal of your system to replicate the output - in some cases you can even clip. Are you after a 'box that works' or something where you get under the hood? Do you want to pay people to calibrate it for you? What's your budget? Where do you want to put it in your signal chain? If you want to experience what correction could or should do I'd do the following: Put out a help request asking for someone to convolve some audio for you Record some white noise or sound sweeps using a calibrated mic (this you'll need whatever you do) Send the recording to someone that will help, along with a favourite test track Have them send back a linearised 'corrected' version of the track given a response function generated from the reference sound sample This will give you an idea of what's possible at minimum cost. You might even be able to borrow the microphone test kit. Then if you like what you hear, step into it. Hope this helps.
  2. rmpfyf

    Great system cooling fans

    AC Infinity are a gold standard. TOTL no question. Silenx a bit quieter (splitting hairs) though they don't package quite as nicely. That AC Infinity kit is super nice.
  3. rmpfyf

    Great system cooling fans

    http://www.silenx.com/index.asp These are amazing. Used them in a few builds though need to get them overseas.
  4. 125dbmonster (no longer on the forums) has posted about this extensively, having built systems to suit. There are other examples globally usually using low voltage inverters, Victron and the like. I used to have a tube buffer that ran just like this (Gary Dodd design). Voltage sensitivity is important here, remember a DC system unless having something to regulate voltage in the middle (which usually increases noise) will drop voltage across battery SoC.
  5. I think you'll be right. I have a Thor DRM95, it's a nice unit, Australian made, though with the mains supply better sorted I think there's a solid chance end up sounding better without it...
  6. Agreed Would note that proper power measuring equipment is super expensive. $20k+. Ears are much cheaper. And if you don't hear any difference when filtered or not PV'd, ignorance or bliss or both are cheaper still My rig sounds best at night though I'd always put that down to voltage rise during the day, which used to be largely because of everyone else's PV system before I went ahead and got my own. A/B would be very hard as even when my system is off we still have sometimes mains at 255VAC, which some of my kit doesn't like so much. This was before an improvement in mains supply by the distributor wherein we'd drop below 190VAC at times (eep). And to complicate things further we're on an old mains which I'm about to upgrade with nice new thick-wire earthing, and I'm going to bet it'll sound somewhat different accordingly. Again. I have a filter inline that's great though it changes impedance somewhat, and I'm going to have some sockets installed that bypass the filter entirely to see what's what. This should be amusing. In PV-landia it gets interesting with big iron inverters, transformerless inverters, new HD-wave style inverters (SolarEdge I don't think is the only firm pushing these) and microinverters (though do a small army of inverters all sync'd together 'sound' any different to one big one? Hmmm). Give me a month or two and I might have some data for one of the cases, pending home owner permission.
  7. It's an interesting discussion. I was once picked on - correctly as it turns out - for suggesting that if it wasn't a mains disturbance I could hear then it wasn't worth hearing. The patient person at the other end pointed out that particularly high frequency line noise, which not at frequencies I (or a dolphin) could hear, did have a measurable effect on the electrical qualities of components in an amplifier. This was at a company designing electric car bits, and there was a test bench, and so the person in question proceeded to demonstrate just as much successfully. In that particular context we were designing filters and other bits that would ensure our components, being downstream of other things capable of generating noise amidst their core functions, were able to at a minimum function irrespective of what the upstream stuff was reasonably doing even when operating in a failure mode of sorts. I would think the same analogies apply to an audio amplifier. Years before that I was a (younger) engineer working in aeroacoustics and we had, in our then lab, a binaural head. These are fantastic devices - literally shaped like a human head and upper torso with some (horrifically) expensive microphones and electronics to simulate how we hear what we do. And you can imagine these things get stuck in cars and test everything from 'does the audio system sound as it should' to 'bloody hell that A-pillar wind noise is annoying'. One does end up computing a lot of cross-spectral phase measurements, more simply 'does this sound hit my ears at the same time or am I hearing some sort of difference in time alignment'. I wasn't into audio back then the same way I am now, or I'd have been one of the engineers in the team that kept nicking the binaural head to calibrate their audio systems at home though when I hear people remonstrate over 'the sound stage falls flat' between this and some other work around DSP I've been involved in these are the closest analogies I can find that piece together some semblance of logic; the notion that single-channel frequency spectra might look amenable, but small differences (and they only need to be small) introduced by any random function affecting multichannel output really can make a significant difference in a perceived sound stage. A lot of my work is in timing audio out of PCs (because it's cheap to do, and because I'm inherently cheap ) and measurements of jitter and feelings of 'this is practically holographic, the singer is right here with me' correlate very well. Accordingly I go a different way on the notion that if something simple can't measure it, it's not there. We see some interesting attempts to simplify measurements in this space that miss a lot - Ethan Winer's null tester is a good example (e.g. a net difference in leads below audible SPL does not equal effectively zero difference at audible SPL). To be fair, much snake oil too. In any instance the notion of minimum performance specifications hold true, as do 'perfect conditions' and a performance differential therein. It really depends what's being chased. Yes, some components our there have some amazing power stages... you pay for them too! Equally true some houses have some horrific mains power quality. As it is true that most pro recording studios engage power systems consultancies to get the very best mains reticulation they can afford. Balanced isolation transformers, the lot are the norm. By the same token I wouldn't rush to call aby recording made plugged into mains 'unlistenable' in any intrinsic way; we're chasing definitions of degree here relative to circumstance. I'm particularly interested in this topic as we're starting to see some spread in the market as to how solar inverters work, and there's potential for different solutions here with differing effects. And to you. That suggestion is sometimes valid, as is the one not to go there
  8. Some big assumptions here of the 'if you can hear it, it should be able to be recorded and played back' variety. You'd need perfect clocking, for one, and binaural acoustic replication one of which doesn't exist and the other is hellishly expensive. So no, a 'reasonable quality portable audio recorder' doesn't cut it, unless the error from ideal in the results you're seeking to acquire are statistically greater than the compound of those of both the recording device and any playback system used to replicate them. It's not as if power engineers characterise power quality issues by tasting cake batter made with a blender connected to supposedly errant mains - this is a fair analogy for what you're suggesting. More likely is sampling at considerably higher frequencies at the amplifier output terminals and characterising that data. You'd need a surprisingly high frequency to do so, just as modern laboratory-grade power analysers do, which is practically what you're looking for. Much easier to sit in your listening positing and flip your solar as you listen. A few slow V pp isn't the entirety problem, it's radiated noise, switching speed, etc. If it were a few V then everything in a house would be fine with ripple control because it's code-compliant, right? (Everything isn't). Sure (though I'd argue they're not ignorant of the variety in mains quality rather than solar PV power conversion specifically, I'm not sure designer x in Taiwan is looking at solar uptake charts in Australia and scratching their chin accordingly). Most are selling or advocating you-beaut power supplies accordingly.
  9. Actually you'd be really surprised. The power systems consultancies engaged to manage power in e.g. medical and laboratory environments are particularly specialist, and focus both on mains reticulation system design for beyond-standard levels of power quality, and then apply an entire industry's worth of solutions to cut down on noise through most any means (whether a direct or indirect function of application). There is almost zero assumption that 'Standards compliant' mains is good enough. (...Thankful to report that the particular consultancy I engage most is run by an audiophile...) A bunch of that stuff occasionally pops up in audiophile-landia. I wouldn't call it fault, though there's room for that too (cue my mains limiting at 255VAC the last three weeks). In many cases it's just that mains can be improved yielding a relative improvement in audio device output beyond what minimum performance specification is guaranteed otherwise. Would think it a difficult thing to record given the recording devices would usually be on the same mains Though listening with/without is usually good enough, no? I don't mean to get into arguments of 'its unmeasurable' vs 'yes it is' - we're likely talking very high frequency components here, which is lab-grade equipment on sampling frequency (not necessarily accuracy). A 512 sample/cycle power analyser (usual fitment at some utility generators) costs around $9k, will pick up the 63rd harmonic, and is not suitable for characterising switch mode noise - not fast enough. That's not to say it can't be done, more that it's lab-grade equipment. It's neither to say that it's not audible. (And please don't mention a null tester either here.) Not least because a solar PV inverter is inverting as little as a few hundred Watts and as much as 10kW off a fast-switching square wave into your mains, because you'd be surprised how little design difference there is between some inverters of differing power, and the designer of a DAC sucking down a few Watts here and there probably wasn't thinking bout the DC off your roof getting chopped into a square wave going BOOMBOOMBOOMBOOMBOOMBOOM, or the industry going broadly transformerless to save cost, or the manufacturers involved realising they could save a few bucks by playing at the upper end of AS4777's tolerances for local quality, or... ...if your perspective is a few hundred Watts of your load, the rest isn't so obvious, possibly.
  10. I get where you're going with this though pause a minute to take in what you're suggesting here. You're suggesting a device should, with regards to mains quality, provide a consistent level of performance for any 'reasonably foreseeable circumstances in which the device could be expected to be used in Australia'. Electrically speaking that's a fairly broad swathe. Our standards allow for a very, very wide variety of performance - which the grid often sees in many characteristics of mains quality, some to degrees not seen elsewhere - and you're suggesting that anything within this as an input should equal same quality output output because consumer Sales of Goods Act. A binary argument for an extremely analogue characteristic, there. Let me assure you a few things here: Whilst what you're demanding isn't physically possible, if it were possible all manufacturers of audiophile equipment would do so. So would all manufacturers of medical equipment, laboratory equipment and many other domains similarly affected. Australian Standards are not an absolute, and do not capture every factor salient to mains quality that might affect e.g. an amplifier. Whilst development processes are long, Standards are not developed with representation from all possible factors having veto rights on any issue. The technical committee driving AS 4777, for instance, isn't ultimately swayed be delegates from the hifi industry's concerns around noise getting into mains. They'll come at their work from a completely different perspective which may run counter to these interests but that is consistent with allowable performance in mains supply, even if on balance their contribution skews net performance in some dimension. Sure, and you've made this point a few times, though that's not the game we're playing here. If it sounds 'great' with your reverse-cycle aircon running on the same circuit and 'superb' when it's off, then you have a change in performance by definition of degree, not a defective product. And you would accordingly have a very hard time with your refund if all you can prove is that minimum performance was met but under some circumstances it performed better. This is why we have minimum performance specifications, and this entire forum dedicated to what while goose chases we send ourselves on to find, understand and recreate circumstances for sometimes-elusive maximum performance. Read above. In most instances you'd have zero grounds to go legal.
  11. rmpfyf

    No more riding for me

    Get well mate.
  12. 'Audiophile' isn't a fit purpose in any Australian Standard concerning mains reticulation, power conversion, inverting devices or any related activity. That my Google Home Mini runs on Standards-compliant mains supply and is capable of recreating beyond-recognisable noises via a litany of streaming services doesn't mean that the ultimate fidelity out of my audio rig should be similarly insensitive to mains quality.
  13. Sure - and why does equipment have to be tolerant? Some is, some isn’t. Just understand and engineer. The understanding bit is tricky... there’s so much to get right to secure a good level of mains quality. Just even earthing varies home to home. At a minimum best practices apply. As for what inverters actually do for audio, TBD. I mentioned a SolarEdge design initially because it’s different in how it generates mains. Different good or bad for audio? TBD. Can start with measuring mains harmonics with it on/off and talk more afterwards. Can repeat with other inverters. No, I don’t think typical harmonics will tell the full story either! But as to a measurable difference in basic inverter designs, we might get a start.
  14. I’d say listen closer. Respectfully. We’ve some of the most variant mains power quality in the western world. Yes, some equipment is more or less tolerant. Effects don’t go to zero though.
  15. They all generate hash. Relative impedance means you get it before your neighbours do. There are some interesting differences in how - whether one big square, three levels, eight levels, etc. Relevant variables are frequency and amplitude of the noise generated, then any filtering applied. Would not expect the microinverters to be any different to a big single. (Enphase’s financials are a bit worrying). You’d be surprised just how much some batteries can contribute owing to cell balancing activities. Some generate some very very aggressive noise. Filtering is important. @Assisi has Enphase, nice kit, though he also has a SolarIQ box - which among other things is a whacking great big filter for your solar geration back into your mains. Nice bit of kit where warranted. Best AC power you’ll get is from a large, mechanical synchronous generator all to yourself though to be honest the noise inherent probably rules out audiophile work. So so in essence it’s always a compromise. I have a few homes running some very high grade metering and could probably do some power analysis over the coming months - some SolarEdge, some Fronius (mine included) and something others. Would be interesting to asses mains quality - if anyone’s got a design of experiment they could suggest I’d be happy to take some data. If there’s an Enphase house out there may do similarly.