Popular Post BrownMagic Posted May 10, 2021 Popular Post Share Posted May 10, 2021 Thanks a lot @GoldenOne! I hope the Holo Spring 3 gets the PLL. If it does then I guess it will become a formidable choice for those who cant step up to the May but can get a Spring 3 + RPI to get a good digital setup! Exocer, Jean Paul D and 87mpi 2 1 Link to comment
scintilla Posted May 10, 2021 Share Posted May 10, 2021 So, that new 180g Kerosene Hat got a play tonight (actually still playing). I had to get a bit tricky to try this a new way TASCAM 192k to macbook air for Purevinyl RIAA, then out to the TASCAM and into the other TASCAM AES/EBU and into iMac and HQP for DSD 512. I am applying my RC in HQP convolution instead of Purevinyl in this setup and running up to 24Mhz. I adjusted my gains in Purevinyl and have plenty of overhead now to avoid digital overs while staying well below -3b in HQP. Finally enough gain for my DL-S1. A little hum, but theres 60db of gain on the preamp. This sounds most excellent. Lots of depth and transient snap. Definitely digging the May. Link to comment
mrkoven Posted May 10, 2021 Share Posted May 10, 2021 so with such an effective PLL, laptop/motherboard USB into May is all you really need? is the PLL so good that it negates the benefit of any ddc/reclockers/usb decrapifying solutions? if so thats good news for our wallets Link to comment
scintilla Posted May 10, 2021 Share Posted May 10, 2021 I suspect it is. Prior measurements of the PLL demonstrate its effectiveness in most input settings, including a computer usb bus. Link to comment
Popular Post GoldenOne Posted May 10, 2021 Popular Post Share Posted May 10, 2021 4 minutes ago, mrkoven said: so with such an effective PLL, laptop/motherboard USB into May is all you really need? is the PLL so good that it negates the benefit of any ddc/reclockers/usb decrapifying solutions? if so thats good news for our wallets The pll addresses jitter. The other concern would be electrical noise which is a genuine concern especially if the source is something like a beefy gaming pc. It's a horrible device to have in an audio chain. Most USB cleanup devices will filter the 5v line to reduce noise, and some might disconnect the ground connection as well in order to prevent ground loops. Some others, like the ifi Igalvanic or using a streamer, will offer full galvanic isolation. Completely disconnecting the pc and ensuring there is no direct electrical connection. But.... The may takes care of this too. It has galvanic isolation built in. You can actually see this on the USB module pcb. The part on the bottom left that actually connects to the dac is physically gapped from the rest of the USB circuitry and connected via optical isolators. So yep, basically any source should work with the may and be absolutely ideal. It truly is a remarkably and wonderfully over-engineered product. Hiker, Gavin1977 and mrkoven 1 1 1 https://youtube.com/goldensound Roon -> HQPlayer -> SMS200 Ultra/SPS500 -> Holo Audio May (Wildism Edition) -> Holo Audio Serene (Wildism Edition) -> Benchmark AHB2 -> Hifiman Susvara Link to comment
BrownMagic Posted May 10, 2021 Share Posted May 10, 2021 4 hours ago, barrows said: also, are you avoiding stacking the May DAC directly on top of the power supply? I would definitely separate them if so, as the filed strength of most transformers is going to be enough in the vertical to potentially create interference with the DAC via magnetic coupling in the vertical plane. @GoldenOne: What is your take on this? Considering the May is over engineered, my guess is that the designer would have thought about this two before offering a two chasis design and it is not logistically convenient if we cannot stack these too. Have you made any observations of interference during your tests? Thanks Link to comment
lpost Posted May 10, 2021 Share Posted May 10, 2021 'So yep, basically any source should work with the may and be absolutely ideal. It truly is a remarkably and wonderfully over-engineered product. ' I suppose this explains why I hear zero difference when changing power supplies that feed my UP gateway NAA computer (Intel Cherry Trail SoC) that connects to the May via USB. I have my May and PS separated only because I built the shelves and could easily separate them some. The DC cable is generously long allowing for this so why not if able. It certainly cannot hurt to have separation though it also may not matter/help anything audible and at the end of the day audible is all that really matters (to me). Link to comment
Extreme_Boky Posted May 10, 2021 Share Posted May 10, 2021 I have May DAC L2 and Aleph J with no AC coupling (sound-coupling/DC blocking) capacitors - apart from the tweeter X-over capacitor inside speaker cabinets. Balanced silver ribbon interconnects, silver alloy speaker cables, solid core gold plated pure copper power cables. Very revealing. The best sound I can achieve is with native sampling rates (44.1/88.2/96/192 kHz). No oversampling in software, nor hardware (inside May). This gives me completely natural sound, relaxed presentation and depth perception (layering) that is simply beautiful. The native 44.1kHz material, in NOS, should be a no-no... but it sounds amazing with May. I also prefer to underclock the CPU (minimum frequency AND voltage), with a skeleton number of background OS processes, and I run my laptop on battery power <- all of which contribute by a large amount to this natural 3D presentation. I know many will disagree. I also have NUC11 and HQ Player, to impress friends (especially when May starts showing that 1.536Mhz number!). But people who play natural instruments/musicians in particular, easily go for that battery powered laptop sound, with starved CPU and native, non-oversampled material. I am not sure if anyone else tried what I'm suggesting here..? It would be good to see what others think... Hiker 1 Link to comment
John Hughes Posted May 10, 2021 Share Posted May 10, 2021 12 minutes ago, Extreme_Boky said: I would agree with you 100%, though I prefer to use an high quality endpoint as an additional firewall from the noise in the PC. 12 minutes ago, Extreme_Boky said: I have May DAC L2 and Aleph J with no AC coupling (sound-coupling/DC blocking) capacitors - apart from the tweeter X-over capacitor inside speaker cabinets. Balanced silver ribbon interconnects, silver alloy speaker cables, solid core gold plated pure copper power cables. Very revealing. The best sound I can achieve is with native sampling rates (44.1/88.2/96/192 kHz). No oversampling in software, nor hardware (inside May). This gives me completely natural sound, relaxed presentation and depth perception (layering) that is simply beautiful. The native 44.1kHz material, in NOS, should be a no-no... but it sounds amazing with May. I also prefer to underclock the CPU (minimum frequency AND voltage), with a skeleton number of background OS processes, and I run my laptop on battery power <- all of which contribute by a large amount to this natural 3D presentation. I know many will disagree. I also have NUC11 and HQ Player, to impress friends (especially when May starts showing that 1.536Mhz number!). But people who play natural instruments/musicians in particular, easily go for that battery powered laptop sound, with starved CPU and native, non-oversampled material. I am not sure if anyone else tried what I'm suggesting here..? It would be good to see what others think... Link to comment
happybob Posted May 10, 2021 Share Posted May 10, 2021 3 hours ago, GoldenOne said: PLL applies to the digital inputs. It's active on all inputs including I2S. So with PLL on, no matter which input/source you use, it should end up basically "perfect". To be clear though. There are good and bad PLLs. Lots of other DACs will have a PLL which will attenuate jitter a little bit. But it's because the may's PLL is so ridiculously good that you can use any source and have it come out basically perfect. It's my understanding that usually an asynchronous USB transfer will not use a PLL, rather it receives the input with the transmitted data rate / clock and then sends it onward (into the DAC in this case) with a completely separate clock that's in the DAC. Am I missing something? Why does the May have a PLL at all if it can completely isolate the input and output clocks for the USB stage? opus101 1 Link to comment
John Hughes Posted May 10, 2021 Share Posted May 10, 2021 13 minutes ago, happybob said: It's my understanding that usually an asynchronous USB transfer will not use a PLL, rather it receives the input with the transmitted data rate / clock and then sends it onward (into the DAC in this case) with a completely separate clock that's in the DAC. Am I missing something? Why does the May have a PLL at all if it can completely isolate the input and output clocks for the USB stage? It has been established by the manufacturer earlier in this thread that the May's USB input does use PLL. Another way the May is pretty unique. Link to comment
happybob Posted May 10, 2021 Share Posted May 10, 2021 3 minutes ago, John Hughes said: It has been established by the manufacturer earlier in this thread that the May's USB input does use PLL. Another way the May is pretty unique. Thanks for clarifying! Was there a reason given as to why a PLL is used in the May? I'd rather not slog through 40 pages of post responses to find it if it's even stated. Thanks! Link to comment
opus101 Posted May 10, 2021 Share Posted May 10, 2021 15 minutes ago, John Hughes said: It has been established by the manufacturer earlier in this thread that the May's USB input does use PLL. Another way the May is pretty unique. Did the manufacturer say whether the USB transfer mode was isochronous or adaptive? Only those two modes require a PLL, asynchronous does not. Link to comment
GoldenOne Posted May 10, 2021 Share Posted May 10, 2021 Tim from kitsune said that the pll was active when using the USB. I'm not entirely sure if that's the case though for two reasons: 1) the jitter performance doesn't change at all regardless of whether pll is on or off 2) there is 0 locking time with usb even with pll on https://youtube.com/goldensound Roon -> HQPlayer -> SMS200 Ultra/SPS500 -> Holo Audio May (Wildism Edition) -> Holo Audio Serene (Wildism Edition) -> Benchmark AHB2 -> Hifiman Susvara Link to comment
The Computer Audiophile Posted May 10, 2021 Share Posted May 10, 2021 Hi @GoldenOne It would be really cool to keep measurements of this DAC over time. The big issue that manufacturers, who don't offer R2R DACs, bring up is that R2R DACs can lose accuracy over time because of all the resistors. I have no idea how big of a deal this is, how much time we are talking about, or if this is still an issue. BrownMagic 1 Founder of Audiophile Style | My Audio Systems Link to comment
GoldenOne Posted May 10, 2021 Share Posted May 10, 2021 4 minutes ago, The Computer Audiophile said: Hi @GoldenOne It would be really cool to keep measurements of this DAC over time. The big issue that manufacturers, who don't offer R2R DACs, bring up is that R2R DACs can lose accuracy over time because of all the resistors. I have no idea how big of a deal this is, how much time we are talking about, or if this is still an issue. I'd be happy to do that. Fwiw this dac is over a year old and measures exactly the same as when I got it. Im not sure why the performance would change over time though cause afaik resistors don't degrade over time? (unless you're running it above its rated power spec which wouldn't be the case here cause everything is line level at most) The only thing I'd be concerned about degrading I think would be the capacitors but then that's the same for just about any product regardless of design. https://youtube.com/goldensound Roon -> HQPlayer -> SMS200 Ultra/SPS500 -> Holo Audio May (Wildism Edition) -> Holo Audio Serene (Wildism Edition) -> Benchmark AHB2 -> Hifiman Susvara Link to comment
GoldenOne Posted May 10, 2021 Share Posted May 10, 2021 Also just thought I'd post here in regards to USB sources for 1.536mhz for the may. I was previously using an SMS200 Ultra, but couldn't do 1.536mhz from HQPlayer with it. You have to use an intel device in order to reliably get 1.536mhz working. (Some other PCIe cards work on and off but intel is the only thing that works reliably). This limits you to not massively ideal devices in terms of your source. I bought a NUC as I could then run a roon core on it, and also use it as a source for the may. BUT, if you're wanting a really good USB source that DOES do 1.536mhz. Get an intel machine (cheapest options are the intel compute stick or one of the mini-PC options on amazon for ~£100 with the atom x5 cpu) and an ifi iGalvanic 3.0. This provides full galvanic isolation, an output noise floor of <5uV and also does reclocking etc. It's 'transparent' from a digital perspective. Meaning the system does not see it as a USB hub or controller. It looks to the PC like a cable pretty much. And I can confirm that 1.536mhz is working flawlessly with NUC -> ifi iGalvanic -> May https://youtube.com/goldensound Roon -> HQPlayer -> SMS200 Ultra/SPS500 -> Holo Audio May (Wildism Edition) -> Holo Audio Serene (Wildism Edition) -> Benchmark AHB2 -> Hifiman Susvara Link to comment
scintilla Posted May 10, 2021 Share Posted May 10, 2021 In a fixed-resistor network, I could see some drift over time. However, here the design affords two different ways to combat that: 1) dynamic compensation and 2) dual-mono, differential networks. Jeff uses the two best way to assure constant linearity and lowest noise floor. Resistor drift is unlikely here but compensation is going to ensure that it isn't going to degrade. As to the PLL, it could be employed between the USB receiver implementation and the resistor network rather than to corral in the data at the input receiver that is using an asynchronous isochronous protocol. We don't have a lot of details on the Titanis code, but since it is proprietary it is possible Jeff uses the USB receiver as a way-station for the data and locks the data output with the PLL for the highest quality clocked-out data. Link to comment
ted_b Posted May 10, 2021 Share Posted May 10, 2021 On 5/4/2021 at 8:56 AM, GoldenOne said: Honestly the may usb is so good, and has galvanic isolation built in anyway, that I've found that many usb cleanup options make no difference to it. (subjectively or objectively) Even though they do make a difference on other dacs. Golden One, you said the above quote last week, but then today said (in a great post about USB sources doing 1.5Mhz PCM) 13 minutes ago, GoldenOne said: BUT, if you're wanting a really good USB source that DOES do 1.536mhz. Get an intel machine (cheapest options are the intel compute stick or one of the mini-PC options on amazon for ~£100 with the atom x5 cpu) and an ifi iGalvanic 3.0. This provides full galvanic isolation, an output noise floor of <5uV and also does reclocking etc. What is the purpose of the iGalvanic if the May USB is already there? Also, I assume those intel compute sticks or Intel NUCs would run Linux NAA as opposed to Windows NAA (I ask cuz I found Windows has high frequency hash at 32fs and 20 bit; 18 bit seems to tame it). "We're all bozos on this bus"....F.T. My JRIver tutorial videos Actual JRIver tutorial MP4 video links My eleven yr old SACD Ripping Guide for PS3 (needs updating but still works) US Technical Advisor, NativeDSD.com Link to comment
GoldenOne Posted May 10, 2021 Share Posted May 10, 2021 6 minutes ago, ted_b said: Golden One, you said the above quote last week, but then today said (in a great post about USB sources doing 1.5Mhz PCM) What is the purpose of the iGalvanic if the May USB is already there? Also, I assume those intel compute sticks or Intel NUCs would run Linux NAA as opposed to Windows NAA (I ask cuz I found Windows has high frequency hash at 32fs and 20 bit; 18 bit seems to tame it). Honestly, nothing but nervosa :P I don't think there is a difference but I wasn't using the Igalvanic anyway and figured some people may be looking for a 'just in case' solution. The nuc/compute stick can run whatever you like in terms of os. If you're getting high frequency hash though that might be your machine. I have that on my Amd pc but the Intel machine is perfect. (also double check you've got the latest drivers installed, are using asio, and the windows volume isn't anything below 100%) https://youtube.com/goldensound Roon -> HQPlayer -> SMS200 Ultra/SPS500 -> Holo Audio May (Wildism Edition) -> Holo Audio Serene (Wildism Edition) -> Benchmark AHB2 -> Hifiman Susvara Link to comment
ted_b Posted May 10, 2021 Share Posted May 10, 2021 Thx. Your description of your new NUC sounds like you are using it as BOTH a roon core and an NAA. Is this at the same time (NUC -> HQPlayer server -> NUC -> May), or can be used either (the same time thing has me baffled :) ). ? "We're all bozos on this bus"....F.T. My JRIver tutorial videos Actual JRIver tutorial MP4 video links My eleven yr old SACD Ripping Guide for PS3 (needs updating but still works) US Technical Advisor, NativeDSD.com Link to comment
Toni-Mang Posted May 10, 2021 Share Posted May 10, 2021 That is a good discussion, maybe a DDC or an islolator can add an additional noise performance to the Titanis section...(the section befor the isolated handover to the Dac)...as mentioned above PLL should handle the jitter... In my case the DI20HE sounded different compared to a direct connection via USB. I am using a noisy and old Laptop with Jriver. The direct USB connection is wider in soundstage, but has not the dephts as the di20he via I2S...thats why i think about a galvanic isolator to compare...maybe the DI is obsolete than... Papas Gear... Link to comment
Quadman Posted May 10, 2021 Share Posted May 10, 2021 23 minutes ago, ted_b said: Windows has high frequency hash at 32fs and 20 bit; 18 bit seems to tame it). Ted, The PGGB website under dacs has this info... It appears they recommend 16 bit at 32fs. Holo Audio May DAC (16FS/20bits** or 32FS/16bits***) Link to comment
Popular Post barrows Posted May 10, 2021 Popular Post Share Posted May 10, 2021 A couple of things to understand in general regarding the use of PLLs and DPLLs. The use of a PLL to reduce incoming jitter is nothing new, as PLLs and DPLLS have been used for a long time to reduce jitter in DACs. Additionally, there is nothing unusual about having a PLL circuit be active for USB inputs, this is common as well. For example, the ESS chips have built in DPLLs (a DPLL is a dual PLL, which operates at a wider bandwidth than a single PLL) and this is what ESS refers to as part of its "jitter eliminator" circuitry. What a PLL does is it adjusts the rate of the DAC clock, slowly, over the long term, to synchronize with the incoming data rate. By adjusting the clock rate over the long term, it reduces the short term jitter (the audible jitter). PLLs have good and bad points, and how well they work depends entirely on their implementation. The May measurements clearly show that its PLL is well implemented, and should be engaged to reduce the jitter of sources with their own clock (SPDIF/AES and I2S). If you have an incoming data stream with no considerable amount of jitter, as would be the case with a very good asynchronous USB receiver, the PLL can be "on" but it will actually, basically, be doing nothing, as there is no incoming jitter for it to reduce (the DAC clock and the USB clock are the same and are synchronous). As the USB interface in the Holo May DAC appears to be quite well implemented, with a good master clock source (same as the DAC clock), there should be no difference when using the USB interface with the PLL "on" or "off", as the clock source is the same in both cases and is already synchronous. 87mpi and scintilla 2 SO/ROON/HQPe: DSD 512-Sonore opticalModuleDeluxe-Signature Rendu optical with Well Tempered Clock--DIY DSC-2 DAC with SC Pure Clock--DIY Purifi Amplifier-Focus Audio FS888 speakers-JL E 112 sub-Nordost Tyr USB, DIY EventHorizon AC cables, Iconoclast XLR & speaker cables, Synergistic Purple Fuses, Spacetime system clarifiers. ISOAcoustics Oreas footers. SONORE computer audio | opticalRendu | ultraRendu | microRendu | Signature Rendu SE | Accessories | Software | Link to comment
barrows Posted May 10, 2021 Share Posted May 10, 2021 23 minutes ago, GoldenOne said: Honestly, nothing but nervosa :P I don't think there is a difference but I wasn't using the Igalvanic anyway and figured some people may be looking for a 'just in case' solution. I would suggest that things are not quite so absolute as this sounds. "Isolation" like this is effective at reducing noise getting across the barrier, and is used in most good DAC's USB interfaces, and many very good USB sources as well. But, this isolation is not 100 percent effective, there will always be some noise which makes it across the "moat" via capacitive coupling-this is likely why many persons using DAC's with well "isolated" USB inputs still hear some improvements with low noise USB sources. Each level of isolation will still make for some additional (although perhaps minute) degree of noise reduction. What is also important to remember, is that processor noise sources like an XMOS chip, or the upstream computer processor, produce noise at very high frequency-which means that some of that noise is going to be airborne, and able to cross galvanic isolation and interfere with DAC circuitry to the other side. It is still going to be a good idea to use low noise sources with DAC's which feature "isolated" USB inputs. SO/ROON/HQPe: DSD 512-Sonore opticalModuleDeluxe-Signature Rendu optical with Well Tempered Clock--DIY DSC-2 DAC with SC Pure Clock--DIY Purifi Amplifier-Focus Audio FS888 speakers-JL E 112 sub-Nordost Tyr USB, DIY EventHorizon AC cables, Iconoclast XLR & speaker cables, Synergistic Purple Fuses, Spacetime system clarifiers. ISOAcoustics Oreas footers. SONORE computer audio | opticalRendu | ultraRendu | microRendu | Signature Rendu SE | Accessories | Software | Link to comment
Recommended Posts
Create an account or sign in to comment
You need to be a member in order to leave a comment
Create an account
Sign up for a new account in our community. It's easy!
Register a new accountSign in
Already have an account? Sign in here.
Sign In Now