johnjen
Headphoneus Supremus
I measured the Mutec 3+ at 0.56Amps which is ≈ 1/2 of the RN3.
JJ
JJ
The DIY'rs Cookbook
- Thread starter johnjen
- Start date
johnjen
Headphoneus Supremus
LPS strikes another blow for Realness
or
If you have ANY SMPS’s in your system, you haven’t truly heard what your system is capable of.
er sumpt’n
So this wild hair-brained idea goes PoP in my head.
Make a true, dual output, completely isolated, LPS.
And not just isolated from one channel to the other, but from ground as well,
using the ground connection made at the load (Mutec 3+ in this case).
Why?
Because this way I can provide power to the desired digital device and minimize spreading its ground plane noise, using the common ac power ground plane.
Which in turn can help keep it from ‘getting into’ the analog gear,
where it doesn’t belong,
at all.
That was my thinking which led me to configure a ‘robust’, well for our measly power needs anyway, LPS with sufficiently low noise and capacity.
And I realized it costs about the same as my previous LPS installed in my RN3.
Case included, but there were a few ‘incidental’ costs associated with an external box vs internal installation.
Which in turn added ≈ $50 more for switch, fuse block, terminal block, gland fittings hardware and wire.
And this would allow me to see if this isolation technique would be of any value.
And I can at any time connect the, already in the box, ac power supplied ground.
Just to see.
So I ordered all the stuff and waited for it all to miraculously just appear upon my doorstep.
Which it did, all the way from China.
And since I wanted 2 separate ‘channels’ it means each one of the dual secondaries was going to be used by itself, so a ‘larger’ capacity transformer was needed.
So here’s the transformer.
http://www.ebay.com/itm/141826005951
pic of transformer
And since I wanted a true dual channel regulator not just 2 outputs from the same regulator,
here’s the dual channel regulator board
http://www.ebay.com/itm/182209407730
pic of reg board
The box was a bit larger than needed but not by a wasteful amount, as its extra size made it easier to ‘build out’.
So the box to house it all in.
http://www.ebay.com/itm/162202150332
pic of box finished
And here again these numbers aren’t exact but are meant for clarity and simplicity, and are ‘good enough’…
65VA seemed like it should suffice which means each channel gets 32.5VA which = 2.7Amps at 12Vac.
So at 5Vdc x 2.7Amps = 13.5Watts.
And as it turns out the power draw from my Mutec 3+ is 0.56Amps = 2.8Watts.
This leaves plenty of headroom.
But that’s getting ahead of this story.
So I built it out and took a few pics of the before, during, and after.
externl mntg
ac wiring
extrnl frnt
extrnl rear
intrnl frnt open
intrnl rear open
bottom
And of course I took a few measurements of its noise, using the same scope I’ve used previously.
And it quickly became apparent that the output from this regulator board was pushing the limits of my scope in terms of resolution.
So again here are the 3 different means of measuring the noise for the ‘left’ and ‘right’ channel outputs into a 7.5K resistor (a minuscule load).
L channel R channel
Norm 1mv 1mv
Average 0.3mv 0.2mv
Envelope 3.7mv (6.7mv with all spikes) 4.8mv (6.0mv w/spikes)
There were very low level, low duration, hi freq spikes at ≈ .18µs (5.5MHz), .118µs (8.5MHz), 0.08µs (12.5MHz)
How they got created and from where, I can only guess.
Even so, these levels of noise are below what I measured from my RN3 when it was being powered by my external LPS.
link to post http://www.head-fi.org/t/781268/the-diyrs-cookbook/780#post_13218514
And I’ll follow up with measurements while running, from both the RN3 and my Mutec 3+ while being fed by their new LPS’s, in due course.
I first wanted to determine if this course of action was going to be ‘Better’ or not and it has exceeded my expectations and in ways I never suspected, but has done so much to my delight.
My take away is the use of SMPS’s mask inner detail and obscure meaningful musical information, which reduces our involvement in the music.
At least that is what I have come to understand as I hear these results in my playback system.
JJ
or
If you have ANY SMPS’s in your system, you haven’t truly heard what your system is capable of.
er sumpt’n
So this wild hair-brained idea goes PoP in my head.
Make a true, dual output, completely isolated, LPS.
And not just isolated from one channel to the other, but from ground as well,
using the ground connection made at the load (Mutec 3+ in this case).
Why?
Because this way I can provide power to the desired digital device and minimize spreading its ground plane noise, using the common ac power ground plane.
Which in turn can help keep it from ‘getting into’ the analog gear,
where it doesn’t belong,
at all.
That was my thinking which led me to configure a ‘robust’, well for our measly power needs anyway, LPS with sufficiently low noise and capacity.
And I realized it costs about the same as my previous LPS installed in my RN3.
Case included, but there were a few ‘incidental’ costs associated with an external box vs internal installation.
Which in turn added ≈ $50 more for switch, fuse block, terminal block, gland fittings hardware and wire.
And this would allow me to see if this isolation technique would be of any value.
And I can at any time connect the, already in the box, ac power supplied ground.
Just to see.
So I ordered all the stuff and waited for it all to miraculously just appear upon my doorstep.
Which it did, all the way from China.
And since I wanted 2 separate ‘channels’ it means each one of the dual secondaries was going to be used by itself, so a ‘larger’ capacity transformer was needed.
So here’s the transformer.
http://www.ebay.com/itm/141826005951
pic of transformer
And since I wanted a true dual channel regulator not just 2 outputs from the same regulator,
here’s the dual channel regulator board
http://www.ebay.com/itm/182209407730
pic of reg board
The box was a bit larger than needed but not by a wasteful amount, as its extra size made it easier to ‘build out’.
So the box to house it all in.
http://www.ebay.com/itm/162202150332
pic of box finished
And here again these numbers aren’t exact but are meant for clarity and simplicity, and are ‘good enough’…
65VA seemed like it should suffice which means each channel gets 32.5VA which = 2.7Amps at 12Vac.
So at 5Vdc x 2.7Amps = 13.5Watts.
And as it turns out the power draw from my Mutec 3+ is 0.56Amps = 2.8Watts.
This leaves plenty of headroom.
But that’s getting ahead of this story.
So I built it out and took a few pics of the before, during, and after.
externl mntg
ac wiring
extrnl frnt
extrnl rear
intrnl frnt open
intrnl rear open
bottom
And of course I took a few measurements of its noise, using the same scope I’ve used previously.
And it quickly became apparent that the output from this regulator board was pushing the limits of my scope in terms of resolution.
So again here are the 3 different means of measuring the noise for the ‘left’ and ‘right’ channel outputs into a 7.5K resistor (a minuscule load).
L channel R channel
Norm 1mv 1mv
Average 0.3mv 0.2mv
Envelope 3.7mv (6.7mv with all spikes) 4.8mv (6.0mv w/spikes)
There were very low level, low duration, hi freq spikes at ≈ .18µs (5.5MHz), .118µs (8.5MHz), 0.08µs (12.5MHz)
How they got created and from where, I can only guess.
Even so, these levels of noise are below what I measured from my RN3 when it was being powered by my external LPS.
link to post http://www.head-fi.org/t/781268/the-diyrs-cookbook/780#post_13218514
And I’ll follow up with measurements while running, from both the RN3 and my Mutec 3+ while being fed by their new LPS’s, in due course.
I first wanted to determine if this course of action was going to be ‘Better’ or not and it has exceeded my expectations and in ways I never suspected, but has done so much to my delight.
My take away is the use of SMPS’s mask inner detail and obscure meaningful musical information, which reduces our involvement in the music.
At least that is what I have come to understand as I hear these results in my playback system.
JJ
astrostar59
BannedMember of the Trade: Aries Cerat Espana, Auriculares High-EndAKA Headstage, headphoneweekend
Nice write up JJ! I think SMPS tech is flawed for audio. I don't understand all the tech but I understand one of the biggest negative besides noise and polluting other gear in the system is, it has high ripple distortion at the top an bottom of the sine wave. This has to affect audio quality IMO.
Which begs the question, why do so many audio products fit them? It must be the ability to accept 110 - 230 AC maybe? So mains voltage issues in various country and less hassle stocking different versions. That and the fact SMPS's are possibly cheaper and can be standardised to an extent. Many of the Rednet boxes have the same unit it seems.
The other thought is space and heat. In a DAC with a slim case, I can imagine that would suit an SMPS.
Having said that, once you get to 1K+ audio products, there should be no reason a decent manufacturer should need to use an SMPS?
On this subject, I did the Uptone Audio Mac Mini mod to rip out the SMPS and feed it 12V DC. That had a BIG change for the better.
The DC feed goes in where the headphone socket was. The mod is very fiddly as small parts and little space. It is also difficult the take apart. But well worth the effort.
I must say, I dig these kind of mods, they are relatively cheap especially in audio terms, you do it yourself and then enjoy the positives, what can be better than that?
BTW spot the cut outs I did on the iMac stand base, it allows me to sit the Mac Mini on the monitor base and send the cables straight through the back of the stand.
Which begs the question, why do so many audio products fit them? It must be the ability to accept 110 - 230 AC maybe? So mains voltage issues in various country and less hassle stocking different versions. That and the fact SMPS's are possibly cheaper and can be standardised to an extent. Many of the Rednet boxes have the same unit it seems.
The other thought is space and heat. In a DAC with a slim case, I can imagine that would suit an SMPS.
Having said that, once you get to 1K+ audio products, there should be no reason a decent manufacturer should need to use an SMPS?
On this subject, I did the Uptone Audio Mac Mini mod to rip out the SMPS and feed it 12V DC. That had a BIG change for the better.
The DC feed goes in where the headphone socket was. The mod is very fiddly as small parts and little space. It is also difficult the take apart. But well worth the effort.
I must say, I dig these kind of mods, they are relatively cheap especially in audio terms, you do it yourself and then enjoy the positives, what can be better than that?
BTW spot the cut outs I did on the iMac stand base, it allows me to sit the Mac Mini on the monitor base and send the cables straight through the back of the stand.
johnjen
Headphoneus Supremus
Thanks!
SMPS designs offer a number of advantages to designers and equipment manufacturers, but none for hi-end audio, at least as far as I can see.
Especially since they generate so much noise.
And I figure that designers justify their use because the noise they generate is 'out of band', so it shouldn't matter, except it can and does.
Especially when it propagates into the ground that analog gear is also connected to.
And from strictly a design POV if you don't create noise in the first place, you don't have to 'compensate' for it elsewhere.
The KISS principle in action.
This as I see it, is in keeping with our desired goal of achieving fully 120dB capable playback systems.
Which means the noise floor of ground should be at least -120dB.
This is a HUGE technical hurdle, one we won't be solving quickly.
It would be really kewl if we had a inverse MOV type of device, one that sucked up low voltage hi freq spikes, which is sorta what the Akiko sticks/cannisters do but they seem to be most effective at very high frequencies and don't seem to reach down even to 100KHz (that was what I noticed from the FR graphs I saw).
As for the modding of our gear to better suit our individual needs and the satisfaction we gain by doing so, well one of our rewards is HMMAIAA (Hearing MY Music As If Anew, Again).
Along with the added insights and learning we gain from the experience.
Besides I'm really dig'n the results I'm hearing.
And now that I see the Mac Mini I can appreciate what you did there.
NICE!
Yeah that is a 'sano' install for sure.
And for another,
here is Musiqboys setup using a windows NµC
It feeds the RN3 and the rest of his digital stack, which now includes a few LPS's that have been added to the mix.
Very Nice.
JJ
SMPS designs offer a number of advantages to designers and equipment manufacturers, but none for hi-end audio, at least as far as I can see.
Especially since they generate so much noise.
And I figure that designers justify their use because the noise they generate is 'out of band', so it shouldn't matter, except it can and does.
Especially when it propagates into the ground that analog gear is also connected to.
And from strictly a design POV if you don't create noise in the first place, you don't have to 'compensate' for it elsewhere.
The KISS principle in action.
This as I see it, is in keeping with our desired goal of achieving fully 120dB capable playback systems.
Which means the noise floor of ground should be at least -120dB.
This is a HUGE technical hurdle, one we won't be solving quickly.
It would be really kewl if we had a inverse MOV type of device, one that sucked up low voltage hi freq spikes, which is sorta what the Akiko sticks/cannisters do but they seem to be most effective at very high frequencies and don't seem to reach down even to 100KHz (that was what I noticed from the FR graphs I saw).
As for the modding of our gear to better suit our individual needs and the satisfaction we gain by doing so, well one of our rewards is HMMAIAA (Hearing MY Music As If Anew, Again).
Along with the added insights and learning we gain from the experience.
Besides I'm really dig'n the results I'm hearing.
And now that I see the Mac Mini I can appreciate what you did there.
NICE!
Yeah that is a 'sano' install for sure.
And for another,
here is Musiqboys setup using a windows NµC
It feeds the RN3 and the rest of his digital stack, which now includes a few LPS's that have been added to the mix.
Very Nice.
JJ
Muziqboy
500+ Head-Fier
The Talema transformer and the PC board for it arrived today but no LT3042 yet so the RN3 mod will have to wait. But the LPS for the Antelope Live Clock Reference Master Clock should be here by Monday.
And right now with the LPS's for the 3 Mutec's running, it has revealed more details that was hidden before by the noise of the SMPS's.
Yeah! I am diggin' this step up yet again in the SQ. SuperDuperGlue experience indeed.
If you remember JJ, during the last meet I told you that I will just mod the PSU of the Mutec usb since it is the last device before the DAC and it did added to the increase in SQ but I was surprised by another SQ step up with the LPS mod on the 2 Mutec non-usb. So I'm wondering what the LPS for the Live Clock will bring to the table.
Guess I'll find out on Monday. Can't stop listening right now.
And right now with the LPS's for the 3 Mutec's running, it has revealed more details that was hidden before by the noise of the SMPS's.
Yeah! I am diggin' this step up yet again in the SQ. SuperDuperGlue experience indeed.
If you remember JJ, during the last meet I told you that I will just mod the PSU of the Mutec usb since it is the last device before the DAC and it did added to the increase in SQ but I was surprised by another SQ step up with the LPS mod on the 2 Mutec non-usb. So I'm wondering what the LPS for the Live Clock will bring to the table.
Guess I'll find out on Monday. Can't stop listening right now.
johnjen
Headphoneus Supremus
Yeah last night was a late night here as well.
SuperDuperGlue makes it REALLY difficult to stop listening.
And today I upgraded the ac power cable to the dual LPS and took some measurements while it was playing music.
I'll write up those results here in a bit.
And I still haven't hit that SQ peak yet and to a certain extent it doesn't matter because right now the acoustic presentation is compelling in and of itself.
JJ
SuperDuperGlue makes it REALLY difficult to stop listening.
And today I upgraded the ac power cable to the dual LPS and took some measurements while it was playing music.
I'll write up those results here in a bit.
And I still haven't hit that SQ peak yet and to a certain extent it doesn't matter because right now the acoustic presentation is compelling in and of itself.
JJ
johnjen
Headphoneus Supremus
Voltage and power measurements of the dual LPS in operation.
So while I was upgrading the ac power cable from a Gen-1 to a Gen-4 cable on my dual LPS which feeds my Mutec 3+, I figured I take some measurements as well.
Input current at 120Vac = 0.135A with a peak of 0.14Amps, which makes it 16.8 Watts of input power.
And using the previously measured draw of 0.56Amps x 5Vdc = 2.8Watts of power being used by the Mutec.
Which makes it 16.7% efficient and is ‘wasting’ 14Watts of power, mostly as heat.
And using the Tek scope, I measured the noise in 3 ways…
R Channel R Channel Under load
Norm 1mv 7.5mv (no/spikes) - 13.5mv (w/spikes)
Average 0.2mv 1.2mv (no/spikes) - 2.5mv (w/spikes)
Envelope 4.8mv (6.0mv w/spikes) 19mv (no/spikes) - 22mv (w/spikes)
I also measured the noise and frequency of the few ’non-random’ portions of the noise.
I saw…
0.95µS = 1.05MHz
1.1µS = 909KHz
19mS = 52Hz
The amplitude of these are reflected in the noise measurements above.
And as you can see the increase in noise contributed by the active digital circuit, while minimal, is still 4-6 times greater than what the LPS ‘brings to the party'.
Next up I’ll measure the RN3’s LPS noise using resistors and the active digital load, and make power usage measurements as well.
JJ
So while I was upgrading the ac power cable from a Gen-1 to a Gen-4 cable on my dual LPS which feeds my Mutec 3+, I figured I take some measurements as well.
Input current at 120Vac = 0.135A with a peak of 0.14Amps, which makes it 16.8 Watts of input power.
And using the previously measured draw of 0.56Amps x 5Vdc = 2.8Watts of power being used by the Mutec.
Which makes it 16.7% efficient and is ‘wasting’ 14Watts of power, mostly as heat.
And using the Tek scope, I measured the noise in 3 ways…
R Channel R Channel Under load
Norm 1mv 7.5mv (no/spikes) - 13.5mv (w/spikes)
Average 0.2mv 1.2mv (no/spikes) - 2.5mv (w/spikes)
Envelope 4.8mv (6.0mv w/spikes) 19mv (no/spikes) - 22mv (w/spikes)
I also measured the noise and frequency of the few ’non-random’ portions of the noise.
I saw…
0.95µS = 1.05MHz
1.1µS = 909KHz
19mS = 52Hz
The amplitude of these are reflected in the noise measurements above.
And as you can see the increase in noise contributed by the active digital circuit, while minimal, is still 4-6 times greater than what the LPS ‘brings to the party'.
Next up I’ll measure the RN3’s LPS noise using resistors and the active digital load, and make power usage measurements as well.
JJ
Muziqboy
500+ Head-Fier
LPS for the Antelope Live Clock arrived this morning so I hooked it up and let the voltage stabilize for a bit.
OMG! I can't believe the increase in Dynamics, more inner details that I have not heard was being revealed.
REALISM took another bump up.
JJ, this really proves at least to me that those clocks are really sensitive to the quality of power that you feed it.
I'm waiting for the postman to drop off my LT3042 then I can begin modding the PSU in the RN3 but I'm afraid after I do that, I will never want to take the headphones off my head ever. LOL.
Cheers.
OMG! I can't believe the increase in Dynamics, more inner details that I have not heard was being revealed.
REALISM took another bump up.
JJ, this really proves at least to me that those clocks are really sensitive to the quality of power that you feed it.
I'm waiting for the postman to drop off my LT3042 then I can begin modding the PSU in the RN3 but I'm afraid after I do that, I will never want to take the headphones off my head ever. LOL.
Cheers.
astrostar59
BannedMember of the Trade: Aries Cerat Espana, Auriculares High-EndAKA Headstage, headphoneweekend
I too am waiting for my LT3042 LPS kit from Ebay. Ordered it on the 22nd Feb, should be soon now. Lets keep in touch.
johnjen
Headphoneus Supremus
Yeah greatly reducing or killing all of that upper frequency noise and those fast and intermittent spikes coming from the SMPS's, really does pay off.
And not generating them in the first place so all that noise can't spread to other gear may play as significant a role in the SQ increases we are hearing as well.
At least that is my thinking at this point.
And we may have to come up with a replacement term above and beyond SuperDuperGlue.
What it might be I'm not sure yet, but I figure it'll make itself known in due course.
Enjoy HMMAIAA, as the system settles down.
JJ
Muziqboy
500+ Head-Fier
Well after about 8 hours of surgery, I'm happy to report that the lps install in the RN3 was a success.
I was going to take some pics but forgot about it and put it all back in a hurry to do some listening. Clarity and articulation took another bump up. I can clearly hear every word being said. I'm afraid this will be another long listening session.
I was going to take some pics but forgot about it and put it all back in a hurry to do some listening. Clarity and articulation took another bump up. I can clearly hear every word being said. I'm afraid this will be another long listening session.
astrostar59
BannedMember of the Trade: Aries Cerat Espana, Auriculares High-EndAKA Headstage, headphoneweekend
Nice. 8 hours, wow, maybe this mod is more complicated than I thought? Did you have any problems with the instal?
From past new gear experiences, I would think you may see further improvements yup to a week at least (on 24/7). The caps in the supply need to burn in.
I look forward to your review, inc the affect of the LPS on the Antelope clock. My chain is much more basic than yours, I will probably stop at the LPS mod on the Rednet, as no funds for a Mutec.
Muziqboy
500+ Head-Fier
It took me that long since this is the first time I modded an internal psu. I was taking my time and being meticulous about it plus I was taking a lot of breaks in-between.
If you've done internal psu install before, it may take you only half of that time or less. I'll open it back up later and take some pics.
astrostar59
BannedMember of the Trade: Aries Cerat Espana, Auriculares High-EndAKA Headstage, headphoneweekend
Excellent!
And indeed clears up that bit of confusion.
I do wonder why the usb version runs at the higher voltage though.
A bit of a curiosity, that!
JJ
I read the Mutec+ USB has a different board and clock, so maybe the reason?
astrostar59
BannedMember of the Trade: Aries Cerat Espana, Auriculares High-EndAKA Headstage, headphoneweekend
JJ
A quick question. My LPS is due to arrive soon. I am unsure on the wiring at the plastic box connection at the SMPS and PCB. Are there only 3 wires there? i.e. 2 x 5VDC and one ground?
I have got the replacement top box for that connection and will solder new push fit ends to then make a press fit connection as opposed to hard wired. Got some spares as well incase I bust any with the crimping pliers. Can't wait for the mod.
A quick question. My LPS is due to arrive soon. I am unsure on the wiring at the plastic box connection at the SMPS and PCB. Are there only 3 wires there? i.e. 2 x 5VDC and one ground?
I have got the replacement top box for that connection and will solder new push fit ends to then make a press fit connection as opposed to hard wired. Got some spares as well incase I bust any with the crimping pliers. Can't wait for the mod.
Users who are viewing this thread
