The true experimental tweak thread

[tapatrick]

1 hour ago, Speed Racer said:

If you guys follow Tubelover2's path you will be on a never ending journey of buying new stuff that always result in SQ improvement. BY now his chain should sound 10 times better than live music......

LOL! You are saying that on Cornan's thread. Thought that was the whole point. 

[Speed Racer]

7 minutes ago, tapatrick said:

LOL! You are saying that on Cornan's thread. Thought that was the whole point. 

 

I thought the whole point was the best sound quality....

[Cornan]

57 minutes ago, octaviars said:

 

All connected devices will share the ground at the single PSU as the ground passes right through the regulator cards so you might not end up with the ultimate solution regarding noise if that is what you are seeking. 

 

I only see one option for you buy more Ghoperts and many regulator boards

 

Unfortunately I do agree with you initially...but I still want to see how it turns out. I have pretty much mixed results using a single floating Gophert powering several devices via a DC terminal strip. In some cases it seems to be beneficial and in some cases not. It troubles me a bit. Using the Regen Amber with a Y-split cable did improve things that could, atleast in my personal theory, have to do something about the voltage regulator (and that they might block leakage loops from entering the device). Now the DC terminal strip will power 3-4 5v devices simultaniously with a floating psu connected to a floating & balanced IT (=dual floating grounds). In theory the voltage regulators could benefit isolation making this into a solid improvement instead of a mixed one. That theory is enough for me to be interested enough to give it a try. If it fails I can still buy more Gopherts and still use the LT3045s. Win win!

 

 

[Cornan]

41 minutes ago, Speed Racer said:

 

I thought the whole point was the best sound quality....

 

Well, it is ofcourse. However, I would be very surpriced to find two people here with identical setups. It is all about sharing both progress and failures in the name of good SQ and all about the free mind to try or avoid things based on others peoples experiances. No-one can try everything, so we are left to read about what other people beleive in and make personal choises what to try based on our own knowledge, experiance and specific setup. This thread is all about experimental tweaking and for those who enjoy it. No guarantees that it will be fun for everyone.

 

[sandyk]

FWIW, there are heaps of LT3045 based voltage regulator PCBs available on ebay, but don't forget that the maximum current is of an LT3045 is 500mA, and they will require substantially more heat sinking than almost all of them have, to be suitable for anywhere near 500mA.

[Cornan]

13 minutes ago, sandyk said:

FWIW, there are heaps of LT3045 based voltage regulator PCBs available on ebay, but don't forget that the maximum current is of an LT3045 is 500mA, and they will require substantially more heat sinking than almost all of them have, to be suitable for anywhere near 500mA.

 

Thanks Alex! Do you have a firm recommendation for another ultra low noise voltage regulator that works slightly above the 1A range and up to 15-18v DC?

 

This one http://www.ebay.com/itm/201209828911?_trksid=p2060353.m570.l5999&_trkparms=gh1g%3DI201209828911.N36.S2.R2.TR4 is already on my list thanks to @TubeMan

 

Sjostrom SSR02 is the best option so far, but it have max 1A so I´ll need to measure my Aries Mini before I choose that option.

[sandyk]

Hi Cornan

I posted about that module quite a while back, as another CA member and myself have tried them. A member reported back at the time about obtaining an improvement with it adjusted to +5V and used for USB.

It's still not as affective as the JLH though.

Alex

[octaviars]

1 hour ago, Cornan said:

Unfortunately I do agree with you initially...but I still want to see how it turns out

 

I get that no worries

 

I will go directly to one R-Core transformer per PSU bord or atleast have seperate windings for each PSU that is my plan right now just need to sort some other things out before I go down that path.

[Cornan]

1 minute ago, octaviars said:

 

I get that no worries

 

I will go directly to one R-Core transformer per PSU bord or atleast have seperate windings for each PSU that is my plan right now just need to sort some other things out before I go down that path.

 

Great to hear!

Using R-core transformers through-out is a great plan, but as you might have noticed I am all stuck in the Gophert swamp!

 

[Cornan]

5 minutes ago, sandyk said:

Hi Cornan

I posted about that module quite a while back, as another CA member and myself have tried them. A member reported back at the time about obtaining an improvement with it adjusted to +5V and used for USB.

It's still not as affective as the JLH though.

Alex

 

I actually think you posted the very same voltage regulator to me as well a long time ago, but I could´nt be sure since I never put it on my eBay wish list (bad boy!). Hmmm, I need to get myself a cup of coffee and study that JLH. Maybe Alex at ATL Hi-Fi can make one for me? (fingers crossed) 

 

[sandyk]

5 minutes ago, octaviars said:

 

I get that no worries

 

I will go directly to one R-Core transformer per PSU bord or atleast have seperate windings for each PSU that is my plan right now just need to sort some other things out before I go down that path.

 

 R-core transformers are good, but when using them to power USB devices I have found that connecting their shield wire to mains earth causes a small amount of degradation  of SQ.

USB doesn't like any capacitive coupling to mains earth in their external power supplies .

Alex

 

[sandyk]

4 minutes ago, Cornan said:

 

I actually think you posted the very same voltage regulator to me as well a long time ago, but I could´nt be sure since I never put it on my eBay wish list (bad boy!). Hmmm, I need to get myself a cup of coffee and study that JLH. Maybe Alex at ATL Hi-Fi can make one for me? (fingers crossed) 

 

 

Incidentally, when the JLH is used at low voltages such as +5V, I have found that the tiny (and cheap on ebay) , 2SC5707 can be used  instead of the Toshiba 2SC5171. It has a typical HFE of >240 which results in an even  higher simulated capacitance value .

 

e.g. HFE x 4,700Uf + 4,700uF. That puts us well into the FARAD area compared with the original design.

 

Alex

[Cornan]

1 minute ago, sandyk said:

 

Incidentally, when the JLH is used at low voltages such as +5V, I have found that the tiny (and cheap on ebay) , 2SC5707 can be used  instead of the Toshiba 2SC5171. It has a typical HFE of >240 which results in an even  higher simulated capacitance value .

 

e.g. HFE x 4,700Uf + 4,700uF. That puts us well into the FARAD area compared with the original design.

 

Alex

 

Thanks Alex!

 

I will e-mail to ALT Hi-Fi and see if he can make a JHL Ripple Eater for me. I will make sure that he gets this additional info as well.

 

Micael

[Solid-State]

2 hours ago, sandyk said:

FWIW, there are heaps of LT3045 based voltage regulator PCBs available on ebay, but don't forget that the maximum current is of an LT3045 is 500mA, and they will require substantially more heat sinking than almost all of them have, to be suitable for anywhere near 500mA.

 

But if the are on a big heatsink,,,,

[sandyk]

41 minutes ago, TubeMan said:

 

But if the are on a big heatsink,,,,

 

With the majority of those ebay items it would be very hard, if not impossible to fit an adequately sized heat sink.

https://www.ebay.com.au/sch/i.html?_nkw=LT3045&_trksid=p2484217.m4323.l4&_trkparms=5926%3AKEYWORD|5079%3A0

[tapatrick]

I don't think all the boards available on eBay are the same quality. These are the ones I have been recommended and which I purchased from Alexey at LDOVR . http://stores.ebay.co.uk/LDOVR?_trksid=p2047675.l2563

 I did try some other boards with an LT3045 chip and they weren't as well made.

 

There are 500ma but also 1a versions too. I have not had any heat issues (yet!) using the 500ma and 1a versions in series as I posted above as long as the drop down voltages are correct.

As has been mentioned and from the data sheet below, they can also be paralleled to achieve 2amp or more with even lower noise profiles. Not sure how far you can take it, but max input voltage is 20v and max output voltage is 15v.... Maybe @sandyk or @octaviarscan shed more light as its not my forte but fun to explore...

[Cornan]

23 minutes ago, tapatrick said:

These are the ones I have been recommended and which I purchased from Alexey at LDOVR

 

Yepp, this was the seller I bought my two initial ones from too. I am getting old, so I just double checked! 

 

[octaviars]

As I get it it is all about how much voltage the regulator needs to drop the voltage as it will turn into heat. With LT3045 I would try to supply it as low voltage as possible but enough to still be able to regulate.

If you look at the scematics with 4 parallel LT3045 it says 300mV dropout which means that if you need 5v out you will need 5,3v in but that is perhaps on the edge so 5,5v may be on the safe side. You could always put a pre regulator board with a LT1083 or something like that with adjustable DC output to set it at right level before entering the LT3045.

 

 

[octaviars]

5 hours ago, sandyk said:

USB doesn't like any capacitive coupling to mains earth in their external power supplies .

Alex

 

What did you do with the sheild?

 

I had a plan to perhaps use a balanced iso transformer and isolate the R-cores from the ground in the cabinet and connect the shield to the center tap on the iso transformer but perhaps it is just easier to use unsheilded R-cores......

 

 

[lmitche]

1 hour ago, octaviars said:

As I get it it is all about how much voltage the regulator needs to drop the voltage as it will turn into heat. With LT3045 I would try to supply it as low voltage as possible but enough to still be able to regulate.

If you look at the scematics with 4 parallel LT3045 it says 300mV dropout which means that if you need 5v out you will need 5,3v in but that is perhaps on the edge so 5,5v may be on the safe side. You could always put a pre regulator board with a LT1083 or something like that with adjustable DC output to set it at right level before entering the LT3045.

 

 

I have two 5 volt 1 amp ldovr lt3045 cards running with 7 volt input from my lps-1s for the past week with no heatsink and no problems.

[octaviars]

46 minutes ago, lmitche said:

I have two 5 volt 1 amp ldovr lt3045 cards running with 7 volt input from my lps-1s for the past week with no heatsink and no problems.

 

That sounds good but do you know how much power the stuff you have connected draws?

[lmitche]

5 minutes ago, octaviars said:

 

That sounds good but do you know how much power the stuff you have connected draws?

I'm powering a USB NIC, Adnaco USB box and HDD in an enclosure on one, Microidsd DAC and ISO REGEN on another,  last an Amber Regen. The LPS-1 on the HDD is the hottest. I don't have measurements of the current draw.

[Cornan]

I thought it could be fun to continue my recycling path. This time the hate-loved Audioquest Jitterbug! ?

 

I installed it on the other end of my DIY Canare 4S6 Y-split DC cable. Same leg powering the Regen Amber but close to the Gophert.

 

 

Massive bass attack listening to Zhu "Genesis Series"! Not quite a smooth listening to No-Man "Schoolyard Ghosts" and a bit fatiguing. Darker sound stage overall. Still very detailed and ambient. I do not feel as relaxed as with the Regen Amber alone in front of the ISO Regen. Interesting to try. Could be someone else's cup of tea perhaps, but a no go for me. 

 

It is back into my Aqvox switch port 5 again, where it do clear up the high frequencies a tiny bit. ?

 

[sandyk]

6 hours ago, octaviars said:

 

What did you do with the sheild?

 

I had a plan to perhaps use a balanced iso transformer and isolate the R-cores from the ground in the cabinet and connect the shield to the center tap on the iso transformer but perhaps it is just easier to use unsheilded R-cores......

 

 

I fitted a toggle switch on the rear of the case to enable quick switching of the shield wire of the transformer's shield wire to mains earth. I always preferred the sound without the wire connected to mains earth.

 

With my PSUs I always load test them with suitable value load resistors, as we don't always know what the loading will be.

With SSDs for example, many earlier ones drew far more current than many newer ones.

What is adequate cooling for short term use may not be adequate for lengthy sessions writing to  USB memory for example.

 

Alex

[tapatrick]

15 hours ago, octaviars said:

As I get it it is all about how much voltage the regulator needs to drop the voltage as it will turn into heat. With LT3045 I would try to supply it as low voltage as possible but enough to still be able to regulate.

If you look at the scematics with 4 parallel LT3045 it says 300mV dropout which means that if you need 5v out you will need 5,3v in but that is perhaps on the edge so 5,5v may be on the safe side. You could always put a pre regulator board with a LT1083 or something like that with adjustable DC output to set it at right level before entering the LT3045.

 

 

The drop down range is quite wide. I believe it can be a few volts. As I mentioned I used 8v from my Teradak psu into a 7.5v LT3045 500ma  then into a 5v LT3045 500ma. This means the second board is getting a very clean stable current. In series like this gives the best SQ. I also do the same with 7.2 and 5v 1amp boards powering Aqvox switch. Great combo all round.

 

I've also tried this with LPS-1 set to 7v out into Y cable with one LT3045 5v at each end of the branch of the cable running my Wavio board and ISO Regen. The first setup with dual in series LT3045s powered by modded Teradak DC30 W is better SQ in my system.