DIY DC power cables

[sandyk]

10 minutes ago, JohnSwenson said:

You CAN use the resistance of the interconnect cable as the ballast resistor, but if you do that the output voltages of the two boards have to be very closely matched. The boards need to be within at least 10mv of each other if you are going to use the cable length as the resistor. And that is 10mv at load. That is hard to achieve unless you have a pot setting at least one of them. Then you have to worry about the resistance of the pot changing as the board warms up. Its a very touchy thing to do.

 

You are MUCH better off getting a board designed to handle the load on one board.

 

John S.

 

Agreed.

But IMO, anyone playing around like this should at least have a decent resolution DMM to check things like this if they don't want to risk damage to their new voltage regulator PCBs.

[agladstone]

2 hours ago, gstew said:

Cornan,

 

Have I got a solution for you! A fellow in Germany made up some paralleled LT3045 boards and is selling them on DIYAudio here:

 

http://www.diyaudio.com/forums/vendor-s-bazaar/310402-fs-ultra-low-noise-power-supply-lt3045-based-pcb-post5227003.html

 

He has both boards with 2 sections of 3 paralleled LT3045 each (1.5A / section) and boards with a single section of 6 paralleled LT3045 for 3A. He'll sell raw boards (what I got), boards with the LT3045 already installed, or finished & tested boards

 

I just built one of the dual section boards up and it is working swimmingly powering my SDTrans384 SD Card player (which really only pulls 400mA max). I used .1% resistors to set the V-outs instead of the pots he shows and the 2 sides measured a measly .01V difference (4.92V & 4.93V). Each side is fed from an LPS-1 & are paralleled for 3A output max, overkill in this application.

 

Problem solved!

 

Greg in Mississippi

Greg:

I will start by saying that I don’t understand at all really what these LT3045 boards do at all and I’m completely confused by all the talk here about them! 

However, I THINK what you’re saying in your post is that by using 2 LPS-1’s and your LT3045 boards as configured, that you’re able to get 3A current output? Is that correct? 

So if I’m correct, the next question, is there a way (perhaps it would require 3 or 4 LPS-1’s?)

to use these boards in conjunction with multiple LPS-1’s to get 12V 3A ?? (I’m thinking this would be an incredible power supply for my Mytek Brooklyn DAC)? (Although if 4 LPS-1’s are required plus the boards, it may not end up being a good cost to value ?

[Cornan]

8 hours ago, sandyk said:

 John could answer this far better than I can, but provided that the well matched PCBs used identical length connecting leads to the joining point of the cables, it may not matter ? 

If one had a slightly lower voltage output, perhaps decreasing the length of it's output lead a little may help to ensure improved current output matching ?

 I defer to John's wealth of experience in this area though. 

I am here to learn too !

 

In any case the output leads and in-between the boards are in the range of 2-3cm when you put them in the DC cable path. I am very surprised if this could affect the voltage/current. 

[sandyk]

10 minutes ago, Cornan said:

 

In any case the output leads and in-between the boards are in the range of 2-3cm when you put them in the DC cable path. I am very surprised if this could affect the voltage/current. 

 That may be the case with regulator PCBs in series, but not likely when in parallel.

Quote

The two boards must be matched from the seller (Alexey). Then use two of this cable (input/output) where DC screw terminal plugs is the LT3045s side-by-side.

The lengths of 2 of the cables in your photo suggests otherwise.  

[Cornan]

7 hours ago, gstew said:

Cornan,

 

Have I got a solution for you! A fellow in Germany made up some paralleled LT3045 boards and is selling them on DIYAudio here:

 

http://www.diyaudio.com/forums/vendor-s-bazaar/310402-fs-ultra-low-noise-power-supply-lt3045-based-pcb-post5227003.html

 

He has both boards with 2 sections of 3 paralleled LT3045 each (1.5A / section) and boards with a single section of 6 paralleled LT3045 for 3A. He'll sell raw boards (what I got), boards with the LT3045 already installed, or finished & tested boards

 

I just built one of the dual section boards up and it is working swimmingly powering my SDTrans384 SD Card player (which really only pulls 400mA max). I used .1% resistors to set the V-outs instead of the pots he shows and the 2 sides measured a measly .01V difference (4.92V & 4.93V). Each side is fed from an LPS-1 & are paralleled for 3A output max, overkill in this application.

 

Problem solved!

 

Greg in Mississippi

 

Thanks Greg! @tapatrick have showed me this beautiful board a couple of times before on my "True experimental tweak thread". I am definately interested to give it a try on my Brooklyn DAC which reguores 12v/~1.5A if I not DIY my own parallel board. Do you have a picture of your parallel boards you could share? It would be very interesting to see a picture! ?

[Cornan]

5 hours ago, agladstone said:

(I’m thinking this would be an incredible power supply for my Mytek Brooklyn DAC)?

 

I am planning to use the LT3045 post Brooklyn DAC as well. However, it does'nt really require more than 2A. To be sure I'll recommend that you monitor the current draw during 10-15 min incl power on/off. My bet is that it will only draw max 1.5A.

[Cornan]

2 hours ago, sandyk said:

 That may be the case with regulator PCBs in series, but not likely when in parallel.

The lengths of 2 of the cables in your photo suggests otherwise.  

Not sure what you mean that it is the case with regular PCB in series but not in parallel? Why is there a difference? Could really 2-3cm make such a difference in voltage/current between the boards?

The picture was just to show how it should be wired. The cable lenght should ALWAYS be as short as possible.

[jean-michel6]

18 hours ago, Cornan said:

 

Hi jean-michel6! 

 

Here is my answers with numbers! ?

 

1. Gophert csp-3205II floating lab SMPS

 

2. The sound get more full with improved details and blacker background.

 

3. Serial LT3045 equals even less noise. Fixed voltages. I set mine to 0.7-1v drop down. IME 0.7v drop-down is optimal in my setup. For example powering my Aries Mini I set Gophert to 15.7v into LT3045 (1A) 15v into LT3045 (1A) 14.3v into Aries Mini.

 

4. No there is 1A boards around as well. If you put two 1A boards in parallel you'll get 2A. Measure the actual current drawn during 15 minutes (incl. Power on/off) to find out which amperage you'll need.

 

5. Two benefits. EMI/RFI reduction and a slim case/protection for the dual boards. 

 

I hope that helps! ?

 

 

 

 

Hi Cornan 

 

Thank's for the detailed answer , I will definitely try that in my system.  

Have you tried to add different Elna or other capacity on your LT boards ?

Were do you purchase the Entreq sleeves , easier for you because it is a Swedish company. 

I am located in france. 

[Cornan]

Just now, jean-michel6 said:

Hi Cornan 

 

Thank's for the detailed answer , I will definitely try that in my system.  

Have you tried to add different Elna or other capacity on your LT boards ?

Were do you purchase the Entreq sleeves , easier for you because it is a Swedish company. 

I am located in france. 

 

I have only tried ELNA silmic on my LM317 board. IMO the combo could'nt compete with LT3045 without ELNA. With that said several people have reported an improvement with ELNA silmic on LT3045. I am personally not that interested in that path though since I want to ensure best possible isolation without adding too much sound signature to the mix. I have currently a couple of shunt regulators with Kelvin starquad cables and plugs on order. As soon as they have arrived (next year) I will compare them to the LT3045. Should be interesting! ?

 

I bought my Entreq mini wraps from a local Swedish dealer. According to Entreq's website you should get them here in France http://www.pi-music.fr 

 

Happy New year! ?

[jean-michel6]

9 hours ago, gstew said:

Cornan,

 

Have I got a solution for you! A fellow in Germany made up some paralleled LT3045 boards and is selling them on DIYAudio here:

 

http://www.diyaudio.com/forums/vendor-s-bazaar/310402-fs-ultra-low-noise-power-supply-lt3045-based-pcb-post5227003.html

 

He has both boards with 2 sections of 3 paralleled LT3045 each (1.5A / section) and boards with a single section of 6 paralleled LT3045 for 3A. He'll sell raw boards (what I got), boards with the LT3045 already installed, or finished & tested boards

 

I just built one of the dual section boards up and it is working swimmingly powering my SDTrans384 SD Card player (which really only pulls 400mA max). I used .1% resistors to set the V-outs instead of the pots he shows and the 2 sides measured a measly .01V difference (4.92V & 4.93V). Each side is fed from an LPS-1 & are paralleled for 3A output max, overkill in this application.

 

Problem solved!

 

Greg in Mississippi

Hi Greg ,

 

 

Thank's for the good info .

 

Several questions :

 

LPS-1 output 1.1 A max so two in parallell is 2.2A max ? Correct ?

 

The German DIY shows only the AC-DC version , does he has also DC-DC version , which is needed to work with LPS-1 ?

( question for you and John S ) This LPS-1 is already a very good PS , what improvement bring the LT board ?

[jean-michel6]

4 minutes ago, Cornan said:

 

I have only tried ELNA silmic on my LM317 board. IMO the combo could'nt compete with LT3045 without ELNA. With that said several people have reported an improvement with ELNA silmic on LT3045. I am personally not that interested in that path though since I want to ensure best possible isolation without adding too much sound signature to the mix. I have currently a couple of shunt regulators with Kelvin starquad cables and plugs on order. As soon as they have arrived (next year) I will compare them to the LT3045. Should be interesting! ?

 

I bought my Entreq mini wraps from a local Swedish dealer. According to Entreq's website you should get them here in France http://www.pi-music.fr 

 

Happy New year! ?

You are super fast to answer ?

 

Where have you been able to source shunt regulators as they are supposed to be among the very best for PS. 

 

Happy new year too !

[Cornan]

35 minutes ago, jean-michel6 said:

You are super fast to answer ?

 

Where have you been able to source shunt regulators as they are supposed to be among the very best for PS. 

 

Happy new year too !

 

Thanks! ?

I have ordered it from Alex at ATL Hi-Fi https://www.atlhifi.com

They are still under construction though. Not expecting delivery until Q2, 2018.

Mind you that so far the shunt regs is just possible for max 400mA load. Alex is working on a 1.5A version, but I am not sure yet if/when it will be finalized.

[Cornan]

14 hours ago, JohnSwenson said:

Putting more than one together ON THE SAME BOARD, guarantees the output voltages will be VERY close, so only very small ballast resistors are needed.  (literally a half inch of trace on the board) BUT that only works when they are on the same board. You don't get that advantage when paralleling two boards.

 

How about using 4 separate boards? For example two matched 13v parallel boards in series into two 12v parallel boards powering a Brooklyn DAC (12v)? PSU set to 14v. 

[Cornan]

14 hours ago, gstew said:

Cornan,

 

Have I got a solution for you! A fellow in Germany made up some paralleled LT3045 boards and is selling them on DIYAudio here:

 

http://www.diyaudio.com/forums/vendor-s-bazaar/310402-fs-ultra-low-noise-power-supply-lt3045-based-pcb-post5227003.html

 

He has both boards with 2 sections of 3 paralleled LT3045 each (1.5A / section) and boards with a single section of 6 paralleled LT3045 for 3A. He'll sell raw boards (what I got), boards with the LT3045 already installed, or finished & tested boards

 

I just built one of the dual section boards up and it is working swimmingly powering my SDTrans384 SD Card player (which really only pulls 400mA max). I used .1% resistors to set the V-outs instead of the pots he shows and the 2 sides measured a measly .01V difference (4.92V & 4.93V). Each side is fed from an LPS-1 & are paralleled for 3A output max, overkill in this application.

 

Problem solved!

 

Greg in Mississippi

 

Thanks for the reminder of this one Greg! ? I have just sent a pm to Stammering that I want to order a single ps. I will use it for my Brooklyn DAC powered by Gophert csp-3205II. I have also a couple of matched 12v LT3045 boards on the way from another member. I will use try them out for my coming 12v main router where I think they will make a great addition.

[gstew]

 

Jean-Michel,

 

You are very welcome. & I had planned to start out this set of replies with an apology on implying that one could get 3A out of 2 LPS-1's in this setup. You are very correct... and I apologize for not being clearer. The dual 3||LT3045 boards each have a max output of 1.5A and in parallel (when tightly matched) will provide 3A. BUT the system current will be limited by the LPS-1's 1.1A max each, so total out available is 2.2A. All very sufficient for a device that draws less than 500mA, but I'm an overkill kind of audiophile! 

 

As for improvement, I have AFAIK the only pair of LPS-1s that have been output-matched for paralleling (thanks again, Superdad!).  I tried them in parallel to power either a Raspberry Pi I2S-output renderer or an SDTrans384 SD card player. I describe my initial experiences here: 

and an update after trying a dual 4||LT3042 board (800mA max per side, 1.6A max when parallelled)  with each side powered by an LPS-1. That has been my reference for both of these applications since then.

 

My gut feel is that the lower noise provided by these boards along with the additional energy storage (there's about 36,000uf on each board) are what make them better than raw LPS-1s for this application.

 

I should add that in some other applications (such as Ian Canada's R-Pi Isolator board and Allo.com's R-Pi Isolator board and Kali reclocker board), I have tried both an Acko AKR75 regulators (his previous version using an ADM715x chip... next best chip regulators I know of after the LT3042/45s) and an older generation Belleson regulator between the LPS-1s and the load board. In these applications, I preferred directly powering them from the LPS-1 without the intervening regulator boards. That MIGHT change if I try an LT3045-based regulator board, but that is my experience so far.

 

Pix of a couple of the AKR75 reg boards attached...

 

 

12 hours ago, agladstone said:

Greg:

I will start by saying that I don’t understand at all really what these LT3045 boards do at all and I’m completely confused by all the talk here about them! 

However, I THINK what you’re saying in your post is that by using 2 LPS-1’s and your LT3045 boards as configured, that you’re able to get 3A current output? Is that correct? 

So if I’m correct, the next question, is there a way (perhaps it would require 3 or 4 LPS-1’s?)

to use these boards in conjunction with multiple LPS-1’s to get 12V 3A ?? (I’m thinking this would be an incredible power supply for my Mytek Brooklyn DAC)? (Although if 4 LPS-1’s are required plus the boards, it may not end up being a good cost to value ?

 

agladstone,

 

See my comments and apologies above. Powered from 2 LPS-1s, the max current you get from that setup is 2.2A. AND it takes care to match the board outputs (and should be done under load, as John Swenson said). So not easily today. For what is out there now, I'd look at the Vinnie Rossi supplies. Hopefully we'll see some new products soon that will make this more easily done.

 

Also see Cornan's comments below on the real-world current draw of that DAC.

 

I should add that I run a DIY Soekris DAM1021 DAC from the I2S output of the SDTrans384 SD Card Player. It is my reference setup. I have set it up to be powered either from a pair of very good DIY'd linear power supplies OR a pair of LPS-1's set to 7V. While the LPS-1's exhibit some fine qualities in this setup, I've stuck with the linear supplies due to slightly greater dynamic impact and shading. You have to understand I LOVE LPS-1's and have 10 of them around here currently. They are ground-breakingly great supplies. AND I have found some situations where I either find them better augmented (such as feeding the dual 3||LT3045 or dual 4||LT3042 supplies) or where I prefer a different solution. Still, in MOST cases, I can't beat them!

 

7 hours ago, Cornan said:

 

Thanks Greg! @tapatrick have showed me this beautiful board a couple of times before on my "True experimental tweak thread". I am definately interested to give it a try on my Brooklyn DAC which reguores 12v/~1.5A if I not DIY my own parallel board. Do you have a picture of your parallel boards you could share? It would be very interesting to see a picture! ?

 

7 hours ago, Cornan said:

I am planning to use the LT3045 post Brooklyn DAC as well. However, it does'nt really require more than 2A. To be sure I'll recommend that you monitor the current draw during 10-15 min incl power on/off. My bet is that it will only draw max 1.5A.

 

Pictures attached...

 

 

I should add that I use different input and output capacitors than what is listed in the BOM... these are the same line as those used in the dual 4||LT3042 boards that have also worked very well where I've used them, but alas are no longer available. I also:

 

1. Added an additional 5,600uF capacitor at the output of each side.

2. Omitted the rectifier diodes and jumpered as appropriate to make them DC input.

3. Left out the heat sinks for this application due to both the low current draw of the driven device AND a fairly low input voltage (7V out of the LPS-1s).

4. Used .1% resistors instead of a pot to set the output voltage for each side.

 

6 hours ago, jean-michel6 said:

Hi Greg ,

 

 

Thank's for the good info .

 

Several questions :

 

LPS-1 output 1.1 A max so two in parallell is 2.2A max ? Correct ?

 

The German DIY shows only the AC-DC version , does he has also DC-DC version , which is needed to work with LPS-1 ?

( question for you and John S ) This LPS-1 is already a very good PS , what improvement bring the LT board ?

 

Jean-Michel, I left out the Shottky rectifiers & jumpered across traces to make these DC input... see pix of my unit.

 

Greg in Mississippi

[gstew]

14 hours ago, sandyk said:

Hi Greg

 You need to be careful here. The text book implementations of these ensures a very low ESR, which favours upper HF detail. Unless you change a few input capacitor types and values you may end up with overly etched HF detail.

 

This shouldn't  happen with Digital, but it does ,  even with SATA filters. The later SATA filters use much larger value electros which help to ensure a more balanced sound. (Previously I had a good play around with DIY SATA filters using L & C.)

 

Regards

Alex 

 

 

Alex,I totally agree with your point. While I didn't have that in mind when I first tried one of the paralleled LT304x boards, I suspect the reason they've worked so well is that they are really designed as small power supply units with AC input from a transformer and a significant amount of capacitance (~18,000uF) on each half. AND they do retain good small hi-frequency bypass capacitors before and after the regs as suggested by the datasheets. 

 

Thanks for pointing this out!

 

Greg in Mississippi

 

[Cornan]

37 minutes ago, gstew said:

Pictures attached...

 

 

I should add that I use different input and output capacitors than what is listed in the BOM... these are the same line as those used in the dual 4||LT3042 boards that have also worked very well where I've used them, but alas are no longer available. I also:

 

1. Added an additional 5,600uF capacitor at the output of each side.

2. Omitted the rectifier diodes and jumpered as appropriate to make them DC input.

3. Left out the heat sinks for this application due to both the low current draw of the driven device AND a fairly low input voltage (7V out of the LPS-1s).

4. Used .1% resistors instead of a pot to set the output voltage for each side.

 

Great stuff! Thanks for sharing and  a happy new year! ?

[tapatrick]

On 12/30/2017 at 6:22 PM, jean-michel6 said:

Hi ,

Several questions about the use of the LT3045s

- which power supplies are you using upstream of the LT modules. 

- what is the main benefit of using those LT modules on the sound 

- why several modules in series ? Those modules are fixed voltage or adjustable ?How do you set them to have the 0.3 v drop to insure the regulator is working properly ?

- The amp accepted by the module is only 0,5 amp , which is quite low 

- was is the benefit of using the Entreq sleeves ?

 

Many questions , thank for your help. 

 

 

Hi @jean-michel6 Looks like Xmas is over and this thread has sprung to life again..:)

 

To add to what @Cornan has already replied to you:

The LT3045s have an Ultralow RMS Noise: 0.8µVRMS (10Hz to 100kHz), plus Ultralow Spot Noise: 2nV/√Hz at 10kHz, PLUS Ultrahigh PSRR: 76dB at 1MHz. This makes them superb for powering end of chain low power devices, so the 500ma is adequate for these devices. There is much discussion about achieving 1-2 and 3amps but I have not needed that much.

 

I power sets of LT3045s with a modified Teradak DC-30W (replaced caps with Nichicon HWs) into an Aqvox switch/USB 5v line/LPS-1 and finally into a Bluewave USB to Spdif board (which Micael/cornan and I have raved about many times).

 

The already low 0.8µVRMS drops even more when put in series. I use for instance 7.7v boards into 5v into 3.3v. The main benefit of the in series boards is increased enjoyment! - deeper bass, smoother treble, cleaner spacious sound and expansive soundstage among other qualities. The drop out can be a couple of volts but must be more than 260mv and I recommend the fixed voltage ones.

 

Highly recommend them if they fit your requirements.

 

[tapatrick]

7 hours ago, jean-michel6 said:

Hi Cornan 

 

Thank's for the detailed answer , I will definitely try that in my system.  

Have you tried to add different Elna or other capacity on your LT boards ?

Were do you purchase the Entreq sleeves , easier for you because it is a Swedish company. 

I am located in france. 

I tinkered around with both adding to and replacing the Tantalum input caps with Elna silmics which were recommended by @sandyk and @TubeMan. Adds an extra smoothness...

[tapatrick]

1 hour ago, gstew said:

I tried them in parallel to power either a Raspberry Pi I2S-output renderer or an SDTrans384 SD card player. I describe my initial experiences here: 

Hi @gstew over Xmas I came across some of your posting on Tirnahifi and DIYaudio as I found myself once again getting curious about SD card players, but the volume and length of the threads overwhelmed me.. . Without derailing this thread can you point me to latest info on finding or building a SOTA SD player? can go to PM if better for you.. Many thanks in advance.

[sandyk]

25 minutes ago, tapatrick said:

The already low 0.8µVRMS drops even more when put in series.

 

 That is unlikely to be correct. You will be limited by the inherent noise level of the LT3045 .

 A parallel configuration as used with input transistors in some phone moving coil designs should reduce noise further.

[Cornan]

1 hour ago, tapatrick said:

 

I power sets of LT3045s with a modified Teradak DC-30W (replaced caps with Nichicon HWs) into an Aqvox switch/USB 5v line/LPS-1 and finally into a Bluewave USB to Spdif board (which Micael/cornan and I have raved about many times).

 

I cannot rave enough for that combo. It makes music for angels! 

A Happy New Year Patrick! ?

[tapatrick]

1 hour ago, Cornan said:

 

I cannot rave enough for that combo. It makes music for angels! 

A Happy New Year Patrick! ?

Happy New Year Micael!

[tapatrick]

3 hours ago, sandyk said:

 

 That is unlikely to be correct. You will be limited by the inherent noise level of the LT3045 .

 A parallel configuration as used with input transistors in some phone moving coil designs should reduce noise further.

Thanks @sandyk, You have more knowledge in this area. I read that somewhere so I will try to find the thread where this was stated to see if it was a measurement or just an opinion.

 

In terms of listening impressions though I can confirm that adding the second board in series in my system added a subtle but magical quality, not as dramatic as adding the first one but significant enough to increase my enjoyment.

[sandyk]

20 hours ago, gstew said:

I just built one of the dual section boards up and it is working swimmingly powering my SDTrans384 SD Card player (which really only pulls 400mA max). I used .1% resistors to set the V-outs instead of the pots he shows and the 2 sides measured a measly .01V difference (4.92V & 4.93V).

 

Hi Greg

 

Not bad !
Were they Naked Bulk Metal Foil Resistors, or a less expensive item ?
For the less experienced members, the results are of course governed by the actual reference voltage of the voltage regulator I.C. used, not so much by the actual measured resistance values of the close tolerance resistors (1% or better) when used with a less precise voltage regulator. (Most DMMs don't have enough resolution to accurately measure close tolerance low value resistors)

Kind Regards

Alex

Warning - Information Overload.
(An  experienced DIY person can  also further reduce the voltage difference of the higher voltage output regulator by retro fitting a parallel metal film resistor of > 100 times the value across the higher value resistor of the 2 voltage setting resistors, usually under the PCB.

If the Voltage regulator uses through hole resistors, it is possible to slightly increase the output voltage of a voltage regulator by replacing the lower value of the 2 resistors used, with a 1% resistor that measures a fraction lower than the original.

You can also reduce the output voltage slightly by slightly increasing the measured value of this resistor, or decreasing slightly the value of the higher value resistor of the 2 voltage setting resistors. With the latter, you can  also use a very high value (>100 times) resistor in parallel with it instead of replacing it.)