The true experimental tweak thread

[Forehaven]

10 hours ago, TubeMan said:

and remember,
higher voltage rating better quality,

even for low voltage application

Ok, I need 25V 1500uf and 25V 3300uf caps.  Unfort. it's a no go on Elna Silmic's.  Nothing in the rating I need:  http://www.elna.co.jp/en/capacitor/alumi/catalog/pdf/rfs_e.pdf

I didn't see these in Mundorf MLytic either:  

 

7 hours ago, tapatrick said:

Yes its good to have sucker, flux etc but I didn't need to use them when adding the Elna caps. I added a dot of solder on the pads when I removed the 47uf ceramic tantalum as they are very small, then bedded the Elna legs which I tinned into the added spots of solder. It is fairly easy but always be careful not to over heat board or touch any other components.

 

I replaced the supplied cables with Ghent audio ones, these can be ordered in any way you need from Ghent, Y cables,  GX12 etc.. They are very good.

So I'll have to follow you tapatrick with using Nichicon UHW's.  And the 47uf ceramic tantalum, are you also referring to the DW30?  If so, you mean the blue ones below?

 

And if this is my only choice, then are these two the right ones?

https://www.mouser.com/ProductDetail/Nichicon/UHW1E152MHD/?qs=Ajmft%2bTTukE7%2FUDuqMyoRQ%3D%3D

https://www.mouser.com/ProductDetail/Nichicon/UHW1E332MHD/?qs=Ajmft%2bTTukH%2FQVH6VQFcEg%3D%3D

[Solid-State]

7 hours ago, sandyk said:

 

Has anybody tried soldering  something like a 100nF MKT capacitor under the new Elnas etc.  ?

The capacitors you are using may result in an increase in Impedance at very high frequencies, which is not what the I.C. designer intended, and why they used a ceramic capacitor there.

if you want to try this,
you dont ned to solder it on the pcb,
You can simply put on the input cable, terminal
much easier,

[sandyk]

44 minutes ago, TubeMan said:

if you want to try this,
you dont ned to solder it on the pcb,
You can simply put on the input cable, terminal
much easier,

 

 That depends on the gauge of the leads that you use to connect between modules.

You may have trouble fitting both in the same screw terminal holes.

[tapatrick]

3 hours ago, Forehaven said:

Ok, I need 25V 1500uf and 25V 3300uf caps.  Unfort. it's a no go on Elna Silmic's.  Nothing in the rating I need:  http://www.elna.co.jp/en/capacitor/alumi/catalog/pdf/rfs_e.pdf

I didn't see these in Mundorf MLytic either:  

 

So I'll have to follow you tapatrick with using Nichicon UHW's.  And the 47uf ceramic tantalum, are you also referring to the DW30?  If so, you mean the blue ones below?

 

And if this is my only choice, then are these two the right ones?

https://www.mouser.com/ProductDetail/Nichicon/UHW1E152MHD/?qs=Ajmft%2bTTukE7%2FUDuqMyoRQ%3D%3D

https://www.mouser.com/ProductDetail/Nichicon/UHW1E332MHD/?qs=Ajmft%2bTTukH%2FQVH6VQFcEg%3D%3D

Yes those are the Nichicons  I got they have a long lifespan.

 

I didn't change the 47uf ceramic caps on the Teradak DC30W only on the LT3045 boards. Good luck, be careful with the Teradak AC voltage and all that and have fun!

[tapatrick]

12 hours ago, sandyk said:

 

Has anybody tried soldering  something like a 100nF MKT capacitor under the new Elnas etc.  ?

The capacitors you are using may result in an increase in Impedance at very high frequencies, which is not what the I.C. designer intended, and why they used a ceramic capacitor there.

Alex, What would be the reason for and likely effect of adding a 100nF MKT capacitor as well?

[tapatrick]

15 hours ago, Forehaven said:

1.  Holy Shit!  LOL!! What's happened to us guys?? hehe  #sosimilar

 

 

Ha! never imagined I'd end up here for sure but one thing led to another...:) and having tremendous enjoyment from listening to music these days..

[tapatrick]

41 minutes ago, tapatrick said:

Alex, What would be the reason for and likely effect of adding a 100nF MKT capacitor as well?

I am not talking about problems, I am suggesting that they may even add a little more "air" or space between instruments etc. without unbalancing the overall tonal qualities. You have already further improved the low end .

Ah  Alex I see you answered this earlier.... thx

[BigGuy]

I have a router for various PCs I use but the only connection between it and my dedicated music server PC is when I need the connection to use JRemote to control playback.  I do no streaming, etc.  The SMPS for the router is plugged into a Tripplite power center that is separate from the APC UPS powering the music server other than they are both sourced from a shared AC duplex.

 

Would it be worth adding  a grounding umbilical to the router SMPS?

 

On another note, the APC UPS is connected to the music server PC via USB to allow diagnostic communication between the two. 
Based on recommendation from "tubelover2" (nee "rb2013"), I purchased an ART PB4x4 Pro for final filtration.  Bob had indicated putting the PB4x4 downstream of the UPS to power the PC and monitor.  However, when I tried this connectivity, the PC would  power up but not boot.  As soon as I moved the power plug for the PC to the UPS, system booted normally.

 

Any thoughts on this?

 

 

[sandyk]

6 hours ago, tapatrick said:

Alex, What would be the reason for and likely effect of adding a 100nF MKT capacitor as well?

 The designers deliberately chose a  low ESR capacitor there. The much higher value replacement electro will have a much higher ESR at 100kHZ for example.

 You MAY  retain a little more of the "air" around instruments and separation by doing this.

The original capacitor appeared to increase HF detail a little, but at the expense of the rest of the audible band.

[Cornan]

4 hours ago, BigGuy said:

I have a router for various PCs I use but the only connection between it and my dedicated music server PC is when I need the connection to use JRemote to control playback.  I do no streaming, etc.  The SMPS for the router is plugged into a Tripplite power center that is separate from the APC UPS powering the music server other than they are both sourced from a shared AC duplex.

 

Would it be worth adding  a grounding umbilical to the router SMPS?

 

On another note, the APC UPS is connected to the music server PC via USB to allow diagnostic communication between the two. 
Based on recommendation from "tubelover2" (nee "rb2013"), I purchased an ART PB4x4 Pro for final filtration.  Bob had indicated putting the PB4x4 downstream of the UPS to power the PC and monitor.  However, when I tried this connectivity, the PC would  power up but not boot.  As soon as I moved the power plug for the PC to the UPS, system booted normally.

 

Any thoughts on this?

 

 

 

It is definately worth grounding the DC- output of the SMPS for the router since otherwise the leakage can find its way via Ethernet cables into connected network devices as well.

 

Unfortunately I do not know what is going on with the ART PB4x4 Pro, but try removing loads or change the socket for the PC to another one to see if that resolved your issue.

[Cornan]

I have a question regarding changing out the 47uF ceramic capacitor on the LT3045. I did'nt see anyone recommending Jensen caps.

 

How would it be if I just change out the 47uF ceramic caps to these 47uF Jensen? https://www.ebay.com/itm/Jensen-47uf-450V-Axial-Electrolytic-Capacitor-For-Audio-/180789359450

 

 

Bad idea?

 

Also, how is RIFA caps?

 

[tapatrick]

2 hours ago, sandyk said:

 The designers deliberately chose a  low ESR capacitor there. The much higher value replacement electro will have a much higher ESR at 100kHZ for example.

 You MAY  retain a little more of the "air" around instruments and separation by doing this.

The original capacitor appeared to increase HF detail a little, but at the expense of the rest of the audible band.

Great, thanks for clarifying

[BigGuy]

3 hours ago, Cornan said:

 

It is definately worth grounding the DC- output of the SMPS for the router since otherwise the leakage can find its way via Ethernet cables into connected network devices as well.

 

Unfortunately I do not know what is going on with the ART PB4x4 Pro, but try removing loads or change the socket for the PC to another one to see if that resolved your issue.

Thanks, Cornan.

 

Will ground the DC output of the router SMPS per your recommendation.

 

I will experiment more fully with the PB4x4 Pro...seems strange!

 

Have another question after reading so much about adding LT3045 regulator(s) to SMPS setups.  Would like to increase the current output of my LPS from 1.0A now to ~2.0A.  Is there a prebuilt board with a regulator that has noise specs comparable to that of the LT3045?

This board  was suggested...  <https://www.ebay.com/itm/LM317-AC-DC-Adjustable-Voltage-Regulator-Step-down-DC-Power-Supply-Module-LED-2A/111791252026?epid=26004199188&hash=item1a0747223a:g:LzcAAOSwWnFWClIJ >

 

but have read that it has actually been used as a noise source for testing!   There are threads re modifying capacitors, etc. but even these mods do not get down as low as the LT3045.

[Cornan]

Here is a very interesting reply from JS. Especially #3. That pretty much confirms that the LT3045 do block leakage in the DC-DC path. Surely one reason why SQ is so great with those LT3045's now when my ISO Regen grounding issue is sorted out! ?

There is a lot more interesting info there as well for both LPS and SMPS.

 

 

[sandyk]

2 minutes ago, Cornan said:

 

Here is a very interesting reply from JS. Especially #3. That pretty much confirms that the LT3045 do block leakage in the DC-DC path. Surely one reason why SQ is so great with those LT3045's now when my ISO Regen grounding issue is sorted out! ?

There is a lot more interesting info there as well for both LPS and SMPS.

 

 

 

According to this line of thinking, then even a lowly LM317T regulator should do the same !

One of the causes of leakage with a Linear PSU is via the capacitance between the transformer's primary and secondary windings, then to mains earth, even without a 3 wire mains plug, as in many countries, Neutral is connected to Earth at the premises. HF Switching noise of the rectifier diodes is also coupled back through the transformer, with Schottky diodes being much better in this respect.

[Cornan]

10 hours ago, BigGuy said:

Thanks, Cornan.

 

Will ground the DC output of the router SMPS per your recommendation.

 

I will experiment more fully with the PB4x4 Pro...seems strange!

 

Have another question after reading so much about adding LT3045 regulator(s) to SMPS setups.  Would like to increase the current output of my LPS from 1.0A now to ~2.0A.  Is there a prebuilt board with a regulator that has noise specs comparable to that of the LT3045?

This board  was suggested...  <https://www.ebay.com/itm/LM317-AC-DC-Adjustable-Voltage-Regulator-Step-down-DC-Power-Supply-Module-LED-2A/111791252026?epid=26004199188&hash=item1a0747223a:g:LzcAAOSwWnFWClIJ >

 

but have read that it has actually been used as a noise source for testing!   There are threads re modifying capacitors, etc. but even these mods do not get down as low as the LT3045.

 

Your welcome BigGuy! ?

 

Since the LT3045's needs to match the voltage perfectly when used in parallel I would actually ask Alexey (eBay seller) to make me a 2A board with two 1A boards. He is a very helpful seller https://www.ebay.com/itm/LT3045-A-Ultralow-noise-0-8-Vrms-LDO-linear-regulator-0V-15V-1A-Fixed-out-/253115156739?_ul=DO

 

As you can see on my previous post. JS have confirmed that LT3045s and LT3042s is useful for the purpose of blocking DC-DC noises just like we have found here.

 

I have a LM317 board with ELNA silmic II 1000uF at the input/output. So far the LT3045s is superior in my setup. I am also going to compare the LT3045s in series with shunt regulators with Kelvin output cable and connectors in a while.

 

[Cornan]

3 minutes ago, sandyk said:

 

According to this line of thinking, then even a lowly LM317T regulator should do the same !

One of the causes of leakage with a Linear PSU is via the capacitance between the transformer's primary and secondary windings, then to mains earth, even without a 3 wire mains plug, as in many countries, Neutral is connected to Earth at the premises. HF Switching noise of the rectifier diodes is also coupled back through the transformer, with Schottky diodes being much better in this respect.

 

The LM317 + Elna is surely better than without. In a couple of days I hope to be able to make a head-to-head challenge between the LM317 + Elna and 1A LT3045s in series. So far though the LT3045s make a bigger impact on SQ overall compared to LM317 + burn-in Elna, but that could also be that there is more leakage/noises into IR and Aqvox compared to Aries Mini. The shunt regulators will probably take a while before they arrive. It will be interesting to try them as well, but it looks like they will have a brutal challenge ahead! ?

[sandyk]

I think that you may have misinterpreted what John has said ?. 

[Cornan]

1 minute ago, sandyk said:

I think that you may have misinterpreted what John has said ?. 

 In what way?

Quote:

"Putting a good linear reg after the DC-DC converter can essentially eliminate the noise on the output. In particular the LT3042/LT3045 are particularly effective in getting rid of noise from switching converter. The combination is a very useful tool, the switching converter can handle large differences in input and output voltages, producing a voltage just a little bit above the output voltage and the linear reg can clean up the output without generating too much heat."

[sandyk]

19 minutes ago, Cornan said:

In particular the LT3042/LT3045 are particularly effective in getting rid of noise from switching converter.

 

 That's because the LT3045 has a much wider bandwidth than earlier generic voltage regulators, which weren't very effective at even removing all of the ripple from SMPS which could have a switching frequency around 25kHZ or higher with some SMPS. It's also undoubtedly part of the reason why the LT3045 uses ceramic capacitors at it's input, instead of normal electros with their higher ESR at these frequencies.

To improve HF ripple rejection with earlier voltage regulators, a suitably sized choke could have been inserted in series with the input to the voltage regulator.

 It doesn't mean that an LT3045 will prevent rubbish getting back into the AC mains supply, only that it is more effective at preventing HF rubbish sailing right through the voltage regulator into the load.

[Cornan]

20 minutes ago, sandyk said:

 

 That's because the LT3045 has a much wider bandwidth than earlier generic voltage regulators, which weren't very effective at even removing all of the ripple from SMPS which could have a switching frequency around 25kHZ or higher with some SMPS.

To improve ripple rejection with earlier voltage regulators, a suitably sized choke could have been inserted in series with the input to the voltage regulator.

 It doesn't mean that an LT3045 will prevent rubbish getting back into the AC mains supply, only that it is more effective at preventing HF rubbish getting through the voltage regulator into the load.

 

I have never thought that the LT3045 could block 100% of the noises. My goal with the LT3045s in series have only been to limit the noises coming from the SMPS, through the DC cable into the powered device. Not to limit AC mains noises from going back into the SMPS. My Gopherts are all floating and they are connected to a star-earth wired PSD connected to a IT which is balanced and floating. I do not know how much damage those AC mains noises going back is causing, but I don't think it is much damage it could do when heading back. Do enlight me if you know better.

 

However, while writing this I just got an idea. How about using the two LT3045s in each end of the DC cable instead of directly in series? So, for example Gophert 7v> LT3045 6v> DC cable> LT3045 5v> Aqvox switch.

[sandyk]

An evil thought.

 

If you weren't using SMPS to power your devices, you wouldn't have to worry so much about HF ripple getting into the load ! LPS outputs mainly low level  multiples of the A.C. mains frequency along with the regulated DC voltage.

[Cornan]

Just now, sandyk said:

An evil thought.

 

If you weren't using SMPS to power your devices, you wouldn't have to worry so much about HF ripple getting into the load !LPS outputs mainly low level  multiples of the A.C. mains frequency.

 

I will surely realize that evil thought and see what comes out of it! ?

 

If those Gophert was'nt sounding so great I would have left them a long time ago. I wish I could pm you the sound waves so you could here them yourself. I am really over the moon how good it sounds right now, but I want to explore if there is some more juice to squeeze out of them. When I think that I have hit the roof I will make the final test and power my Brooklyn DAC with a Gophert as well with the same mix.

So yes, I do not leave things that I love. I care for them! ?

 

[sandyk]

Try powering your LT3045s from A 12v Li Ion battery.

[Solid-State]

or a really good LPS !!