[kevin gilmore]

the current version is the kgsshv-carbon which uses a silicon carbide output jfet with a cascaded current source. better and simpler. you can find the schematic on your own.

The GRLV is definitely NOT cheaper. extra stage at the input that subtracts out diode switching noise and a $10 voltage reference. 30 to 40 nanovolts of peak to peak output noise and voltage stability of better than .01%

I don't design based on parts price.

 

[abartels]

I understand the switching of capacitor banks, but how to make the switching glitchless, and the resulting regulation after that also bumpless seems very hard.

Just using a bypass cap for each voltage rail to damp the glitch, I use some 3300uF ones because I had them laying around. Capacitance of course depends on how much mA circuitry draws.
Most circuitry though have enough buffer caps at input I think to avoid glitch when switching between banks.

There is no so called regulation, it is just a set input voltage (i.o.w. start output voltage) and a timer. Just create enough buffer for a specific runtime.

 

[coinmaster]

I don't design based on parts price.

Me neither but for the purpose of the experiment I don't see the point in spending extra money when jungs regulator is already redundantly clean and dirt cheap. If super caps do sound better it won't be because of their superior measured performance, it will be something else.
But then again that's up to Abartels because I don't have any money to spend on this right now.

 

[abartels]

Our ultracap psu's do not measure superior. They have much higher output impedance than your jungs.
Adding ESR of ultracaps, which in our case is around 12mOhm per cell, to the relay contacts,
you end up with much higher impedance values, varying from around 10mOhm to 50mOhm,
depending on psu design and relay contact performance.

That "something else" what we are talking about is the absence of mains influence. You have to test it, hear it, to believe and understand / comprehend it.

And no, it is not very expensive, at least, not for less power hungry low voltage applications.

 

[coinmaster]

That "something else" what we are talking about is the absence of mains influence.

Half the point of a regulator is to negate mains influence anyway so that can't be it.

For starters you have a transformer between mains and the prefilter, then the prefilter (or pre-reg) smooths it out, then the regulator largely ignores whats left at the prefilter and actively smooths out the voltage at the output.

To even use a regulator you need to have reasonable voltage headroom before the regulator as well so any ripple left will have minimal impact and regs typically have pretty high psrr anyway.

I mean you could use a 1 to 1 transformer at the start of the chain and get rid of most of the mains noise just by doing that.

I once thought of maybe using LEDs strapped against a solar panel to power an amp to see how it sounded, turns out panels don't work like batteries in that they sag, but that's not necessarily a bad thing. I never ended up trying it although I do have a large stack of cells I could use if I wanted.

Actually thinking about it, it could be interesting to vary the LEDs brightness as a sort of active regulation.
Probably a pointless exercise in the end since I can't think of any benefits but interesting non the less.

If I were to take a wild guess at why super caps might sound better it might be the combination of voltage sag in phase with the signal in combination with the capacity to supply current at any reasonable frequency or duration of the signal. Something not typically found together in regulated or non regulated supplies.
But then again dacs don't need that much current.

Of course speculation is pointless until it's actually tried.

 

[abartels]

I really think this discussion is leading us nowhere, so, please lets stop this. If you want to keep spreading around the above statements without at least having tested the ultracap psu's and listened to them, this is going nowhere.
You stated before you don't have the money to test those ultracap psu's, so please, let it be.

Btw, the so called Class-A psu's from AGD are nothing more and less than superregs too, which Balazs did compare with his ultracap designs months ago.
I think you are aware of the out comings.

Balazs shared his experiences with us, which I still am very grateful for. This ultracap principle is a complete game changer, at least, for me.

 

[coinmaster]

We are telling you that the absence of mains DOES have a BIG influence. Why keeping denying this???

Um, that's what I said? One of us is confused.

If you want to keep spreading around the above statements without at least having
tested the ultracap psu's and listened to them, this is going nowhere.

The "above statements" are technical fact, I've already stated a variety of times that better measured performance doesn't necessarily mean better sound and that practical testing is all that matters in the end. Which is why I asked if you would compare the two since it's more efficient then having me replicate your project to confirm.
I'm not sure why are you getting angry?

You stated before you don't have the money to test those ultracap psu's, so please, let it be and keep using superregs in this kind of applications.

Here is a quote from you below

"If you can find me a superreg which has an infinite PSRR, and a solution to galvanic isolate that reg's
output from it's input, for example, then I will look into it, promised........."

Don't be mad if I spent a few posts giving you what you asked for.. I never said you had to do it either, you are being overly sensitive.

 

[abartels]

That was not my intention, sorry, had a rough day, but you keep pointing in your own directions instead of diving into ultracap stuff, which doesn't lead into anything at all.


Like said, galvanic isolation, from mains, that is what all is about in this case, and that absolutely isn't possible with any reg. And that exactly is what you are denying.

Edit: complete disconnected from mains instead of galvanic isolated.

Lets keep the discussion clean and ultracap related, ok?

 

[coinmaster]

Lets keep the discussion clean and ultracap related, ok?

Um, but I am. The discussion of regulators compared to capacitors is entirely relevant because the point of discussion is to reach the truth of the matter that is being discussed, not to perpetuate biases, at least if you want the best results.
I'll take it that you don't want to try the super-reg and that's fine.

I also want to hear your side of why you think galvanic isolation is needed.

Galvanic isolation is only required for safety or because the circuit requires it in order function for some specific purpose, such as connecting floating or high voltage circuits to other circuits.
Like I said before, a proper regulator flushes out any noise from the mains. The only way this is not a fact is if there is some undiscovered scientific "thing" that somehow get's through, which is not entirely impossible but highly unlikely.

Also, in line with your desire for galvanic isolation of the mains, that's exactly what a transformer does.

The transformer you use to convert mains electricity to usable voltage for your DAC is by nature galvanic isolation. Capacitors aren't galvanic isolation.

Plugging in your entire dac including its own transformer to a 1:1 transformer instead of the wall will provide double galvanic isolation and I don't think I've ever heard of an instance where this wasn't enough eliminate mains noise.

Not to mention using a mains rated cap between the windings of the transformer will short out high frequencies from mains ignoring whatever other cap you might have in the filter.

My suggestion to use a solar panel was also another method of eliminating mains from the equation.

 

[abartels]

I know transformer isolates galvanic. Still there's a lot of HF influencing the circuitry.
Maybe I shouldn't mention galvanic isolation at all, complete disconnected from mains, that is what I meant to say. Galvanic isolation in relation to regs, yes, that is important too but not existent. This would mean infinite psrr which I think does the trick.

My previous DAC was based on a VERY extended psu section (look at my pics). 10 R-Cores, 11 emi/rfi filters, all LT3042 regs with pi filters etc.
I know LT3042 is not compareable with jungs, but still very good on its own with very high psrr and very low noise figures.
Replacing the whole psu section with ultracaps was a H U G E improvement, and one can not say that I powered my dac the wrong way before,
when used a dedicated R-Core transformer - emi/rfi filter and lt3042 psu with pi filters for EACH voltage rail (6 pieces) of the dac.

It absolutely is NOT in my intention to prove or investigate or whatever, why ultracaps in these kind of designs DO sound superior.
I have some ideas, but it is not in my intention to prove to others why this is the case. If you want prove, investigate it yourself in your LAB.

I tried all kind of power supplies before, all linear, some discrete some not. For source components this, in my opinion, is the best, by far.
Why? I don't have a clue. Does it matter to me? Not really. The only thing that matters is how it sounds, and how to improve its performance.
All other stuff related to this subject isn't of any interest to me. There will be others who can and will explain why ultracaps are the way to go.
I'll bet that you can find enough information on the Uptone page, i can imagine John Swenson is explaining why his LPS-1 does sound so good.
Or, ask Vinnie Rossi about his designs, probably there's enough info on his page too which could explain why ultracaps are very interesting to use in audio products.

The bottomline is, I don't have to explain why, I don't have to investigate, for me it just is, and because it is, it stays and will be improved over time.

 

[coinmaster]

Well to be fair I wouldn't want to listen to a LT3042 either, those types of regulators are known to sound bad. All those filters don't mean much when feeding an IC reg. It's a good idea to stay away from IC regulators in hifi.

On the topic of capacitors in the power supply, I have a bit of something you may with to try.

My original amp I bought before I decided to learn electronic design used a voltage stabilizer (not a regulator, just a zener fed transistor) and it used 300uf electrolytic output capacitors. I replaced them with 275uf film capacitors shown here

http://www.mouser.com/Search/Produc...virtualkey58110000virtualkey581-FFVI6J2756KJE they have 0.9 miliohm impedance. The transformation in sound was gigantic, especially in the bass region. I did a little math trying to figure out why but I couldn't figure it out, the impedance and capacity shouldn't have created such a huge change in sound, especially in the bass give that I reduced capacity. I went as low as 100uf and it still provides the benefits which is interesting.

Even using other film caps provides the same type of benefit. From what I understand super caps are similar to lytic caps so you may want to throw a few big film caps on there and see what it does.

That being said, a super reg sounded as good or better when I swapped it, not that I did any extensive listening comparisons between the two and it may not have been the only factor contributing at the time.

Once I get some power supplies that don't kill themselves every other day I'll be able to do these kinds of tests without it being a huge p.i.t.a.

Why? I don't have a clue. Does it matter to me? Not really. The only thing that matters is how it sounds, and how to improve its performance.

Oh, well... you can't improve without understanding. From someone who supposedly designed dacs in the 90s I thought you might agree. I guess it's up to you what you want to do.

There will be others who can and will explain why ultracaps are the way to go.

I doubt it, it's like the infamous "do caps sound different" or " do cables sound different".
No one knows why they do but they just do, no matter how much you try to reason with it.

I've personally experienced differences in "cable" sound but I wouldn't say the results were conclusive. However I have no doubt that caps do make a difference in sound, it's the entire reason I got into electronic design.

I've already read the explanation on the super cap switcher supply a while ago I don't remember it being anything more than another way to skin a cat.
Id be surprised if someone discovered why super caps would sound better in such a supply since there should be no known scientific basis for it unless I've been missing out on some study on the effects super caps play on linearity in audio supplies in which case please point me in that direction I'd love to be proven wrong.

 

[bballas]

trafo- lt 3042 yes,sound bad !!
trafo- audio gd reg. much better!
ultracap- lt 3042 much better than audio gd!!!!
u.cap-lt3042 Vs. pure u.cap - not much diffrence
conclusion:
Stay away from ac mains in hifi

super regs: big , old , hot and expensive
lt 3042 : small , new , cold , a third of the price

 

[coinmaster]

conclusion:
Stay away from ac mains in hifi

Bro, that is not the conclusion you draw from that lol.
Any regulator that is not complete crap in a power supply designed with any competency is going to negate AC mains noise. This is just a scientific fact. If it sounds better it is due to another cause.
If you don't understand this then I doubt the integrity of your comparative tests can be trusted to begin with.
Clearly neither of you have the willingness, understanding, or reasonability required to move this subject further. I'm out.

 

[bballas]

somebody want to die if hear R2r 7 better than Yggy. -So funny!!!!!
Alex and me happy with ultracaps.
does it hurt you?

 

[abartels]

Hi all,

I got questions per private messages which I think I can better answer here so everybody can read this and profit from it.

Considerations before designing ultracap psu's

1- Doesn't draw the consumer circuit too much current for this kind of psu?
2- If not, investigate the upper and lower voltage threshold of the circuitry
3- Use this tool to calculate the needed capacity https://tools.eatonelectronics.com/tools/supercapcalculator
4- Build psu's, set voltage of dc-dc converters accordingly and set timer for switch time accordingly
5- Last but definitely not least, don't use more then 3 cells in series without balancing boards or balancing resistors to avoid over voltage

The basics:

Ultracaps and supercaps can deliver a constant current but do drop in voltage immediately after delivering current.
Depending on the capacity of the ultracap and the needed amperage of the consumer it does drop in voltage slower or faster.
Most ultracaps can be used at 2.7V max. Keep this in mind. Lifespan DRASTICALLY decreases when over voltage is applied, and if voltage is too high they will EXPLODE.
If you want the best of the best out of ultracaps within audio designs it is best NOT using regulators behind them, thus let them delivering directly without regulating them.
The drawback of this is that you need a big capacitance so voltage drop keeps within the margins of your consumer.

How to handle voltage drop:

Measure the current flowing thru your circuitry and note it. Investigate it's voltage thresholds (upper and lower). Some circuitry use on board regulators, thus having a very wide voltage threshold.

For example, my first ultracap psu I build was for feeding a Kali reclocker from Allo.com This Kali reclocker has regulators on board and regulates voltage down to 3.3V.
Knowing this we can feed the Kali with much higher voltages, lets say the max of 2 ultracaps in series, 5.4V The lower threshold at which the regulators still work normally would be around 4V
I tested this and Kali stil worked perfectly at 3.6 V. So, when feeding Kali reclocker with ultracap psu, starting at 5.3V (to be sure the ultracaps don't get overloaded) and let voltage drop to
3.6V, this took about 15 minutes. To be on the (very) safe side I set switching time at 10 minutes and did set starting voltage of 5.3V a little lower, at 5,1V to spare the ultracaps and extend their lifespan.

The above story shows and tells us that every ultracap psu is specific made and set for a certain consumer.

Second example:

Directly feeding voltage rails of a dac. This is a little more crucial and more difficult to manage.

The first time I created ultracap psu's for a dac was for the AKM AK4495SEQ. Since I used this dac with LT3042 psu's first, it was easy to measure their current so I could calculate needed ultracaps in relation to runtime.
This dac had 6 voltage rails consisting of 4 Analog 5V rails (which do perform best at 7V), one digital 3.3V rail and a 3.3V analog rail. To create 7V I had to put 3 ultracaps in series, thus every 7V psu consisted of 6 ultracaps.
The 3.3V rails had 2 ultracaps in series, thus 4 ultracaps per psu.
To define the voltage thresholds I consulted the AK4495SEQ whitepapers which lead me to minimum and maximum voltages. I kept all voltages between those thresholds and managed to keep runtime at 15 minutes.
Finally I did set it to 10 minutes so runtime was same as Kali reclocker and I could switch the psu's at the same time.

The second time creating ultracap psu's was with the DA-M1 modules. Since Kingwa provided me with different mA draw ratings, I wasn't sure how the end result would be and I calculated worst case scenario.
Looking at the Xilinx whitepapers, I thought this was the chip used in the DA-M1 modules which consumed the most amperage. I created ultracap psu's of 400F for these, so 4 ultracaps in parallel.
In the end this seems to result in HUGELY overkilled 1.8V psu's. I can run them at least for half an hour without seeing any voltage drop.....
The 3.3V Analog and Digital rails are fed with 100F psu's, 2 parallel sets of 2 ultrcaps in series. The 3.3V rails I start at 3.45V and saw them NEVER drop below 3.4V even after half an hour runtime.



Final word.

Why do ultracap based power supplies perform that well in audio applications? Well, it definitely is related to the absence of mains influences such as HF EMI/RFI. Further, ultracaps can deliver PURE DC,
and last but not least, they do deliver WITHOUT NOISE. Even a battery has noise because of chemical reaction, ultracaps do not have this attribute and do NOT suffering from creating a noise level when charging or discharging.

About keeping output impedance as low as possible: Many ultracaps have very low ESR figures nowadays. It is possible to create psu stages with the lowest possible ESR values, but keep in mind, if using relays or other devices
like mosfets to switch between the capacitor banks, they all inherit some sort of resistance. You have to add that resistance to your ESR values of the ultracaps. Keep this in mind when designing this kind of psu's.
To keep impedance low I did choose for dedicated relays for each voltage line, and also for each ground line. Further, I did take use of bifurcated double throw relays, and paralleled the contact pairs to half the resistance of the contacts.

Please keep in mind that relay contacts have a max current. If designing this kind of psu's, be aware of the destructive force these ultracaps have. For example, the 100F powerstor ultracaps we use have a shorting current of 225A.
No, this isn't a typo, they can deliver 225 AMPS shortly. Their pulse current performance exceeds 60 Amps!!!

That said, can you imagine what happens when i short an output or input cable at one of my four 1.8V ultracap psu's? Those are 400F, 4 cells in parallel. They can deliver a pulse current around 250 Amps and have a shorting current of 900 Amperes!!!!!!!

When tweaking my psu's inside my ΛΓienStreamer I made mistakes, several times. At the beginning I wasn't aware of the BIG danger soldering and screwing around in those while the ultracaps were charged.
Well, this did cost me about 20 relays ($4,50 each.....). While screwing around and shorting some cables, as be it VERY short, those relay contacts some sort of incinerated immediately.... they weren't there anymore, hahahaha
Keep that in mind if soldering and modifying while they are charged!

What I learned of this is: When designing ultracap psu's, and when working with relays, DO take use of DIL sockets to put those relays in (I GLADLY DID, pfew, hahaha). This makes it VERY convenient to replace defective relays.
What I forgot to keep in mind when designing ultracap psu's was: DO PLACE ALL RELAYS AT PLACES WHERE YOU CAN EASILY REPLACE THEM....... My psu's ALSO have relays INSIDE the psu's and not only on top.
Hmm, you get the point, hahahahaha

Having any questions? Please feel free to ask them.

Cheers,
Alex