Super caps Ultra caps

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.

PRR

Member
Joined 2003
Paid Member
Supercaps are optimum for long storage; minutes to months.

Wall-power, we get a jolt 100/120 times a second, no long storage needed. Plain old electrolytics are still good for this.

Classic Supercaps were 5V max and you could not get the energy out in a hurry (high ESR). These limits are changing and I may be behind the curve here.
 
Good you mentionned charge balancing.
I know this about lead acid cells with car and RV batteries. That can be a serious issue.
This is better taken care by manufacturing rather than DIYing.
I wonder wether there are super caps at 15v and 48v, I am thinking about a portable supply for a microphone preamp where I need +15V -15V op amp supply and +48V phantom voltage.
Fast charging thanks to super caps would be a nice feature.
 
Last edited:
New problems for the one example manufacturer quoted above in-thread - apparently several ..!
Here you go: selling point is the 'Supercap' supply, yet utterly massive sidebands of mains noise in every measurement:
Vinnie Rossi LIO modular integrated amplifier Measurements | Stereophile.com

- it seems to me there must be some massive/primitive common-impedance errors in the basic layout.

tl;dr: It doesn't matter how 'good' your psu scheme is, if the rest of the design and pcb/psu layout is incompetent. A basic, but well-considered jellybean 3-pin reg based supply can do 30-40dB better everywhere than this example.
 
Last edited:
: LIO Technology for the Audiophile : mentionned in post #2
Is this dual super cap bank power supply an interesting technique ?

I wonder how is done the bank switch over ? How can one make a clean output at switch-over ?

The idea of isolating mains 50Hz/60Hz power with some sort of double buffering seems good, but is it actually effective ?

In a usual power supply there is a short time charging the reservoir caps ( when the rectifier diodes provide charging ) then the load is isolated from mains power ( the rectifier diodes are off ) while the caps discharge in the load; Until the next charging.
The issue is the charging phase which is dirty with high current and ringing.
One can imagine an auxiliary reservoir that is isolated from the load when recharged. Then this auxilliary reservoir would be disconnected from the dirty charger and connected to smoothly recharge the main reservoir.
This dual reservoir technique could be implemented with two banks of super caps or with two banks of electroctrolytics, as well.

I wonder if such a dual bank technique to isolate the load from the 50Hz/60Hz mains, has been tried.

I wonder wether the LIO dual super cap technique is not just a commercial gimmick.
Is there really something new ?
New about dual banks, new about super caps versus electrolytics?
 
Last edited:
New problems for the one example manufacturer quoted above in-thread - apparently several ..!
Here you go: selling point is the 'Supercap' supply, yet utterly massive sidebands of mains noise in every measurement:
Vinnie Rossi LIO modular integrated amplifier Measurements | Stereophile.com

- it seems to me there must be some massive/primitive common-impedance errors in the basic layout.

tl;dr: It doesn't matter how 'good' your psu scheme is, if the rest of the design and pcb/psu layout is incompetent. A basic, but well-considered jellybean 3-pin reg based supply can do 30-40dB better everywhere than this example.

+1. The incompetence in the design is very clear here. But that doesn’t matter for most audiophiles...unfortunately.

Best,
Anand.
 
PS with Super Caps.

In a classic power supply, is there an advantage using Super Caps instead of electrolytics.
This asks for very high rating at the diodes because of a long charging time; This is easily affordable. Then, what about ripple and ringing ?

Would it be interesting in a CRC filtering to use both technologies: Electrolytic caps and super caps.

In any case, I am aware, ground layout is of a paramount importance in high quality power supplies.
 
In a classic power supply, is there an advantage using Super Caps instead of electrolytics.
This asks for very high rating at the diodes because of a long charging time; This is easily affordable. Then, what about ripple and ringing ?

You don't consider the long charging time to be a problem other than for the rectifier diodes?

If we take the ESR of a typical 10,000uF 63V cap as the target (say 10mohm) then its jolly hard to match that with supercaps as we need a long string of them (say 20 of the 3V kind). Which means each cap's ESR cannot exceed 0.5mohm. To get that low an ESR we'll be looking at 1000F and up for each supercap. The energy stored in a 3V/1000F supercap is 500*9 = 4500J. With 20 that's a total energy storage of 90kJ.

Let's take a 500VA transformer, if we assume it can short-term put out 500W (irrespective of power factor) then we can charge our array of 20 supercaps in about 3mins. However this neglects that charging a cap through series resistance loses as much power in the R as is stored in the C. So that gives us a 6mins charge-up time. In 6 mins perhaps our trafo has gone over temperature so we probably need some kind of electronic charging circuit (similar to a power factor correction boost circuit) to get better charging efficiency and keep the transformer within its thermal SOA. The electronics also needs to maintain charge balance amongst the cap array.

The supercap supply is going to be a huge amount better in the bass than a 10mF 'lytic, but at what cost?
 
Good point, I had seen the issue of initial charging, but did not seen beyond the stress on the rectifier.
As you point, the drawbacks are:
Start up will take a long time.
If not taken care, the transformer will have time to overheat up to destruction or thermal fuse blown.
Same about the rectifier.
So this asks for a soft start and user patience.
This may be acceptable for a high power supply, then a sleep mode be handy for a quick warm start. The cold start would stay as obnoxious as a Windows at boot time.

Who knows wether, that is worth the effort and the extra cost ?
 
To compare the cost it might be worth specifying an equivalent performance supply built with normal 'lytics. If impedance down to 40Hz is the reference basis then we'd need 0.4F for the same impedance (10mohm) as a bank of supercaps.

Mouser wants $44 each for 1500F/2.7V so 20 of these comes to $880. Whereas 20 * 22,000uF @ $15 each (Chemicon) comes to only $300.

Not really any competition.
 
Its possible that the supercap version might be a little less bulky. But no, I can't see any other advantage for it. Presumably neither can poweramp manufacturers since no-one apart from Vinnie Rossi has adopted them, and he's doing it for the isolation not for getting a lower noise supply. For myself I think there might be easier ways to achieve the isolation - an air-cored transformer would be my first guess.
 
With air core you are sure of a linear inductor.
Unfortunately, when a large inductance is needed, it is likely a soft steel core is needed. Because of saturation, the core is cut to insure a gap to improve this. However, the best core material and air gap cannot make a good linear inductor.
Only air core can do and unfortunately cannot give a large inductance of a small size.
They are very easy to make and calculate. You will find many on line calculators, use it and see what I mean about inductor values and sizes.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.