Develop ultra capacitor power supply and LiFePO4 battery power supply

Member
Joined 2009
Paid Member
For the fifo pi clean side use universlal 3.3v conditioner by bypasing On board 5v to 3.3v regulator.
For the clean side of FifoPi I am thinking to use a ConditionerPi (or UcConditioner5V) to "pre-buffer" separate local 3V3 regs (x4) on FifoPi (x1 for XO) and DAC HAT (x3).
I prefer local separate regulation so a 3V3 conditioner is not useful to my purpose.
 
@NicMac

As I understand it the UcConditioner has the advantage of larger capacitance, while the ConditionerPi has the advantage of being closer to the load (part of the stack).

Yes, Correct!

Could a second ConditionerPi be used to buffer the 5V rail for the clean side of FifoPi and DAC HAT?

It's good to use ConditionerPi to power a RPi through GPIO.
But for FifoPi clean side, I suggest use an UcConditioner


Ian
 
As AKM DAC user, what is the ideal way to use LiFePo for 5V? Are we stucking using a regulator?

I am considering using NiCd cells, at 1.2V each this can get within the recommended supply range of 4.8V.
Thankfully NiCd seem to be some of the better performing cells after Lithium, not the ultimate battery supply but supplemented with a range of caps from ucaps to small film caps it should be be pretty close .
 
Last edited:
LinearPi 5V/3.3V 2A 0.8uV ultra-low noise PSU

Design ideas:

1. Linear full SMT 0.8uV Ultra-low noise blade powre supply
2. 5V/3.3V 2A heavy duty high current design
3. High speed soft recovery Schottky Rectifiers with option shunt cpapcitors
4. Seamless upgrading with UcConditioner for ultra capacitor power supply
5. Bult-in on/off control logic with on-board or external power switch
6. More LinearPis can work together as group with isolated master/slave control chain
7. Optional Kelvin sense input for better possible performance
8. 2oz double thickness pcb copper layer for lower ESR
9. 6V AC or 7-9V DC power input
10. Great to work with Raspberry audio application or other DAC


LinearPi1
by Ian, on Flickr

More LinearPis can be easily integrated together with master/slave on/off control

LinearPi3
by Ian, on Flickr

Upgrade with UcConditioner seamlessly

LinearPi4
by Ian, on Flickr

Ian
 
Last edited:
LinearPi is based on 4 parallel LT3045, like this one below? The price should not exceed 25USD.

O1CN01jWkwIW1gwqpKUrOIm_!!4129314207.jpg
 
@Ian,

You misunderstand me. I mean if LinearPi is based on 4 parallel LT3045, don't sell more than 25USD. I believe LinearPi's design will be better, but the details are not mentioned in your introduction, making it difficult for members to judge. I would hesitate to buy LinearPi because the details are unknown. The noise of LT3045 is 0.8uV at 500mA, and the current of LinearPi is 2A, so I guess it is based on 4 parallel LT3045.
 
Last edited:
@Studley,

Since the current of the audio device is provided by ultra capacitor, not the current provided by the charging device in front-end, no matter whether it is a battery, a bad LDO board or a great LDO board, in theory, there will be no difference between them. Don’t forget that ultra capacitor with huge capacitance has excellent filtering ability, and the noise of the charging device is unlikely to cause negative effects.

The capacitor was originally used for filtering, but the capacitance of the electrolytic capacitor is much less than ultra capacitor, which makes us no feel that the capacitor can make us ignore the difference in the front-end charging device. However, now that we have all used ultra capacitors, the front-end devices need not be too concerned.
 
Last edited:
@Studley,

Since the current of the audio device is provided by ultra capacitor, not the current provided by the charging device in front-end, no matter whether it is a battery, a bad LDO board or a great LDO board, in theory, there will be no difference between them. Don’t forget that ultra capacitor with huge capacitance has excellent filtering ability, and the noise of the charging device is unlikely to cause negative effects.

The capacitor was originally used for filtering, but the capacitance of the electrolytic capacitor is much less than ultra capacitor, which makes us no feel that the capacitor can make us ignore the difference in the front-end charging device. However, now that we have all used ultra capacitors, the front-end devices need not be too concerned.

On the basis of that logic you could use an SMPS, but Ian has already said that the UCC would not be able to filter all the noise from such a PS. Conclusion: there is a threshold level of noise for maximising the potential of the UCC. I’m assuming the LinearPi exceeds (ie betters) that threshold but it would be good to have confirmation from Ian . . .
 
Member
Joined 2009
Paid Member
@Studley,

Since the current of the audio device is provided by ultra capacitor, not the current provided by the charging device in front-end, no matter whether it is a battery, a bad LDO board or a great LDO board, in theory, there will be no difference between them. Don’t forget that ultra capacitor with huge capacitance has excellent filtering ability, and the noise of the charging device is unlikely to cause negative effects.

Are you sure of this "excellent filtering ability"?

Because of its huge capacity I would assume the the UC is an excellent DC buffer, but a very lousy filter for frequencies above just a few Hz. Therefore I assume that any noise from the PSU will flow virtually unaffected to the load if both are connected to the UC at the same time.

Does anybody have data to show on the filtering capacity of the UC on AC signals in the audio range?
 
I'd really like to order UcMateConditioner for my LifePO4 MKII board to provide clean power for Rpi4. As I keep RPI always on, I'm interested if and when will it possible to have continuous DC 5V mode enabled for MKII?
It's been mentioned that firmware update can be done. Would really appreciate to have more information before then next GB.

Ian, if you read this, please let us know.
 
On the basis of that logic you could use an SMPS, but Ian has already said that the UCC would not be able to filter all the noise from such a PS. Conclusion: there is a threshold level of noise for maximising the potential of the UCC. I’m assuming the LinearPi exceeds (ie betters) that threshold but it would be good to have confirmation from Ian . . .

@Studley,

Yes, you are right.

UcConditioner is doing very good job to eliminate the 1/f noise form the voltage rail, which was very difficult to deal with by a normal voltage regulator and also has big impact to the sound quality.

Theoretically UcConditioner can turn it to a perfect ideal power supply. But we live in a real world, we have to face the reality. The ESR of an ultra capacitor is not constant. It increases along the frequency. Though the ultra capacitor can still remove the high frequency noise(based on real test), but will be not as perfect as low frequency. In this case, good low noise linear power supplies still make difference.

Besides the ultra low noise performance, LinearPi has very good dynamic load response. That will be really great to work with UcConditioner to set up a best possible power supply for ROi and other application. I highly recommend using the on-board on/off control with keeping the AC input always connected for your audio applications.

Many other linear power supplies are also very good. You can try different ones with UcConditioner to see which is the best for your system.

Ian
 
Last edited:
Another question Ian; I want to power an USBridge Sig (the specs of which say it needs up to 3A) with the UC Conditioner. If I use the LinearPi with its 2A output there is a risk that the UCC will eventually run out of charge and drop out of conditioning mode. (This assumes the USBridge is consuming more than 2A in typical operation, but that may not be true for all I know!)

Is my understanding correct re the LinearPi?