Develop ultra capacitor power supply and LiFePO4 battery power supply

Ian, why not just do Regulator -> LifePO4 -> Super capacitor?
The regulator will keep the battery and the capacitor charged 100% of the time and the battery+capacitor act as low impedance power delivery while also filtering out the remaining noise from the regulator.
The best of all worlds.
I never understood why people want to place a battery or super cap before the regulator.

1. LifePO4 and ultra capacitor are much better power supplies for DAC and I/V stage than any active regulator with feedback. You can google the principle;

2. Any regulator or LDO after LifePO4 or ultra capacitor will degrade the power supply performance;

3. Floating charging has two problems here:One is that the battery or ultra capacitor will work as decoupling capacitor in this case, they will switch between charging and dis-charging rapidly for dynamic load current demand. Switching between two states will cause ripple thus add noise to voltage rails. The other problem was mentioned above, yes, it's the isolation.

Regards,
Ian
 
Hi Ian - I am looking forward to your progress on this. Just as a reference, I built a quick and dirty LiPO supply for my Allo isolator and Boss DAC. The Raspberry Pi is powered by a 5 volt linear supply (which also powers the battery charging module). 2 relays are used to switch the LiPOs between parallel for charging and series for powering the isolator. A third relay switches the power out to the isolator between battery and unregulated power, so that the isolator and DAC are totally galvanically isolated from the Pi.

The relays are controlled by a GPIO pin which goes high when the player is active (it is running PiCorePlayer). When the player is quiescent, the batteries are switched into charge mode. This works well enough so that the batteries always have enough charge for extended listening.

This is pretty simplistic, so for my next project, I would like to implement your idea.

Pictures (one is the custom PCB I made for this) and schematic attached for anyone who wants to duplicate.

Regards

Whit

Hi Wtnh,

Your PCB looks very good. I think you are very happy with your battery power supply and the sound quality.

I did some thing similar years ago. But this time, I have to take a lot of things into design consideration.

Regards,
Ian
 

Attachments

  • 19_BatteryManagement.JPG
    19_BatteryManagement.JPG
    207.5 KB · Views: 3,285
  • BatteryManagementV2.5.pdf
    163 KB · Views: 445
1. LifePO4 and ultra capacitor are much better power supplies for DAC and I/V stage than any active regulator with feedback. You can google the principle;

2. Any regulator or LDO after LifePO4 or ultra capacitor will degrade the power supply performance;
I meant that you use a regulator before the battery/capacitor, not after.
3. Floating charging has two problems here:One is that the battery or ultra capacitor will work as decoupling capacitor in this case, they will switch between charging and dis-charging rapidly for dynamic load current demand. Switching between two states will cause ripple thus add noise to voltage rails.

Since the battery and/or super capacitor act as a decoupling capacitor with <1miliohm impedance how can they cause ripple?

That would be like saying that a bypass capacitor on the output of any regulator will cause ripple. It doesn't make sense.
 
Last edited:
@hellokitty123

In my design, no regulator is used before LifePO4 battery or ultra capacitor. The output voltage rails are pure and isolated.

Both LifePO4 and ultra capacitor are pretty different to normal capacitors.

LifePO4 has different charging and dis-charging curve. So, output voltage will drop or rise when switched in between.

Ultra capacitor also has slow issue when switches between charging and dis-charging, because of the material.

Designing this kind of power supply would be a bit crazy. A lot of issues have to be taken into consideration to avoid any degrading to the performance.

Regards,
Ian
 
Using a 18650 battery to charge supercapacitor and then power it directly from the supercapacitor without LDOs. This is much simpler, anyone can DIY without the PCB.

According to experience, it is better to use multiple supercapacitors in parallel.

Supercapacitors have poor voltage endurance performance. If you use a general DC power supply, even if you use LDOs, there is still a risk of damage to the supercapacitor. If LDOs are not specifically protected, LDOs are easily destroyed by the discharge of supercapacitors. Therefore, using a stable output battery to charge supercapacitor is the most ideal method.

mobile01-d926d80b7877a0e7bff84de0e284dec0.jpg
 
@sontero

The whole idea is to use LifePO4 battery directly in order to get rid of any LDO/regulator after it.

There is no problem if you DAC runs at 3.3V but with LDO which has 5V input. LDO will be bypassed automatically if the input is lower than the dropout voltage. In this case, you can still feed the 3.3V LifePO4 power into it.

However If your DAC chip runs at 5V originally, you have to use ultra capacitor power supply which will have pure 5V output.

Regards,
Ian

Can BUFFALO-IIISE PRO be used directly?

The Buffalo-IIIse Pro (Stereo Edition) 2-Channel DAC
 
Using a 18650 battery to charge supercapacitor and then power it directly from the supercapacitor without LDOs. This is much simpler, anyone can DIY without the PCB.

According to experience, it is better to use multiple supercapacitors in parallel.

Supercapacitors have poor voltage endurance performance. If you use a general DC power supply, even if you use LDOs, there is still a risk of damage to the supercapacitor. If LDOs are not specifically protected, LDOs are easily destroyed by the discharge of supercapacitors. Therefore, using a stable output battery to charge supercapacitor is the most ideal method.

View attachment 703421
Not bad. I like it.
I still have to wonder why batteries or super capacitors are better than a good regulator though. All 3 of them will be useless @ RF no? The output impedance @ audio frequencies is also similar between them. What makes batteries or super caps superior?
 
Last edited:
Not bad. I like it.
I still have to wonder why batteries or super capacitors are better than a good regulator though. All 3 of them will be useless @ RF no? The output impedance @ audio frequencies is also similar between them. What makes batteries or super caps superior?

There does seem to be a bit of assumption going on in this thread as regards the audible differences between these different types of PSU. Where's the measurements boys?
 
"There does seem to be a bit of assumption going on in this thread as regards the audible differences between these different types of PSU. Where's the measurements boys?"

Let me be the first and say it plainly: Don't need no stink'n measurements. What do you think this is, science? :)
 
Hi Wtnh,

Your PCB looks very good. I think you are very happy with your battery power supply and the sound quality.

I did some thing similar years ago. But this time, I have to take a lot of things into design consideration.

Regards,
Ian

Hi Ian - have you worked out the logic flow for the controller yet?

I started a design about a year ago to have dual LifoPO4 banks - one charging and one supplying - with a small Pic controller with relay isolation. It quickly became way too complex, particularly with having an isolated sense circuit (using a comparator) to help the logic decide when to switch banks. So I went with the simple approach I posted here, which is "good enough" for a little Allo-based DAC player.

But - I look forward to your solution because I want to build a much higher end DAC with batteries or caps for the clocks and DAC (possibly using your new Pi Fifo - keep me in mind for a GB).

It seems like the design should have some options, like music sensing to control the switching in addition to sensing battery or cap charge levels.

Cheers

Whit
 
@Whit

I'm working on the controller FW now. It's really a bit complicated. But everything is under control so far.

Yes, the charging voltage will also be monitored and managed. Not only for extending LifePO4 battery life time, but also for some "sweet spot" voltages of particular audio applications. For example, you can set the charge ending voltage at 3.30V, 3.55V or something in between for the best preferred sound quality.

How to protect relay contacts would be also part of consideration.

It will be very well integrated and I think it will be good for your high-end DAC project as an all in one power supply solution.

Regards,
Ian
 
Theredoes seem to be a bit of assumption going on in this thread as regards theaudible differences between these different types of PSU. Where's themeasurements boys?
SQ, I'm an objectivist too. That being said, have you read John Walton's (diyAudio member jackinnj) article in Linear Audio, A Comparative Overview of Power Supply Regulator Designs with Listening Tests? While the various regulators measured differently and, in listening tests, sounded differently, Walton had a difficult time correlating the measurements to the listening tests. It's a fascinating article.
In any event, and FWIW, I prefer the sound of my DAC with the master clock powered by a LiFePO4 battery rather than a regulator.


 
@Whit

I'm working on the controller FW now. It's really a bit complicated. But everything is under control so far.

Yes, the charging voltage will also be monitored and managed. Not only for extending LifePO4 battery life time, but also for some "sweet spot" voltages of particular audio applications. For example, you can set the charge ending voltage at 3.30V, 3.55V or something in between for the best preferred sound quality.

How to protect relay contacts would be also part of consideration.

It will be very well integrated and I think it will be good for your high-end DAC project as an all in one power supply solution.

Regards,
Ian

I had been thinking about the relay contact current problem also. On the charging side, using a constant current/constant voltage charging IC should limit the current through the contacts. The supply side is trickier since the downstream decoupling caps could be discharged and easily arc the contacts when they close. One solution could be a second relay in series with a low value current limiting resistor. After a suitable delay, the main relay would bypass the resistor and provide a low impedance path.

Another approach could be a low Rds-on mosfet in series with the contacts to limit the inrush. Similar to special power distribution ICs which are used in hot-swap applications (although most of these have charge pumps which would inject noise). The mosfet could provide a soft start function.

Cheers

Whit
 
I had been thinking about the relay contact current problem also. On the charging side, using a constant current/constant voltage charging IC should limit the current through the contacts. The supply side is trickier since the downstream decoupling caps could be discharged and easily arc the contacts when they close. One solution could be a second relay in series with a low value current limiting resistor. After a suitable delay, the main relay would bypass the resistor and provide a low impedance path.

Another approach could be a low Rds-on mosfet in series with the contacts to limit the inrush. Similar to special power distribution ICs which are used in hot-swap applications (although most of these have charge pumps which would inject noise). The mosfet could provide a soft start function.

Cheers

Whit

For the mosfet case, perhaps something like the attached....

Cheers

Whit
 

Attachments

  • Batt Relay Current Limit.png
    Batt Relay Current Limit.png
    79.9 KB · Views: 1,785
@Whit

I like your idea of the MOSFET solution. Very smart!

I would be glad to implement it in my design if I didn't have the relay array. Now, the array has delay relays included. The delay could be around 300ms to 1s. High current charger also has soft start feature.

Maybe I can implement your idea in my ultra capacitor power supply design later on.

Thank you so much and have a good weekend.
Ian
 
@Whit

I like your idea of the MOSFET solution. Very smart!

I would be glad to implement it in my design if I didn't have the relay array. Now, the array has delay relays included. The delay could be around 300ms to 1s. High current charger also has soft start feature.

Maybe I can implement your idea in my ultra capacitor power supply design later on.

Thank you so much and have a good weekend.
Ian

Glad you like it. The CSD17312 seems like a good choice for this because the typical Rdson (with Vgs of 3 volts) is less than 2mOhms, a lot less than typical relay contacts. The Vgsth is only 1.1 volts and it can handle 30 amps. The datasheet is here. Also, it comes in a Q5 SMD package, so not much PCB real estate needed.

Cheers

Whit