Develop ultra capacitor power supply and LiFePO4 battery power supply

[Greg Stewart]

In Ian's GB thread, starting at post #3415, @supersurfer wrote:

Originally Posted by Supersurfer

Hi Ian,

Is it possible to modify the non isolated power supply on the battery board so it is always on?

This would be very helpfull for running distributions like Moode/volumio without the risk that the power supply is stopped on low battery and corrupting the sd card.

And a few posts later, Ian replied:

Originally Posted by iancanada

Hi Supersurfer,

It is possible,

Hardware solution would be easier. Software solution also good be need more time.

Please let me know.

Have a good weekend.
Ian

Let's continue that discussion here.

Greg in Mississippi

[iancanada]

Originally Posted by Greg Stewart

In Ian's GB thread, starting at post #3415, @supersurfer wrote:



And a few posts later, Ian replied:



Let's continue that discussion here.

Greg in Mississippi

I'll post solutions later this evening.

Regards,
Ian

[Greg Stewart]

Ian,

I was surprised when you said there might be a software solution to for 'always-on' version. Though I haven't seen the schematic, one thing I think I have figured out on your LiFePO4 supply is that each cell has it's own relay which either connects it to the charging side or to the output side... no in-between. SO I haven't yet figured out a SW solution with a single board. BUT because you have configured the boards for easy ganging, there is one software solution I can see, though not without risks... have 2 board with the controls ganged AND the outputs paralleled. Configure them so that when you turn the pair on, only one board is connected to the output while the other is left to charge if needed. Then at a pre-set voltage level for the lowest cell of the on-board, they swap. The dangerous part of this configuration is if a relay sticks connected to the output side on one board, it could have a fully charged cell trying to charge a nearly discharged one, where the charged cell would put a LOT of current into the other cell. I have some other concerns about this setup, but that is the most serious one.

AND of course the cost of a 2nd board and cells!

I can also see an always-on solution that uses a separate HW 10-relay board and provides a means to 'float-charge' the cells ala' the TirNaHiFi method. One drawback of this configuration is you need 10 separate 3.3V power supplies to keep all of the cells seperately float charged. This separate relay board would also gang to the existing battery board HW, with a connection at each cell's + & -. When you turn on the supply and connect the cells to the load, the separate relay board would connect the float-charger-supplies to each cell to keep it topped up. Of course as Ian outlined in post #334, there are some drawbacks to this. OTOH, the TirNaHiFi guys have many times commented in their setups they don't detect a sonic hit with the float-charging supplies connected. As I haven't tried this with LiFePO4 cells yet, I can't offer an opinion. I DO the float-charging setup with some of Ian's Ultracap developmental boards using Uptone Audio LPS-1.2 units as chargers. I hear only a very small amount of degradation with these, but using a $400 USD supply for each rail is more than most would want to invest.

Curious what your solutions are, Ian!

Greg in Mississippi

[Greg Stewart]

Originally Posted by bigpandahk

Ian

I am connecting the Buffalo III to your battery supply. What should I do with the onboard Placid regulators? Especially the one for DVCC/AVCC/GND/VD, thanks.

@bigpandahk,

What I believe Ian did when he configured one of these supplies to work with a BuffIII setup was to install jumpers at each of the 3.3V Trident positions to routed the supplied 3.3V directly to those areas of the board. Then for the 1.2V Trident, he left that one installed. AND the 13V rails were used to power the output stages.

Ian, am I correct?

Greg in Mississippi

[ofswitched]

Originally Posted by LinuxGeek

Who cares ? Ians PSU is not just 4 connected batteries, it features WAY more and ofcourse all the features this batteries offer... What if you stop spamming this thread with taobao stuff, that nobody wants.

It seems that you have a strong hostility towards Chinese products. However, many of the A123 batteries you have purchased are made in China...

Our life can no longer be separated from China.

In addition, the battery pack I mentioned, it supplies about 500,000 or more per year, and is widely used in cars or motorcycles which is harsh environment. In other words, it evolved over a long period of time, about 5 years, and the version that is now available is already the result of numerous modifications. Regarding the products that involve the battery, I only like to buy versions that have been used by others and have been modified several times.

I wish you luck for using new products.

[ofswitched]

Originally Posted by Greg Stewart

Ian,

I was surprised when you said there might be a software solution to for 'always-on' version. Though I haven't seen the schematic, one thing I think I have figured out on your LiFePO4 supply is that each cell has it's own relay which either connects it to the charging side or to the output side... no in-between. SO I haven't yet figured out a SW solution with a single board. BUT because you have configured the boards for easy ganging, there is one software solution I can see, though not without risks... have 2 board with the controls ganged AND the outputs paralleled. Configure them so that when you turn the pair on, only one board is connected to the output while the other is left to charge if needed. Then at a pre-set voltage level for the lowest cell of the on-board, they swap. The dangerous part of this configuration is if a relay sticks connected to the output side on one board, it could have a fully charged cell trying to charge a nearly discharged one, where the charged cell would put a LOT of current into the other cell. I have some other concerns about this setup, but that is the most serious one.

What you mean is that it would be safer if the entire system was divided into all charging or discharging. In other words, let the user decide the timing of charging and discharging, is it a safer design? In this case, it is impossible to happen that the charged cell would put a LOT of current into the other cell because the current from the external charging has an upper limit,
that the protection mechanism can also control charging accidents ,unlike a power battery can instantaneously supply up to 800A, that the protection mechanism is impossible to control, it will only produces high heat and fire.

When all charging, the user does not need to remove the battery, just simply plug the power cord into the board and press the charging button.

If Ian are more sure for the design, then evolve to automatic charging, maybe it will be better?

[ofswitched]

I am not going to criticize Ian's design. I have bought many Ian products. I love it very much. I just don't want to see the good designer Ian's reputation damaged. Because of Ian's design with the battery module in the first, an accident has caused Ian to no longer release the product. This is a very serious loss for everyone.

[iancanada]

Originally Posted by iancanada

I'll post solutions later this evening.

Regards,
Ian

To always enable the 5V DC output, there are two solutions

Hardware solution:
Cut the wire between TP16 and ARM, then connect TP16 with any pin of L9 at bottom side of PCB.

Software solution:
I'll introduce this new feature into LifePO4 mkII firmware. So, you will be able to set 5V DC output to Disable/Enable/always Enable

I'm trying to make the new firmware working with LifePO4 I also. But the software solution will need more time.

Please let me know if there are more questions.

Regards,
Ian

[iancanada]

Originally Posted by Greg Stewart

@bigpandahk,

What I believe Ian did when he configured one of these supplies to work with a BuffIII setup was to install jumpers at each of the 3.3V Trident positions to routed the supplied 3.3V directly to those areas of the board. Then for the 1.2V Trident, he left that one installed. AND the 13V rails were used to power the output stages.

Ian, am I correct?

Greg in Mississippi

@Greg,

That's correct.

Regards,
Ian

[iancanada]

Originally Posted by ofswitched

I am not going to criticize Ian's design. I have bought many Ian products. I love it very much. I just don't want to see the good designer Ian's reputation damaged. Because of Ian's design with the battery module in the first, an accident has caused Ian to no longer release the product. This is a very serious loss for everyone.

My LifePO4 power supply design is pretty safe.It has a lot of safety features. I myself have been used it for more than five months I don't have any problem. I use three of them everyday in my three different systems. The only time need to be very careful is the time of assembling/installing battery cells. Metal tweezers can cause short circuit. Caution needs to pay.

However the coming LifePO4 mkII can work with standard 26650 battery holders, so there will be no risk at all to solder them on to the PCB. And installing battery cells would be much safer/easier just by plugging in. Tabs are no longer in need.

But any way, safety would be always the first priority. That's why I put the safety notes in the first section of user's manual.

Be safe and have fun with your diy projects.
Ian