Adding higher-powered MOSFETs to picoUPS-100 by Mini-Box - recommendations?

Thread Starter

summersab

Joined Apr 8, 2010
161
I am wondering if it would be possible to swap out the MOSFETs on the picoUPS-100 by Mini-Box for higher-powered ICs. I know it would void the warranty, but it would be great to get a little more juice out of the device:
https://www.mini-box.com/picoUPS-100-12V-DC-micro-UPS-system-battery-backup-system

I actually have an inexpensive hot air rework gun, so I know HOW to modify the board. However, my electronics education stopped with basic passives, so spec'ing out MOSFETs is a little beyond me. What properties (beyond package type) would I need to consider when looking at possible replacement MOSFETS? Would I need to change any passives on the board to support the higher power?

Also, would there be anything wrong with connecting two of these boards in parallel?

Thank you!
 

LowQCab

Joined Nov 6, 2012
4,026
Buy a larger UPS, or a second "Mini-Box", if You need more Output-Power.

If all You are looking for is longer "Run-Time",
just install a much larger Battery.
Nothing will give You longer Run-Time except for a larger Battery.
.
.
.
 

Thread Starter

summersab

Joined Apr 8, 2010
161
Buy a larger UPS, or a second "Mini-Box", if You need more Output-Power.

If all You are looking for is longer "Run-Time",
just install a much larger Battery.
Nothing will give You longer Run-Time except for a larger Battery.
.
.
.
The problem is that I haven't been able to find a comparable product (at least, not at a reasonable price-point). I'm not looking for a longer runtime - I understand that depends upon the capacity of the battery. I'm looking for a higher current output.

The picoUPS is designed for those small 12V pico ATX PSUs, but there are some 12V PSUs that are rated at 330 or even 500W. I COULD chain several picoUPS boards together, but I imagine it would be far less expensive (and less cluttered) to swap out the MOSFETs on a single board.

So, while I appreciate the suggestion, I go back to my original question: what properties do I need to consider when selecting an alternate MOSFET, and would I need to change out other passives on the board?
 

LowQCab

Joined Nov 6, 2012
4,026
Changing the Switching-FETs will make ZERO difference in Output Power.

Buy a larger UPS, or a second "Mini-Box", if You need more Output-Power.

You can't get more Power from the Batteries that You now have.

More or bigger Batteries .............
there is no way to "cheat" the Math-Equations.
.
.
.
 

PhilTilson

Joined Nov 29, 2009
131
Changing the Switching-FETs will make ZERO difference in Output Power.

Buy a larger UPS, or a second "Mini-Box", if You need more Output-Power.

You can't get more Power from the Batteries that You now have.

More or bigger Batteries .............
there is no way to "cheat" the Math-Equations.
I'm not sure that I agree with the third statement, but I do agree with the first two!

Output power, battery capacity and run-time are the three interdependent factors here; in theory, the same battery should be able to provide more power for a shorter time. HOWEVER...

To attempt to increase the power output by changing the FETs or by paralleling up boards is sure to fail. Heed the advice - buy a larger UPS!
 

DickCappels

Joined Aug 21, 2008
10,152
Power supplies are very complicated - much like an iceberg there is way more than meets the eye.

The engineer who designed the power supply probably spent a lot of time optimizing the design and checking real life performance against expectations for reliability. Anything you touch in a power supply design is likely to ripple through the rest of the design and in the end, the power supply has lower quality than the designer intended. That's why warranties become void once designs are modified.

Swapping out "bigger" parts for more performance is likely to add to the stress on other components at some point. The best overall solution is to buy a power supply that is designed for your load and application.

Written as one who designed a few power supplies for products.
 

RPLaJeunesse

Joined Jul 29, 2018
252
You might be able to swap out the FETs without cooking adjacent parts, but likely not. Bigger FETs might not switch as fast so there may not be an instantaneous transfer, with the FETs overheating during the switchover. Hopefully there is current sensing involved, that overcurrent limit needs to change. But now you can probably handle enough current to exceed the PCB trace capability. Try it and maybe watch the board catch fire!
 

Thread Starter

summersab

Joined Apr 8, 2010
161
Kinda feeling like no one actually read what I wrote in my second post. I'm not trying to increase run time. I'm trying to increase power output. It is rated for 100W, and that is currently limited by the ratings of the ICs on the board (namely, the two FDD6690A MOSFETs).

I also understand that simply swapping out the MOSFETs isn't going to be enough and will stress other components. This is why I asked what other parts I would need to change.

On the product page, the company provides a block diagram and CAD files. The picture of the board is a little low res and hard to make out what ICs are used. Would it help if I bought one and provided a diagram with a full BOM?

Also, I addressed the suggestion of "buy a bigger UPS" in the first sentence of my second post:
The problem is that I haven't been able to find a comparable product (at least, not at a reasonable price-point).
Does anyone know of a reasonably-priced product that does the same thing and is rated for 300-500W? I've looked all over, and I haven't found one, so that suggestion really isn't helpful. I wouldn't have posted here asking for help if I hadn't done any research.

EDIT:
If it helps, my application is providing a battery backup for something like one of these:
http://www.itxpower.net/en/itx300wh.htm

Most UPS solutions on the market are AC, and I'm sure you can all agree that those are stupidly inefficient: 120VAC (mains) -> 12VDC (batteries) -> 120VAC (inverter) -> ATX PSU (12VDC, 5VDC, 3.3VDC outputs) results in a lot of losses. If everything was in 12VDC, it would be far more efficient.
 
Last edited:

DickCappels

Joined Aug 21, 2008
10,152
All you can do it swap out the MOSFETs then on paper look at the gate drive requirements and the gate drive circuitry. Then look at the power supply and what components need to be made bigger (spec. wise), and then when you get everything you can think of upgraded look at the actual performance, component temperatures, currents and voltages and then take remedial action as indicated. As I said, one change can ripple through the entire design.
 

Thread Starter

summersab

Joined Apr 8, 2010
161
All you can do it swap out the MOSFETs then on paper look at the gate drive requirements and the gate drive circuitry. Then look at the power supply and what components need to be made bigger (spec. wise), and then when you get everything you can think of upgraded look at the actual performance, component temperatures, currents and voltages and then take remedial action as indicated. As I said, one change can ripple through the entire design.
Are you able to make out the markings on the gate drivers in this image?
https://resources.mini-box.com/online/PWR-PicoUPS-100/moreimages/picoUPS-100-big2.jpg

To me, it looks like 82 2413 or 8Z 2413, but that isn't turning up any parts when I search. I don't know enough about common gate driver models to recognize the markings. I can recognize the LM317 and LM2904 ICs because I'm familiar with them, though.

EDIT:
Maybe a TPS2413? That's coming up with results.

I don't know enough about power MOSFETs and gate drivers to know what specs to look for when considering alternate ICs. Any thoughts?
 
Last edited:

LowQCab

Joined Nov 6, 2012
4,026
How much Money is ~40-hours of your Time worth ?

Go to your local Thrift-Store/Goodwill-Store, or Computer-Repair-Shop and
pickup a used UPS with a dead-Battery,
then go to Walmart and buy a big Car-Battery .........

Now You have an instant, very High-Power, UPS,
that will run your Load for ~2 to ~5, or maybe even ~8 hours.
All this for around ~$75.oo total.
For a more polished look, get a Plastic "Battery-Box" for the Battery from an Auto-Parts-Store.

I've gone through 4-Car-Batteries doing this over the last ~25-years, with zero down-time.

If You want to spend some big Money, ( but well worth the cost ) ................
Get an "Optima" brand, Automotive Spiral-Wound-Sealed-Lead-Acid-Battery,
it will last at least ~6-years or more. ( around ~$250.oo each )
For another ~$50.oo, they make a "Deep-Cycle" version too,
( that would be the "Yellow-Case" version ),
which would be an even better match for UPS-Service.
.
.
.
 

Thread Starter

summersab

Joined Apr 8, 2010
161
How much Money is ~40-hours of your Time worth ?

Go to your local Thrift-Store/Goodwill-Store, or Computer-Repair-Shop and
pickup a used UPS with a dead-Battery,
then go to Walmart and buy a big Car-Battery .........

Now You have an instant, very High-Power, UPS,
that will run your Load for ~2 to ~5, or maybe even ~8 hours.
All this for around ~$75.oo total.
For a more polished look, get a Plastic "Battery-Box" for the Battery from an Auto-Parts-Store.

I've gone through 4-Car-Batteries doing this over the last ~25-years, with zero down-time.

If You want to spend some big Money, ( but well worth the cost ) ................
Get an "Optima" brand, Automotive Spiral-Wound-Sealed-Lead-Acid-Battery,
it will last at least ~6-years or more. ( around ~$250.oo each )
For another ~$50.oo, they make a "Deep-Cycle" version too,
( that would be the "Yellow-Case" version ),
which would be an even better match for UPS-Service.
.
.
.
...respectfully, the fact that you just recommended a car battery for a UPS means I can't take anything you say on this topic very seriously, anymore.

How much is 40 hours of my time worth? Well, considering that I may need about half a dozen of these and I would rather use smaller, less-expensive deep cycle batteries for the same amount of runtime due to the efficiency of the circuit, I'd say it's worthwhile.
 

RPLaJeunesse

Joined Jul 29, 2018
252
Before this discussion continues further I suggest you step back and define the problem. Right now you have defined a solution without giving us a good understanding of the problem. Using 12V DC to power a 500W (your number!) computer means LOTS of current and the attendant I2R losses. Better to go for higher voltage and much lower losses. That's why 48V is a common distribution voltage in some environments. 4x the voltage, 1/4x the current, 1/16x the I2R losses. At that 500W power level safety is another issue, and using approved equipment like a commercial UPS (with car batteries!) makes sense. A friend runs a 24V powered super-critical timing system off a simple 24V battery system, and a quality charger keeps it all alive 24/7/365. Another issue is you have not defined what "12V" means in terms of maximum and minimum, which makes it impossible to determine what is really needed.
 

LowQCab

Joined Nov 6, 2012
4,026
"" the fact that you just recommended a car battery for a UPS means I can't
take anything you say on this topic very seriously, anymore. ""


So why don't You explain how it hasn't worked perfectly for me for over ~25-years.
That's roughly ~220,000-hours without a single glitch.
And much of that was with 2-giant CRT-Monitors, ( well .... ~21" was "giant" back in the day ).

Pretty hard to argue with that kind of performance record.

The Battery was also used for a Bench-Power-Supply at the same time.

I only use Optima SLA Batteries now, ( no gassing or corrosion problems ).
.
.
.
 

Thread Starter

summersab

Joined Apr 8, 2010
161
@LowQCab, oh, dear - you quoted me in bold, red, and italic. Looks like I struck a nerve. Perhaps that wouldn't happen if you would stick to answering the actual question instead of giving alternative solutions that I explicitly explained I don't want. This is a project forum, right? I like doing things the hard way, sometimes. Maybe I want to waste 40 hours only to learn that this won't work - at least I'll have learned something.

@RPLaJeunesse, fair enough. I'd like to build a few medium-sized servers with a UPS. As I explained (repeatedly), off-the-shelf UPS units are big, bulky, and inefficient. They convert AC to DC multiple times to keep the system running, and this results in a reduced runtime and larger batteries compared to a straight DC setup.

I looked at 12VDC because that is the most common low-cost DC ATX PSU available. Yes, some datacenters used 48VDC a decade or so ago, but that has largely disappeared, and 48VDC PSUs are incredibly expensive as a result (starting at $250). There have actually been talks about switching to 380VDC for datacenters, but I don't think anyone has actually done it.

Most of the 12VDC PSUs are 80-120W and intended for people building computers for their car. That is what the picoUPS was designed to support - low-power computers running in vehicles. There are DC ATX PSUs that go up to 400-500W, and some of them support a wide range of voltages from 12-30VDC. This one provides 400W with an input of 16-30VDC, and it's only $100.

However, I can't find a suitable DC UPS. I tried searching for off-the-grid solar UPS systems (they're almost all DC/AC, which I want to avoid), backups for surveillance systems (usually low-wattage), etc - nothing fit the bill.

I figured that the picoUPS is a simple design that already provides a decent amount of power. 100W is impressive for a board like that, so perhaps modifying it to fit my needs would be an option. If nothing else, it seemed like a good place to start since it isn't a complex design. I've built and soldered my own boards before (including a custom SMD PCB), so if I could use the picoUPS as an initial design and tweak it for my needs, that could be an option.

That's where I'm at. Any thoughts?
 
Top