I have built a dual 5v 3a and 12v 1a PSU with battery backup w/ auto-switch to battery power in case of main power failure. the system has a single output line for each (5v 12V), to prevent bleedback from either supply (5v PSU [5v5a], 5v3a battery)i used diodes (Schotty which cause .7 volt drop). All this is to run a RPi4 w/4 piplates on stack. When I started this project (testing each output seperatly) i had no real problem, then when I added the diodes the output voltage(5v) dropped to 4.98. I have added a buck-boost to bring it back up but the voltage floats (4.8 - 5.2) depending on the particular experiments. Can I cure this voltage float problem with capacitors, and if so what size, how many?
Help Please, I'm almost desperate!
Bo

 

RPi4 is spec'ed for 4.75 to 5.25 volts, so you haven't taken it beyond operating parameters. If you are supplying it with 4.98 instead of 5 that is a droop of about 0.5%, quite negligible.

The optimum is 5.1V apparently, so maybe you just need to increase your 5V regulated supply by a small amount to compensate for diode loss. A following boost converter seems a complicated solution.

You might improve transient supply response with large capacitors but its not really ideal except for very transient loads. Better to improve the supply.

 

Would you want to consider using MOSFETs to do the isolation to eliminate the voltage drop of the diodes?

 

I am trying to understand what the benefit of those added diodes is supposed to be, and also just exactly where in the power supply circuit they are located. My guess, based on what I think I understand from the description, is that somehow they are to prevent "bleedback"????? some place.
So I am asking about where the diodes are in the circuit and just what they are supposed to be preventing, because presentlyit is not clear to me.

 

I am trying to understand what the benefit of those added diodes is supposed to be, and also just exactly where in the power supply circuit they are located. My guess, based on what I think I understand from the description, is that somehow they are to prevent "bleedback"????? some place.
So I am asking about where the diodes are in the circuit and just what they are supposed to be preventing, because presentlyit is not clear to me.

Basically I have 2 power supplies connected to 1 output (one is from PSU and one from battery pack), the diodes prevent voltages feeding into other supply creating excess drain.

 

Would you want to consider using MOSFETs to do the isolation to eliminate the voltage drop of the diodes?

What MOSFET's would you suggest and how many would I need as this problem only seems to be on the main PSU circuit (primary) the other supply is 2 18650 batteries which seem to not have a problem keeping steady output (stays constant @ 5.2v).

 

If you could provide a circuit for your design it would help understand where the problem might lie.

From what you describe (battery backup steady, main PSU drops under load) it sounds very like your PSU is unable to supply the needed current at an acceptably regulated voltage. Since you built the PSU you would be best to reevaluate it's design with a view to improving behaviour under load.

MOSFETs as diodes will have a very low voltage drop but this is not addressing the real issue - the voltage will still vary under load if the power supply regulation is inadequate.

 

What MOSFET's would you suggest and how many would I need

I thought your problem was from diode voltage drop, but if that's not the case, then MOSFETs won't help.

 

If you could provide a circuit for your design it would help understand where the problem might lie.

From what you describe (battery backup steady, main PSU drops under load) it sounds very like your PSU is unable to supply the needed current at an acceptably regulated voltage. Since you built the PSU you would be best to reevaluate it's design with a view to improving behaviour under load.

MOSFETs as diodes will have a very low voltage drop but this is not addressing the real issue - the voltage will still vary under load if the power supply regulation is inadequate.

The PSU portion of this setup is an open-frame 5volt 5Amp. It has a .4+/- volt adjust pot. So now I'm thinking I have a connection problem Some days I only have a .2v varience!
Thank all you guys for your input, I'll let you know what I find.
Bo (IT@Large)

 

It sounds like you are desperate to fix a problem that is probably not a problem.

Why do you feel the need to make the regulation of your 5V supply even tighter Thani it is?

 

I have seen power supplies built so that they could be used with remote sensing, with the sense terminals connected to the output terminals by a low value resistor. In one instance a tech accidentally connected the external load to the external sense terminal. The functioning was not satisfactory, and the regulation was "rather goofy". Fortunately there was no damage to the system or the supply. My point being that an inadequate sense connection will cause problems.

 

Here is a rough sketch of my system...

 

My battery charging circuit is only drawing 1.3A 5V and my actual requirements (output max) are 5volts @ 2.8Amps. Have not found any connection problems.

 

From what your posts, I understand that the 5V open-frame psu is a purchased item.

Every supply has a voltage compliance behaviour, and it seems that the one you obtained has more regulation drop under load than you like. Is that the problem you are describing? You might need to source a better supply (ie a more expensive one, probably)

Do you have a datasheet for the supply or can you give the manufacturer and part number for it?

 

I see the battery voltage claimed as 4.7 volts. Are those cells in series?? or in parallel?? A 4.7 volt battery will not "force power" back into a five volt power supply while the supply is active. In the inactive state there may be an issue.
And we have no hint as to what that "voltage supply switching" block contains, thus no way to provide any comment as to it's functions. So really, the diodes are quite redundant, unless there is some unstated purpose. The standard scheme is to have EITHER an automated change-over scheme OR diode isolation with no separate change over system.

 

Certainly the series diodes are redundant, and possibly the TS understood that. And the several blocks with no hint as to the internals make any more detailed suggestions to be pure guesses.
In addition, my mind reading skills at a distance are even worse than before.
And it is almost a week since the TS has appeared.