DC/DC Converter Help

[alec_t]

The small load is about 250 mA, and the output drops from 16V to 0V in 15-20 minutes.(Very estimated guess!!)

That suggests the cut-off actually trips at 16V, but some cap in the load itself is holding charge. A test would be to monitor the supply current, or the load current, before and after the 16V point is reached.

 

[LukeKnepp]

I tested that the day before yesterday, I am testing it again now, with the added components, but I have to wait for the battery to drain that far first...

I will let you know late tonight or early Monday morning if I can

I had to add a 220 uF capacitor in order to start up the load I wanted, and then the board started smoking, I believe the MOSFET was smoking...

However, the problem is that now it will not shut off anything, so I am wondering what I ruined...

Was it the MOSFET, probaby? And if it was, is there a way to convert an N-MOSFET to a P-MOSFET?

Because I ruined my only P-MOSFET...

Thank you!!!

 

[alec_t]

I had to add a 220 uF capacitor in order to start up the load I wanted

Where did you add it? If C2 (cut-off #1) or C1 (cut-off #2) is too big the cut-off circuit won't cut off, or might oscillate (which could cause the FET to overheat).

now it will not shut off anything

As above, or perhaps the FET has a drain-source short .
And no, you can't convert a N-FET to a P-FET.

 

[LukeKnepp]

Where did you add it?

I added it in parallel to C2 in cutoff #1, so that is why it blew the MOSFET or does not shut off...

So is there another place to add it, or can I only go so large??

And no, you can't convert a N-FET to a P-FET.

I ordered some new ones that should be here in 2 days...

 

[alec_t]

So is there another place to add it, or can I only go so large?

Run some LTspice sims to find out. It will be cheaper than trying random mods to the real-world circuit and destroying FETs etc in the process .

 

[LukeKnepp]

On LTSpice, everything works great, so I cannot tell when it will actually work to turn on the load, and when it will not...

If C2 cannot be very high, then what else can I change to make so it does not shut off when the battery drops the cutoff voltage for a few milliseconds or so?
Will an input capacitor work for this, or a capacitor across the MOSFET, or anything like that?
With this 100nF C2, it cannot even start up a 75mA load, much less the 6-7A range I am looking for...

 

[alec_t]

On LTSpice, everything works great, so I cannot tell when it will actually work to turn on the load, and when it will not.

That's a failing of simple simulations. Nine times out of ten they tally with the real world; but there's always that tenth time when they don't, because component tolerances/unknowns and model approximations can't be taken fully into account .
Try C2=1uF.

 

[LukeKnepp]

That's a failing of simple simulations. Nine times out of ten they tally with the real world; but there's always that tenth time when they don't, because component tolerances/unknowns and model approximations can't be taken fully into account .

No software is perfect!!!

Try C2=1uF.

Okay, Thanks!!!

 

[LukeKnepp]

I have this circuit built, one with all through hole components, and one with the large current path in through hole, and the small resistors and capacitors are surface mount components, with the same IC.

The one with all through hole components is working like a charm!!!

However, when I apply a load to the surface mount version, then the output voltage drops to 20V...

On the surface mount version, C1, C3, L1, M1, R1, U1, and D1, are the same pieces as the working version.


So my question is, what components are not bulky enough to handle the load?

Thanks!!!

 

[Nigel Goodwin]

View attachment 120595

I have this circuit built, one with all through hole components, and one with the large current path in through hole, and the small resistors and capacitors are surface mount components, with the same IC.

The one with all through hole components is working like a charm!!!

However, when I apply a load to the surface mount version, then the output voltage drops to 20V...

On the surface mount version, C1, C3, L1, M1, R1, U1, and D1, are the same pieces as the working version.


So my question is, what components are not bulky enough to handle the load?

None, you've made a mistake somewhere.

 

[LukeKnepp]

None, you've made a mistake somewhere

Could it be that my traces are not large enough???

Or radio interferance, since I have two other converters on the same board?

 

[alec_t]

Check that R4 and R5 are the correct values and that R1 has very secure connections.
Can you put a 'scope on R1 to check that the FET is actually switching when the converter output is loaded?

 

[LukeKnepp]

Check that R4 and R5 are the correct values and that R1 has very secure connections.

Check, check, and check!

Can you put a 'scope on R1 to check that the FET is actually switching when the converter output is loaded?

I put a scope on it, this is with a few milliamps draw -

This is with 30W draw -

This is with 100W draw -

 

[LukeKnepp]

It all looks good to me here...

 

[LukeKnepp]

None, you've made a mistake somewhere.

What kind of mistake?

 

[LukeKnepp]

I GOT IT!!
R7 and C8 were not large enough!!!!!!!!!