(HELP) On-chip regulator can not start-up ,

[zarric]

A On-chip regulator is designed with input volatage:12-25V ,Vout=5V. Current Limit=120mA.
The schematic is shows below,



Vout is connected to outside. Vin=12V, then connect a resistor RL=100ohm to Vout. Regulator can not start-up, Vout only 17mv. RL=200omh, Vout=36mv....If RL changed to 500ohm. then regulator can normally start . IF Vin=20V. then RL=300ohm, can start.

The pass device and current sense transistor all use thick oxide high volatage nmos.threshold is about 3V. pass device:60*80u/3u, sense device:7u/3u.

 

[dick_freebird]

That schematic is illegible to me, at least the device info,
with that baby-blue-on-wite scheme.

What sets up your vbias? Does -that- start up?

That whole mess at lower center of the schematic looks
like it could kill the output device gate if anything leaked.
To me, 3um is a bit iffy for a 12V supply. I don't recognize
what that stuff is for and and it looks like a lot of potential
for misbehaving positive feedback.

If this is all just simulation, I'd cut that out. The regulator
ought to boot right off VREF and VBIAS.

If it's already silicon, if you have a cutting laser I'd
blow away the drain connection of the NFET that hangs
off the diff pair right-hand drain / output gate node.
Just to see.

I use a much simpler version of this often and don't
see why this one has all the extra junk.

 

[zarric]

That schematic is illegible to me, at least the device info,
with that baby-blue-on-wite scheme.

What sets up your vbias? Does -that- start up?

The schematic is legible ,you shoule click it and zoom in or save the image and then open it .

Connect a 100ohm resitor load ,the regulator cann't start , if then disconnect the load , regulator then start. i think vbias is work well, only regulator has bug.

 

[dick_freebird]

Zoom in on it yourself. It's not. Not even in black & white.

For someone asking advice, you don't seem very keen
on taking it.

And it doesn't matter what you think. Tell us what you
know.

Here's one last bit of advice. You have data at various
loads. Plot the I-V curve of the output, with V being
VDD-VOUT. See what that tells you. A low Vout does
not necessarily imply failure to start. In fact I bet load
has nothing to do with startup at all. Some of that mess
in the center looks like current limit circuitry.

In general you would do well to challenge what you
imagine about the circuit, rather than defending it.
Presuming you want to succeed as a designer.