Electronics > Projects, Designs, and Technical Stuff
The TLV760xx regulators are a nice piece [s]of crap[/s]
floobydust:
I would say almost every voltage regulator experiences overcurrent on start up while charging the output and load capacitance.
Are we saying this IC shouldn't have to protect itself? It seems to be using die temperature for current limit...
I've had some SMPS IC's fail due to input dV/dt being too fast. Like ignition key-on in a car, or plugging a banana jack to a bench power supply, it's a fast rise.
Sometimes I wonder why yet another linear voltage regulator IC rolls out. Do you really need 573 parts in your portfolio? :palm:
Linear Tech App Note 88 Ceramic Input Capacitors Can Cause Overvoltage Transients. On a 24V wall-wart, spike is over 40V, and TVS, snubber etc. tried.
tszaboo:
--- Quote from: bloguetronica on February 10, 2019, 05:30:09 pm ---Well, I don't want LDOs for 1 cent, but I want LDOs that won't damage parts.
--- End quote ---
FYI, the LDOs dont damage the parts. The lack of input protection damages the part, and the lack of inrush current control.
TI tells you this, there is nothing wrong with the part. The way you connect it that is the problem.
http://www.ti.com/lit/an/slva670a/slva670a.pdf
If the 10uF capacitor has for example 0.1 Ohm internal resistance, then your regulator has to source 3.3A, which is a little bit more than it is capable of.
bloguetronica:
--- Quote from: floobydust on February 10, 2019, 10:31:00 pm ---...
Are we saying this IC shouldn't have to protect itself?
...
--- End quote ---
Completely out of the context. Precisely the opposite, it should protect itself and the load. When I see comments like this, I don't even bother reading the rest. However, the graphs were insightful.
--- Quote from: NANDBlog on February 10, 2019, 10:44:10 pm ---
--- Quote from: bloguetronica on February 10, 2019, 05:30:09 pm ---Well, I don't want LDOs for 1 cent, but I want LDOs that won't damage parts.
--- End quote ---
FYI, the LDOs dont damage the parts. The lack of input protection damages the part, and the lack of inrush current control.
TI tells you this, there is nothing wrong with the part. The way you connect it that is the problem.
http://www.ti.com/lit/an/slva670a/slva670a.pdf
If the 10uF capacitor has for example 0.1 Ohm internal resistance, then your regulator has to source 3.3A, which is a little bit more than it is capable of.
--- End quote ---
I can confirm that a 22 \$\Omega\$ resistor in series and a 1uF capacitor in parallel after that seems to do the trick. Even a 4.7 \$\Omega\$ resistor and the same capacitor will help protect the regulator. It is a fact that this board, unlike other boards I've made, doesn't have a big electrolytic capacitor to help filter the spikes (I can't have one in the final application).
Kind regards, Samuel Lourenço
T3sl4co1l:
--- Quote from: bloguetronica on February 10, 2019, 11:07:49 pm ---Completely out of the context. Precisely the opposite, it should protect itself and the load. When I see comments like this, I don't even bother reading the rest. However, the graphs were insightful.
--- End quote ---
Quite in context -- if you want the regulator to do everything, you'll need to find one that's rated to do everything.
If you just need a teensy little point-of-load regulator, you'd choose the one in question here.
If you want something more robust, consider an old fashioned regulator (the power rating is higher, and the reliability is well understood), or something automotive qualified (often rated for high peak voltages, negative input and so on, and generally high reliability -- again, within ratings at least).
No one part is suitable for everything, don't try to force fit something that doesn't belong. Just keep shopping, read datasheets very carefully. :)
Tim
bloguetronica:
Ok, it seems that the resistor in series with the input does the trick. It limits the inrush current going to the regulator. The short circuit protection also works.
Although, in the final application, this section will be supplied by a 8W DC-DC converter, I'll keep the 22 \$\Omega\$ resistor for two reasons:
- The pass transistor Q2 is prone to cause transients when the load is a variable CC load;
- The DC-DC converter will cause transients if it enters hiccup mode for some reason.
The schematic for the test module is attached. I hope that there is no other phenomenon, like currents going from the OPA2180 input upsetting the AD5551, or the voltage regulator, and causing it to fail.
As a final note, I stand corrected and the TLV760 series regulators are not a piece of crap. The inrush current to that 10uF niobium cap was taking its toll, along with the transients.
Kind regards, Samuel Lourenço
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version