Can I Rely on Zener for Over Voltage Protection in This Circuit?

[Mechatrommer]

i'm controlling LM317 at high voltage level more than its safe 37V rating, the trick is as everybody knows is to ensure floating Vi - Vo <= 37V, so the control to ADJ pin must not go below something that will exceed this. i'm setting the safe threshold as 20V, so if the input voltage = 50V, ADJ pin must not go below 30V. while this safe threshold specification is implemented in Vset through various monitor circuit, there is possibility the opamp (blackbox) will swing out of range during if the current limit circuit is triggered (even above the Vin rail). so out of naive, i'm building active circuit limiter for the opamp output, but i find it requiring more components, not good. so i'm thinking how to simplify, zener popped in my mind! so as the title said, can i rely on a single zener to do the job? what other issue i need to take of? minimum breakdown current etc? iirc from my last play up with zener, if breakdown current is too small, it cannot ensure its zener voltage, correct? so i have to provide enough current during zener breakdown right? i hope this is practical then i can save alot of pcb space.

ps: for simplicity, please rule out "what if Vout is shorted", i'll manage that (its in another circuit already) and will ask in other thread like this if i got confused and fwiw, Vin 50V is actually variable tracking preregulator.

thanks in advance,
best regards...

[MosherIV]

Hi

Have you seen Dave's most recent video on the zener diode?

That said, I do not understand how the circuit you posted is suppose to work.
The normal way to limit the input voltage seen by the LM317 is to use a voltage divider between the ground reference and the output. I cannot understand how this works when replaced with an op-amp and R4. I cannot work out how the LM317 is supoose to achieve voltage regulation in this configuration (though it might just be me not understanding).

I would put the zener across C2, limiting the max positive output of the op-amp thereby not allowing the ground of the LM317 to be greater than 20V, so the LM317 can only see a max of 40V input voltage.

It is actually harder to try and adapt voltage regulators to have current limit as well as voltage regulation. It is easier to design and build a voltage regulator and a current limit out of 2 op-amps and OR their outputs together.

[singapol]

Have a look at this to get some ideas:

http://www.ti.com/lit/an/snva583/snva583.pdf

[Mechatrommer]

Have you seen Dave's most recent video on the zener diode?

what a coincidence. he included the protection efficiency characteristic of it, now i more confidence thanks.

That said, I do not understand how the circuit you posted is suppose to work.
The normal way to limit the input voltage seen by the LM317 is to use a voltage divider between the ground reference and the output. I cannot understand how this works when replaced with an op-amp and R4. I cannot work out how the LM317 is supoose to achieve voltage regulation in this configuration (though it might just be me not understanding).

Have you seen Dave's older video on the diy bench psu? series 224 in particular..

check minute 1. i downloaded all his series on this subject for my main reference... spanned throughout #221 - #329 and recently there was poster showing dual tracking psu from AoE 3rd Ed that shows similar setup (opamp output driving lm337 adj pin, -ve rail tracking +ve rail) i dont own AoE 3rd, so i cant say more. dave circuit is using 3080 but i tested the LM317 taking 60V and output 40-50V, regulating no problem. how it does that? i dont know either other than it will regulate Vout at anything on the adj pin + 1.25V, that all we know.

I would put the zener across C2, limiting the max positive output of the op-amp thereby not allowing the ground of the LM317 to be greater than 20V, so the LM317 can only see a max of 40V input voltage.

i believe this will defeat the purpose, the purpose above is so that adj pin will never go down far relative to Vinput of LM317, the condition (i assume) that will damage the IC. i believe with your suggestion, LM317 will smoke if Vinput is anything greater than 57V (37 - 20) relative to GND.

It is actually harder to try and adapt voltage regulators to have current limit as well as voltage regulation. It is easier to design and build a voltage regulator and a current limit out of 2 op-amps and OR their outputs together.

since you mentioned it, my current designed, and pseudo circuit i drew above, evolved from dave's original tutorial, is somewhat like you said, but not exactly ORed. now looks something like this... but it will change, feel free to comment if you spot any mistake...

[Mechatrommer]

Have a look at this to get some ideas:
http://www.ti.com/lit/an/snva583/snva583.pdf

thanks for the link, but that relies on bjts working in linear mode to bring Vin down from 170V to safe level. my psu spec is 10-20A at lower volt like 1V, that is somewhere (60 - 1) * (20) = 1200W, more than the transformer rating already. in my setup, i'll be using pre (switching) regulator to variably change Vin relative to Vset.

[Zero999]

I don't see why that won't work. I've done a similar thing before with an LM317 used as a float charger for SLAs connected to some solar panels. Although I haven't tested it under over-voltage conditions, the LM317 hasn't failed since. Before adding the zener the LM317 would often fail, presumably due to over voltage.

[singapol]

Have a look at this to get some ideas:
http://www.ti.com/lit/an/snva583/snva583.pdf

thanks for the link, but that relies on bjts working in linear mode to bring Vin down from 170V to safe level. my psu spec is 10-20A at lower volt like 1V, that is somewhere (60 - 1) * (20) = 1200W, more than the transformer rating already. in my setup, i'll be using pre (switching) regulator to variably change Vin relative to Vset.

You will need more than 2 TIP142 to get 10A.Look at safe area operation graph Fig.6 and power derating of Fig.10 of attached datasheet. Also need  one very big heatsink.

http://www.onsemi.com/pub_link/Collateral/TIP140-D.PDF

[David Hess]

The zener diode will work fine as shown to do what you want.

That said, I do not understand how the circuit you posted is suppose to work.

The normal way to limit the input voltage seen by the LM317 is to use a voltage divider between the ground reference and the output. I cannot understand how this works when replaced with an op-amp and R4. I cannot work out how the LM317 is supoose to achieve voltage regulation in this configuration (though it might just be me not understanding).

The LM317 output voltage is maintained at 1.25 volts above the adjustment pin voltage.  So the output becomes 1.25+(adjustment pin voltage) and the errors (drift, noise, whatever) in both add together.

LM317s and other regulators make fine pass elements when used like this and output stability is increased when the offset voltage at the adjust pin is more stable than the internal reference but usually feedback from the output is used which removes the influence of the internal reference.

It is actually harder to try and adapt voltage regulators to have current limit as well as voltage regulation. It is easier to design and build a voltage regulator and a current limit out of 2 op-amps and OR their outputs together.

I think it is only more difficult because the quiescent current, which may be quite high, of the integrated voltage regulator has to be taken into account and either the adjustment or output has a minimum current requirement.  If you measure the actual output current then this is not a problem.  For example the ground pin current of a 7805 type regulator can be 10 milliamps; in a 317, this minimum current is on the output and normally handled by the voltage divider.