78L05 overvoltage at low current output?

[F1GWR]

Hi,

In my application I power two "74HC" ICs (nominal 6 uA supply for both when not switching nor current sinking at outputs). These drive two LEDs (7 mA while tying down) in a two-phase scheme LED1 25%, OFF 25%, LED2 25%, OFF 25 %. Therefore the 78L05 load jumps alternatively 6 uA and 7 mA every 6 ms. Power is from a stable +13 V rail with a 78L05 SMD down to 5 V (recommended 0.33 and 0.1 uF caps implemented nearby).
As per now 2+1 "74HC" ICs went dead (I suppose by exceeding absolute max voltage). Common cause may appear to be the regulator.

I found this:
https://www.mornsun-power.com/public/uploads/migrationfile/supportfile/Fixed%20Input%20Unregulated%20Output%20Series%20Product.pdf
especially that not sufficiently loaded three terminal regulators may output overvoltage?

Anyone could confirm this ?

[Siwastaja]

IIRC 78xx regulators do not have minimum load specified and I have certainly used them without any load (except a 10Mohm multimeter). Some other three-terminal regulators, like LM317, do have this minimum load spec and will output significant overvoltage if designer does not take this into account (e.g., in case of LM317, by using small enough feedback divider resistor values).

What is the source of your regulators, are you sure they are genuine? Ali/Ebay 78xx regulators are 99% guaranteed outright fakes or no-name parts without datasheet.

[F1GWR]

Thank you for your answer. Did you read the link, which says the contrary?

My supplier is reichelt.com, therefore quite trustable source.

[Chalcogenide]

The document attached does not appear to relate in any way to 78L05 regulators (the document name is "fixed input unregulated output ..." which suggest some kind of unregulated PSU module). 74HC series should be quite rugged to a bit of overvoltage; maybe the cause of death is something else (have you connected all unused inputs to ground or VDD?).

[pcprogrammer]

Thank you for your answer. Did you read the link, which says the contrary?

My supplier is reichelt.com, therefore quite trustable source.

The document from Mornsun is about their DC/DC converter modules and has nothing to do with what a 78L05 does. It only refers to linear regulators like the 78L05 as being a possible cause of problems, when the linear regulator does not have the proper decoupling capacitors.

Your problem might be improper decoupling of the 74HC IC's themself. Even with the slow on/off switching of the loads there might be ringing on the supply caused by it's own instability and inductance in the traces connecting the supply.

Often a schematic and images of the setup tells us more then a little story only telling half the tale.

[dustooff]

From the TI datasheet, page 9
https://datasheet.octopart.com/UA78L05ACDR-Texas-Instruments-datasheet-22123723.pdf
Reverse-Bias Protection
Occasionally, the input voltage to the regulator can collapse faster than the output voltage. This can occur, for
example, when the input supply is crowbarred during an output overvoltage condition. If the output voltage is
greater than approximately 7 V, the emitter-base junction of the series-pass element (internal or external) could
break down and be damaged. To prevent this, a diode shunt can be employed as shown in Figure 7.

[Siwastaja]

From the TI datasheet, page 9
https://datasheet.octopart.com/UA78L05ACDR-Texas-Instruments-datasheet-22123723.pdf
Reverse-Bias Protection
Occasionally, the input voltage to the regulator can collapse faster than the output voltage. This can occur, for
example, when the input supply is crowbarred during an output overvoltage condition. If the output voltage is
greater than approximately 7 V, the emitter-base junction of the series-pass element (internal or external) could
break down and be damaged. To prevent this, a diode shunt can be employed as shown in Figure 7.

Good point, these regulators can die as a result of input short, or large amount of output capacitance combined with other loads on input side. This is why I suggest adding said reverse current protection diode by default unless you are sure it is not needed.

[F1GWR]

Thanks to all contributors.

OK for the Morsun document. I will ignore it then.

As for the regulator, I used a 0.33 uF at the input and 0.1 uF at the output as suggested in the datasheet. Loads are:
- a 74HC132 with one Nand being used as a multivibrator (R=17 kOhms C=47 nF), one Nand as a buffer, and two Nands each driving a LED with sink current = 7 mA,
- and a 74HC112 in the middle.
That's all.
Each 74HC Vcc is decoupled by a 0.1 uF with short connections.
PCB is small (about 30x40 mm): almost no stray inductance.

the emitter-base junction of the series-pass element (internal or external) could
break down and be damaged

This is not the case as there is no significant residual voltage behind the regulator at shutdown. Anyway, even if it would, that would not explain why the 74HC's die and not the regulator.

Friends of mine suggested to place a TVS diode at the 78L05 output. I will order a 5.2 Vbr SMD and solder it on top the output cap.

Meanwhile I soldered a 6.2 V zener at the regulator output with very short leads. Crossing fingers, 48 h burn-in did not lead to any failure.

Finally my guess is: when I switch off the lab PSU (for testing purpose) or the host equipment PSU, the transformers release a back EMF that passes through the rectifier bridge and the 78L05. Maybe the TVS should then be better located at the 78L05 input?

[Siwastaja]

6.2V zener at the output has no chance to do anything useful. It would clamp somewhere around 6-9V and at that point any 5V rated IC would be damaged. If you want to protect against regulator short circuit (only worth if if you have super expensive special parts downstream), you need to use some kind of accurately controlled crowbar circuit.

TVS on input is a great idea if you expect short input overvoltage peaks.

There are crap scam "lab PSUs" which can do anything, like randomly output 30V for no reason whatsoever and kill most circuits. If yours is such way, it belongs in a dumpster. Concept of some "transformer back-EMF" causing voltage spike is made-up and no basis on reality. Transformers don't do that, and no sanely designed circuit do that kind of voltage spike.

I still suspect counterfeit 7805. Where you are buying them?