78L05 woes

[commathe]

Hey everyone,

I'm having some trouble with some 78L05s that I bought. I bought a bag of about 20, and while I haven't tested them all I have tested three and they've all given me the exact same results:
- With the pin out correct they put out 1.097v (yes, they are that consistent!)
- With the in and out pins swapped they get very hot and put out about 0.6v below my battery or power supply (this seems to fit with what I have heard about there being an internal diode)
- Looked at it on an oscilloscope and it is a very, very clean 1.097v
- I tried adding a large capacitance to the input (220u) and a smaller one (47u) at the out - small drop in output voltage to about 0.9v

What's going on? I feel like it has to be something on my end because the results are consistent between the three I tested. I'm basically following this schematic but without the diode:

[GibsonGM]

Very weird!  And you say you have no external diode on your setup...

What are you feeding it for voltage, and what load do you have on it?    (just open, measuring the terminal voltage at the output?)

You have your power supply "IN" also connected to ground, right?  Very important!  <<<

These usually work very simply, so I think you will come across a mistake pretty soon that will fix your problem

[commathe]

Yeah, power supply is connected to ground, the same ground as the ground pin of the 78L05. Measuring across the power supply I get 9.05v. I also tried loading the 78L05 (all three of them) with a 2k resistor (maybe I could try something else?) but there was no change in the voltage put out.

[greaser_au]

A couple of suggestions for things to check: 
1: triple check the pinout you are using (I went to have a look and I have just seen seen 2 pinout diagrams that will definitely cause confusion!)
2: are you 100% sure you are using the right range (DC volts) on your DMM and your DMM is OK?  (within 1 millivolt is far too consistent!!  measure a known quantity like a 9V battery to test)
and:
3: isolate/discard the ones you connected backwards (if the pinout is right, they are most likely dead now)

david

[R.G.]

I'm having some trouble with some 78L05s that I bought. I bought a bag of about 20,

Where did you buy them?

[commathe]

A couple of suggestions for things to check:  
1: triple check the pinout you are using (I went to have a look and I have just seen seen 2 pinout diagrams that will definitely cause confusion!)
2: are you 100% sure you are using the right range (DC volts) on your DMM and your DMM is OK?  (within 1 millivolt is far too consistent!!  measure a known quantity like a 9V battery to test)
and:
3: isolate/discard the ones you connected backwards (if the pinout is right, they are most likely dead now)

david

1: I cross checked against 4 different data sheets. I think I know the one you are talking about, it's looking up from the bottom! That's actually where the original pin out I put along side them in the little plastic baggie was from and I wrote it down the wrong way. Definitely put the first one in the wrong way round.
2: Yes. I measured my power supply, which I know to put out 9.05V when I have my DMM set to 20V. I was getting 1.09 on the same setting so I dropped it down out of curiosity.
3: This may be the culprit. I could actually have tested each new one backwards first, I didn't control for that.

Where did you buy them?

China. But I live here so I don't have much choice. It would also be the first time in the 5 years I've lived here that I've gotten a fake part. Unbelievable I know, but I speculate that distributors unload the fake stock they receive on overseas buyers through eBay to avoid the repercussions of selling it here. I haven't ruled out that they are fakes, but I bought them a while back and it might not have been from one of my usual vendors. Will double check this also. Such a ubiquitous part would have been easy to find from a large, well-reviewed vendor though so I don't think I bought it from anyone sketchy.

[R.G.]

Where did you buy them?

China. But I live here so I don't have much choice. It would also be the first time in the 5 years I've lived here that I've gotten a fake part. Unbelievable I know, but I speculate that distributors unload the fake stock they receive on overseas buyers through eBay to avoid the repercussions of selling it here. I haven't ruled out that they are fakes, but I bought them a while back and it might not have been from one of my usual vendors. Will double check this also. Such a ubiquitous part would have been easy to find from a large, well-reviewed vendor though so I don't think I bought it from anyone sketchy.

I think you're right - buying parts in China is probably less risky than buying them from ebay.

You're on to something - looking at the bottom versus top view of the pins might be the ticket.

[commathe]

I'm thinking what I have done is tested all of them backwards first after triple-checking the pin out and fired every one I have tested. Luckily my tested ones are still on the breadboard so when I get back I know to rule those three out and start with a fresh one while making sure once more that I have the pin-out right.

[Seljer]

Though the ground pin is in the center, so if you've flipped it around, you've covered both options regarding looking at the data sheet wrong.

Unless for some reason the pinout is wrong so its like the negative regulators LM79** where the ground pin is at the side (or like the LM317 where the adjust pin is at the side, might be worth a try to see if running it with two feedback resistors installed like with the LM317 does anything).

[commathe]

Ground pin is indeed centre. Turned out I had fried all of the ones I tested originally by putting them in the wrong way first. Now I have it right they are all working fine. Glad someone pointed out I was possibly frying them as I hadn't thought of that originally. I knew they can dissipate a ton of heat so I thought I had them the right way round when they were getting hot!

[greaser_au]

The Vo>Vin=dead thing is fairly well known- I knew not to do that in highschool,  and then proceeded to prove it by blowing up the variable PSU I built as a project.

Interestingly the 3 modern 78LXX datasheets (TI,NS & FC)  and the modern TI datasheet for 78XX do not mention the need for reverse-bias protection. I had to go to an old TI 78XX (2003) https://www.sparkfun.com/datasheets/Components/LM7805.pdf   and a 1993 NEC 78L05 to find the external reverse diode.

david

[Seljer]

Ground pin is indeed centre. Turned out I had fried all of the ones I tested originally by putting them in the wrong way first. Now I have it right they are all working fine. Glad someone pointed out I was possibly frying them as I hadn't thought of that originally. I knew they can dissipate a ton of heat so I thought I had them the right way round when they were getting hot!

For all these regulators the datasheet usually specify "...with adequate heat sinking". And they do have an electronic limiter inside to save themselves from short circuits on the output but that doesn't work in reverse.

Even with bigger TO220 cased 7805 regulator, without a heatsink you'd probably have to derate that 1.5A output current to 200mA or so (you have to do the math on how much power the regulator is actually soaking up / thermal resistances of the junction and case to see what your limits are)

[R.G.]

The Vo>Vin=dead thing is fairly well known- I knew not to do that in highschool,  and then proceeded to prove it by blowing up the variable PSU I built as a project.

Interestingly the 3 modern 78LXX datasheets (TI,NS & FC)  and the modern TI datasheet for 78XX do not mention the need for reverse-bias protection. I had to go to an old TI 78XX (2003) https://www.sparkfun.com/datasheets/Components/LM7805.pdf   and a 1993 NEC 78L05 to find the external reverse diode.

It's worth noting that if you have large electrolytic bypass caps on the output of a working regulator and short the INPUT, the output caps create the reverse bias condition. If the caps on the output have enough stored energy, they then kill the regulator just like reversed connections do. This happens in pedalboards fairly frequently.