7805 from different brands ?

[JanJansen]

Hi, i have a 7805T from ON-semi,
there are a lot of brands available, what is the best 7805 to buy ?

What brand would run coolest tempratures ?, what brand do you always buy and why ?
Should i not go for the cheapest brand ( i say brand not ebay ).

Can a 7805 break ?, doest it get very hot when broken / does it still work when broken ?
How do you break a 7805 ?

thanks in advance

[karoru]

It's a simple chip that has been made by nearly every semiconductor manufacturer since years (although I remember seeing some die photos on this forum showing that some manufacturers put something resembling LM340 as 7805, but it doesn't matter). Just go with whatever brand cheapest on digikey/mouser/farnell/tme/your dealer of choice.

Power dissipation on 7805 is difference between input and output voltage times current, and thermal resistance doesn't really depend on manufacturer - TO220 packaging is pretty standard process.

Yes, it can break of course, as any chip. It will for example emite reasonable of magic smoke when put on reverse polarity;) On more serious note, it has thermal protection so it won't fail horribly in case of overload, but it shortens the life span of a regulator.

[SeanB]

Temperature depends on current flowing through it, and voltage across it, and just how big a heatsink you have it attached to. Nothing else, and these are all the same for all manufacturers devices within small limits. It has to dissipate the power somehow, and there is no difference between brands as to how that is made, it is V differential times I.

However, there are differences in a generic "7805" from different manufacturers, many are limited in how much power they can dissipate, simply from the method used to attach the die inside to the copper tab, and others are limited in the maximum input voltage they can handle ( often under 12V for some lesser known manufacturers) and in the action they have under overcurrent conditions, some limit to a maximum current, some fold back to a lower current and some just blow up with a shorted output. As well they often have different behaviour when they get too hot, some work correctly and shut down, others just keep on cooking till they go short circuit input to output and others just blow up after going short circuit .

All this is mentioned in the datasheets, so best is to buy reputable manufacturer ones from a reputable dealer, not a low feedback new eBay account seller with a some "ST", "Onsemi", "NXP" or other brand part that most likely is either a counterfeit, reject or otherwise mostly defective part.

As to how to kill them it is simple. apply too much differential voltage ( ie, 35V input and want 5v output at 0.2A which will overheat any of them), draw too much current ( 1.5A from a 1A part with 10V input for long period) or overheat them ( 5V at 1A, input 12V, with a heatsink of 20C/W rating which is most small clip on ones) so that the chip runs at over 150C inside the package, or apply a voltage to the output and nothing to the input and leave the 3 diodes recommended in the datasheet out. Reverse voltage also kills them fast.

[JanJansen]

Ok, thanks i will keep buying the cheapest.

I have build my project, only the 7805 gets to hot.
I build my project twice and the 7805 stays normal.
There must be some broken part in the first build, or should i replace the 7805 ?, can it be broken, or wont it work at all when broken ?

[Zero999]

Check the connections, especially to the LM7805.

Measure the output voltage.

Then, measure the input and output current to the LM7805: if the input current is more than 10mA greater than the output current, then the LM7805 is faulty. If the output current is high enough to cause the LM7805 to get too hot and shut down, then either you need a bigger heat sink or your project is faulty.

[Bloch]

Hi JanJansen

I have build my project, only the 7805 gets to hot.
I build my project twice and the 7805 stays normal.
There must be some broken part in the first build, or should i replace the 7805 ?

It sound like a problem i myself did run into for about 20 years ago 

I did not use decoupling cabs close before and after the regulator 7805.

If not used the IC can oscillate it self that it get so hot that you cant touch it.

Try 330nF on the input and 100nF on the output.

[JanJansen]

Hi, i have the capacitors, only not very close to the IC,
where the power comes in i have a 2200uF + 100nF, then 10 or 20 cm nothing, then the 7805.
I will try adding capacitors very close, only it is the same as my second build project, need to try something at least.

And how do you measure input and output current without desoldering ?, can you measure that just like volts with proper equipment ?

[The Soulman]

You could add in 1 resistor and measure the voltage drop across it with your multimeter, 1mV equals 1mA (assuming a perfect 1 resistor).
Also don't exceed the resistor power rating typically 0,25 W for those little generic standard ones, so (absolute) max 0,5 A for 1 0,25W
You could use a lower value for less power dissipation or multiple's paralleled to be able to dissipate the power, or simply a resistor with higher power rating.
You do the math! (ohms law) 

[DimitriP]

Even each brand may have more than one manufacturing plant and even within the same plant there could be a batch that is marginal.
To put your own mind at ease, you could order from a reputable vendor a few from different brands and do your own testing.

[David Hess]

(although I remember seeing some die photos on this forum showing that some manufacturers put something resembling LM340 as 7805, but it doesn't matter).

This was an excellent discussion of the 7805 and LM340 variations with die photos but of course Google destroyed the images:

http://www.righto.com/2014/09/reverse-engineering-counterfeit-7805.html

[Fisher77]

I get mine from taydaelectronics.com. The TO-220 package is $0.23 cents , and the TO-92 package is $0.09 cents.  I have used several of them and other parts from tayda, and have had no problems yet.

[jitter]

Ok, thanks i will keep buying the cheapest.

Uhmm, even a reputable brand (ST, OnSemi, Fairchild, National) should be cheap enough to not want to buy an unknown quality second supplier. I mean, would you trust one of these in your device?:
Care to guess the brand and era it was made in?

[CJay]

This was an excellent discussion of the 7805 and LM340 variations with die photos but of course Google destroyed the images:

http://www.righto.com/2014/09/reverse-engineering-counterfeit-7805.html

Very impressed with that page, especially the interactive die image.

[CJay]

Hi JanJansen

I have build my project, only the 7805 gets to hot.
I build my project twice and the 7805 stays normal.
There must be some broken part in the first build, or should i replace the 7805 ?

It sound like a problem i myself did run into for about 20 years ago 

I did not use decoupling cabs close before and after the regulator 7805.

If not used the IC can oscillate it self that it get so hot that you cant touch it.

Try 330nF on the input and 100nF on the output.

I've run across problems with 78xx devices oscillating a few times (even though some datasheets suggest you don't need capacitors IIRC), one that springs to mind was in a 24V-12V 'dropper' from a batch of 50 or so that came in for repair.

When that one dropper was powered up it wiped out the VHF FM radio in the workshop and, weirdly, measuring the output voltage with a digital meter gave different readings depending where on the output leads you measured indicating it was producing RF, enough to blank out the unfortunate truckers radio.

Turns out the 7812 that was being used to drive some 2N3055s (or similar, it was a *long* time ago) was oscillating and producing all sorts of rubbish. Couple of caps soldered to the 7812 leads cured it but changing out the chip for a different brand made the problem go away without the capacitors.

We did, of course, fit capacitors to the whole batch.

[tatus1969]

Ok, thanks i will keep buying the cheapest.

I have build my project, only the 7805 gets to hot.
I build my project twice and the 7805 stays normal.
There must be some broken part in the first build, or should i replace the 7805 ?, can it be broken, or wont it work at all when broken ?

Unless you have picked a faked or broken regulator, it does only produce very little heat on its own. When you have two copies of the same board and one of them has a hot 7805, then this is an  indication that the rest of your circuit is drawing excessive current on the one board. I would check current consumption of both with a multimeter, between output of 7805 and your circuit.

[JanJansen]

Hi, thanks for the replys all.
I have this new PSU, wich shows the mAh usage,
i hooked it up, and shows the same mAh as the breadboard, so it must be those capacitors that are missing.

Do i realy need 330nF at the input ?, or can i have a 100nF instead ?
Also when using heatsink, there is no way to place those caps close, else i have to place it behind the 7805, does this matters ?

[TimFox]

The original data sheets specified minimum 330 nF.  I always used 470 nF X7R ceramic, since we had lots of them in engineering stock.  My colleagues who omitted the capacitor often encountered oscillation (somewhere between 100 kHz and 1 MHz).

[Siwastaja]

I always use a combination of a small 100nF to 220nF X7R ceramic very close to the input, and an electrolytic capacitor in range of 47 uF to 2200 uF, depending on current, voltage drop across the regulator, expected cable inductance and expected ripple of the supply supplying the 7805. ESR of the electrolytic capacitor provides damping for the parasitic inductance, while the small ceramic offers very low inductance.

I have never had a 7805 oscillate, but I have never omitted the capacitors.

When in doubt, follow the datasheet recommendations, even though they are not always optimal.

With 78xx regulators, never ever use Ebay/Alibaba/similar sources, even if you like taking that risk with some other parts. With these parts, the fake rate is very close to 100%, and they often do work, but are inferior in specs. Also, these regulators are extremely cheap (around $0.10-$0.20 each in small volumes) even when you buy genuine ones from proper distributors.

[JanJansen]

I placed 2 extra 100nF very close in front and behind the regulator, also 100nF on the output.
Thing still getting very hot, i have no clue, measured everything, i looked for hours.

The weird thing is everything works perfect, so i cannot find the problem.

[Monkeh]

Thing still getting very hot, i have no clue, measured everything, i looked for hours.

What is your input voltage and current?

[KJ-90]

Does anyone out there know why the metal tabs on 78xx and 79xx series voltage regulators are getting so thin? When I was in high school about a decade and a half ago, the tab was about 1.2mm thick and you could see the copper base on the side. It was a National Semiconductor part, another T.I. But today, the tab on the standard known brands is only about 0.5mm! Is the die designed to run so hot safely now- or is this about cost efficiency instead? Yet, all of my 78 and 79 series from the late 70s-90s all had the thicker tab. Any ideas?

[Monkeh]

Does anyone out there know why the metal tabs on 78xx and 79xx series voltage regulators are getting so thin? When I was in high school about a decade and a half ago, the tab was about 1.2mm thick and you could see the copper base on the side. It was a National Semiconductor part, another T.I. But today, the tab on the standard known brands is only about 0.5mm! Is the die designed to run so hot safely now- or is this about cost efficiency instead? Yet, all of my 78 and 79 series from the late 70s-90s all had the thicker tab. Any ideas?

It's cheaper and not a problem thermally, only mechanically. They are to be used with heatsinks, remember..

[rainsee]

Then, measure the input and output current to the LM7805: if the input current is more than 10mA greater than the output current, then the LM7805 is faulty. If the output current is high enough to cause the LM7805 to get too hot and shut down, then either you need a bigger heat sink or your project is faulty.

[JanJansen]

Thing still getting very hot, i have no clue, measured everything, i looked for hours.

What is your input voltage and current?

The voltage is 12, the milliampere usage is 83 for the whole thing, including a 4 x opamp package on 12 volt.

Then, measure the input and output current to the LM7805: if the input current is more than 10mA greater than the output current, then the LM7805 is faulty. If the output current is high enough to cause the LM7805 to get too hot and shut down, then either you need a bigger heat sink or your project is faulty.

I build it twice, the other one stays cool,
i cannot measure because it is soldered already, and i dont have 2 multimeters, i could remove the 4x opamp to read the PSU, and measure.
Only the milliampere usage is the same as in breadboard and the other soldered thing.

I will keep it laying in the house until i found whats happening, measuring the (-) from start to end gives 0,02 ohm, that is also normal for the other project.

+ i am sure i have all floating inputs on the 7414 connected to (-).
i should try remove the optocoupler if i can get it loose, only if it works, why would it be broken ?

Project goes :

12v input to 2200 Uf elko + 100n
goes to 4 x opamp
goes to 5v regulator with 2 more 100n very close

5v goes to 7414 & optocoupler
5v goes to 3,3v regulator

3,3v goes to DSPIC & rail2rail opamp.

thats it

[digsys]

Does anyone out there know why the metal tabs on 78xx and 79xx series voltage regulators are getting so thin? ...

We got caught out on that a few months ago. You can get them in either thickness tab, but the thicker tab IS becoming harder to find !!
In our case, we started finding shorts. Problem was, we use a silpad type thermal insulator, on a tapped hole. The tab was so flimsy, it was being pulled into the
slight chamfer of the hole, and the silpad couldn't stop it. We had to go back to mica, to use up all the thin tabbed regs.
Due to the diminishing availability of the thicker tabs, I've switched regulator type, just not worth the risk in production.