7805 Voltage source .Should I use a backup Zener?

[april]

I found when designing a circuit that when for some reason the 7805 is not grounded properly it output the input voltage?
I fed it with 12 Volts in and intended the feed of 5 Volts for an ATTiny85.

There was a break in my circuit back to ground and 12 V hit the ATTiny85 destroying it .

Would a safer approach be to put a 5.5Zener across from the 7805's output to ground . Of course this will not work if it is not close to the Negative board connection . putting it across near the 7805 would not achieve much.

Are there any other approaches to use when using 12V alongside 5V chips?

 

[geobabu]

Whether you are planning to keep the Zener for the entire 5V VCC line or just for the Microcontroller. Because based on the current consumption, you need to select the appropriate zener wattage, other wise it will burn.
But any way , my suggestion is , just make broken GND path clear, and use the 7805 itself.It will be more reliable than the Zener.

Geo

 

[picbits]

If you are using it on say a development board, a 100ma fuse with a 5.5 volt zener after would offer protection. Just putting the zener on the output of the 7805 would work but you'd need a high wattage one (as stated above) and even then depending on what power supply you had driving the circuit may find it doesn't last very long.

You can get a similar problem on the LM317 when you use a pot to adjust it and have dirty wiper contacts, the output can have nasty spikes killing the circuit attached to it.

 

[Mike odom]

The main problem here is your ground layout. I take it the 5v ground goes back before the regulator ground, that's why when the regulator lost it's ground your uC got hit with 12V. If you do use a zener, put it closer to the uC than the regulator, if it's by the regulator, they might lose common ground and then you'll have the same problem.

A better approach, and one that will serve you well, is to layout your grounds better. Any grounds on the input side of the 5v regulator can be tied together. Any 5v grounds, should not be tied to ground before the regulator circuit. The practice I use is power/ground coming into the board go straight to the bulk cap on the input side of the regulator. The bulk cap power/ground go to the regulator. The output/ground of the regulator go to the bulk cap on the output of the regulator. Then, the 5v/ground to power all the circuits tie together and back to the bulk output cap only. Even though it is modern practice to just show the grounds as a symbol since they are all connected, how you connect them does make a difference. Digital circuitry is more tolerant of ground noise than analog, but learning a good power/ground layout practice will serve you well. This is the case whether you are laying out a pcb or wiring the circuit with discrete wires. This way, if your regulator chip loses ground, so does everything 'downstream' and you won't have this problem. It will be easier to troubleshoot as well.

 

[april]

Thanks , burnt out a third uC tonight. Had enough of that

 

[Inquisitive]

Could we see a schematic layout of how you are hooking it up, would make troubleshooting suggestions easier for the group?

 

[MrAl]

Hello,

There is a very simple solution for this problem. The problem stems from the fact that the input power ground stays connected to the circuit ground even when the IC chip ground becomes disconnected. So to solve this it's just a matter of making sure that when EITHER ground becomes disconnected the power turns off to the circuit. If we make one ground common to the IC and either power or circuit, we can do this quite easily. The main idea is to keep the power and circuit grounds physically separated using two different IC terminations, one for each.

With the TO3 package this is as simple as connecting the power ground to one bolt hole, the circuit ground to the other bolt hole. Do not use the same bolt hole for both grounds. This requires two small bolts (and nuts) and two lug terminals and two star type lock washers. This is more difficult to do if the package is mounted to a heat siink because one part of the packages has to be isolated from the heat sink electrically, so see the TO220 package below. An alternate method is to use TWO SEPARATE lugs and only one bolt with one lug on top of the TO3 hole (top of heat sink) and one lug on bottom of the heat sink, but that's not as sure as the two bolt method because if the bolt becomes loose it could cause the same problem.

With the TO220 and similar packages, use the center leg terminal for the circuit ground and use the bolt hole for the power ground. Do not use the center leg terminal for both grounds, nor the bolt hole for both grounds.

In either case now if one ground (or both) becomes disconnected the power turns off completely and doesnt bother the circuit at all.

For IC packages with only one lead for the ground pin (like the small TO92 three leg packages) there is no way to do this that effectively covers all possible failures because there is no way to prevent the third leg from breaking which would leave the circuit connected to power without an IC ground connection, so the solution here is DONT use regulators that only have three legs and no tab.

For IC packages with a solder tab instead of a bolt hole tab, solder the power ground to the tab and use the ground leg for the circuit ground.

For any other packages that have two separate ground terminals that are not covered above, use one terminal for power ground and the other terminal for circuit ground, and avoid using packages that do not have two separate ground terminals.

 

[Mr RB]

Thanks , burnt out a third uC tonight. Had enough of that

I have a nice 5v PSU that has a current limiting switch selecting 20/50/100mA. All you need to build a PSU like that is a 12v DC plugpack, a LM317, a 7805 regulator, and a few resistors and a switch etc.

It is handy for testing LEDs (on the 20mA current limited range) and wonderful for testing 5v microcontroller circuits thrown together in a hurry as the current limit stops anything that might fry the micro or the other parts.

Maybe you should look at making something like that, it is well worth the couple of hours needed to make it up and will last you for many years.

 

[april]

[MODNOTE]Insulting other members is against the rules....[/MODNOTE]

This circuit requires 1 amp at 12 Volts to run. I dont have a current limiting option

 

[picbits]

I often wonder why people continue to try and help you on this forum.

https://www.electro-tech-online.com/profile.php?do=addlist&userlist=ignore&u=195838

 

[WTP Pepper]

So you have burnt out three micros because of a dodgy ground. Who's to say you won't do the same with a zener and burn out another three where common sense may help.

 

[MrAl]

I often wonder why you people post.
It must be a self gratification thing. Possibly you have nothing else better to do . It must be an enjoyable passtime for old codgers with nowhere else to go
This is a convenient place to roost.
Forum Roosters or perhaps FROOSTERS

This circuit requires 1 amp at 12 Volts to run. I dont have a current limiting option

On the technical side...
Check out post #7 which gives a pretty complete picture of how to solve this.
In addition, with a TO3 package one ground can be connected to one bolt hole on the top of the package and the other ground to the other bolt hole also on the top of the package and this works even if a large heat sink is used for the TO3 package. The TO220 package is cake. The only packages this doesnt work with is the packages with ONLY three leads because there is no way to make sure the ground lead does not break near the package body, except maybe to enclose all three leads with epoxy.

And on the psychological side...
You wont understand the activities of many others until you are a bit older and wiser. I find many people of all ages like to add to discussions and many also like to learn more by reading the replies. And the older folks often have some very good advice because they've worked with a lot of different things. There's more to it than that, but that's good to start with.

 

[Mr RB]

I often wonder why you people post.
It must be a self gratification thing. Possibly you have nothing else better to do . It must be an enjoyable passtime for old codgers with nowhere else to go
This is a convenient place to roost.
Forum Roosters or perhaps FROOSTERS

What you are getting from these froosters are words of wisdom from people with so much more experience than you that they can see (and say) things that would help you even when you can't see (or have not yet seen) those things.

As for "enjoying" helping you, well you make that a little difficult.

This circuit requires 1 amp at 12 Volts to run. I dont have a current limiting option

With more experience, you might have provided that information earlier on. The information you did provide was this;
"I fed it with 12 Volts in and intended the feed of 5 Volts for an ATTiny85.

Leading people to suggest options to you like using a 100mA fuse, or a constant current source for testing a mcirocontroller circuit.

If you are going to continue being a member here and asking people with much greater experience than you to use their time to help you, then you NEED to improve your manners.

 

[april]

Rubbish-Wake up and admit the truth . Stop trying to hide it behind a veil of reasonablness and fantasy

Gee ,its even easier to burn them out when you put 24 Volts on them. Make that 4 now $10 worth They smoaoke and get hot really nicely.
This is the next stage to run them safely in a 24V 1 Amp environment

I'm thinking the best way might be to separate the boards altogether and zener across the exit point of the 24V board after the 7805.
Wonder why they don't design the 7805 to not put out when the ground is unconnected ?
Maybe there is a device available already?
Seems to me a small fet or mosfet on the output source would fix it only conducting if there is current flowing through to ground.
That would not be too hard to design surely?
Nobody else has offered any solution! Just wind ups

 

[MrAl]

Hello again,

April you need to read ALL of the posts, not just some of them
If you continue to ignore i wont help you with any circuits any more.

Cam on ban da post bai, chuc mot tuan moi vui ve va hanh phuc.
Dien dan chung ta se ngay mot manh me hon.

Rough translation:
"Thanks for the post. Wishing you a good week and good luck with the forum as we continue to learn more."

Hi,

You're welcome, and same to you, but i think you should post in English on this forum as most readers read English.
Also, you might want to say a little bit more next time

 

[ChrisP58]

Rubbish-Wake up and admit the truth . Stop trying to hide it behind a veil of reasonablness and fantasy

Gee ,its even easier to burn them out when you put 24 Volts on them. Make that 4 now $10 worth They smoaoke and get hot really nicely.
This is the next stage to run them safely in a 24V 1 Amp environment

I'm thinking the best way might be to separate the boards altogether and zener across the exit point of the 24V board after the 7805.
Wonder why they don't design the 7805 to not put out when the ground is unconnected ?
Maybe there is a device available already?
Seems to me a small fet or mosfet on the output source would fix it only conducting if there is current flowing through to ground.
That would not be too hard to design surely?
Nobody else has offered any solution! Just wind ups

The 7805 is designed to regulate it's output voltge to be 5 volts With Respect To it's Ground Pin. If if sees more than 5V, it turns down the drive to the pass transistor the output. If less than 5V, it turns the drive up.

If the ground pin is not connected to ground, then the 7805 sees less than 5V across the output, so it cranks up the output as high as it can.

The manufacturers data sheet shows how a part should be used in order for the component to work properly. If it's not hooked up that way, then you can't expect it to work as intended.

Troubleshooting of circuits is quite often a matter of finding what's not hooked up right. Broken or missing connections, shorts, or swapped wires, are always possible causes of a circuit not working. Sometimes it just 'doesn't work.' Other times, the failure is more exciting.

As far as adding a zener, a fuse, or some other form of secondary protection, that is never a bad idea. The decision is one of tradeoffs of the cost of the extra protection vs the potential cost of damaged circutry. But even those protection components need to be properly used, or they won't provide the protection expected.

 

[april]

If the ground pin is not connected to ground, then the 7805 sees less than 5V across the output, so it cranks up the output as high as it can. The manufacturers data sheet shows how a part should be used in order for the component to work properly. If it's not hooked up that way, then you can't expect it to work as intended.

Yes fair enough I'll go study it some more but that seems to be the reverse of what might be expected.
It should be looking at voltage on the input and voltage on the output and ensuring the difference never goes above its rating . I would see that as its purpose.! A comparator or some such might do it . I don't know but obviously if you put a 5 Volt supply IC it is unacceptable to get more than that out of it . Of course you are right but I am just looking to see if it could be different

 

[ChrisP58]

Yes fair enough I'll go study it some more but that seems to be the reverse of what might be expected.
It should be looking at voltage on the input and voltage on the output and ensuring the difference never goes above its rating . I would see that as its purpose.! A comparator or some such might do it . I don't know but obviously if you put a 5 Volt supply IC it is unacceptable to get more than that out of it . Of course you are right but I am just looking to see if it could be different

Without a good connection to ground, how can the regulator know what the input voltage is? It would be like trying to measure a voltage with your multimeter, with only one probe.

 

[MrAl]

Hi,

The problem is not really a good connection to ground, it's when the regulator ground lead (if only a single lead is used) comes loose. This does happen and when it does it blows out the external circuits because they can only take 5v. We dont care if the ground comes loose and it stops output a current, but the trouble is it doesnt stop it raises the voltage. So the question of how to avoid this failure mode comes up. Note that this is a failure mode analysis not a regular circuit analysis. The regulator connection(s) fail but we want to try to avoid a big failure.

To get around this it is so simple to tie the power input ground to one regulator terminal and the output circuit ground to the other ground terminal. That way if either or both grounds come off the circuit just shuts down but doesnt get destroyed. So the failure mode is made a little better. I explained this several posts back but apparently nobody understands this simple rule yet.

 

[april]

That way if either or both grounds come off the circuit just shuts down but doesnt get destroyed.

I dont understand what you are getting at.If one comes off fine if both come off you have the same situation don't you?

I am thinking of ways to be certain it does not happen . A buck converter may be my answer but if it loses earth does it shut down?

nobody mentioned these devices either LMZ14201 https://www.ti.com/product/LMZ14201
Not that cheap ,$5 odd and with so many pins and this diagram I'll have to work out what happens if it looses earth