Strange behaviour of 78xx reg.

[Staigen1]

Hi all

Tony_Stoynov have encountered some strange behaviour in some 78xx regulators. I dont recognize
this, but maybee someone else have had this problem too. We let Tony describe the problems he
had, and he's remedys. Maybee someone also know why.

//Staigen

 

[Tony_Stoynov]

Hi to all,
yes, for my practice i see a several stranges in 78xx behavior.
1. This i think all must know, must have a good frequncy capacitor, ceramic 47 or 100 nF is good, this capacitor must be in input and utput, to prewent oscilation process.

2. The temperatute problem, in NS or ST documentation can see that the operating range is up to 150 C. I can sugest to calculate the working condition up to the maximum, but i have the problem at a considerable low temperature, no more that 80 C. I not say that i see this promblem with NS or ST chip, but i bay chip from distributor, put in the bord and start the device. The result is that after several minutes,i see the problem with a LCD. The first thing in a minde is for a software problems, becouse the RS comunication work correctly, and the LED indicator work. I check all soft, and finally i open the device case to change the display .. may be .. no matter how i finde the problem, but finally the problem was unstable voltage. When i connect the scope, and close the device box, after several minutes i see pulsations, may be around 1 - 0.5 Hz. I mesure the 78xx temperature with closed box, and if i remember right, the temperature was around 80 C. I solve this problem with a small heatsink, but on other device i have similar power over 78xx, but i not see this problem. The my conclusion is that some chips, not wrk very good on hight temperature.

3. Input voltage, agin acording to NS or ST documentatin, the max input voltage for 7805 is 35 V, for 20,24 series in t documentation the max input voltage is 40 V. But ... this is deend from the manufacture for example, in some east europian parts, this voltage is a 12 V. I find that this is not only in some of east europian part, like a Bulgarian version or MA7805 version /i think this part not in the production now/ but in some other chips, i not write the number, but i buy this device from big distributor.

4. The noise attenuation. If have a some interference on the power line, and if this is a high frequency interference, speccially pulse interference, is a very possible this interference going directly to the output, whit a very low attenuation. The one of the big problem that i see in this aspect, was wit a device that controled 220 V realy, and the efect is that eery time when i switch off the relay, the pic going to reset. When i put the scope to power line, i see a strong impusl, that restat may processor, no matter that i have a 1 k from reset to Vcc and i am use the internal reset ciruit. I tray to use the LC filter, but the effect is to same, i solve the prblem in prototype with LC filter befor 220 V transformer, but i the serial production dvice i change 7805 with a 2575-5.0, in standart schem, without any extra filters.

Ok this is some of stranges in behavior. For the first one, you can see info in application note,this is easy to be solved. Second, and third problem, i know that this is depen from the manufacture, now i buy only this device that i have documentation for, and that hve a good quality. The last stranges, actually ths is not a stranges, this is a chip parameter, but many people look at data sheet, and the see that ST give around 60 dB attenuation for 100 k, or NS give 45 dB at to same frequency, but this is attenuation for sin signal,not for impuls with a steep front.

 

[ante1]

Hi Tony,

The capacitor at the input prevents oscillation but the one at the output is supposed to take care of load variations. Both of these must be connected close to the regulator or they will be less efficient or useless. If you still have problems with oscillation try to add a 0.33u tantalum across the input (short leads).

If you got a surface temperature of 80C your chip inside is at a much higher temperature! I consider 80C as an unhealthy temperature for any circuit, try at least to get below 70C or better. The pulsation you mention might be cured with a low ESR cap at the regulators output.

You should use a mains filter for the incoming power and a suppressor link on any relay.

A circuit diagram and a layout of your devise would be helpful if you need more help.

 

[Tony_Stoynov]

Hi,
thanks for th post, i know all this, but i just what to share some my  experiance.
About temperature, yes the crystal temperatur is higer, but this is temperature for junction in data sheet, the thermal resistanse junction-case fot TO-220 is a 5 C/W. When the package is 80 C, the junction temp not be very high, specially in the case that i not have a heathsink. The power over device is no more that 2-3 W, this give the 15 C different between case temeprature and juction temperature.
This problem i think is connect to chip manufacture, sombody else can be have a similar problems, i just share some strangese from my experience.
About noise problem, ESR not sole the problem, actually the low ESR electo have a too big inductance for this signl, i put a tantal capacito, put a several eramic, but this not solve the problem, only LC filter befor transformer.
Yes, i not write this, but i has a suprestor on the relay, Epcos suprestor, 380 V  .... but this not help me.
About main filter, yes, how i say this solved the problem, but this was a more costlly decision, i just change the regulator fom linear to switchin mode.

 

[audioguru2]

With a 30 degrees C ambient temperature, at a power dissipation of only 1.75W, a 78xx regulator in a TO-220 case without a heatsink is at it max allowable operating temp of 125 degrees C for the chip and 118 degrees C for its case.
With a power dissipation of 2.22W then the chip reaches 150 degrees C and causes the regulator to have thermal cutoff.

 

[Tony_Stoynov]

Hi,
yes, you a right, but i give this 2-3 W as a value to clculate eventual different between case temperature and crystal temperature, im sure for the case temperature, becose i use a thermometer. The chip separate some power via the PCB. And total power over the chip .... i not remember, in this device i have one PIC at 20 Mhz, one MAX232 and one LCD without backlight .. and several LED, let say that led have a 10 mA consumation. The PIC, MAX and LCD not have mre that 40-50 the result is 60, the input vltage was 24, then power over the regulator is (24-5)*0.05= 0.95 W. I dont know what is the PCB termal resistance, it's no very small, but now i thnk that 80 C is a normal temperature, in the case the chip must be aroun 85, that si twice less that maxium.
I know abut thermal cutoff, but in this case i think have not a reason for this, i create this project many yaer ago, and have not a pcb to check again the chip temperature, no matter, i think this is a not a normal chip behavior. I have this problem only one time, but this can cost a time, after this i chek not only for the bulgarian chips, becouse they not in the production, but still can find it in stock, i look for chip from good company.

But thanks for the answer, this can be very helpfully for sombody, i remebmer .. may be 15 or more yars agoo i made a audio amplifer with a OU and BD139, then i los some time to understand what is th powr, termar resistance, and how they reflected over the ca temperature , from this moment, to today, i made a thermal calculation for every circuit, and if  more complex with a active cooling, use the FlowWorks soft.

 

[ante1]

Hi Tony,

When I build something I like to be able to touch any part of it with my fingers without getting burns. Any excessive heat is a waste of energy and a reduction of lifetime for the circuit. However this “technique” is not applicable for every kind of circuit and then you has to sacrifice the expected lifetime of the circuit for performance. I guess from your experience that you might have been opposed by low-grade parts which have haunted you.   
About the relay problem; I don’t know for sure if you have the “reset” caused by a problem on the coil side or at the contacts of the relay? You should have no problems on the coil side with a proper suppressor link installed here. 

 

[Tony_Stoynov]

Hi Ante,
thanks for the answer, about the temperature yes, if is possible i put the all part in goot temperature condition, if is a possible, i know how temperature reflect over reliability, iuse the Item software for reliability analisys. But sometime, in some case have other reason, price, space .. actually this si s a two generral reason to put some device in not very good temperature condition. I design the different products, and not i give a decision about price, and demention
Abut the relay, yes the problem is from coil, and how i say i have a Epcos 380 V suprestor, but this not solve the problem, may be the speed ov increasing the voltage is was too high, and generate strong interference before time needing to supprestor to switch on.

I am very happy that have a post, i solve all this problems a long ago, but may be some other peopel wil have this, or similar problem, wil be good if my or your experiance can help im to olve the problem more fast.

 

[Staigen1]

Hi

when i build a cirquit i dont try to run the pieces at their maximum ratings or
more, they usally break then, sooner or later. Also, when it comes to regulators
like 78xx, i use bigger than normal electrolytics to avoid noise in the output.

I remember a calculator i once repaired, it was around 1986 or so, a friend claimed
that it have never been reliable, and now it did not work at all. I opened it up, and
found that the regulator was broken, i also found that it was not bolted to the heat-
sink, the bolt and nut was missing. I put in a new regulator and bolted it to the heat-
sink, and it worked as it should after that.

//Staigen

 

[Tony_Stoynov]

Hi,
yes, i try to say above to same, not is good idea for any part to work closely to the maxium parameter, in the case with 78xx the surface temperature was 80 C, this is two time lowwer, and i think is acceptable in some cases.
About the problem with a regulator and heatsink without bolt and nuts, ooo i see this problem no once. Some times i think in fabric no put the bolt when made the assembly, and if they have not a good QA,is easy this product to going  on the market. In some other case, the device is repared from sombody else, and may be they think that this bolt and nut can be missing.

This ia a big problem with a maintenance. Have something other, in a different devices enclouser is used to despare the heat but after maintenance procedure the technician put just 2-3 screw, becouse this is a easy and then the thermal resistance between enclouser and thermal source is a very big, and .... after several hours or day, some times after several month the device did fail.

 

[Staigen1]

Hi Tony

What do you mean? Please, at least read what you have written before you hit
the "post" button. I usally hit the preview first, and read my writing carefully and
make changes and try to correct typos before posting. I also have an
english/swedish/english dictionary handy, if there is any word i dont know.

//Staigen

 

[Tony_Stoynov]

Hi Staigen,
yes, you are right, excues me for this, i have too the dictionary, and translator, but not every time i have time for this, but in the future will try to correct may errors.

p.s And i have one more problem, i use the HP NX7010, this laptop has a some very strange keyboard, if not press the button in the center you not have a simbol ...for this reason when i write fast very often i drop a letter.

 

[Staigen1]

Hi

Its okay, but it is sometimes very hard to read what you write, english is not my
native language, i live in sweden. BTW, where do you live, i see that you are a
very skilled electronics engineer, ooh, i see that you live in Las Vegas. You say
that you have a dictionary, but you are living in USA, and your native language
is then english. I'm sorry, i could never have guessed that you lived there,
according to how you spell.

//Staigen

 

[Tony_Stoynov]

Hi,
mm yes i Living in Vegas, but my native language is not English, i am from Bulgaria, and actually 7 months ago my english was like your Bulgarian language I use Bulgarian dictionary. About the language, i know that my english is not good, i try to improve it. But the syntax error very often is not from ignorance, may style of typing is the reason. The situation in bulgarian language is to same, i make a lot of error when i typing, and after this must correct my errors. And i try to do this and in forum, but how i say, some time i have not time to do this, but i want to help to somebody .. may be better to write my post when i have a time ... i never not think about this.