Overwhelming difficulties with regulators, please help!

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I have two problems whitch are probably easy to solve.

1st problem LM317

I'm trying to replace CD player's LM7805 regulator with simple LM317 regulator circuit:
An externally hosted image should be here but it was not working when we last tested it.


I used 240 ohms for R1 and 750 ohms for R2.
The pin order for LM7805 (in, out, gnd) and LM317 (adj, out, in) is different.
I connected the pins and resistors like this:
http://kotiweb.kotiportti.fi/audiovideo/download/kytkenta2.jpg

I tested the regulator with 1 kohm and 47 ohm load and I had a 9 volt battery as a powersupply. The voltage was first 5.8 and started getting higher. I had a 4,7 uF cap between putput and ground. No other caps.

2nd problem LM7808
I'm building a simple regulator for Tentlabs XO clock. I used this PCB:
http://kotiweb.kotiportti.fi/audiovideo/download/piirilevy3.gif

Again the voltage was first over 5 V (should be 3.3 V) and started getting higher every second.

I've tried everything and searched half the internet for a solution. You people are pretty much my last hope. Please help !! :bawling:
 
LM317 is going to be installed to replace this:

regu2.gif


I add nothing else but the regulator with the two resistors (no diode). I haven't tried it yet in the cd player.

LM7808 is connected to that circuit. There is nothing else in it. PSU is going to be unregulared 16 V DC.
 
Hi,

LM317 oscillates easily if you have capacitance at the output but not at the input. Connect a small ceramic 0.1uF capacitor fron the input pin to the same point where the voltage setting resistor is connected to ground - keep wires short. This should fix it is the problem is due to oscillation. Use an oscilloscope to verify. It might be that you already have there a cap - but I was not able to tell because your spelling putput spelling mistake ;)

Regards,

Christian
 
LM 317...

Refer to the datasheet on www.national.com

You will find there is a cap to ground from adj pin.
Use the given formula and calculate the value of R2.

If you connect a larger cap at the output then the input, the output cap tends to supply voltage to the output of the regulator when the input voltage is switched off and damages the regulator

Every aspect and variants of the application exist at national website.

I suspect you are trying to replace 7805L with LM317T

Also fix a heatsink since your i/p voltage is quite high.

Gajanan Phadte
 
2 DIAR: To probelm #2: It's hard to read the circuit for me just by seeing the board, and no shcematic.
AFAIK the 7808, the mounting in the board should be the following: metal cooling fin should be facing the 100 uF caps (if the metal is facing to the "outer" border of the printed circuit board, IN and OUT are reversed forthe 7808 -- this might be the reason for the voltage change.

Let us know how things turn out, pls.
 
LM7808 schematic can be seen here:

http://www.tentlabs.com/ProductEntities/XO/XO.html


An externally hosted image should be here but it was not working when we last tested it.


I noticed the wrong pin order. I made the pcb but I haven't tested it yet.
http://kotiweb.kotiportti.fi/audiovideo/DIY/kello/piirilevy_V2.jpg


I assembled LM317 regulator but I haven't tested it yet.
http://kotiweb.kotiportti.fi/audiovideo/DIY/regulaattori/valmis_1.jpg

The PCB layout is attached.

I really hope they work this time. I have burned several regulators already :dead:
 

Attachments

  • uusin moduuli_b.png
    uusin moduuli_b.png
    87.4 KB · Views: 606
2 DIAR,

thank you for the website link.

Well, I can't understand why they do so much efforts to propose this circuit for genrating the Supply voltage for the XO:

Datasheet says 5 V +/- 5 %.


So you can simply use 1 7805 regulator, 100uF at input, 10 uF at output -- voila - 5 V with 3 parts.

The original circuit is much more effort, and leaves you with a little bit lower noise level on the +5V rail -- but that's not relvant for the XO.

BTW: If you "burn" the regulator IC's than it is very very shure that there is a problem with your pinout on the board. If I run into this, I let a freind look at the board -- people that are not involved in the project see faults in secounds where I fail for days not seeing these.
 
2 DIAR:

Looks like got a "limit batch" of LM317: Datasheet specifies a param "minimum load", which is 5 mA typ, -- so the 240 ohms @ 1,2 V is alright. 10 mA is maximum

Your LM317 seems to need appr. 8mA (the 5 from the voltage divder and the appr. 3.5 mA fro mthe 700 Ohm).

So you are little bit out of params - no problem, but took you a long time to figure out. congratulations and happy DIYing!
 
I must say, I didn't quite understand everything you were saying :cannotbe:

I suppose that you just made sure that everyhing is in order :confused: (If so, thank you very much :D ).

I checked the data sheet and I believe everything is ok.

Vref is ([5/(240+740)]*240) = 1,22 V (should be 1,2 - 1,3 ?)

Test load was 5V/700 ohms = 7 mA + 5 mA --> 12 mA (from R1 and R2). Minimum load is 10 mA ?

--> Everything is ok :confused:
 
Re: Lm317

gmphadte said:
For LM 317

General practice for values of capacitors around LM317...

LM317 data sheet:

In order to optimise the load regulation, the current set resistor R1 (see fig. 4) should be tied as close as possible to the regulator, while the ground terminal of R2 should be near the ground of the load to provide remote ground sensing.
Performance may be improved with added capacitance as follow:
An input bypass capacitor of 0.1mF
An adjustment terminal to ground 10mF capacitor to improve the ripple rejection of about 15 dB (CADJ).
An 1mF tantalium (or 25mF Aluminium electrolitic)
capacitor on the output to improve transient
response.


I made the thing according to Fig. 5 on the data sheet. The power supply is allready regulated and there is a 1000 uF Elna Silmic in the output.

Would it be a goo idea to add a 10 or 100 uF Os-Con near the DAC chips which the regulator is feeding?
 
2 DIAR:

well, it's ok if you don't understand it what I try to tell: This parameter is quite seldom discussed.


To get get a common basis for discussion, please look at this datasheet: http://www.age.et.tu-dresden.de/datasheets/V-Reg/LM317_ST.pdf
You should now see a LM317 datasheet.

Scroll to page 3, search for parameter I o (min), minimum load current.

check the very rightmost colummn: 10 mA max.

So this means: Typical value is 5 mA ( column left from the 10mA entry, but manufacturer garantees only 10mA.

What consequences has this foryou as the builder of the circuit? Ok, a minimum load of 10 mA has to be maintained to ENSURE regulation. A minimum load of 5mA is TYPICALLY enough.

In your circuit, with your specific components, we checked by your expereiments that you need at least 7 mA. That's why I wrote : You got a limit batch of LM317.

OKAy, what does this mean in GENAERAL?
This means that forthe standard application circuit with the 240 Ohms resitor between output and ADj. pin, only 5mA current flow. This is okay in MOST of the LM317.

But if you want to make SHURE that in every circuit at every time this will work, you have TO CHANGE TO 120 OHM!!

This is is a typo in most of the datasheeet, and -- bad luck -- most of use stumble into this sooner or later. I did too :)

Or, DIAR, for you: NOT EVERYTHING is okay. BUT if you change the 700 Ohm load to a 470, you have a total MINIMUM load of appr. 11 mA. THIS IS OKAY for EVERY situation (temp. change, input voltage variation): regulation will be maintained.

I hope this shades some light in my former posting. In case there are still questions, don't hesitate to ask pls.
 
What an excellent explanation. Thank you very much ! :D

I have absolutely no idea how much current the DAC chips take. I believe it's much more than a few mA but I am certain that it is not near the maximum current that LM317 can feed. Therefore I think it's wise to simply change the two resistors to be on the safe side.

I told earlier that there are two 1000 uF caps in the input. That was wrong (they are situated before the first regulator). The input voltage is regulated and there is only 47 uF cap in the input (and pretty far away). There is one 0,1 uF cap near the LM317 input. Should I use bigger capacitor in the input?

What about the large 1000 uF Elna Silmic in the output. Is it ok or should I replace it with a smaller (100 uF) Os-Con?
 
Yep, changing the resistors is the first step.

The 47 uF is alright because there are bigger caps beforethe first regulator (that already smoothes the 50/60 Hz ripple in your raw DC voltage). A larger cap nearthe lm317 is overkill.

I would personally add a 1 ohm 1/2 W carbon film restor in the output of the lm317 (after the voltage divider ) and then place a 100 uF os-con cap between plus 5V and GND. This is an "el cheapo" filter, smoothing much of the higher frequency noise the lm317 can not regulate.

Well, but as written: That's personal taste. You have to experiment: Listen to the sound. If it sounds more comfartable to you, add the 1 ohm/cap combo. if not, don't use the combo.

There is no general rule now forthe fine details. As soon as your lm317 behaves well, the experiments beginn -- all the best wishes to you!

Let us know what you find out, pls.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.