You are here

Possible issue with L7812 voltage regulator

Customising, building or repairing your own gear? Need help with acoustic treatment or soundproofing? Ask away…

Possible issue with L7812 voltage regulator

Postby dwebb » Mon Oct 08, 2018 12:26 pm

I've just built a Eurorack Power supply (http://www.frequencycentral.co.uk/?page_id=894) which is working well bar one intermittent oddity - every so often when I power it up I don't get any signal from the +12V rail. When it doesn't start up correctly I am seeing 16V DC after the rectifier and smoothing capacitors and around -0.7V after the L7812 regulator. Turning off and on again usually fixes it.

Having looked at the data sheets for the L7812 (http://www.farnell.com/datasheets/2307057.pdf) it looks like there are various protection mechanisms (thermal, short circuit and SOA), so I am guessing that one of these is being tripped intermittently.

I was hoping someone could tell me whether it is likely I have a faulty L7812 or whether there is some other issue I should be looking for. I have put together a basic schematic of the relevant bits of the PSU here:

Image

Thanks in advance!

Dom
dwebb
Regular
Posts: 53
Joined: Wed Aug 26, 2009 12:00 am
 

Re: Possible issue with L7812 voltage regulator

Postby resistorman » Mon Oct 08, 2018 1:38 pm

Does this happen even under no load conditions? It would be cheap and easy enough to replace the 7812 to troubleshoot...
User avatar
resistorman
Frequent Poster
Posts: 591
Joined: Sun Nov 22, 2015 1:00 am
Location: Asheville NC
future26.com

Re: Possible issue with L7812 voltage regulator

Postby James Perrett » Mon Oct 08, 2018 2:25 pm

Either you have a faulty 7812 or one of your modules has a large capacitor on the power supply which is causing the protection to activate when you initially try to charge up that capacitor. When you try switching on again the capacitor will still be partially charged so that module won't draw such a large startup current.
User avatar
James Perrett
Moderator
Posts: 7604
Joined: Mon Sep 10, 2001 12:00 am
Location: The wilds of Hampshire
JRP Music - Audio Mastering and Restoration. JRP Music Facebook Page

Re: Possible issue with L7812 voltage regulator

Postby dwebb » Mon Oct 08, 2018 3:45 pm

It does seem to happen more with more modules attached - though intermittent so hard to be sure. A fair number of the modules do have capacitors on the power supply so that looks like a good suggestion.

In any case I'll try replacing the regulator to see if that fixes it. If not is there a practical way of limiting the current at start-up to give the capacitors in the modules time to charge slowly? I would assume the simplest approach would be to use a resistor in series with the +12V rail, but then that would presumably reduce the overall amount of current the power supply could produce.
dwebb
Regular
Posts: 53
Joined: Wed Aug 26, 2009 12:00 am
 

Re: Possible issue with L7812 voltage regulator

Postby Hugh Robjohns » Mon Oct 08, 2018 4:21 pm

There are several ways you could handle this. One would be to redesign the circuitry to use the regulators to control beefier power transistors to handle the bulk of the current demand.

Another would be to use current-restricting resistors in the outputs of the regulators with a relay (or power transistor) to shunt the resistors after a short time when the start-up surge currents have abated.

Or, run fewer modules off the power supply, and build additional power supplies to power other modules -- thereby reducing the start-up load on each PSU.

H
User avatar
Hugh Robjohns
Moderator
Posts: 21946
Joined: Fri Jul 25, 2003 12:00 am
Location: Worcestershire, UK
Technical Editor, Sound On Sound

Re: Possible issue with L7812 voltage regulator

Postby dwebb » Mon Oct 08, 2018 4:35 pm

Thanks for the suggestions. Looking around at various current limiting devices I have come across NTC thermistor inrush current limiters - from what I have read they appear to offer a higher level of resistance at power on which drops to a negligible level as the devices heat up. One of those in series with the +12V rail would be a very simple modification - could that be a potential solution?
dwebb
Regular
Posts: 53
Joined: Wed Aug 26, 2009 12:00 am
 

Re: Possible issue with L7812 voltage regulator

Postby ef37a » Mon Oct 08, 2018 4:53 pm

IMHO the trouble is further back, the 12V rat transformer.

The "twin 1/2 wave rectifier" dodge is a valid one, I know of a pedal that used it* very successfully to derive two 15V rails at about 200mA bit the supply was 16V at 800mA nominal and the off load AC voltage was 22V so the caps got a good jolt at connection. IIC the input side of the regulators was about 20Volts.

The usual and "proper" way to feed a regulator circuit is indeed a secondary about equal to the final voltage, in this case 12 but! That is with the more efficient full wave bridge PER regulator. So, basically I think the input voltage is boarder line too low. The regulators can cope with at least 30V, especially as the current is well below maximum so I would try a 15V or so rat.

I doubt it is downstream capacitors loading the circuit? The 78/79s have been around a long time, bet they thought of that! A quick and dirty solution however would be to shunt one of the diodes to bleed a few mA across to the output. I also think the 100mfd caps on the outputs are needlessly large? 4.7-10mfd is big enough to keep things stable.

*"used to" because such small AC-AC wall rats and line lumps have been banned by EU regs now. The pedal now uses a 22V SMPSU and a very clever way to generate the -ve supply (which gave me a bit of bother at first!)

Just a cotton pickin' minute! 15,000 mikes to smooth a 500mA supply? Tall African ungulate. I bet you can SEE the supply rising on those caps when you switch on?

I don't know but I bet the supply is simply not rising fast enough and the regulators is having a hissy fit. In any case the power factor for that transformer is appalling. I bet it runs hot and I doubt it will last long. 4,700 per side is plenty. The rule of thumb I think is 2,200 per 1,000mA? Maybe double that for half wave?

Dave.

Dave.
ef37a
Jedi Poster
Posts: 9549
Joined: Mon May 29, 2006 12:00 am
Location: northampton uk

Re: Possible issue with L7812 voltage regulator

Postby dwebb » Mon Oct 08, 2018 5:14 pm

The 12VAC transformer I am using is this one:

https://uk.rs-online.com/web/p/chassis-mounting-transformers/0504672/

I decided to use this rather than the suggested wall wart as it has two secondary outputs allowing me to add an additional power supply in the future. If I understand the specs correctly it offers 25VA per secondary which I think means it should be able to provide around 2A at 12V - I hoped that would be more than enough (or perhaps too much?).

I can't see any problem with the voltage at this stage (though I am possibly not looking correctly). There is a DC voltage of around 15-16V before the regulator and the 12V power rail seems to be extremely stable - even when powering a good number of modules.

When the 12V rail doesn't start up I can measure around 17V DC on the input side of the regulator - so it doesn't appear to be a shortage of voltage to regulate. I would be concerned that in using a higher voltage AC transformer the regulators would just need to generate more heat?
dwebb
Regular
Posts: 53
Joined: Wed Aug 26, 2009 12:00 am
 

Re: Possible issue with L7812 voltage regulator

Postby ef37a » Mon Oct 08, 2018 5:22 pm

Just looked at "my" pedal PSU. Hmm, I reckon that guy got a zero in the wrong place!

The pedal circuit is almost identical but uses 4X 470mfd i.e. 2000 mfd near as. The regulators have the shunt diodes but an additional diode across the output. The cap on the 15V side is just 10mfd. This is of course for 15-0-15 volts.

Dave.
ef37a
Jedi Poster
Posts: 9549
Joined: Mon May 29, 2006 12:00 am
Location: northampton uk

Re: Possible issue with L7812 voltage regulator

Postby ef37a » Mon Oct 08, 2018 5:39 pm

dwebb wrote:The 12VAC transformer I am using is this one:

https://uk.rs-online.com/web/p/chassis-mounting-transformers/0504672/

I decided to use this rather than the suggested wall wart as it has two secondary outputs allowing me to add an additional power supply in the future. If I understand the specs correctly it offers 25VA per secondary which I think means it should be able to provide around 2A at 12V - I hoped that would be more than enough (or perhaps too much?).

I can't see any problem with the voltage at this stage (though I am possibly not looking correctly). There is a DC voltage of around 15-16V before the regulator and the 12V power rail seems to be extremely stable - even when powering a good number of modules.

When the 12V rail doesn't start up I can measure around 17V DC on the input side of the regulator - so it doesn't appear to be a shortage of voltage to regulate. I would be concerned that in using a higher voltage AC transformer the regulators would just need to generate more heat?

Gaghh! If you are using a twin secondary transformer why did you not go for a conventional bridge rectifier circuit? Yes, 17V is easily enough input voltage but I still wonder if the 1/2 wave rectification and the huge capacitance is still causing a startup problem? I guess you will have to change the regulator as it is the easiest thing to try.

Note, the transformer cannot be "too big" in terms of current delivery in fact it is always better to err on the side of caution and go for bigger. The history of electronics is littered with perfectly good bits of kit that failed because the transformer was only just big enough and cost just a bit less than it shoulda! The only exception is the toroid. They are used close to their rating as that is when they are most efficient. don't have too but then no point in paying for a toroid!

Dave.
ef37a
Jedi Poster
Posts: 9549
Joined: Mon May 29, 2006 12:00 am
Location: northampton uk

Re: Possible issue with L7812 voltage regulator

Postby dwebb » Mon Oct 08, 2018 5:54 pm

I got and built the power supply circuit before I got the transformer - originally I was just going to use the suggested wall wart option.

As I don't really know what I'm doing (you've probably gathered that by now) it looked like an easier option to just populate a ready-made circuit board than build a power supply from scratch.

Regardless of my foolish ignorance I am most grateful for the assistance ...
dwebb
Regular
Posts: 53
Joined: Wed Aug 26, 2009 12:00 am
 

Re: Possible issue with L7812 voltage regulator

Postby ef37a » Mon Oct 08, 2018 8:52 pm

dwebb wrote:I got and built the power supply circuit before I got the transformer - originally I was just going to use the suggested wall wart option.

As I don't really know what I'm doing (you've probably gathered that by now) it looked like an easier option to just populate a ready-made circuit board than build a power supply from scratch.

Regardless of my foolish ignorance I am most grateful for the assistance ...

Yes, your inexperience was noted but so long as I thought you were just dealing with sealed mains units and low voltages fine. Now you must have wired that transformer up for mains operation and that raises considerable concerns. There are strict rules as to insulation and earthing arrangements. I guess so long as it is only YOU that has contact with the equipment that is one thing but can you guarantee that will always be the case?

A photograph of how the transformer has been mounted, enclosed, wired, fused! Together with details of the earth connection would set minds at rest (or not!)

Dave.
ef37a
Jedi Poster
Posts: 9549
Joined: Mon May 29, 2006 12:00 am
Location: northampton uk

Re: Possible issue with L7812 voltage regulator

Postby James Perrett » Mon Oct 08, 2018 9:15 pm

ef37a wrote:Just looked at "my" pedal PSU. Hmm, I reckon that guy got a zero in the wrong place!

Having now had a look at the board layout and instructions I'd say that there are a few things in the wrong place!
User avatar
James Perrett
Moderator
Posts: 7604
Joined: Mon Sep 10, 2001 12:00 am
Location: The wilds of Hampshire
JRP Music - Audio Mastering and Restoration. JRP Music Facebook Page

Re: Possible issue with L7812 voltage regulator

Postby Tim Gillett » Tue Oct 09, 2018 6:11 am

With such regulators I thought there was now little need for substantial capacitance on the regulated side, so the 100uF caps struck me as unusual.
Tim Gillett
Frequent Poster
Posts: 1593
Joined: Wed Jan 30, 2013 1:00 am
Location: Perth, Western Australia

Re: Possible issue with L7812 voltage regulator

Postby ef37a » Tue Oct 09, 2018 9:38 am

Tim Gillett wrote:With such regulators I thought there was now little need for substantial capacitance on the regulated side, so the 100uF caps struck me as unusual.

Yes, I believe the data sheet says to use a small cap because the regulators put out a bit of RF at around 10mHz. Mind you, Mr Self uses 100mfd in his Audio Design book. He also cautions about regulators not starting up and shows that simply loading on capacitance does little to reduce ripple whereas separating two 4,700 caps with 2R2 gives several dB of ripple reduction. Gives the traff an easier time as well!

"Small Signal Audio Design". If you wield a solder iron and are into audio, gotta have!

Dave.
ef37a
Jedi Poster
Posts: 9549
Joined: Mon May 29, 2006 12:00 am
Location: northampton uk

Re: Possible issue with L7812 voltage regulator

Postby ef37a » Tue Oct 09, 2018 12:54 pm

Reading Self again he mentions +V rail startup problems. Likely due it seems when there is a capacitor "rail to rail" across the whole regulated supply.

Two diodes reverse biased across the output helps but they also protect the regulators from large downrange capacitances. Peanuts, fit them.

You can still have +V rail start up issues and it seems you need some asymmetry in the rectifier side circuitry. I would remove at least one of those 4,700s and have more in the + side than the -ve. Actually that sounds counter-intuitive! You would think you want to delay the -ve side more than +ve? Cut and try is all I can say!

This really needs sorting because one rail not coming up can burn out the very common NE5532 IC.

Dave.
ef37a
Jedi Poster
Posts: 9549
Joined: Mon May 29, 2006 12:00 am
Location: northampton uk

Re: Possible issue with L7812 voltage regulator

Postby dwebb » Tue Oct 09, 2018 3:25 pm

Re your earlier questions the mains enters the case via a switched and fused IEC socket. I've put the transformer in a sealed box separate from the rest of the case.

To clarify your first suggestion about putting 'two diodes reverse biased across the output' - are those the diodes marked D3 and D4 in the schematic I posted above?

I'll also try removing one of the 4,700uF capacitors on the -ve side and replacing the L7812 in case that is faulty.

Thanks again for your help - much appreciated.
dwebb
Regular
Posts: 53
Joined: Wed Aug 26, 2009 12:00 am
 

Re: Possible issue with L7812 voltage regulator

Postby ef37a » Tue Oct 09, 2018 4:32 pm

" Cased, switched, fused" and EARTHED I trust?

No, the diodes for start up protection need to go across C9,C10 reverse biased of course.

These four diodes are often omitted on self contained equipment because the capacitances are known and won't change (except some folks HAVE to "improve" things!"If 2,200 mfd is good, 10,000 MUST be better!" "Oh! Why does that primary fuse keep blowing?").

Dave.
ef37a
Jedi Poster
Posts: 9549
Joined: Mon May 29, 2006 12:00 am
Location: northampton uk

Re: Possible issue with L7812 voltage regulator

Postby dwebb » Tue Oct 09, 2018 5:34 pm

Understood and thanks again.

PS Yes, earthed.
dwebb
Regular
Posts: 53
Joined: Wed Aug 26, 2009 12:00 am
 

Re: Possible issue with L7812 voltage regulator

Postby dwebb » Wed Nov 07, 2018 11:59 am

... a quick addendum should it be of use to anyone. I replaced the L7812 and all has been working well for a couple of weeks. Thanks again to all who offered assistance.
dwebb
Regular
Posts: 53
Joined: Wed Aug 26, 2009 12:00 am
 


Who is online

Users browsing this forum: No registered users