78 series regulator problem

[Mike Brett]

Help needed please.
I have built a small power supply for 12v using a L7812 2amp. regulator. This is fed from a 9 x 9 transformer followed by a 4 amp 200v bridge rectifier which gives out 28 volts DC. I have a 0.1u elec. on the input pin of the regulator and a 1u elec. on the output pin. There is also a 1N4002 diode across in and out pins. Problem is I am only getting out 11 volts, I have lost a volt somewhere, any ideas please.
Mike

Is that on or off load? A 7812 is a 1 amp regulator AFAIK.

[Cobaltblue]

Hi Mike you don't seem to say how big the main smoothing capacitors are on the input side of the regulator unless that 0.1u is a typo

Cheers

Mike T

[maninashed]

Mike please attach a diagram of your PSU. 7812 heatsink mounted in TO220 case is only good for 1A, if you require 2A you will need to parallel 2 of them with balancing resistors. After your bridge rectifier, you then need a capacitor of about 1000uf min to allow smoothing, then a 0.1 to decouple from the input pin to ground, this should be mounted close as possible to the 7812. This is for stability only.
Bill

[Mike Brett]

I only require a output power of about 180 ma at 12 volt so the L7812 should be fine if it gave me the 12 volts instead of the 11 I am getting. At present I have only tested it of load but it should still read 12 volts I would have thought. The input capacitor is a 63 volt 0.1 uf electroltic and the output capacitor is a 63 volt 1 uf elec. Where exactly should I fit the 1000 uf capacitor, should it be between the input pin and ground and should it be polarized, and will this give me the 12 volts I need.
Cheers Mike

[robjkmannering]

Hi,
Change the input cap after the bridge rectifier for a 1000uF 25V (minimum) electrolytic type. This will be a polarized component. I suspect that the problem you have is because you have un-smoothed dc going to the regulator at present.

Rob

[maninashed]

Mike, Robs advice is correct. Connect the +ve side of the 1000uf capacitor to the +ve end of your bridge rectifier and the neg side to ground. If all's well you should then have your 12v.
Bill

[Mike Brett]

Rob
You sir are an absolute star. I put the 1000uf cap on like you said and bingo, 12 volts spot on. We live and learn, many thanks to all of you though for your help.
Cheers Mike

[Ed_Dinning]

Hi Mike, most of these regulators need at least 3v above the regulated value on the input.

Despite the large e-cap on the bridge it is still advisable to use the 0.1uF caps as close to the reg pins as possible for stability.

Ed

[robjkmannering]

Glad you have a working circuit. As Ed has suggested, 0.1uF cap should be retained for stability.

Rob

[G8HQP Dave]

The capacitors next to the 7812 are for stability. As you found, you still need the reservoir capacitor for the DC supply feeding the 7812. The 7812 datasheet probably won't mention this, as it assumes you already know how to build a DC supply.

[mhennessy]

Watch out for heat. I guess it'll dissipate somewhere around 3 watts, but that'll need more heatsinking that you might think. Keep an eye on the operating temperature of the IC until you're happy it's OK.

If you stumble across one in the future, a 12V transformer will result in cooler running.

Mark

[Herald1360]

Originally Posted by mhennessy

I guess it'll dissipate somewhere around 3 watts, but that'll need more heatsinking that you might think.

Mark

A single TO220 78XX in free air will do about 1W comfortably. A 15C/W heatsink will be OK for 3-4W.

[Top Cap]

The manufacturers also recommend a minimum load of 5mA which is of course no problem if the output is wired permanently to a load. Where I have situations when the load can be disconnected, I usually place a LED on the output. With a 1k series resistor, it not only provides a permanent load of around 10mA but also serves to show the PSU is working. I usually place an electrolytic on the output as well. My Mentor used to say " Rule of thumb - 2000uF per Amp on the input and 200uF per Amp on the output". As you will be drawing much less than half of this, a 1000uF will do very nicely. The reversed diode across the regulator is there to protect the regulator from reversed voltage where an output capacitor is employed.

[mhennessy]

Originally Posted by Herald1360

A single TO220 78XX in free air will do about 1W comfortably. A 15C/W heatsink will be OK for 3-4W.

I am very conservative when it comes to cooling

The total thermal resistance of a bare 7812 from junction to air is in the order of 70C/W, meaning that when dissipating 1 watt, the junction will be at least 70C above ambient. Ambient in an unventilated enclosure could easily be 40C plus. The highest permissible junction temperature for the 7812 is 125C. From this, it's hopefully easy to see that 1 watt with no heatsink is marginal to say the least. I appreciate that plenty of manufacturers did this, but we've all seen the dry joints and charred PCBs that result.

When we put it on a heatsink, we need to add up the total thermal resistance - not just the thermal resistance of the heatsink. From junction to case, it's 5C/W, from case to sink it'll be around 2C/W with proper installation. So, 22 (5+2+15) C/W in total, and a junction temperature of 66C above ambient for 3W. Again, with an ambient of 40C, that's getting close to the danger zone.

Don't forget that the dissipation will shoot up if the output is accidentally short-circuited. The thermal protection built into these things is by no means 100% reliable.

I use an infra-red thermometer to check on these things, but the "finger test" is a good starting point; if you can hold a finger on it for more than 5-10 seconds, then it's below about 60C. As a rule of thumb, that's as hot as I like a TO220 to run.

For this, if the PSU is being built into an aluminium box, that would probably provide enough heatsinking, but obviously it's up to the OP to check this.

Hope this helps,

Mark

Numbers from: www.fairchildsemi.com/ds/LM/LM7805.pdf

[Herald1360]

I just looked at the graph on p22 in the ON semi datasheet attached- a bit lazy perhaps...... they're also abit more generous about Tj max at 150C.

[Radio Wrangler]

Mike,

Chris has posted the ON data sheet and you'll see that there are two grades of the 7812. The plain vanilla one and the 7812A. The plain one has a tolerance of 11.5 to 12.5 Volts on its output, and the A suffix one 11.75 to 12.25.

ON are one of the more reputable makers (they used to be called Motorola before the company broke up). Some 78 series regulators on the market are unbranded and can be worse than the +/- 0.5V on this data sheet.

These are rough-ish regulators for general use, not accurate voltage references.

For a higher performance 3-terminal regulator, I've used the LT1086 family quite a lot, but you'd only need one of these if your 12v accuracy was critical, because they are several times the price of the basic 78 series job.

http://cds.linear.com/docs/Datasheet/1086ffs.pdf

David

[mhennessy]

Originally Posted by Herald1360

I just looked at the graph on p22 in the ON semi datasheet attached- a bit lazy perhaps...... they're also abit more generous about Tj max at 150C.

Yes, the fact that practically everyone makes a version of the 78xx series is a good reason to be conservative. A designer who specifies the On-Semi version could well come a cropper when the purchasing department realises they can save N times 3 pence by choosing the Fairchild version.

Things undoubtedly last longer when running cool. I'm forever replacing these things, and retro-fitting (or upgrading) heatsinks when I find that they're running too hot to touch. What is surprising - but fortunate - is that they tend to fail O/C, so they don't put the unregulated input through to the output when they do give up.

Mark

[Herald1360]

Whilst it's a bit squalid, if the load on your 7812 is fairly constant, you can always put a suitable size WW resistor between the input and the raw supply to take some of the dissipation off the regulator and save the need for any heatsink. It's not very clever if your load has large step changes in its current, though......

[kalee20]

Herald1360's suggestion is very crude, as he admits, but it's good advice.

If the regulator were to fail short-circuit, it'll also give some sort of current limiting and protect the load.

Watch that any capacitors for stability are still connected directly on the regulator however. A 7812 fed from a sizeable resistor only, with no capacitor, might hoot.