Ground lift for 78xx regs

For the three and more legged things
Post Reply
Clive
Old Hand
Posts: 374
Joined: Tue May 22, 2007 9:17 am
Location: Cheshire, England

#1 Ground lift for 78xx regs

Post by Clive »

I need to increase the voltage from a 7824 reg by around 2.6V to 3.0V, the exact voltage is not critical. I could use 4 or 5 IN4001 diodes to connecting the centre leg to ground, is there a more elegant solution, ie a single LED?

Cheers,

Clive
User avatar
Nick
Site Admin
Posts: 15707
Joined: Sun May 06, 2007 10:20 am
Location: West Yorkshire

#2

Post by Nick »

Blue LED or zener could be used.
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
Clive
Old Hand
Posts: 374
Joined: Tue May 22, 2007 9:17 am
Location: Cheshire, England

#3

Post by Clive »

Hi Nick,

Would the LED see much in the way of voltage (ie exceed say a 5V max rating) or due to the way these regs work does it see no real voltage and simply raise ground by around 3.3V?

Is the spec I need to look for "forward voltage" to tell me what voltage the 7824 will be lifted by?

Clive
User avatar
Nick
Site Admin
Posts: 15707
Joined: Sun May 06, 2007 10:20 am
Location: West Yorkshire

#4

Post by Nick »

Yep, the LED will only see its forward voltage at a few mA
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
User avatar
Mike H
Amstrad Tower of Power
Posts: 20157
Joined: Sat Oct 04, 2008 5:38 pm
Location: The Fens
Contact:

#5

Post by Mike H »

Yep datasheet says quiescent current is 5mA, so that will work.

Could also use a 2.2V or 3.3V Zener diode :?:
 
"No matter how fast light travels it finds that the darkness has always got there first, and is waiting for it."
brig001
Old Hand
Posts: 550
Joined: Sat Nov 13, 2010 9:56 pm
Location: Back home in Preston now

#6

Post by brig001 »

AFAIK, these things give a constant current on their ground pin (5mA according to Mike H), so would a resistor not be a better solution?

Brian.
Clive
Old Hand
Posts: 374
Joined: Tue May 22, 2007 9:17 am
Location: Cheshire, England

#7

Post by Clive »

brig001 wrote:AFAIK, these things give a constant current on their ground pin (5mA according to Mike H), so would a resistor not be a better solution?

Brian.
Dunno, there has to be a reason people use diodes in these situations, maybe the 5mA is not reliable? I'm loathe to use a zener and they are prone to pump out a little noise. Maybe I'm mistaken.
User avatar
Nick
Site Admin
Posts: 15707
Joined: Sun May 06, 2007 10:20 am
Location: West Yorkshire

#8

Post by Nick »

Looking at the circuit diag on the spec sheet, there are several non constant paths from In to Gnd, So I would think a low impedance path to ground is important.

I have checked this in the past, if you look at the voltage across the diodes in the connection to ground, you see it mirrors the input signal, so I think a simple resistor would reduce the quality of regulation.

I think the best solution is a variable reg and resistors, but the diodes is a simpler quicker solution.
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
brig001
Old Hand
Posts: 550
Joined: Sat Nov 13, 2010 9:56 pm
Location: Back home in Preston now

#9

Post by brig001 »

In case anyone is really interested to know...

The current does change with input voltage as Nick says.

National recommend two resistors to make the regulator variable - presumably hoping that the current through the resistors on the output swamp the variation in quiescent current.

Fairchild recommend an op amp to make the regulator variable.

Clive, since you don't care what the exact voltage is, any of the suggestions will do, but the diode solutions should give better line regulation and the resistors might give better stability against temperature.

What about an LM317? :twisted:
Brian.
Attachments
Fairchild
Fairchild
National
National
National.PNG (8.51 KiB) Viewed 5891 times
Current variation in a 7805
Current variation in a 7805
Clive
Old Hand
Posts: 374
Joined: Tue May 22, 2007 9:17 am
Location: Cheshire, England

#10

Post by Clive »

brig001 wrote:
Clive, since you don't care what the exact voltage is, any of the suggestions will do, but the diode solutions should give better line regulation and the resistors might give better stability against temperature.

What about an LM317? :twisted:
Brian.
Hi Brian,

The diode / LED is fine for what's needed, the 7824 in an existing pcb, adding an LED is easily done, converting to a 317 less so. Also the 7824 is a pre-reg, the next reg is indeed a 317, all the 7824 does is feed the 317.

Thanks to everyone for the advice.
Clive
Old Hand
Posts: 374
Joined: Tue May 22, 2007 9:17 am
Location: Cheshire, England

#11

Post by Clive »

Tried a LED (or two....). When the reg is run under voltage, ie 20V for the 7824, the LED lights up. When the reg is given 30V at its input the LED expires......most LEDs seem to have a reverse voltage spec of 5V. I suppose I'm exceeding this.
User avatar
Nick
Site Admin
Posts: 15707
Joined: Sun May 06, 2007 10:20 am
Location: West Yorkshire

#12

Post by Nick »

I can't see how the LED is getting any reverse voltage, I think its more likely the forward current limit is being hit.

However looking at the farchild doc, the circuit suggested for raising the output voltage shows rge quescient current with |Qi| which makes me believe it could go -ve. Maybe try the LED with a series diode to prevent reverse voltage.

But if the Gnd pin can go -ve, that makes me question just how valid using diode in that leg is anyway.

The two resistors on page 23 of the spec may be better if possible.
Attachments
0900766b80d25b92.pdf
(1.7 MiB) Downloaded 469 times
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
Clive
Old Hand
Posts: 374
Joined: Tue May 22, 2007 9:17 am
Location: Cheshire, England

#13

Post by Clive »

Thanks Nick, probably I'm best sticking with 4 or 5 1N4001s.
User avatar
Mike H
Amstrad Tower of Power
Posts: 20157
Joined: Sat Oct 04, 2008 5:38 pm
Location: The Fens
Contact:

#14

Post by Mike H »

What's the sequence for the LED to "expire" ? Does it go a funny colour first?

brig001 wrote:National recommend two resistors to make the regulator variable - presumably hoping that the current through the resistors on the output swamp the variation in quiescent current.
That's the other option I thought of, but wasn't sure.

Yes provided the resistor chain current was some factor larger than the ground pin current, then output will rise until the necessary V-drop across the upper resistor is reached, then stabilize.

 
 
"No matter how fast light travels it finds that the darkness has always got there first, and is waiting for it."
Post Reply