Regulator question

[grrr_arrghh]

Hi.

I came accross something on the net that was talking about regulators. It mentioned the whole virtual ground thing, and said that if you have a (for example) 5v reg, you can make it give 6v, but if you have a 6v reg, you can't make it give you 5v.

1) This set me thinking. To get a 5v reg to give you 6v, you give it a ground that is equivelent to 1v, yeah? So if you had a 6v reg, and you gave it a ground equivalent to -1v, would it give out 5v? This is just a hypothetical question.

Oh, and for negative regulators (e.g. 79xx series) does the input have to be negative?

Cheers,

Tim

 

[Nigel Goodwin]

1) This set me thinking. To get a 5v reg to give you 6v, you give it a ground that is equivelent to 1v, yeah? So if you had a 6v reg, and you gave it a ground equivalent to -1v, would it give out 5v? This is just a hypothetical question.

Presumably so, but you have the problem of obtaining the -1V. This sort of technique is often used in variable voltage PSU's, most of which can't go down as low as 0V - by having an internal supply -ve of 0V, you can design supplies that go down to zero.

Oh, and for negative regulators (e.g. 79xx series) does the input have to be negative?

Yes, of course. But like all DC voltages, the polarity has to be referenced to something - not always ground or 0V.

 

[zevon8]

This is something I have done with fixed regulators like the LM78xx series. You can insert a diode in series with the centre ground pin, and alter the output by 0.6 volts. I have put in 2 diodes before, but this "trick" gets unreliable with any more than 2 diodes.

 

[grrr_arrghh]

This is something I have done with fixed regulators like the LM78xx series. You can insert a diode in series with the centre ground pin, and alter the output by 0.6 volts. I have put in 2 diodes before, but this "trick" gets unreliable with any more than 2 diodes.

that would presumably make the output voltage higher though? Because you would still have the voltage drop accross the diode...?

Cheers Nigel,

Tim

 

[zevon8]

Yes, it makes the output rise by the voltage drop on the diode. I did it with 1n914, or similar, when I couldn't get a 6 volt regulator, only had 5 's on hand. Never tried lowering the voltage.

 

[panic mode]

Why not use voltage divider to obtain reference gnd when increasing outptu voltage? It looks just like with LM317...

 

[grrr_arrghh]

Why not use voltage divider to obtain reference gnd when increasing outptu voltage? It looks just like with LM317...

yeah, thats how its usually done, but create a negative output using a voltage divider between +v and gnd. It can't be done, unless you re-lable gnd as -v.

 

[ante]

Zevon,

The “trick” works fine at least with 5 diodes maybe more (I have not tried with more) and if you need higher voltage use a zener of whatever voltage to add to the regulators voltage. Remember to turn the zener the other way.

Ante :roll:

 

[zevon8]

True, Ante, it works with more, but I always had problems when I ran more that 2 or 3. Even the data sheets show this "trick." Perhaps it was the diodes I used, or something. Similar things can be done with the LM317 series, at really high voltages using a zener and a divider network, all that needs to be done is to stay within the i/o differential voltage between the sense and i/o pins.

Good point, thanks for adding it, Ante.