[SOLVED] Voltage regulator equation term question

[d123]

Hi.
In the attached image, what does the lower case letter b represent here? Also, to use a MOSFET instead of a BJT, what do I need to substitute or omit (if anything) in this formula? Thanks.

 

[D.A.(Tony)Stewart]

b, Beta of Q

L78xx supports 1A with 2V drop min. so if L7812 then Vin must be just over 14V. and >7V for L7805.
Then up to 1.5A with more input drop V, but if drop is too high, the Power * Thermal resistance thermal rise may reach thermal cutout.

All L78 type regulators are emitter followers with comparator drive on voltage sensing against internal Vref.
The bypass transistor pulls up by sensing current to act as current limiter for the regulator while supplying as much as load demands but without OC detection.

To use MOSFET , a low RdsOn such as 50 mOhm at e.g. Vgs = -1V to 1.5V drop at max desired current and power dissipation in Q. This choice will affect min. Vdrop overall. by design to higher or lower and may latchup or be unstable.

Consider https://www.onsemi.com/pub_link/Collateral/AN1686-D.PDF

Again the bypass device does not have any thermal or current limit protection unless both mounted on same heatsink.

The BJT starts to share the load when R current shunt reaches saturation of Vbe. (~0.7)

 

[d123]

Power supply ripple voltage affecting transistor turn-on doubt

Back like a bad penny, my apologies...

To implement the current bypass (load carrying) and short circuit protection transistors around a 7805 using MOSFETS, I am worried that the 2V drop which turns them on may be caused by the power supply ripple voltage (because I admittedly don't fully understand how this circuit will function beyond bread-boarding it at 12V and a low current load, and checking each MOSFET I'll use does turn on at approx. 2V below the rail voltage).

The transformer puts out 16VAC, the rectifier is a full bridge of 1N5408s followed by a 0.5Ohm resistor in series with the positive DC before the filter capacitors.

Based on the ripple calculation in the image, and the information kindly given by the transformer manufacturer, I understand that approx. 2000uF per DC Amp for 15% ripple. Not sure if I understand but worry that's too much if the transistors are directly in series with the lightly regulated supply and come before the voltage regulator itself.

I calculated that to get about 1V ripple I'd need approx. 11,000 - 12,000uF filter caps. This seems a lot as I usually see people apply about 5,000uF for this purpose.

a) Is 11,000uF overkill, or actually necessary to prevent the P-channel MOSFETS turning off and on erratically in an unwanted way due to ripple voltage?

b) Will an 11,000uF filter mean that I need to additionally protect surrounding devices from its discharge besides placing protection diodes around the voltage regulator and counting on the MOSFET body diode to protect them?

The schematic is a modified ST 78xx series application schematic, and the ripple formula is from this website:

**broken link removed**

Thanks.



View attachment rsc and bypass.pdf

 

[Audioguru]

Re: Power supply ripple voltage affecting transistor turn-on doubt

You are talking about using a P-channel Mosfet that might begin turning on when its source voltage is 4V higher than its gate voltage but the circuit you posted uses a PNP transistor where its emitter voltage is only about 1V higher than its base voltage.

 

[d123]

Re: Power supply ripple voltage affecting transistor turn-on doubt

Hi,

Correct, you are very observant, I won't try to sell you exchange traded funds or government bonds... The image is just to show the datasheet schematic I'm following for the sake of simplicity and clarity, I'm copying it verbatim except for using MOSFETs instead of PMOS devices. Although at this rate I may just bury the slow progress power supply in the garden at midnight and have done with the problem.

The datasheet said -2 to -4V VGS th, so I checked five IRF9520 gate voltages and source output voltages with a DMM, an LED, using a trimpot to slowly decrease from full rail voltage with the bench supply I have (11.7V) to approx. 9.7V gate voltages for all five, and see when the devices began to output voltage above 0V at the source. One MOSFET of the five I interviewed was too eager, at 1.5V drop turn-on, so that has been excluded from the selection process , the other four do turn on at -2V from rail voltage.

I think I need 2.7 Ohm for the RSC resistor, and 27 Ohms for the R1 resistor for the 0.75A max output.

My concern is that if the unregulated supply voltage has a ripple of 2V, the IRF9520s may turn on when they shouldn't.

So, should I use enough filter capacitance to keep ripple voltage around 1V or lower, or will the ripple have no effect on the PMOS devices as they will respond in a constant manner to the supply rail (unregulated) voltage anyway?

 

[FvM]

Re: Power supply ripple voltage affecting transistor turn-on doubt

Your considerations about turning on the output transistor by voltage ripple seems to be based on an error of reasoning.

The transistor is turned on by current through the shunt resistor rather than voltage. The only way to turn it on inadvertently by an input voltage change would be to use a huge capacitor instead of the 0.33µF capacitor. But you don't want to do this.

I still think that using a PMOS instead of PNP transistor is a bad idea, because it increases the regulator drop-out voltage, depending on the actual Vgs voltage required for full output current.

 

[Audioguru]

The datasheet for the IRF9520 Mosfet shows that the maximum Vgs threshold voltage is 4V for a current of only 250uA which is 0.25mA. Is that what you want or do you want the Mosfet to conduct many Amps?
When the Mosfet conducts many Amps then its maximum Vgs voltage could be as high as 10V but the PNP transistors in the schematic need 0.65V for the current limiter and maybe 1V or 1.5V for the series pass transistor.

Therefore you should not use Mosfets in this circuit.

EDIT: The original circuit does not use any PMOS which are Mosfets. They are PNP transistors.

 

[d123]

Thanks for the clarifications, several important points there, both replies are very helpful and clear, based on this information I'll use PNP. Thank you.