Ground current for a voltage regulator

[Weaver]

Homework Statement

Homework Equations

3. The Attempt at a Solution [/B]
Currently working on the question relating to the ground current on the voltage regulator, 7805 (data sheet included below)

I am very much unsure what to calculate here
I'm assuming the current to ground is the input current - output current

However, I don't know how to calculate the input current

Perhaps do I assume there is an PCB track impedance of 0.1 ohm and 0.2uH (from previous questions)? And take the magnitude of the current?

There is no specification in the LPC1768 for the max current the Vin pin draws. However I found the mbed can produce 400 mA of current on the GPIO, so perhaps I can use that for output current of the regulator.

Does this seem like the right approach?UPDATE: Is this what the quiescent current is?

 

[CWatters]

There are several issues you need to address.. some hints...

The problem statement mentions ripple on the 24V supply. Have studied the design of a transformer based DC power supply (eg transformer, rectifier and capacitor)? Do you know how to select the capacitor to minimise ripple?

1A is a lot of current to have flowing around a PCB with sensitive analogue components. What happens to the voltage of ground traces if there is any resistance/inductance in them?

 

[CWatters]

The current flowing in the reference pin of a 7805 is small. You don't normally need to worry about the ability of a PCB trace to handle it if that's your concern. I'm sure some data sheets will even give you a recommended PCB layout for the 7805 and it's high frequency decoupling capacitors depending on the package.

I think this question is more concerned with the topology of other parts of the circuit.

 

[Weaver]

There are several issues you need to address.. some hints...

The problem statement mentions ripple on the 24V supply. Have studied the design of a transformer based DC power supply (eg transformer, rectifier and capacitor)? Do you know how to select the capacitor to minimise ripple?

1A is a lot of current to have flowing around a PCB with sensitive analogue components. What happens to the voltage of ground traces if there is any resistance/inductance in them?

Thank you for your reply!

We actually haven't studied decoupling capacitors. The lecturer for this subject has assumed we have, which is one of the reason I have posted this.

I found an equation online that suggests using C =I/(Vripple x f) to calculate this.

I have calculated the voltage on the track using the assumption of a wire having 0.1 Ohm and 0.2uH impedance.

I have also calculated the voltage across the PCB track to be 10 + 64pi mV

Z = 0.1 + j2pi(1.6 x 10^6)(0.2 x 10^-6)
= 0.1 + j16/25 pi

V = IZ
V = (100 x 10^-3)(0.1) + j16/25 pi)
V = 10 + j64pi mV

Can I use this as Vripple in the above equation

Or do I need to do consider the components and put a capacitor in parallel and the fine a value of C that reduces the ripple?

An accompanying question asks to find a value of C so that the error is less that 0.8 mV.

Would I consider that by subtracting it from the Vripple value?

The capacitor would then be in parrarell with X1/Q1?

 

[Weaver]

The current flowing in the reference pin of a 7805 is small. You don't normally need to worry about the ability of a PCB trace to handle it if that's your concern. I'm sure some data sheets will even give you a recommended PCB layout for the 7805 and it's high frequency decoupling capacitors depending on the package.

I think this question is more concerned with the topology of other parts of the circuit.

I understand we don't need to worry about it. We have been given questions with this circuit and I think the point of an accompanying questiom asking to find the currentcto ground is to highlight that it isn't a concern compared to other components

In this case, would I just assume the quiescent current or is there some other calculation?

The datasheet does also include the decoupling capacitors so I will use them in my updated design

 

[CWatters]

Another method to work out the capacitance is to use...

Q=CV

Differentiate both sides gives you..

I=C dV/dt

Rearrange to give

C=I dt/dV

Where...
I is the ripple current
dV = the allowed ripple voltage
dt = the period

The frequency is reasonably high so the capacitance required shouldn't be very large meaning there is probably scope to double the calculate value the provide a safety margin.

 

[CWatters]

Ok I'm going to have to look at your other thread again tomorrow on my PC. It's hard to read on this tablet but it looks like they are asking you leading questions that walk you through a design process. My answers above don't take that into account. For example I see an earlier question mentions a track impedance which makes your method more appropriate.

 

[gneill]

Currently working on the question relating to the ground current on the voltage regulator, 7805 (data sheet included below)

I am very much unsure what to calculate here
I'm assuming the current to ground is the input current - output current

Yup.
Your datasheet has an entry for Quiescent Current. That should give you what you want. See, for example:
Understanding the Terms and Definitions of LDO Voltage Regulators (it's a Texas Instruments pdf)

 

[Weaver]

Another method to work out the capacitance is to use...

Q=CV

Differentiate both sides gives you..

I=C dV/dt

Rearrange to give

C=I dt/dV

Where...
I is the ripple current
dV = the allowed ripple voltage
dt = the period

The frequency is reasonably high so the capacitance required shouldn't be very large meaning there is probably scope to double the calculate value the provide a safety margin.

Great. Thank you very much! I think the value works out at about 37 uF

 

[Weaver]

Yup.
Your datasheet has an entry for Quiescent Current. That should give you what you want. See, for example:
Understanding the Terms and Definitions of LDO Voltage Regulators (it's a Texas Instruments pdf)

That's brillant! Thank you. Thanks for the link also!

 

[Weaver]

Ok I'm going to have to look at your other thread again tomorrow on my PC. It's hard to read on this tablet but it looks like they are asking you leading questions that walk you through a design process. My answers above don't take that into account. For example I see an earlier question mentions a track impedance which makes your method more appropriate.

I really appreciate that. I think I have worked out most of the answers now. I know I am stick stick stuck on the ground current flowing out of the op-amp. I'll post updated workings by this evening. I really appreciate all your help. Thank you so much

 

[Weaver]

There are several issues you need to address.. some hints...

The problem statement mentions ripple on the 24V supply. Have studied the design of a transformer based DC power supply (eg transformer, rectifier and capacitor)? Do you know how to select the capacitor to minimise ripple?

1A is a lot of current to have flowing around a PCB with sensitive analogue components. What happens to the voltage of ground traces if there is any resistance/inductance in them?

Upon revision of the design, I am not sure what to do with this 1A. We haven't really covered ways to reduce DC current to ground, only the issues that arise from AC. They only thing I can thing of is make sure this ground wire is connect to the ground node and doesn't' connect to say the ground wire of the op amp until reaching the ground node? In that way, it is isolated?

 

[CWatters]

That's the right idea. Separate ground traces for the high current and sensitive parts.

That 1A flowing through the 0.1Ohm trace resistance causes a voltage drop. It causes a trace that should ideally be at 0V along it's whole length to have a higher voltage at one end. This is commonly called "ground bounce".

Common solution would be to create a "star point" where the -ve of the 24V supply comes into the PCB. Then run separate traces from the star point to each part of the circuit where there would be decoupling capacitors designed for that part.

 

[CWatters]

Just for info.. On multi layer PCBs an inner layer is normally used for ground. This can be divided into different zones, one for high current, one for sensitive etc these would typically be connected together at a common point near where the 0v form the 24V power supply comes into the board.

Another layer might be divided into 24V, 5V and 3.3V zones with the regulators that produce these voltages placed where the zones meet.

 

[Weaver]

Just for info.. On multi layer PCBs an inner layer is normally used for ground. This can be divided into different zones, one for high current, one for sensitive etc these would typically be connected together at a common point near where the 0v form the 24V power supply comes into the board.

Another layer might be divided into 24V, 5V and 3.3V zones with the regulators that produce these voltages placed where the zones meet.

Thanks. I didn't know that. When we did ground planes during a PCB design module, it was just a case of cover all the free space with a ground plane and everything connects to that.

I have to say, I really didn't appreciate the importance of ground before this exercise.

 

[CWatters]

Yes grounding issues can be quite complicated. On one project I worked on we initially had a simple ground plane over the whole pcb. Along one edge were connectors for signals going to other equipment. We had issues with radio frequency noise coming out of the enclosure on these connectors including ground/0V wires. Part of the solution involved cutting the ground plane so there was a strip along that edge of the PCB connected to 0V at just one place where the power came onto the board. Signals that crossed over into this zone to get to the connectors did so via low pass RC or LC filters that used this "clean" ground.