PCB problem with voltage

Thread Starter

camerart

Joined Feb 25, 2013
2,437
Hi J and E,
J, see #31
E, see #38 EDIT.

How important is placing the 100nF caps within 5mm of the VDD-VCC PINS?

There isn't a heat sync under any of the Regs. Do you think I should start again and make another PCB, with the above requirements before moving on?
 

Attachments

jjw

Joined Dec 24, 2013
605
It depends how much current is taken from 3.3V.
It has about 8V over it, so P~ 8xI
It has been proposed before to take 3.3V reg input from 5V reg output.
Btw. where do you need the 12V battery?
 

Thread Starter

camerart

Joined Feb 25, 2013
2,437
It depends how much current is taken from 3.3V.
It has about 8V over it, so P~ 8xI
It has been proposed before to take 3.3V reg input from 5V reg output.
Btw. where do you need the 12V battery?
Hi J,
As there's only a couple of LEDs at the moment, the current should be pretty low. I will eventually use the QEI on it, and it will do calculations, thats all. The 4620 has much more, and has the same REG, and doesn't get hot.

It has about 8V over it, so P~ 8xI What does this mean?

Ok, I'll rewire it to the 5V.

It's for a radio controlled flying machine, that runs on a LI-ON 3x cell12V battery.
C
 

jjw

Joined Dec 24, 2013
605
P=8xI is the wasted power on the 3.3V regulator. Without big enough copper pad on the PCB it may be too much.

I did not notice, that the 5V regulator is of the same type as the 3.3V reg., then the max power of the 5V reg might be exceeded.
 

Thread Starter

camerart

Joined Feb 25, 2013
2,437
P=8xI is the wasted power on the 3.3V regulator. Without big enough copper pad on the PCB it may be too much.

I did not notice, that the 5V regulator is of the same type as the 3.3V reg., then the max power of the 5V reg might be exceeded.
Hi J,
Yes the 5V 1A REG is a concern, this is why I have 2x 3.3V.

I have been trying to calculate the peripheral currents, to estimate the total, so far they are withing 1A TOT, yesterday, I was told that one of the unknowns is the Joystick on the transmitter. (2x PCBs almost the same 1xBASE and 1x REMOTE) My mate explained that the Joystick wouldn't need much current, so perhaps still the total may be <1A.

It looks like I'll need to re-design the PCB which will be PCB7, perhaps I can make the 5V PADS large enough to cater for the next size up 5V REG.
C
 

jjw

Joined Dec 24, 2013
605
Hi J,
Yes the 5V 1A REG is a concern, this is why I have 2x 3.3V.

I have been trying to calculate the peripheral currents, to estimate the total, so far they are withing 1A TOT, yesterday, I was told that one of the unknowns is the Joystick on the transmitter. (2x PCBs almost the same 1xBASE and 1x REMOTE) My mate explained that the Joystick wouldn't need much current, so perhaps still the total may be <1A.

It looks like I'll need to re-design the PCB which will be PCB7, perhaps I can make the 5V PADS large enough to cater for the next size up 5V REG.
C
This 5V regulator can't handle 1A, if Vin is 12V.
It is the power dissipation (Vin-Vout) x current, that matters and it can't exceed the maximum power of the regulator.
 

Thread Starter

camerart

Joined Feb 25, 2013
2,437
hi C,
Use a DVM set for current measurement, measure the current at the 12V +battery terminal, what is it.?
E
Hi E,
Why didn't I think of that:rolleyes:

This PCB (6) (2xPICs) has only 2x LEDs connected, and reads approx 1.4mA at the 10V Battery.

PCB5. I'll attach all of the peripherals, including the Transmitter (joysticks and incremental encoder) later and measure it's current. This works ok, but must have some issue causing me to make PCB6.
C
 
Last edited:

ericgibbs

Joined Jan 29, 2010
12,589
This PCB (6) (2xPICs) has only 2x LEDs connected, and reads approx 1.4mA at the 10V Battery.
hi C,
This current reading cannot be correct, when using the load you have stated.

What DVM are you using and what is it set to Read.?

E
 

Thread Starter

camerart

Joined Feb 25, 2013
2,437
hi C,
This current reading cannot be correct, when using the load you have stated.

What DVM are you using and what is it set to Read.?

E
Hi E,
I was only 1x decimal point out:)
140mA

I have 3x meters that will measure current, and all of them need attention, between them I got the above reading.
EDIT:Just remembered an old analogue meter, that shows 100mA.
C
 
Last edited:

ericgibbs

Joined Jan 29, 2010
12,589
hi C,
OK,
So for that load with a 3.3V reg from a 10vSource.
Drop across the reg 10V -3.3v = 6.7V, that is 6.7V *0.14A =0.91Watts.
Now measure the surface temperature of the 3.3Vreg while it is powering that load, what is it.??

E
 

Thread Starter

camerart

Joined Feb 25, 2013
2,437
hi C,
OK,
So for that load with a 3.3V reg from a 10vSource.
Drop across the reg 10V -3.3v = 6.7V, that is 6.7V *0.14A =0.91Watts.
Now measure the surface temperature of the 3.3Vreg while it is powering that load, what is it.??

E
Hi E,
As mentioned, there are 2x PICs on each board, plus 2x LEDS on the offending 4431.

I can cut the 4431 track and measure there?
C
 

Thread Starter

camerart

Joined Feb 25, 2013
2,437
The maximum dissipation power Pd is about 1.4W with a 30x30mm copper area, minimum ~ 0.6W.
Hi J,
Your graph needs me to get some education, as some of the terms, I don't understand.

Having said that, I can add copper areas, on PCB7, also it looks like all of the REGs need updating.
C
 
Last edited:

Thread Starter

camerart

Joined Feb 25, 2013
2,437

ericgibbs

Joined Jan 29, 2010
12,589
hi,
Do you have a bench variable power supply,?
Set it for 3.3v out, remove the onboard 3.3v reg on your PCB, use the Power supply to provide the the 3.3v
This should highlight any incorrect pcb track layout.!

E
 

Thread Starter

camerart

Joined Feb 25, 2013
2,437
hi,
Do you have a bench variable power supply,?
Set it for 3.3v out, remove the onboard 3.3v reg on your PCB, use the Power supply to provide the the 3.3v
This should highlight any incorrect pcb track layout.!

E
Hi E,
I have a variable power supply.
There won't be much time before the 4431 goes the same was as the last 3x, and may be faulty already, so what will happen to highlight the track fault?
c
 
Top