Maker Pro
Maker Pro

What diode is used in this schematic?

Kenneth Tan

May 22, 2015
58
Joined
May 22, 2015
Messages
58
Thanks for the tip Adam. I'll update the schematic.
The powertraces are 1.6mm wide.
A gutt feeling tells me that they aren't wide enough. But I assume that also depends on the pcb quality/type.
Haven't etched a single board before, but this moment will come soon. Have to order some peroxide and hydrochloric acid.
Also I just remembered I don't have a single pcb drill. :)
Ordered the Youyoue 858d+ Hot air smd rework station today from amazon.
 

Arouse1973

Adam
Dec 18, 2013
5,178
Joined
Dec 18, 2013
Messages
5,178
Dependant on the acceptable temperature rise and voltage drop you may have to increase the width of the trace. I am assuming your using 35 um base copper with no plating? With a 25.4 mm trace length on an external layer, 3 Amps will raise the temperature of the trace by about 8.5 degrees C, with a 0.01 Volt drop.
Adam
 

Arouse1973

Adam
Dec 18, 2013
5,178
Joined
Dec 18, 2013
Messages
5,178
Here is the formula if interested

External traces: I = 0.048 x dT^0.44 x A^0.725

A = cross-sectional area in mils^2
dT = temperature rise above ambient in degrees centigrade
I = current in Amps

Adam
 

TedA

Sep 26, 2011
156
Joined
Sep 26, 2011
Messages
156
Kenneth,

Have you considered using your power switch transistor to switch the power-up current? If you do this, the manual switch, and any isolation diode, need only carry very small currents.

I'm reasonably sure that the processor I/O pin over and under voltage maximum current limit ratings are static values. However, it may be too much to expect the device to function while a pin's input protection is absorbing current. Particularly, any nearby analog inputs may not function accurately.

If you are only making one, or even 10, examples of a PCB, you can solve a trace width problem by adding a bare copper bus wire to the trace.

Ted
 

Kenneth Tan

May 22, 2015
58
Joined
May 22, 2015
Messages
58
Actually Ted, that's the same thing I was thinking of as I was lying in bed last night.
Why not add another small p-mosfet just to power the arduino ( or other processor) and have the irf9540, after power-on, enable the power rail for the real load.
That would also solve the diode problem and the need for a "no-wait" bootloader for arduino.
I guess I'm going back to the drawingboard and build a V3.0 today.
I'm loving this puzzle.
Thanks!
 

Kenneth Tan

May 22, 2015
58
Joined
May 22, 2015
Messages
58
I've found this small p-chan mosfet.
AO3401. It has a 30V Vds and a 0,9V Vgs. It's in a nice small sot-23 package.

Is 0.9V too small to function as a logic level switch? Too sensitive?
Or would a pullup resistor be sufficient to clamp the off state to 0V?
 

TedA

Sep 26, 2011
156
Joined
Sep 26, 2011
Messages
156
Kenneth,

It's not clear what you might intend to do with the AO3401. Perhaps you can fill us in.

And, if you have a link to an online schematic, that would save the rest of us some time.

Going back to the manual on / software off application, I think you can easily modify the posted circuit to use the manual switch to turn on the existing FET(s) to start the system.

Ted
 

Kenneth Tan

May 22, 2015
58
Joined
May 22, 2015
Messages
58
Yes I realize that my question wasn't that clear.
(I will post the new schematic tomorrow - had a busy day in my photostudio)

I'm thinking of changing the setup, where currently a powerfull p-chan mosfet is feeding both the load and the processor sub-power rail.
(Standard) Arduino's have a bootup fase that takes almost 2 seconds and in that period I have no control over the main power rail that during this time is powered on.

I'm thinking of replacing the p-chan mosfet for a small smd package p-chan mosfet for just powering the arduino, and let the arduino enable the powerful mosfet.
With this method, I can use any unmodified arduino or esp-8266 that have these power-on boot delays.
(There are bootloaders available for flashing that have these power on delays removed. "No-Wait" bootloaders)
The schematic is getting larger, but it's a learning process towards getting what I want it to do.

And parts of the schematic can still be used for other purposes.

The AO3401 seems to tolerate -30V Vds but has a Vgs of -1V. is this good enough to work as a logic level mosfet? In the schematics a saw on google images, it seems to be used for that purpose often.

I'm just starting with electronics so I've got loads to learn.
 

TedA

Sep 26, 2011
156
Joined
Sep 26, 2011
Messages
156
Kenneth,

I said "And, if you have a link to an online schematic, that would save the rest of us some time.", when I meant to say "data sheet" for the new P/N FET!

Vgs is a condition, not a characteristic of the device. The -1V figure "could" be a breakdown voltage, the voltage where the drain current hits a particular number, or ???

On all of your FET circuits, you want to be sure that the maximum Vgs ratings cannot be exceeded. While you are messing with the part on the bench, a zener soldered across the part's leads is often a good idea.

You might want to add a capacitor to hold the power on briefly after the manual switch is actuated, to give the arduino time to wake up. To avoid having to hold the switch for seconds.

Ted
 

Kenneth Tan

May 22, 2015
58
Joined
May 22, 2015
Messages
58
Ok, here's V3.0
I've replaced the p-chan power mosfet with a AO3401 p-chan mosfet. It delivers enough current to power an arduino and some parts. Also can handle up to 20V
The power diode (Schottky) is replaced with an 1n4007 (have a bunch of them)
Now I've added the same circuit again with the IRF9540 power p-chan mosfet. But his part will be controlled by the arduino.
All parts are now small smd part, except for the irf mosfet, tho caps and the switch.

This way the arduino gets powered up, the load remains off and the software can control what will happen next with the load.

If there are any errors in the schematic, I hope some of you find them.
Soft latch power switch CPU controlled V3 - Schema - SMD.jpg
 

Attachments

  • Soft latch power switch CPU controlled V3 - PCB - TOP - SMD.jpg
    Soft latch power switch CPU controlled V3 - PCB - TOP - SMD.jpg
    61.1 KB · Views: 211
  • Soft latch power switch CPU controlled V3 - PCB - BOTTOM - SMD.jpg
    Soft latch power switch CPU controlled V3 - PCB - BOTTOM - SMD.jpg
    52.3 KB · Views: 198

TedA

Sep 26, 2011
156
Joined
Sep 26, 2011
Messages
156
Kenneth,

Thanks for the link to a data sheet. For some reason it didn't work, but it did ID the manufacturer, Alpha-Omega, and I found a datasheet for the AO3401 on their site.

Unless you have located a particularly handy source for these parts, you might want to consider something you can get from Mouser or Digi-Key, instead.

The AO3401 has a very low gate turn-on voltage that is likely not needed for this project. This goes with a pretty low 12 volt max rating for Vgs.

By the way, what is your Vin, on pin 1 of J1? This voltage, and its possible range, will have some bearing on the design of this circuit.

Ted
 

Kenneth Tan

May 22, 2015
58
Joined
May 22, 2015
Messages
58
Ted, I want the input voltage aa flexible as possible so the circuit can be used for a wide range of applications.
The input voltage has a lower limit of 6,5V because of the Lm1117-5 cpu power rail. and currently a maximum of 20V also because of the lm1117-5.
The AO3401 is a very cheap component. The question remains if it could be used as a logic level mosfet?
 

TedA

Sep 26, 2011
156
Joined
Sep 26, 2011
Messages
156
Kenneth,

The last circuit posted will apply the full Vin voltage to the gate of the AO3041. Better keep it well under the 12V maximum rating of the part.

Some protection on the gate's a good idea in any case, and 20V rated parts are not hard to find and cheap enough.

I'm not sure what you mean by using the FET as a logic level mosfet. It qualifies as such, due to the low threshold voltage, and the low on resistance at a low gate voltage. ( 85 milliohm at -2.5V )

But parts rated for operation with 4.5 or 5 V drive will work fine, since you have at least 6.5V gate drive.

Do you have a figure for the maximum current you will need to switch with the part? We really need to know this...

Ted
 

Kenneth Tan

May 22, 2015
58
Joined
May 22, 2015
Messages
58
Indeed Ted, When the 2n7002 switches on the full Vgs has to be 12V max.
I looked at the Vds voltage.
I need to search for a new mosfet.
Problem is, that I really would like that the main power rail would support 35V. The voltage required for most powerleds.
But then I need to redesign the 5V circuit.
Thanks for the heads up, I go look for a better mosfet.

Question: Isn't it possible to put a 10K resistor between the gate of the p-chan and the drain of the n-chan mosfet. Wouldn't that work like a voltage divider?
 
Last edited:

TedA

Sep 26, 2011
156
Joined
Sep 26, 2011
Messages
156
Kenneth,

You can use a resistive voltage divider to reduce the gate voltage to some fraction of the input supply voltage, but then the gate voltage may be too small at lower supply voltages.

If you add a resistor to limit the current, and a zener diode to limit the gate voltage, the circuit can safely work over a wider range of input voltages.

Ted
 
Top