All I’m wanting to do is drive this strip below with a micro controller.
I’ve got a 5v 3A power supply which is hard connected to my bread board.
It’s directly powering the strip and I’d like to power my micro controller too.
I guess then I’m looking for a board that can be powered off the pins.
The Xiao ESP32S3 looks promising. Here it talks about powering the it through the 5v Line.
5V [pin] - This is 5v out from the USB port. You can also use this as a voltage input but you must have some sort of diode (schottky, signal, power) between your external power source and this pin with anode to battery, cathode to 5V pin.
Suits me, I have heaps of diodes. In particular I own a bunch of IN4148s which I believe to be signal diodes. Would they work?
Any idea why this diode is required at all?
The diode is there in order to ensure that, if you have a USB cable connected at the same time as the external power, then the USB input will not attempt to provide current to the 5V pin in the event that it sees that pin at a lower potential. Considering that the external power supply is also driving a heavy load in the LEDS, a nominal 5V external supply might drop quite low. The important ratings in this case are the reverse voltage rating which is easily met at 75V, and the forward continuous current which is marginal for that MCU board at 300mA. A 1A/50V diode such as 1N4001 would be a better choice.
Just to throw another horse in the race, the Pico has really good documentation on powering it including how to add that diode. Starts at 4.5 Powering the Pico.
It has the same need for a diode as @Jeff105671 mentioned. Can get away without it as long as you don’t power it from both sources at once, but that’s just an accident waiting to happen so it might be worth the $1 diode.
Think what your saying here is that the 3a power supply on the anode does not need to be accounted for, rather, It’s the 300ma~ on the cathode at the chip that I need to spec for.
Did I understand that right?
Hi Pix
As Jeff points out the 1N4148 has a modest current capability and @ 300mA i think that would be pretty max maximum. The forward voltage drop is quite high so a schottky would be a better choice. The 1N5819 @ 1A or the 1N5822 @ 3A would be a better choice with the slightly lower forward voltage. 0.6V for the 1N5819 and 0.5V for 1N5822. but the higher current capability is a decided advantage. The 1N4148 is classified as a small signal diode and are very good in this use. A bit of trivia, used to be known as 1N914 and are still listed as such sometimes.
Interestingly I have some 1N5819s and on a multimeter diode test they measure a forward voltage drop ov 100 to 200mV. The figures above are at 1A and 3A respectively.
Cheers Bob
Seems like the Raspberry Pi Datasheet is keen on a schottky diode
I also like what it’s saying about the Mosfet but I don’t have a vbus port on this chip.
Something I want to check here.
I’ve made an assumption here that there is no way I could power the Xiao off USB with a 3a raspberry pi power supply and then the strip off the 5v pin because the maximum current that 5v pin would allow is like… at most 700ma.
However, when I read the documentation closely it doesn’t seem to mentions anything about a maximum current… It just calls it USB Power Thru which suggests to me that it’s passing power unmitigated. Am I missing something?
The 300mA specification is the continuous forward current. That is the current flow through the device - it’s the same at the cathode and the anode, and is barely enough to power the ESP32S with WiFi enabled. Some data sheets quote 200mA which would be more than the ESP32S with WiFi enabled would draw, and the diode would be destroyed.
If you have a Schottky available it would be preferable due to the lower voltage drop, but check the current rating.
The limiting factor for the Xiao ESP32S is probably the maximum current that the USB port and the traces to the power tab could support. Even if they survived the current there would be a significant voltage drop. But remember that the diode for the external supply is only required if you have USB and the external supply connected at the same time. If there is a clear distinction between a development system, USB connected with LED power from USB and a very small LED string on the one hand, and the production system, with no USB, a large LED string and external power on the other, then there is no need for a diode. That’s how I would do it, but that would require a programming fixture - an interesting project in its own right.
Yeah, 3 Amps might be pushing it. For the Pico, I have only ever pushed 2.5a through it somewhat continuously. Unsure if the traces will hold up under 3 Amps continuously hahaha.
Unless I’ve missed the mark with this one, the easiest way to do what you’re doing would be something similar to the attached photo, I have a few WLED setups at home that use this exact wiring without the cap and the resistor. Otherwise I have made up these little boards specifically for XIAOs to run up to 3A of WS2812 strips, powered by USB-C that I could arrange for one to be put into your next order.
That’s fine as long as you don’t need to connect the USB. But during the development the diode is required when the external power is connected to the MCU 5V, so that the only change needed between development and final use is to unplug the USB.
I guess the other option is to use a USB cable with the 5V disconnected.
I knew there was something that I was missing with this, in my experience messing with this kind of stuff I’ve always made note to not have both power sources active at the same time. I agree with your suggestion of the 1N4001 being plenty up to the task, although I would go with a 1N4007 solely based on price.
Like the wire diagram above, but I used a USB connector with only the 5V wires connected. So when in dev mode and connected the the PC, the PC was powering the ESP32 while the external 5v was powering the LED stripe.
Then in live mode, 5V main power → usb connecter into the USB port on the ESP32
Hi Dan
Yes the only difference here is the reverse voltage, the 1N4007 being 1000V. The forward voltage drop could be a bit of a downer being quoted as 1V @ 1A. But a diode check on a multimeter could measure a bit less.
This is the downside of low voltage systems. At 5V you might just get away with it (maybe, depending on application) but would be a no go for a 3.3V system.
There are some quite low forward voltage diodes around but once again as always there is some trade off. As the PN junction is designed for a low voltage so the reverse leakage current usually goes up in relation. That is the lower the forward voltage the greater the reverse leakage. I don’t think you can have both low voltage and low leakage.
I think there are some Mosfets and circuits appearing to overcome some of this. But if we wait long enough some bright spark will fix all of this.
Cheers Bob
VIN is only used for the 3.3V regulator which has a dropout maximum of 1V, so it can cope with a 1V drop in the 5V supply without a problem. If the external 5V was much below nominal then the diode type might become relevant.
Hi Jeff
Surely if you are planning on something which has a good chance of working without too much drama you should consider worst case.
In your case you have a 3.3V regulator which has a max dropout of 1V.
Then a diode which could have a forward voltage drop of 1V @ 1A.
Then you have a bloke called Murphy who will make sure that in some cases all these tolerances will go the wrong way.
Now 3.3 + 1 = 4.3.
5 - 1 = 4.
Granted most of the time you will get away with this if all the tolerances behave. But there is always the chance that there will be occasions where this not the case and all components are within their respective specs and the project or whatever falls over due to Mr Murphy. The experienced person might get on top if this problem quickly but this Forum is populated with newcomers to the field and other not so experienced people and this sort of thing becomes a nightmare. Particularly where person 1 builds a circuit that works fine, then newcomer person 2 builds exactly the same thing and fails miserably. How you explain to him (or her) that the problem is not considering worst case tolerances and he just got unlucky.
The classic case here is the practice of powering 5V relays with 3.3V. Most of them worked but there were some reports of not working at 3.3V and OK at 5V. I think everybody has woken up to this one now.
Cheers Bob
I’ve been thinking about this more.
So far I’ve been talking about my production setup when all the code is done.
But while I’m developing the actual pattern I was just going to have the Xiao connected to my computer via USB, and seperately power the strip off the power supply. Then I was going to have them share a common ground.
By connecting the ground of the Xiao to the ground of the power supply isn’t it possible that the return path of all 3amps of current might choose to flow through the xiao, and then my computers usb port, on it’s path back to earth damaging them both?
Or, will the xiao have a higher resistance due to the small traces, and perhaps all current will prefer to go through the relatively thicker cables of my power supply?
Hi Pix
The current will take the path of least resistance. As will any fluid. Sometimes it helps to compare electrical current with the action of water.The head of water supplying pressure = Voltage, The pipe as wire containing some resistance to flow like the pipe interior and a tap as a variable resistor.
In your case the current will share depending upon the resistance as you say. You can minimise this by taking the heavy current part of the circuit directly back to the supply and making your common connection at that point.
Cheers Bob
