W5100 Current Consumption Problem

Hi,

I am using the circuit as attached. It appears something on the 3V3 supply of pin 2 is drawing too much current such that it pulls down the 3V3 supply to somewhere around 1.7V. When I remove L1 the digital supply goes to 3V3.W5100 Module.pdf (34.6 KB)

Imagine the current flowing through this ferrite bead!

There are only two possibilities - current flowing through pin 2 of the chip - thus chip is defective - or current flowing through magjack circuitry. I would start with second possibility, checking magjack circuit and looking for short circuits, for example one of C19 or C20 shorted thus making current flowing through magjack coil and 49.9 Ohm resistors.

Thank you for the quick response.
I disconnected power supply to Pin to of the W5100 (by cutting a track) and the power supply now measures 3V3 as expected. Something must be wrong then with the chip however also tried on another PCB and have the same problem.
Any ideas ?

Did you check magjack circuitry anyway? What is the part number of the ferrite bead you use?

Ok fixed cut and wire so power back on pin 2. Then removed connector with magjack circuitry completely. Seems good power is 3V3 as expected. The circuit in schematic draws 60mA though. Seems too much as datasheet for W5100 says 5mA and all LED’s are off.
Any ideas ?
Also on magjack connector ?
Using magjack:
https://www.digikey.com/products/en?keywords=732-4495-5-ND

Of course. I asked you to ensure there’s no shortage on caps C19 and C20. Did you do it? When you remove magjack the current pack from +3V3A to ground is removed, thus either magjack is issue, or most probably there’s a shortage.

Ok removed C19, C20 and even C22, R9 and R16.
So problem still exists and is around R10 R12 and R13 R14 and magjack connector as you said.

Looking to your circuit diagram and magjack part number I would say that it should not cause such problem. Did you wire magjack properly on the board? Is your magjack genuine? I also use WE magjack, it costs me more than $10 per piece, however in the past I was proposed for below $3 from China with comment “quality is good”, and of course I refused because saving $7 per assembly I risk getting returns of whole assembly.

Another question - is power supply able to supply required current? I do not see typical supply current in the datasheet, but guess it should be around 100 mA (looking into W5500’s datasheet). Probability that this issue is the cause is very low, but anyway worth checking.

Yes magjack is wired on the board the same way as per schematic. I assume the magjack is genuine yes, bought from Digi-Key:
https://www.digikey.com/products/en?keywords=732-4495-5-ND

Regulator rated 150mA. However datasheet for W5100 says 5mA and all LED’s off. Maybe I’m wrong on that and the W5100 together with magjack needs more than 150mA, or perhaps rather only the magjack as everything was powering up when removed the magjack.
When I remove the magjack the board draws 56mA current. With magjack in place the board briefly draws around 140mA with regulator at 3.3V then voltage drops to around 1.8V and current measures around 80mA.
Putting it straight to a bench power supply at 3V3 draws way more current. Think it must be on pins 4 and 7 (3V3A) on magjack else something would surely have blown. Also hear singing when power on.

Where did you find it? There’s section about absolute maximum ratings, it says

I-in - DC input current - ±5 mA

And it is about input current to the input pin, it is not about supply current.

This would be equivalent of (3.3-1.8)/0.08=18.75 Ohm of real impedance; there’re no such resistor devices in the circuit, two 49.9 in parallel give about 25 Ohms.

Can you put another regulator rated 300 mA (for example)? Looking to other circuits I see they use at least 800 mA regulator; input voltage you feed your regulator with also matters, and its voltage dropout specification also matters.

Yes under number 7. page 64 it says DC input current +/- 5mA. Is this wrong ?

Sorry I made a mistake. Swapped the wires. Placed the board on a bench power supply and it draws 146mA in total.
So it must be the regulator not having enough juice. Does this amount of current consumption sound right ?

I would say that under normal operating conditions chip heats considerably given its small size.

W5500 consumes 132 mA when active, thus I would consider W5100 taking 146 mA as acceptable. Never measured how much it eats anyway.

Thanks a lot for all your help. Will replace regulator and see how the programming goes…

Hi,

I have a new issue on some of the PCB’s. Everything works fine for a a while and then stops. All LED’s etc goes off.

This is only on 2 out of 7 PCB’s so far.
When I then press a multi meter positive terminal onto the 3V3 digital supply eveything coms to life and is happy while a hold the probe there. The negative probe is not even grounded.

Pic attached.
new regulator used: NCP1117LPST33T3G

Please help ?