Diode Test

[Per Hansson]

Hi, I've got a fun project going just for the heck of it
Trying to repair a D-Link 10/100 Switch, DES-1008D

At first it did not even blink it's lights, I thought it was the cheap DC/DC converter that had failed so I swapped that for a much more higher end ST Microelectronics MC34063EC, the original was from Add Microtech AMC34063

That did not help, issue was the same, only 0.5v seen on the output...
So I started lifting things and cutting traces, I found out that the big diode, a 1N5822 was what stopped it from working

With it lifted the LED's come on, but I noticed something burning, so I started hunting it down with my finger, bad idea, I found it lol, the small IC in the middle of the pic, think it says STM9435 but it's very difficult to make out... It becomes like hotter than my soldering iron without the diode lol

Anyway, so I measured the 1N5822 diode while it was lifted, of course since it seems to be the problem I would change it but I don't have a spare... I read nothing in the direction you should not be able to read anything
And in the "forward" direction I read 0.17 with my DMM
What does this value mean? Is the diode ok or not?

[ratdude747]

Re: Diode Test

more importantly what does it say in the reverse direction. a bit of resistance seems normal testing a few i have around.

[starfury1]

Re: Diode Test

I hate DMM's
it should read open circuit one direct (Very high resistance)
near short the other
(actually I think your reading a voltage drop across it or something..forgotten now on resistance ranges..but dont quote me on that)

Diodes mostly Fail completely
but sure there are cases of some aspect cause problems in circuit

Type is High speed rectifier its still a diode PDF

Here the pdf on the STM9435 Appears to be a FET

Sound like that might be your problem
but wait for better feedback from hopefully one of those here thats familiar with them mate (I haven't got a clue on them)

Sounds right to me for the Diode
but hey my brain hasn't been on electronics for a while Per

Cheers

[Per Hansson]

Re: Diode Test

In the reverse direction I have infinite resistance / no connection just like it should be (sorry if this wasn't clear in my first post)

And in the direction it is supposed to lead current I get a reading of 0.17, whatever that 0.17 means...

I also seem to have forgot to add the pic, so here it is, 2:nd attempt

EDIT; Reading the PDF of the diode it says this;
"Electrical Characteristics; Forward Voltage @ 3.0 A = 525mV"

Does that mean I should read 0.525 with the multimeter and not 0.17?

[gg1978]

Re: Diode Test

I've got a Fluke 175 here in front of me at work, and according to it, it's measuring volts DC .. For a 1N4007, it gets about 0.56V in the forward direction, and "OL" reverse biased.. I assume the DMM is just trying to inject a small amount of current through the junction of the device, and simply measures the voltage dropped over the junction based on the current.. Hence, it gives a voltage when forward biased, and no volts, indicates "OL" when reverse biased.. The forward voltage that i've measured is very close to the nominal 0.6 to 0.7V forward voltage for a standard diode..

Since a DMM obviously can't shoot 3.0A through your diode, i would expect the voltage indication in the forward direction to be lower, as the current is lower.

[gg1978]

Re: Diode Test

Google for the 1N5822 datasheet, on the second page, look at the "forward voltage characteristics".. I think that might answer the question fully.. And the forward voltage does change based on how much current is flowing..

[Gianni]

Re: Diode Test

gg1978 is right, look at the characteristics in the datasheet.

Does that mean I should read 0.525 with the multimeter and not 0.17?

With the DMM no, but if you can test the diode with current at 3A and Tamb=25ºC than yes.

Schottky diodes have very low forward voltage to reduce power dissipation (V X I), they are used in all PSU. In the past they were available only for low voltage (40-60V) but nowadays there are also high voltage schottky (>150V).
So it is normal to read 0.17V with a DMM because the current is very low, probably 1mA. The problem with schottky diodes is the reverse current, it is very high compared to fast-ultrafast-standard diodes and can be in the range of some mA. It depends on the applied reverse voltage and temperature.

Edit:

For the fet you can use this one STS5PF30L if you need to replace it and you can't find the original replacement.


Ciao
Gianni

[PlainBill]

Re: Diode Test

In the reverse direction I have infinite resistance / no connection just like it should be (sorry if this wasn't clear in my first post)

And in the direction it is supposed to lead current I get a reading of 0.17, whatever that 0.17 means...

I also seem to have forgot to add the pic, so here it is, 2:nd attempt

EDIT; Reading the PDF of the diode it says this;
"Electrical Characteristics; Forward Voltage @ 3.0 A = 525mV"

Does that mean I should read 0.525 with the multimeter and not 0.17?

The ohms function on a DMM usually works by injecting a low voltage into the test leads and measuring the current flow. That is translated into ohms using ohms law. ( If they use .1 volts and the current is 1milliamp the resistance must be - 100 ohms). That voltage is typically less than .25 volts so the reading won't be affected by any diodes or transistors.

Most also have a diode test range. In that range they use a higher voltage (typically greater than 2 volts), and often display the voltage across the junction. Of course, if the junction is reversed biased, the meter will show it as open.

Without knowing the characteristics of a meter and what range you have it set on, the numbers are meaningless. HOWEVER, the 1N5822 is a Schottky diode, with a low forward voltage. The data sheet shows about .2V at 10 Ma, so .17 is acceptable, especially since you are getting open in the reverse direction.

The STM9435 is a FET, but even more interesting are the inductor (toroidal transformer?) and U3 (CHN063EC?) next to it. Odds are this is a bucking regulator and the diode was to protect Q2 by absorbing the voltage spike when the FET turned off.

Alas, diode was not present, FET is now puddle of molten silicon.

PlainBill

[Gianni]

Re: Diode Test

Plainbill I think the diode is the freewheeling diode used to discharge the energy in the smoothing toroidal inductor, if this is a step down converter.

The IC CHN063 is a power IC and can be used as step-up step down so I think the small fet is not use for power conversion.

@ Per Hansson: maybe you can follow the tracks and draw the electrical diagram to better understand what the small Fet does in the circuit.

Ciao
Gianni

[Per Hansson]

Re: Diode Test

PlainBill; I checked the diode using both a Fluke 113 and a CHY 17 multimeter, both of them in the diode test mode...

U3 on the PCB is a ST Microelectronics MC34063EC, this one I already suspected was bad so I replaced it, originally there was a Add Microtech AMC34063...

Gianni; You shold be correct that this is a step down converter, it's input is ca 10v and output I am unsure what it should be, but around 2 or 3 volts I'd say for the D-Link switching chip that does all the ethernet magic

Your help is much valued guys, I am doing this only as a learning excercise of course, the switch itself is worth like nothing

I do have a spare Si4410DY N-Channel mosfet but I guess that wont work to replace the STM9435 P-Channel mosfet, though I am not in any way certain that there is anything wrong with the STM9435, I only had the thing running without the diode for a few seconds before I figured out what it was that was hot, and the specs for it says it is able to operate at up to 150°C!

[Gianni]

Re: Diode Test

Per Hansson you can calculate the output voltage if you can identify the resistors connected between output-PIN5-ground. The reference is 1.25V on PIN5 so you can divide it by the resistor connected to ground, than multiply the current for the resistor connected to the output and add to the calculated voltage the reference voltage.

In the datasheet step-down example FIG15:
R1=1.2Kohm
R2=3.6Kohm

Iref = Vref/R1 = 1.25/1200 =1.0416mA

VR2 = R2 x Iref = 3600 x 0.0010416 = 3.744V

Vout = VR2 + Vref = 3.749 + 1.25 = 4.999V

Is it possible that the small Fet is used as a switch or linear regulator? Maybe there's a short circuit after it and the current is so high to run it very hot.

Ciao
Gianni

[Per Hansson]

Re: Diode Test

Well I think I know what the voltage is supposed to be
Because as I said after I lifted the diode the LED's and all come on and the thing looked to be working

Only the small fet became like 150°C hot...
The voltage measured into the switching chip that does the ethernet magic is ca 3.2v, this sounds very reasonable to me...

What is interesting is that now when I put the diode back into the circuit the thing still "works" i.e. the LED's come on just fine and voltage stays at 3.2v
But the small fet still becomes very very very hot...

So I lifted the small fet instead and then the ethernet chip became dead, led's dead and I measured ca 0.5v like when I started on the ethernet chip

I did check the voltage into the small fet before I lifted it and it was around 4v in and 3.2v out, with it lifted there is 10v on the in side which is lifted...

[Per Hansson]

Re: Diode Test

R67 = 0.84k ohm
R68 = 0.75k ohm

Iref = Vref/R67 = 1.25/840 =1.488mA

VR2 = R2 x Iref = 750 x 0,0014880 = 1.12V

Vout = VR2 + Vref = 1.12 + 1.25 = 2.37V

But actually now with the small fet lifted I am seeing 0.5v here, hmm
Do note that I didn't lift the resistors, I measured them in circuit

[Gianni]

Re: Diode Test

Per Hansson: I think you have measured wrong values, in the attached pic I see R67 has 3 digit number like 122, so it should be 1200ohm.
Looking at the tracks it seems to me that the feedback resistor is R66 not R68.
I can't identify R66, perhaps you can post the value because.

R68 is connected between FET gate and ground.

Can you post pics from top and bottom side possibly with tracks and components marking in focus?

Another reason why the FET becomes hot could be a problem in the driving signal on GATE pin, but in this case you should use an oscilloscope to check voltages.

Ciao
Gianni

[Per Hansson]

Re: Diode Test

Ah, sorry, I mixed up the numbers
I meant R66 but wrote R68...

I lifted them now, R66=1000ohm & R67=1200ohm

Iref = Vref/R67 = 1.25/1200 =1.042mA

VR2 = R66 x Iref = 1000 x 0.0010417 = 1.0417V

Vout = VR2 + Vref = 1.0417 + 1.25 = 2.2917V

[Gianni]

Re: Diode Test

Per Hansson for what I can understand the circuit should be like my attachment.
So the FET is the step-down POWER SWITCH and the MC34063 is used only as controller.

You can check with the DMM if the circuit is like your board.

Now it is very late and I must get up within 3 hours... see you later

Ciao
Gianni

[Gianni]

Re: Diode Test

I did check the voltage into the small fet before I lifted it and it was around 4v in and 3.2v out, with it lifted there is 10v on the in side which is lifted...

Well 0.8V drop on the FET can justify the high temperature when working in linear mode with output current of 0.6~0.8A (Pwasted 0.48~0.64W).
Since the FET is connected to a switching controller, maybe the voltage drop you measured is not real but only "what" the DMM can read on a square waveform (Drain pin) at switching frequency (KHz).

with it lifted there is 10v on the in side which is lifted...

It looks to me like an overload or bad ac adapter: the MC34063 "can operate" at 3V but if you read the datasheet the parameters are always specified at Vin=5V so an input voltage of 4V is too low IMO. Moreover you have to drive the FET which need at least -4.5V on GATE to work at low RDSon (85mOhm). With a VCC at 4V and considering the voltage drop on the controller (between PIN 2 and ground), probably the voltage driving the FET on GATE pin is too low (-3~-3.6V).

Has the ac adapter feeding the board a label with output power/voltage/current?
Can you check it with an external load and verify if it can sustain the rated load/output?
Can you check the output ripple?

If the ac adapter is OK, I would disconnect the step-down output (if possible) and I would check it with an external load to understand how much power it can deliver.
At the calculated voltage (2.29V) I would expect an output current in the range of 2~3A with a temperature FET not higher than 80~90ºC.

Let see what you can do because to troubleshoot some circuits it is not easy without an oscilloscope, lab PSU and loads.

Ciao
Gianni

[Per Hansson]

Re: Diode Test

I cracked open the AC adapter first thing I did, it's very simple with only a transformer, a cap, one resistor and 4 diodes to form DC current
I checked the capacitor and it was ok, both ESR and capacitance...

I also tested the D-Link switch with another known good adapter and behaviour was the same

Unfortunantey I think I may have killed it now with my testing, before as I said without the diode or FET it would actually power up and blink it's lights
It did a real diagnostics test, first power turned on then it blinked each port's light to let you know it was working...

But now just a number of LED's get lit when I power it on, and they stay that way, no matter if I remove the diode or not... And the switching chip becomes very hot, so I think I cooked it

[Gianni]

Re: Diode Test

I'm sorry it died.

I did a quick test on a step down board in my lab: I have removed the freewheeling diode.
Without load I had the output voltage but as soon as I connected the load (a resistor) the IC has died
My board is different from your circuit, it has a monolithic IC, so I don't know if the behavior it is the same.
For sure it is not recommended to remove the diode because the energy stored in the inductor must find another path to discharge, so you don't know what it causes in the circuit.

What I can suggest you for future opportunities is to try to test the power supply or step-down converter isolating it from the board if it is possible.
Another thing I would do is to feed the board with an external PS, if you have calculated the voltage you can feed the circuit and check if it works.
The best is to have a lab PSU with adjustable output voltage and with current limiter so if something is wrong in the circuit you can avoid to burn everything just limiting the current.
Proceeding in this way you can understand if the problem is on the board or on the step-down converter.

Ciao
Gianni

[Gianni]

Re: Diode Test

Per Hansson ignore the above message... due to "edit time limit" I have lost some words... here below the original post:

I'm sorry it died.

If the power switch become hot probably it is shorted.
What it is strange to me is that the board power up with the FET lifted: does it have two different supply voltage? I mean the 2.29V seems to me a low voltage for logic or micro-controller but probably on the Ethernet port the voltage is higher, I'm not an expert about them.
So again if the only voltage feeding the board is generated by the step-down converter, removing the FET is like opening the input voltage so nothing should be powered up.

I did a quick test on a step down board in my lab: I have removed the freewheeling diode.
Without load I had the output voltage but as soon as I connected the load (a resistor) the IC has died
My board is different from your circuit, it has a monolithic IC, so I don't know if the behavior it is the same.
For sure it is not recommended to remove the diode because the energy stored in the inductor must find another path to discharge, so you don't know what it causes in the circuit.

What I can suggest you for future opportunities is to try to test the power supply or step-down converter isolating it from the board if it is possible.
Another thing I would do is to feed the board with an external PS, if you have calculated the voltage you can feed the circuit and check if it works.
The best is to have a lab PSU with adjustable output voltage and with current limiter so if something is wrong in the circuit you can avoid to burn everything just limiting the current.
Proceeding in this way you can understand if the problem is on the board or on the step-down converter.

Ciao
Gianni