How to best change orientation of diode, and other errors?

[JacobW]

@straubm

but according to KiCad Calculator 10A would require ~7.5mm trace width. For 7A your 4.7mm are ok.

What… 7.5?

When I calculated with EE Toolkit on Iphone I got 3.83?

How did you find the leftover traces?

[Fabio_De_Bernardi]

You have different settings :
DeltaT = 20
Thickness cu 43um

If I set this data in the result calculation it is almost the same as yours.

[JacobW]

What is “standard”?

What should I use?

I found on JLCPCB, witch I will be using, that they use 1oz witch seams to be 34.1 µm. That seems to be standard at most pcb manufacturer, wonder why EE Toolkit used 43. I have not put in 43 so that must be defult.

After taking a look att Altum design I found what 1 Oz is 34.1 µm.

If I use this in EE Toolkit I gett this results.

8.2A in a 4.6mm track, witch seems to be the widest track I can get.

I can laborate with temp, don’t know what is ok?

[straubm]

Environmental temperature of 25°C is unrealistic. The board is not in free air, it is in an enclosure, so around 30-40° are more likely. A temperature rise of 20° is ok.
But again: you said you go for 7A/channel. There are 8 channels, so you have 56A in total, not 7 or 10.

[straubm]

Btw, your track length is some 25cm, not 20cm.

[JacobW]

Maximum it must have is 150W in 5V on a chanels, preferably devided by 7 Chanel. I can not devide the load 150W witch is 30A at 5V with 7. Witch is 4.3 something. I can only use 2.5A. So if I want I can have 2.52=5A on a Chanel or 2.53=7.5A

But I should ame for som headroom.

The amount of current that the load will probably have is 2.5A. I want to be able to run 3 on each output up to approx. 150W in total, but of course I want to have excess capacity. when I started drawing I thought that there can be no problem with 80A, but it turned out completely crazy values. I can’t solve it. But anyway. It will not be in any particular enclosure, I would say it will be in the open air. It will at least have several liters of air around it.

But back to the diods. How do I in a simple way fix them?

Why can I draw them correctly in schematics ant then they appear wrong in pcb?

And what will happen with the return current threw ground when so many tracks are blocking the return path?

[Fabio_De_Bernardi]

@JacobW

I was locked at the design of this relay card: https://www.electrokit.com/produkt/8x-relakort-seriellt-ek006/

It looks like it only have one diod to protection for all relays. I hade diodes in parallel with all relays but decided to use only one.

But you’re not using pin10 which performs this function on the ULN2803A chip.
I was unable to see on the scheme where the diode performing this function is located.
A diode alone cannot do this function for every relay.

@JacobW

At the moment most diods on the pcb are in wrong direction. Only Chanel one on the output is corect. So most of them would create a short circuit.

For example look at the diods after the d-sub. They are corect in the schematics but not correct on the pcb.

You have the electric symbols of the diodes with the numbers on the pads reversed from the convention.

@JacobW

I want the board to handel about 150W at 5 volt.

Let’s say 160W, = 32A at 5V

I also would prefer if each output could handel 7,5A per Chanel.
2.5*3=7.5A per Chanel.

So they don’t need to handel 10A per Chanel but it is good with head room.

All chanels should be able to handle 7.5A but all don’t have to do it at once.

But what I am most worried abut is the return current. I don’t have a clear ground path for return current. A lot of tracks in the way of the current on the ground plane.

As for the current associated with + 5V, the slopes are fine if I don’t see badly.

For the + 12V and + 28V they are too small.

The real problem is the return to GND, it seems to me that there are very close sections and it is difficult to control the PCB for this important problem.
In my opinion you should review this part in order to have a larger and more compact GND plan.

I hope I haven’t written any nonsense, it’s not easy to analyze the whole project.

[EDIT] Sorry for the mess in the post, I made a mess with the translation.

[JacobW]

@Fabio_De_Bernardi

used SMA diodes and for the current values involved they are not good, at least to a first general assessment.

What diods should I use?

[JacobW]

@Fabio_De_Bernardi

A diode alone cannot do this function for every relay.

Why is it not enough with one diod?

So they how have done the relay card have constructed it wrong?

[Naib]

@JacobW

But back to the diods. How do I in a simple way fix them?

the eeschema symbol you are using is the SPICE symbol and not the DIODE symbol. SPICE is notorious for flipping what pin1 and 2 are w.r.t. ANODE-CATHODE of diode.

The correct diode is from the DEVICE library not SPICE library.

Two ways todo this

1. Stick with the SPICE symbol but use the “Alternative Pin Assignments” UI to swap A-K. Its a manual job for each one
   

   

2. use the “Symbol library reference” to change the diode from SPICE to DEVICE
   

   

@JacobW

The amount of current that the load will probably have is 2.5A. I want to be able to run 3 on each output up to approx. 150W in total, but of course I want to have excess capacity. when I started drawing I thought that there can be no problem with 80A, but it turned out completely crazy values. I can’t solve it. But anyway. It will not be in any particular enclosure, I would say it will be in the open air. It will at least have several liters of air around it.

MAke sure you get this right… Personally I am never happy with > 5C temprise. Does it drive wider tracks and thicker copper? of course but its needed.
One solution is to apply a mask to the tracks in question and post-apply solder to thicken the traces for the higher current. Also not all sections need to be thick, only the common part before each coil branches off.

[JacobW]

@Fabio_De_Bernardi

slopes

What does that mean?

[straubm]

Protection diodes should be behind a fuse or a reverse voltage will eventually kill them.
I use SS34 as diodes.

[Fabio_De_Bernardi]

Problem with the translator, I am a donkey with English.
I mean the + 5V track as a size is correct.

[Fabio_De_Bernardi]

@JacobW

@Fabio_De_Bernardi

Un diodo da solo non può svolgere questa funzione per ogni relè.

Perché non è abbastanza con un diodo?

Quindi, come hanno fatto la scheda relè, hanno sbagliato?

They used ULN2803A pin10.
[EDIT] this way you have a diode for each relay.

[straubm]

@JacobW As for the ULN and the relais: connect ULN pin 10 to 5V and all is well.

[Fabio_De_Bernardi]

@straubm

Protection diodes should be behind a fuse or a reverse voltage will eventually kill them.
I use SS34 as diodes.

He wants to use them as suppressor diodes and not as protection for polarity reversals.
With currents in the fuses so high I don’t think a fuse diode solution can work properly.
You risk burning the diode first with respect to the fuse.
It would be a nice test to do.
I have never used this functionality with currents greater than 1, 2 A .

[straubm]

Ah, yes, makes sense.

[straubm]

I’m quitting here. Everything above 24V and/or 3A is dangerous, life-threatening technology to me . Well, I had to use 170V in my Nixie clock project, but with sweaty palms…

[Fabio_De_Bernardi]

I arrived at 380V cc for a 3Kw led lamp and at the first moment ignition I get a heart attack.

Everything went well since I’m still here to write.

[Fabio_De_Bernardi]

@JacobW

@Fabio_De_Bernardi

used SMA diodes and for the current values involved they are not good, at least to a first general assessment.

What diods should I use?

If the diode function is as a suppressor (TVS) you could use this type: SMF5.0A
It is just an example of the type of diode to refer to to perform this function, you have to evaluate its characteristics based on your project.

SMF5.0A can be used on + 5V .

The choice of TVS diodes must be made on the basis of any disturbances you have on the power supply and this is obtained from the tests.