Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

USB 2.0 tracking

Rajinder1268

Member level 1
Joined
Mar 20, 2021
Messages
38
Helped
0
Reputation
0
Reaction score
0
Trophy points
6
Activity points
270
Hi all,
I have attached a copy of a board which has USB tracking of D+ and D- signals from a USB connector. I am going to do something similar. Looking at the layout design in the picture, Is there going to be an issue with tracking the D+/D- lines as shown in terms of getting the same length? Or is it best to turn the microcontroller by 90 degree clockwise and then adjust its position to get the length same?

My second question is that I am going to connect the shield of the connector to GND via a 0R resistor. Do I connect all the shield pins on the USB connector with a thick track 2.54mm or similar and then to the 0R and then the 0R connected to the main GND.

I was thinking of having a solid GND on the BOTTOM layer, the Top layer being signal tracks, tracks for VCC. Is it a good idea to have a polygon plane for the VCC or will tracks be sufficient.

Lastly, i was thinking of having additional GND on the remaining of the Top PCB - which will be stitched to the BOTTOM layer GND through VIAS.

Does my approach seem ok?

Thanks in advance


1619977421152.png
 

aminpix

Member level 3
Joined
Sep 30, 2008
Messages
54
Helped
1
Reputation
2
Reaction score
2
Trophy points
1,288
Activity points
1,767
The length of the two traces should be identical, but since your traces are short, I think it will work. Just make sure the trace impedance are 90ohm.
I recommend to connect the shield to the ground plane directly, not via any resistor.
Finally the ground plane at the top is useless for your design (since you have ground plane at the bottom layer). If you want to add ground plane at the top layer, make sure they have enough distance from traces (Check this link).
 

Rajinder1268

Member level 1
Joined
Mar 20, 2021
Messages
38
Helped
0
Reputation
0
Reaction score
0
Trophy points
6
Activity points
270
The length of the two traces should be identical, but since your traces are short, I think it will work. Just make sure the trace impedance are 90ohm.
I recommend to connect the shield to the ground plane directly, not via any resistor.
Finally the ground plane at the top is useless for your design (since you have ground plane at the bottom layer). If you want to add ground plane at the top layer, make sure they have enough distance from traces (Check this link).
Hi, thanks for your reply. I have looked at the USB/FTDi recommendation for the USB shield. They recommend a low impedance from the screen to OV, the 0R can be replaced by a capacitor or ferrite? I do have GND on the bottom, what shall I do with spaces between tracking on the Top Layer?

I am also using a QFN20 device, which has a centre pad i.e. pad 21. This needs to be connected to the 0V. What would be the best way? Could I connect the centre pad to the pins (VSS) on the device? Or is it best to place a few VIAs on the centre of the pad?

Thanks in advance.
 

KlausST

Super Moderator
Staff member
Joined
Apr 17, 2014
Messages
19,443
Helped
4,299
Reputation
8,603
Reaction score
4,260
Trophy points
1,393
Activity points
128,956
Hi,

length differnce is length difference ... and does not depend on total trace length.
I don´t know how much difference is allowed.
I guess it should be O.K. ... but to be sure: read specification.

Impedance does only matter when trace length is let´s say 1/10 of minimum wavelength or more. No problem for these couple of millimiters.

I´d say the impedance is less the problem than length difference.

Klaus

added:
having a solid GND on the BOTTOM layer
I see pads and traces in the blue layer. If this is the bottom layer then you don´t have a "solid" GND plane ... It´s cut into many pieces...

And especially below the USB_data traces there is no GND at all.

Klaus
 
Last edited:

aminpix

Member level 3
Joined
Sep 30, 2008
Messages
54
Helped
1
Reputation
2
Reaction score
2
Trophy points
1,288
Activity points
1,767
Basically You should connect the shield pins to the chassis and ground plug to avoid ground loop; We all know it is not always practical so we connect them to ground.
In most of my designs I directly connected the USB shield to the ground plane; however I have seen some design connecting the shield to the ground through a 0ohm resistor or a ferrite. Just keep the resistor as close as possible to the shield.

As Klaus mentioned you don't have a solid ground under your USB traces. I highly recommend to keep a solid ground under the traces without any cut.

USB/FDTI won't get hot therefore you don't have temperature concern. You can either connect the center pad to the ground through via or connect to the ground pins.

Sometimes we fill the space between the traces and components with ground because of the EMI. but your board has only two layers and it wont help.
Making a ground plane in one side (bottom layer) is enough for two layer board.
 

Rajinder1268

Member level 1
Joined
Mar 20, 2021
Messages
38
Helped
0
Reputation
0
Reaction score
0
Trophy points
6
Activity points
270
Basically You should connect the shield pins to the chassis and ground plug to avoid ground loop; We all know it is not always practical so we connect them to ground.
In most of my designs I directly connected the USB shield to the ground plane; however I have seen some design connecting the shield to the ground through a 0ohm resistor or a ferrite. Just keep the resistor as close as possible to the shield.

As Klaus mentioned you don't have a solid ground under your USB traces. I highly recommend to keep a solid ground under the traces without any cut.

USB/FDTI won't get hot therefore you don't have temperature concern. You can either connect the center pad to the ground through via or connect to the ground pins.

Sometimes we fill the space between the traces and components with ground because of the EMI. but your board has only two layers and it wont help.
Making a ground plane in one side (bottom layer) is enough for two layer board.
Thank you. I will sort out the GND on the bottom layer.
--- Updated ---

Hi,

length differnce is length difference ... and does not depend on total trace length.
I don´t know how much difference is allowed.
I guess it should be O.K. ... but to be sure: read specification.

Impedance does only matter when trace length is let´s say 1/10 of minimum wavelength or more. No problem for these couple of millimiters.

I´d say the impedance is less the problem than length difference.

Klaus

added:

I see pads and traces in the blue layer. If this is the bottom layer then you don´t have a "solid" GND plane ... It´s cut into many pieces...

And especially below the USB_data traces there is no GND at all.

Klaus
Thanks, Is will sort out a GND underneath the tracks, I will have a solid GND on the bottom layer.
--- Updated ---

Thank you. I will sort out the GND on the bottom layer.
--- Updated ---


Thanks, Is will sort out a GND underneath the tracks, I will have a solid GND on the bottom layer.
My other question is regarding the USB shield. I am going to connect the 4 pins together with a thick track then to the 0R which will then connect to the bottom layer GND. There will be no GND underneath the USB connector but a thick track to connect the shield. I will keep the 0R close to the USB shield. The GND will start underneath the 5 smd pads of connector, does that seem ok. My only concern is if the 0R fails, then there will be no connection to 0V. Would it be advisable to have another 0V in parallel? Thanks in advance.
 
Last edited:

Aussie Susan

Advanced Member level 4
Joined
Jan 5, 2015
Messages
1,247
Helped
377
Reputation
754
Reaction score
375
Trophy points
83
Activity points
13,504
What is the value you hope to gain by using a 0R resistor? It is basically just a piece of wire so why not just connect to the ground plane directly. Adding 2 solder joints would increase the chance of a problem due to a faulty join (in my opinion).
I really doubt if the 0R resistor will fail unless it receives a mechanical shock and the solder breaks. What sort of thing are you worried about that would cause the 0R resistor to fail?
Susan
 

Rajinder1268

Member level 1
Joined
Mar 20, 2021
Messages
38
Helped
0
Reputation
0
Reaction score
0
Trophy points
6
Activity points
270
What is the value you hope to gain by using a 0R resistor? It is basically just a piece of wire so why not just connect to the ground plane directly. Adding 2 solder joints would increase the chance of a problem due to a faulty join (in my opinion).
I really doubt if the 0R resistor will fail unless it receives a mechanical shock and the solder breaks. What sort of thing are you worried about that would cause the 0R resistor to fail?
Susan
Hi, the 0R resistor is to provide a low impedance path. It can also be replaced by a ferrite or capacitor if we run into EMI issues. I have seen the USB guidelines, that say connect the shield to GND via a capacitor. In terms of it failing, in production, bad soldering etc.
 

LaTeX Commands Quick-Menu:

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Top