Chassis and Signal Ground Connections

[Steve Haynal]

Hi Group,

What are people's thoughts and experiences on where and how chassis ground should connect to signal ground? A rule of thumb is to connect them at one point near the IO. On the HL2, which can slide into an enclosure, there are the following potential options:

1. Connect chassis to signal ground at the four mounting holes. Currently the mounting holes are not connected to ground on the PCB.

2. Connect chassis to signal ground along the side edges of the PCB with one long thin pad that makes physical and electrical contact with the enclosure when it slides in. This is the current option, but only in the board edge around the PA as it serves for thermal dissipation too. There is one nut/bolt to provide some down pressure of the PCB in the enclosure slot at this area.

Combining both options is also possible. Option 2 can be just around the PA, or extended to the full length of one or two sides.

73,

Steve

KF7O

[in3otd]

Hello,
I think that the internal board GND connection to the chassis GND is less important *if* the proper GND connections are made at the connector entry points of the chassis. Easier said than done, of course.
Every cable having a "ground" connection which enters the metal box should have its "ground" firmly connected to the chassis, first. The idea is to avoid having any external ground currents flowing thru the board from one connector to the other. For example, assume that the ground of your antenna is not at the same potential of the ground of your power supply: the current flowing between these two grounds should go thru the metal chassis and not having to traverse the PCB, potentially causing (tiny) voltage drops across the PCB ground path.
Less easy is the case when there is no cable shield but there is anyway a "ground" connection which needs to go to the PCB. Following the example above, better will be to have a connection between the PCB GND and the chassis as near as possible to the point where this "ground" is connected to the PCB, so any ground current that does need to go to the chassis travels the least possible distance on the PCB ground.

So, in short, I vote for having the PCB ground connected to chassis ground as often as possible.

73 de Claudio, IN3OTD / DK1CG

[Graeme Jury]

I completely agree with Claudio. The critical thing is to get all shields grounded to the metal case at point of entry or exit so that ground currents flow through the case from connector to connector. Where this was not followed in industry it was known as the "pin 1 problem". I usually bring the power into the pcb as close as I can to the highest current device (the PA in this case) and ground from there as well which would be the PA fet cases. I have never had trouble with multiple ground points to the case but am not sure is this is best practice or not but if I did use multiple grounds (which I usually do and seem to get away with it) I would definitely earth the corner mounting points for major ground connection points.

73, Graeme zl2apv

[James Ahlstrom]

Hello Steve,

I agree with Graeme and Claudio here.  I always use plated-through grounded pads for mounting holes, and then the mounting holes will ground the PCB ground planes to the case.  I don't trust leaving a bare edge on the PCB and depending on the card guides for a ground because some enclosures are anodized and therefore non-conducting, and the connection is a slip fit and not screwed or soldered down.  The power input and RF output are high current points, so I always run a substantial deliberate ground from them to a suitable ground plane on the board.

Audio circuits often use the single-point-ground you describe, but that is because they are trying to prevent the high power output from reaching a sensitive input.  Our problem is more a case of establishing a low impedance path to ground.  Grounding as often as possible will likely result in success except perhaps in a very high gain amp.

Jim

N2ADR

[Steve Haynal]

Hi All,

Thanks for the feedback. I have grounded the four mounting holes and plan to also add pad strips on the board edges. I am considering only connecting these to the bottom and bottom inner layers to discourage any current from the more sensitive top and top inner ground layers. If there is an enclosure, I'd like to encourage any current to use it rather than the copper on the PCB.

73,

Steve

KF7O