Design question - multiple power sources, ground loops?

[metermannd]

Hopefully this question may have some useful takeaways for others' Arduino projects.

I had originally sought to redo two boards from a piece of 1980s test equipment to make use of Arduino technology, but for the moment, I am moving forward with a non-Arduino implementation, partly to better understand this circuit's operation before I begin developing the Arduino code, do signal analysis, etc.

Anyway, in the original equipment, a pair of ribbon cables bring 5V, ground, and data from the main processor board to the external processor board. The 5V and ground lines are present in each cable, as they plug into sockets installed in place of the NVRAM ICs. Another ribbon cable continues the 5V, ground, and data from the external processor board to the display - I/O board. To power the LEDs and displays, 12V and ground is brought in directly from the main power board.

To rephrase, I have no less than three grounding leads, two 5V leads, and one 12V lead running from the main unit to the two add-on boards (processor and display - I/O).

Also, the main unit generates a 12.5kHz, 200V peak-to-peak signal, so I may have to keep interference issues in the back of my mind.

Should I worry about ground loops or current-carrying issues in the double 5V power connection?

On the new board I'm designing (which consolidates the external processor and display - I/O boards into one), 5V, ground, and data are still carried in through the two ribbon cables. Since I still need 12 volts (to power the displays / LEDs), I have added a connector for the 12V, and while I was at it, I added a pin for 5V as well.

Now that I've hopefully described the setup (if needed, I can make a diagram and post), I want to see what would be the best practice in regards to the incoming DC power to the board.

If I choose to bring in both 5V AND 12V through the power connector, do I need to go back and break the 5V path from the ribbon cables?

Also, should I be looking at breaking the ground path from one or both of the ribbon cables?

[raschemmel]

Are the low voltage circuits SEPARATE circuits or does the 12 voltage circuit interface with the 5V circuits in some way ? (YES/NO)
Are the 5V circuits SEPARATE circuits or do they interface with each other in some way ? (YES/NO)

The 5V and ground lines are present in each cable, as they plug into sockets installed in place of the NVRAM ICs.

Another ribbon cable continues the 5V, ground, and data from the external processor board to the display - I/O board.

To power the LEDs and displays, 12V and ground is brought in directly from the main power board.

To rephrase, I have no less than three grounding leads, two 5V leads, and one 12V lead running from the main unit to the two add-on boards (processor and display - I/O).

Do any of the above mentioned circuits share a ground ? (YES/NO)

Also, the main unit generates a 12.5kHz, 200V peak-to-peak signal, so I may have to keep interference issues in the back of my mind.

How far is this signal from the low voltage circuits ?
Is there any shielding for the HV circuit (enclosure etc)
Can you post a block diagram of your system ?
In general low voltage circuits routed together affect each other as function mutual inductance which directly related current. If none of thee low
voltage circuits conduct high current they can be
considered as independent circuits. If they interface
with each other they should have a common ground. The only concern I see is the level of isolation (shielding) provided for the HV circuit.

[SteveMann]

Do you have a DVM?

[raschemmel]

Or a DMM or an Analog meter or a Dousing Stick ?

[Paul_B]

It is more correctly called a "dowsing" stick.

[raschemmel]

I'm an American. I can't spell 'tire'. (tyre).

[metermannd]

Allllrighty then!

raschemmel - your answers questioned:

The only 'interface' between the 5V and 12V circuits are on opposite sides of the LED / 7-segment buffers (source side: TBD62783, sink: ULN2803A).

In the original unit, there is a single 5V supply, its output going from the P/S board to the logic, then from the logic to the I/O processor via two ribbon cables, the 5V conductor being carried in one conductor (out of 18) apiece. Moot. A subsequent design change added a 'backplane' arrangement for power between all the boards.

Shared ground? Yes. There is a ground pin on in each of the inter-board connectors carrying data and ground. There are also two pins dedicated to ground in the backplane power connection, and one of the mounting holes on each board has a connection to ground.
Hence my ground loop question.

Distance between signal and low-voltage circuits? In the original design, they were on adjacent panels, connected by cabling. In my revised design, they are separated by roughly 1.5 cm / 0.5" and connected by a pair of mating headers.

The transceiver board in my revised design has a ground pour facing the logic in an attempt to act as a shield.

The HV side of the transmitter circuit has been significantly redesigned. AC mains is now limited to SMPS module power and the secondary of the transmitter output transformer. The entire transmitter stage has also been swapped for one requiring only 24VDC.

Block diagram is post #21, thread Project: Converting two boards into Arduino - Project Guidance - Arduino Forum (will revisit that thread soon)

SteveMann: I do have a DVM / DMM. Why?

raschemmel, Paul__b: Re: dowsing stick, if there is a punchline, I must have missed it. :-\

[Paul_B]

metermannd:
raschemmel, Paul__b: Re: dowsing stick, if there is a punchline, I must have missed it. :-\

Did you check my link?

[Grumpy_Mike]

If you adopt star wring then you will not get any ground loops. That means all the common connections are wired to meet at one point.

Do not chain the ground connections as current draw in one part of a chain can mean the ground has a higher potential due to the large current flowing through the wire’s small resistance.

This effect is called ground lift and can cause intermittent operation or inaccurate measurements.

[metermannd]

GrumpyMike: Okay. I'll drop the ground pins from the inter-board headers - the backplane and support post connections should be sufficient (I can always use a full set of nylon posts for the standoffs to omit that ground connection for the upper boards).