2
\$\begingroup\$

I have a battery pack with lithium ion cells. The battery pack has a maximal voltage of 84V (it goes from 84V to ~50V -> depending about the current state of charge).

I use this battery pack to power all sort of things (lights, horns, MCU(BMS), inverter - that then gives power to the eletrical motor...). The inverter draws about 60A of current continuously from a battery pack when the motor is rotating.

For my concept to work I must have different DC voltages on a single PCB:

  • 12 VDC (for lights and horns)
  • 5 VDC (for some sensors...)
  • 3.3 VDC (for MCU, operational amplifiers...)

The idea is to use some sort of step-down converters on the PCB to convert from 84V to 12VDC and to convert from 84V to 5VDC. And then to use a LDO to give me 3.3V from 5V.

The 12VDC must have a power of ~40W, and the 5VDC must have a power of ~10W.

The question is regarding if the step-down converter should have galvanic isolation (like a flyback) so that I can have different grounds for high voltage and low voltage? Or is OK to use a "normal" buck regulator?

I think that for 12VDC that supplies the horns and lights is not so necessary to have different grounds as for 5VDC that supplies most of the electronic ?

Any suggestion?

\$\endgroup\$
  • 1
    \$\begingroup\$ Will the MCU control the motors via the inverter and, if so, won't the connections required make the galvanic isolation null and void? \$\endgroup\$ – Andy aka Feb 1 '16 at 13:19
  • \$\begingroup\$ @Andyaka unless you also isolate those. AtOP: There's things to be said for either scenario full separation or just very good filtering. Long and short of it is, if you are able to design good star-ground schemes and proper filtering/buffering you don't need isolation perse. But if you aren't you are best off in your first design to isolate all signals in and out of the digitals and swallow the (huge) extra cost of that. \$\endgroup\$ – Asmyldof Feb 1 '16 at 13:31
  • \$\begingroup\$ @Asmyldof yeah, I'm aware of that but I want the op to consider this. \$\endgroup\$ – Andy aka Feb 1 '16 at 13:33
  • \$\begingroup\$ @Andyaka I know, but in true European Fashion I want everything to always be fair and balanced. Unless we're talking about Europe or Healthcare. \$\endgroup\$ – Asmyldof Feb 1 '16 at 13:38
  • 1
    \$\begingroup\$ @Asmyldof Ahhh you are dutch - that explains it LOL ;) \$\endgroup\$ – Andy aka Feb 1 '16 at 13:39
1
\$\begingroup\$

The inverter draws about 60A of current continuously from a battery pack when the motor is rotating

Not a small amount of current and, if you don't get your grounding sorted out correctly you might find tens of amps trying to flow through the wrong ground return wires and hurting your MCU.

Therefore there is a case for providing galvanic isolation to your sensors (5V) and MCU (3V3) supplies because, although your MCU may need to connect to ground at the motor controller, having isolation in a potential ground return path could save you lots of heartache and dollars.

\$\endgroup\$
  • \$\begingroup\$ or he can just split the power cords at the battery contacts, making the case impossible.. \$\endgroup\$ – student Feb 1 '16 at 13:46
  • \$\begingroup\$ @student - there will still be a potential return path for current from the inverter control wires through the MCU and back to the battery - to reliably control the inverter, digital connections between MCU and inverter need to be good and this means a solid ground connection from MCU to inverter. \$\endgroup\$ – Andy aka Feb 1 '16 at 13:48
  • \$\begingroup\$ True, but this can be remedied with short and thick cabling. Anyway, I get your point that this requires careful design. \$\endgroup\$ – student Feb 1 '16 at 13:54
1
\$\begingroup\$

Nothing in your description requires the derived voltages to be isolated from the battery ground. Not isolating them will be easier, and it will be easier to get higher efficiency.

Tie all the grounds together.

\$\endgroup\$
  • 1
    \$\begingroup\$ In what circumstances there is a need for a galvanic isolation between the battery pack and the electronics (BMS, VCU(Vehicle control unit)...) ? \$\endgroup\$ – Nibelung26 Feb 1 '16 at 13:39
1
\$\begingroup\$

It depends, if you are careful you can often go without galvanic isolation but using it can make the system design much easier.

In a high power system with both voltage and current significantly higher than in a normal logic system you can develop voltage drop on the wiring both DC and transiant. The transiant volt drops can be especially nasty since while you can easilly reduce the resistance of wires by making them fatter it's very hard to reduce the inductance.

That is ok if the high and low voltage parts of the system don't interconnect at all or if they meet in exactly one place but it becomes problematic if their grounds are interconnected in more than one place. The voltage drops on the power grounds can then cause voltage differences in the signal grounds. If those voltage differences get too big it can cause misbehaviour and/or damage.

I assume your MCU will be connected to the motor controller. If it is and if the motor controller doesn't have galvanica isolation between it's control and power sides then you have already "spent" your single connection between signal and power grounds. So you should use an isolated converter in your circuit.

On the other hand if the motor controllers control input is galvanically isolated then you can get away without isolation in your circuit.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.