1
\$\begingroup\$

I'm looking at the 'layout considerations' for TPS61030. On p.18, it says

Use a common ground node for power ground and a different one for control ground to minimise the effects of ground noise

Aren't the power ground and control ground connected to the negative of the battery anyway? It seems to imply the 2 grounds should be somehow separated.

Thanks

\$\endgroup\$

2 Answers 2

2
\$\begingroup\$

Yes, the grounds do come back to the battery if you follow the traces or conductivity.

But that is not what they are really saying.

What they are are saying is separate the loop currents for power and control. You want to ensure that they are separated.

Notice how the PGND is kept away from everything else ? The high frequency nature of SMPS means that there can be some noisy current paths and by controlling their path, and what they are allowed to interfere with, helps control noise, and EMI.

enter image description here

\$\endgroup\$
5
  • \$\begingroup\$ I see. But there's something about this diagram I don't quite understand. Firstly, the GND pin isn't even connected to the ground, right? There's only the 'Feedback Resistor 2' connected to it. According to the 'Simplified Schematic', GND should be connected to both power ground and earth, right? Secondly, 'Output Capacitor 1' should be connected to the power ground according to that schematic, but it's connected to GND at the bottom right. Am I missing something? \$\endgroup\$
    – Kar
    Commented Aug 30, 2015 at 13:54
  • \$\begingroup\$ @Kar The layout, is an example (as indicated in the title of the section). It only shows the top side. But the vias give you an indication of connections. Notice how PGND vias and GND vias are connected to the ground the plane. \$\endgroup\$
    – efox29
    Commented Aug 30, 2015 at 13:58
  • \$\begingroup\$ IMO the PCB is double layer, if you loook evaluation board TPS61030 EVM it makes sense. \$\endgroup\$ Commented Aug 30, 2015 at 14:05
  • \$\begingroup\$ But it mentions VIA to Ground Plane and VIA to VBAT Plane - wouldn't it need at least 3 layers then? \$\endgroup\$
    – Kar
    Commented Aug 30, 2015 at 14:28
  • \$\begingroup\$ @Kar its an example layout. It's not the "you have to design it this way" layout. Everything comes at a compromise. You can do 2 layers, and it could work just fine for your application. If your requirements are tight, you might have to do a 4 layer board or more. It all depends. Marko stated that the TPS61030 EVM is on a 2 layer board. Look at the schematic and PCB layout for the EVM if its available. Also, you can't really manufacture odd number of layers, fab houses typically only do even. \$\endgroup\$
    – efox29
    Commented Aug 30, 2015 at 14:41
2
\$\begingroup\$

Yes the power ground and control ground are connected to eachother, but allmost all switched mode converters like the TPS61030 need these grounds separated in the sense that the large currents flowing through the power ground must not disturb the control ground. Imagine what happens when a large current flows through power ground, the (small) series resistance of the PCB tracks cause some voltage drop. You do not want this voltage drop to be "seen" by the circuits connected to the control ground. So what you want to do is to give the control ground it's own separate track to a common ground piont. You should connect the ground from your power source also to this point. If you have an old computer power supply lying around, have a look inside ! You will notice that on the PCB there will be a large ground but smaller ground tracks branching off from that ground. It is called a "star ground".

\$\endgroup\$
2
  • \$\begingroup\$ For what's worth it, this idea applies more generally to segregating control signals and power signals. Even the datasheet of something like LM317A (ti.com/lit/ds/symlink/lm117.pdf) recommends separating the traces of the current set resistor and actual load/output, even though they are connected to same pin of the IC. The datasheet of LT1086 (cds.linear.com/docs/en/datasheet/1086ffs.pdf) "Figure 2. Connections for Best Load Regulation" has a nice[r] diagram with explanations. \$\endgroup\$ Commented Aug 30, 2015 at 20:45
  • \$\begingroup\$ Be it Linear regulators or switching regulators: they all need proper grounding and clean reference and feedback signals. The other answer talks about EMI and noise but in my opinion the main reason why you need proper grounding is to keep a clean feedback path ! \$\endgroup\$ Commented Aug 30, 2015 at 20:52

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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