search

openlcb / documents

The OpenLCB specification: standards, recommended practices and other documentation.
3 stars 7 forks source link

In-rush current #2

Open balazsracz opened 7 years ago

balazsracz commented 7 years ago

We should have some statement about the in-rush current, or maximum capacitance applied to the power line at the consumer board level.

Power supplies have to be able to drive the given capacitance and still create some specified voltage ramp-up rate without assuming that the line is shorted.

Saying something like "never exceed the marked current limit, not even at startup" is too simplistic. On the other hand, having a capacitance allowance that is proportional to the marked current limit seems like a potential workaround:

dpharris commented 7 years ago

Good thoughts about this. I think our limitation is the current carrying capacity of the CAT5, so double current inrush should not be a problem. David

On Thu, Jan 5, 2017 at 10:46 AM, Balazs Racz notifications@github.com wrote:

We should have some statement about the in-rush current, or maximum capacitance applied to the power line at the consumer board level.

Power supplies have to be able to drive the given capacitance and still create some specified voltage ramp-up rate without assuming that the line is shorted.

Saying something like "never exceed the marked current limit, not even at startup" is too simplistic. On the other hand, having a capacitance allowance that is proportional to the marked current limit seems like a potential workaround:

  • if a board is marked as X mA of current consumption, it is allowed to have 2*X uF of bulk capacitance on the power line. Example: a board with 30mA of marking may have a 60 uF capacitor.
  • The power source must be able to start up with a bulk capacitor of 2* the mA rating on the power line and still create a voltage ramp-up of e.g. 1 V / msec. Example: a power supply of 500mA rating must be able to start up with 1000 uF of bulk capacitance on the power line.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/openlcb/documents/issues/2, or mute the thread https://github.com/notifications/unsubscribe-auth/AAg4SgDpYfIMbv9f1bIFQ-Hy3Bztp8U3ks5rPTqagaJpZM4LcABF .

balazsracz commented 7 years ago

​note that I purposefully did not say what will be the current upon startup. I only said how much capacitance there can be -- the actual current will depend on how strong the power supply really is during startup. However, since the capacitance limit is quite low, high current can only flow for a few milliseconds before that capacitance gets charged. That should not cause any heat problems.

bobjacobsen commented 1 year ago

What's the underlying problem this is meant to solve?

From inspection, it looks like a Tower-LCC (marked for 20mA in) has a 470uF capacitor. Not all of that might be visible on the input, though, because it seems to be after the power converter.

RR-CirKits commented 1 year ago

Hi, Here is the power input circuit for the Tower LCC. image The 470uf is decoupled from the input line with a 100 ohm resistor. Dick :)

balazsracz commented 1 year ago

RCN-530 talks about this problem in DCC.

balazsracz commented 1 year ago

Conceptually we want to say something like "after X msec at most 2x the rated input current may be drawn".

We might want to test the power supply of the power point (1.2A 15V) on how it behaves at startup and how much startup overload it is willing to take.

Nodes have to operate correctly in the presence of a slowly ramping bus voltage.

balazsracz commented 1 year ago

Next step: