There’s a matrix that shows how each pin type is handled in ERC:
1 Like
Open-Collector is synonymous with Open-Drain as is Open-Emitter is synonymous with Open-Source.
There is zero point putting an option down for all 4 when 2 will do and thus it would be about what is the best label if there is … confusion
Open Collector/Drain
Open Emitter/Source
Would be the best compromise otherwise it would be coming up with a kicad specific ontological notation which would be running tangent to the electronics industry. Meh, I might make a MR to just sed this…
–edit–
3 Likes
Sorry but I feel the need to quibble here:
They may well be close enough for KiCad pin types and it may well be OK to group them together. But as an engineer I feel the need to be precise. Particularly for MOSFET source and bipolar emitter, they are certainly not synonymous. For one thing: A MOSFET Vgs may be 5x or 10x a bipolar Vbe. So if an amplifier IC is powered by +5V, you might be able to get significant output current at 4V from an open NPN emitter. But you will not get much current at 4V from an Nch open source.
While we are on the topic, I wonder whether we ought to call a BJT collector an open collector pin type? Maybe we already do… 
1 Like
NYD and you’re already bored? As mentioned those pin types are more for logic ICs, and perhaps driver ICs, e.g. ULN2803, not for individual transistors since outside of a circuit it’s not yet known which configuration a transistor will use. ERC and pin types are not concerned about Vgs or Vbe.
While you are correct in the general sense, you need to appreciate how Kicad uses this information and it’s the ECR table. As such they are synonymous.
Whether the output stage is a BJT or a FET is extremely important, especially with regards to the LOW voltage while sinking current and compatibility with upstream circuit threaholds, Kicad cannot validate thos.
Bipolar transistors are not obsolete, still bipolar transistors have advantages that MOSFET transistors do not have.
On the main topic,
I agree that it is better to revise the naming of the “Electrical Type” options, for example, instead of “Open Collector”, it is better to write something like “Open Collector/Drain” or a more concise text.
Well yes, but the open collector is just that, and the open emitter is just that.
I am completely illogical or at least do not do a lot of logic design, and I take terms such as “open collector” to mean what they say.
Nothing about the above says that I am bored. And I am certainly not holey, which is what I would be after being bored.
I don’t agree with you.
Open Collector doesn’t mean exactly as it says and also even not tries to be 100% accurate term.
Everyone, ‘since always’ understand what that means and it not means ‘open collector’ because if it would be understand that way than both npn open collector would be understand under this name and pnp open collector also, what no one assumes hearing Open Collector.
I think that if someone ever will use npn-pnp darlington at the IC output with pnp emitter being pin connected to IC output using Open Collector term for this output would also be acceptable like is acceptable for NMOS Open Drain.
The idea of short text descriptions of something is simply opposite to expect that term is totally unambiguous.
If you want to modify it in such direction it should say ‘NPN Open Collector or NMOS Open Drain’ and it is still not enough to cover all possibilities.
If you not accept Open Collector for Open Drain than how you tolerate Passive for Diode and Transistor pins?
1 Like
It seems to me that we may have different understandings…
Even classic “open collector” drivers like the ULN2003 have an output catch diode, so not truly what the term says. Experienced engineers use the term loosely to mean any two state output that is not push-pull.
If you search for “open-collector” devices on many manufacturers’ websites, they will also return parts with open drains. And vice-versa.
Changing the description from “Open Collector” to “Open Collector/Drain” would be cosmetic only. It won’t change the behavior of the ERC because they’re the same thing.
If you want to be more precise, you need to use something like Spice, not ERC, for your checking.
All this talk of whether the collector is truly open if there’s a diode there, what manufacturers classify the pin as, whether single transistor symbols in the library should have open collector and open emitter pins becase they are not connected to anything yet, is missing the point.
The sole purpose of KiCad’s pin conflict detection is to draw attention to problematic and dubious connections in a net. Refer to the pin conflict map and you will see the situations where Open Collector type plays a role.
OC and Output: Error because an OC pin could try to pull down a push-pull pin that’s pulled high.
OC and Power Output: Error because this will destroy the OC pin if this shorts the power rail to ground. (Incidentally this means that if you were to loop say the output of a 7406 OC inverter to GND, this entry would be triggered, even though it’s harmless.)
OC and Tri-state: Warning because Tri-state behaves like push-pull some of the time.
OC and Unspecified: Warning because Unspecified hates all other types of pins except Free.
Open Collector and Open Drain should have the same conflict entries so are functionally equivalent. Maybe with so much fuss about this it should be abbreviated to OCD. 
2 Likes
Good idea. You file the bug report. I’ll watch for the comments; from a great distance, while wearing my tin foil hat.
Can we have SnAl foil hat instead of Sn foil hat to be more inclusive? 
The Mop Up Crew here is still waiting for ANY hat.
1 Like