I must confess, never did occur to me to make use of the NC balls. I was thinking of simply pulling some balls from the footprint and simply route underneath them. This seems better, though. As long as you're not unlucky and some NC pins are actually used (as test points, for example) by the manufacturer.
@@sentineo Thanks for sharing, man. I did check out my favorite mmc's datasheet. Turns out you are correct (ofc you'd be, since you already tried it) - there are NC pins, RFU pins and VSF (vendor special function) pins. All of which are unconnected electrically. I do have one question, if you'd be so kind (since you clearly have more experience than I do). How utterly devastating would you say it would be if I were to place the caps for VDD_CORE (the middle balls) on the same side as the MMC. I'd really like to keep the parts on one side only of my PCB.
Sometimes these NC (No Connect) pins are provided for IC manufacturers to use in factory testing. They may have pull-up and pull-down resistors inside. To use the NC pins, it may be necessary to confirm with the IC manufacturer whether they can be used, otherwise, it may affect signal stability. Additionally, the trace width may vary due to the shape of the PAD, causing the trace width is not smooth.
@@mehmeteyuparslan Always happy to help, trying to teach what I learned by doing things so other people can progress faster. Feel free to reach out if you have any remaining questions!
This dependes on the specific IC. Quite often actually NC pins are not truly unconnected inside the IC, but are used by the IC-manufacturer during production (such as for programming, trimming, or testing), and the pin is only "NC" from the board designer's perspective. In these cases, you should really not be connecting your signals to the pads or pins. Thus this trick in the video needs to be validated on a IC-by-IC basis.
Hi, thanks for your comment! There is indeed some debate whether to use the RFU for this or not. Since it is reserved for future use, typically on this version it shouldn't be harmful. But as mentioned below: this should of course be checked by actually assembling the PCB with the specific chipset. Next to that, if you'll be doing extended boundary scans etc, this method might actually harm these checks. So there is a lot of things to considere here. I have used this in a design that has been thoroughly tested and is now in production, without any issue. We verified with the actual PCBs in a first small batch before scaling it up.
I must confess, never did occur to me to make use of the NC balls. I was thinking of simply pulling some balls from the footprint and simply route underneath them. This seems better, though. As long as you're not unlucky and some NC pins are actually used (as test points, for example) by the manufacturer.
Thanks for your comment, appreciate it.
@@sentineo Thanks for sharing, man. I did check out my favorite mmc's datasheet. Turns out you are correct (ofc you'd be, since you already tried it) - there are NC pins, RFU pins and VSF (vendor special function) pins. All of which are unconnected electrically.
I do have one question, if you'd be so kind (since you clearly have more experience than I do). How utterly devastating would you say it would be if I were to place the caps for VDD_CORE (the middle balls) on the same side as the MMC. I'd really like to keep the parts on one side only of my PCB.
Sometimes these NC (No Connect) pins are provided for IC manufacturers to use in factory testing. They may have pull-up and pull-down resistors inside.
To use the NC pins, it may be necessary to confirm with the IC manufacturer whether they can be used, otherwise, it may affect signal stability.
Additionally, the trace width may vary due to the shape of the PAD, causing the trace width is not smooth.
Thanks!
You're welcome, glad you like it!
Hello, Does this method really work, have you tried it on pcb? Can NC pins be used for this purpose or is it only valid for this TOSHIBA e-MMC.
Yes, I am doing exactly this in a product of mine where we have hundreds in the field working fine for several years now.
@@sentineo I'm glad to hear that. Thank you for the video and reply.
@@mehmeteyuparslan Always happy to help, trying to teach what I learned by doing things so other people can progress faster. Feel free to reach out if you have any remaining questions!
This dependes on the specific IC. Quite often actually NC pins are not truly unconnected inside the IC, but are used by the IC-manufacturer during production (such as for programming, trimming, or testing), and the pin is only "NC" from the board designer's perspective. In these cases, you should really not be connecting your signals to the pads or pins. Thus this trick in the video needs to be validated on a IC-by-IC basis.
@@kapa7197 Datasheet manufacturers often specifiy do not connect rather than just no connect, so it's certainly worth checking the datasheet.
there is one mistake，A7 is RFU
Hi, thanks for your comment! There is indeed some debate whether to use the RFU for this or not. Since it is reserved for future use, typically on this version it shouldn't be harmful. But as mentioned below: this should of course be checked by actually assembling the PCB with the specific chipset. Next to that, if you'll be doing extended boundary scans etc, this method might actually harm these checks. So there is a lot of things to considere here.
I have used this in a design that has been thoroughly tested and is now in production, without any issue. We verified with the actual PCBs in a first small batch before scaling it up.