Masterfully presented! I bet it will save many beginner's asses. I mean assets! 😅 What could be particularly interesting is a more elaborate follow-up on extracting drive strengths not only by eye-balling them from data sheets but also by reading them out from spice and IBIS models. I can see no-one that could explain it sooooo well as you do! And a related question: How on earth to size source termination for a microSD card slot when no card in the world specifies relevant data and even if one card did, the next one could be different. The SD standard does not enforce it. 😢 Anyway, you earned yourself a subscriber and a one that will praise your channel with passion. You deserve it.
Thank you for the kind words! I do have a video on SPICE basics, but I agree having a dedicated video on getting useful info out of models is a good idea - I appreciate the feedback and will jot that down as a possible future video. As far as termination on microSD cards - I don't think you really need them in most cases (ie on PCBs with microSD card slots). I've seen them used in longer runs - for example when the card slot is connected via long wires to a main board. In those cases I think they are primarily there for improved signal integrity and possibly reduced EMI.
@@electronicswithemrys, you're right and it does help putting the card slot close to the MCU, which, unfortunately, I can't really do. After some digging, I only found max values for clock rise and fall times of 10 ns in the uSD specifications (for the default 25 MHz clock, while for high-speped mode it is 3 ns max). If I had an oscilloscope, I'd just try a few typical values for termination, but without one, the only way is the hope-it-works-way...
@@dookshi A decent o-scope is definitely a "must have" item. I spent years without one, so I understand the struggle, but I did make it a priority for my home lab and that $400 has gone to good use for many years now. For your system, I would recommend to start with a series termination at the transmitters of 22 ohms -- with fast signals a smaller resistance is better and placement is key. If the resistor is near the receiving device, it doesn't do much to help. Also having a controlled impedance line will make a big difference in signal integrity if you have to place the SD card far from the controller. Most likely the SD card itself has a weak driver (which saves die cost to the manufacturer), so you probably don't need a resistor going from the SD card back to the controller - but it doesn't hurt to put 0 ohm resistors there and swap to 10 or 22 ohm if you have issues. Good luck!
@@electronicswithemrys, my gosh, do we think alike... I made 0603 footprints with electrically shorted pads (as net ties) specifically for that purpose. Nothing to test the signal with yet, though. I mean, a couple of beers should go a long way in procuring a scope and a thirsty helping hand with it, at least for a couple of hours. You mentioned a very nice point about a uSD card probably not going to drive as strongly as an MCU. I anticipated termination on the MCU side for sure. I can also adjust the actual drive strength for each MCU pin in firmware which is certainly helpful. And a known impedance of 50 ohms (because why not?) goes without saying.
put the gates down, then you will know the power. Im pretty sure, but I havent done for total real yet, but if u do, its something that a programmer needs to get to eventually once hes tired of his computer performance.
Masterfully presented! I bet it will save many beginner's asses. I mean assets! 😅 What could be particularly interesting is a more elaborate follow-up on extracting drive strengths not only by eye-balling them from data sheets but also by reading them out from spice and IBIS models. I can see no-one that could explain it sooooo well as you do! And a related question: How on earth to size source termination for a microSD card slot when no card in the world specifies relevant data and even if one card did, the next one could be different. The SD standard does not enforce it. 😢 Anyway, you earned yourself a subscriber and a one that will praise your channel with passion. You deserve it.
Thank you for the kind words! I do have a video on SPICE basics, but I agree having a dedicated video on getting useful info out of models is a good idea - I appreciate the feedback and will jot that down as a possible future video. As far as termination on microSD cards - I don't think you really need them in most cases (ie on PCBs with microSD card slots). I've seen them used in longer runs - for example when the card slot is connected via long wires to a main board. In those cases I think they are primarily there for improved signal integrity and possibly reduced EMI.
@@electronicswithemrys, you're right and it does help putting the card slot close to the MCU, which, unfortunately, I can't really do. After some digging, I only found max values for clock rise and fall times of 10 ns in the uSD specifications (for the default 25 MHz clock, while for high-speped mode it is 3 ns max). If I had an oscilloscope, I'd just try a few typical values for termination, but without one, the only way is the hope-it-works-way...
@@dookshi A decent o-scope is definitely a "must have" item. I spent years without one, so I understand the struggle, but I did make it a priority for my home lab and that $400 has gone to good use for many years now. For your system, I would recommend to start with a series termination at the transmitters of 22 ohms -- with fast signals a smaller resistance is better and placement is key. If the resistor is near the receiving device, it doesn't do much to help. Also having a controlled impedance line will make a big difference in signal integrity if you have to place the SD card far from the controller. Most likely the SD card itself has a weak driver (which saves die cost to the manufacturer), so you probably don't need a resistor going from the SD card back to the controller - but it doesn't hurt to put 0 ohm resistors there and swap to 10 or 22 ohm if you have issues. Good luck!
@@electronicswithemrys, my gosh, do we think alike... I made 0603 footprints with electrically shorted pads (as net ties) specifically for that purpose. Nothing to test the signal with yet, though. I mean, a couple of beers should go a long way in procuring a scope and a thirsty helping hand with it, at least for a couple of hours. You mentioned a very nice point about a uSD card probably not going to drive as strongly as an MCU. I anticipated termination on the MCU side for sure. I can also adjust the actual drive strength for each MCU pin in firmware which is certainly helpful. And a known impedance of 50 ohms (because why not?) goes without saying.
Amazing!
you need that clamp current 100% bang on or you cant get a boolean factor.
put the gates down, then you will know the power. Im pretty sure, but I havent done for total real yet, but if u do, its something that a programmer needs to get to eventually once hes tired of his computer performance.