Raspberry Pi - Schematics V1.0 Model(A) and (B)

Hi Pete,

Nice job on the schematics. Good careful design, sure reads as if you intended it to work first time.

That does mean that to shave off a few pennies of production costs you can test if all the protections are neccessary, and then consider leaving them off on future production runs. The commercial USB hubs I get have PCB space for the fuses, but then don't populate them by the time I get them.

(most copied over from different places).

My questions about the schematics:

• What is the HD interface? (schematics P4 left top) (With all the pins under the BGA, there is no way I could get at those balls, but just asking).
• What value is typical for the built-in pullup/pulldown? I expect that to be on the order of 50-100k. In that case you can encode 3^4 different board revisions with the 4 board-revision strap resistors: read value. If zero, enable pullup, if one enable pulldown. If value now changes, the strap is left "open", so that's the third value.  (Ah! found them! near the GPIO.).
• There is an error in the schematics. p3 mentions the "USB" part of the LAN9512 as "model A only". That should be "Model B only" of course.
• R22-R25, R28, C28 are missing their "Model B only" tag.
• Why is there a line going from R25 to R22?
• Due to space constraints the 100M led is labeled 10M on the board. How about calling that "FE" for "Fast Ethernet"? That's what 100M was called back in the days.
• What is the purpose of D15, D16? I'm thinking: Protection against static discharges, right? The question is: Wouldn't it be better to connect their kathode to +5V0 instead of +5V0_HDMI? The +5V_HDMI is just a small net that only has the 100nF C75 on it....
• Are the HDMI_SDA and HDMI_SCL pins 5V tolerant? I wouldn't have guessed. Or is it allowed to sink that 1mA into the BCM2835? (We don't have the chip's datasheet, but then again most commercial datasheets don't mention this parameter).
• Why not use a BAV99S (from NXP) for D15 AND D16 (i.e. use the 6 pin, 4 diode version). Same goes for D12, D13. And possibly D4 and D14. (but those are annoyingly far apart I guess).  This could reduce the component count by a few.
• As the GPU is able to read the MAC address from somewhere, why can't it read the board revision from the same place and pass it to the kernel using the same mechanism. Or would this require a ROM upgrade to the SOC? I'm now guessing it is/can be done in "start.elf" which apparently is something that can be recompiled twice in a week. (I have Apr 12 and Apr 18 versions. :-) )
• Would it be possible/wise to make a brownout detector on the board? A few analog components could do the trick. Compare 0.43 times the 5V0 line with the 1.8V (This puts the threshold at 4.2V). When this triggers for longer than a few microseconds, pull the reset line of the SOC AND turn on a LED for a few seconds.
• With the I2C pullups on SDA0 and SCL0 already on the 'pi, these pins are not easily used as GPIO anymore. Wouldn't it have been better to specify: "do those yourself if you want to do I2C"? On the other hand, changing it now is going to be a bigger nightmare....

Suggestions:

• Drop a couple of the LAN9512 status leds and subtitute GPIO leds. Just monitor dmesg for info on 10/100 and FDX.

Question:

• I read somewhere that there are two IO voltage banks. Is this true? (would it be possible to drop the IO voltage to the SD card to 1.8V or not.
• What is the nominal core voltage of the BCM 2835? I'm measuring about 1.3V, or is this changed with the load and CPU speed? (I graduated from the group that "invented" clock scaling and undervolting to save power. At least those guys claimed they invented that there. A few years down the road it was commonly known and done.
• I thought the Polyfuse would need to cool down (i.e. a few minutes) for it to reset. However people claim an hour or more. Do you know how this works? (Littlefuse hasn't answered their EMail yet).

Oddity:

Pete, I was just checking my 5V power, and found that with the Raspberry powered off, there was still about 1.3V on the 5V0 line. Removing cables one by one indicated this was coming from HDMI.

Turns out 5V0_HDMI is 2.7V. and the HDMI_SCL and HDMI_SDA are 3.something.

So apparently my IIYAMA monitor is partially powering the RASPI through the I2C datalines.... This leads to the question: Do the specs say something about WHO has to pull those lines up to 5V? (do the specs say which voltage?) I made the same mistake: I designed a GPIO breakout board and had designed pullup resistors on my board.

(and we can conclude that the whole raspberry pi draws less than 2 mA at 1.3V on the power supply.)

Hi Pete,

Nice job on the schematics. Good careful design, sure reads as if you intended it to work first time.

That does mean that to shave off a few pennies of production costs you can test if all the protections are neccessary, and then consider leaving them off on future production runs. The commercial USB hubs I get have PCB space for the fuses, but then don't populate them by the time I get them.

(most copied over from different places).

My questions about the schematics:

• What is the HD interface? (schematics P4 left top) (With all the pins under the BGA, there is no way I could get at those balls, but just asking).
• What value is typical for the built-in pullup/pulldown? I expect that to be on the order of 50-100k. In that case you can encode 3^4 different board revisions with the 4 board-revision strap resistors: read value. If zero, enable pullup, if one enable pulldown. If value now changes, the strap is left "open", so that's the third value.  (Ah! found them! near the GPIO.).
• There is an error in the schematics. p3 mentions the "USB" part of the LAN9512 as "model A only". That should be "Model B only" of course.
• R22-R25, R28, C28 are missing their "Model B only" tag.
• Why is there a line going from R25 to R22?
• Due to space constraints the 100M led is labeled 10M on the board. How about calling that "FE" for "Fast Ethernet"? That's what 100M was called back in the days.
• What is the purpose of D15, D16? I'm thinking: Protection against static discharges, right? The question is: Wouldn't it be better to connect their kathode to +5V0 instead of +5V0_HDMI? The +5V_HDMI is just a small net that only has the 100nF C75 on it....
• Are the HDMI_SDA and HDMI_SCL pins 5V tolerant? I wouldn't have guessed. Or is it allowed to sink that 1mA into the BCM2835? (We don't have the chip's datasheet, but then again most commercial datasheets don't mention this parameter).
• Why not use a BAV99S (from NXP) for D15 AND D16 (i.e. use the 6 pin, 4 diode version). Same goes for D12, D13. And possibly D4 and D14. (but those are annoyingly far apart I guess).  This could reduce the component count by a few.
• As the GPU is able to read the MAC address from somewhere, why can't it read the board revision from the same place and pass it to the kernel using the same mechanism. Or would this require a ROM upgrade to the SOC? I'm now guessing it is/can be done in "start.elf" which apparently is something that can be recompiled twice in a week. (I have Apr 12 and Apr 18 versions. :-) )
• Would it be possible/wise to make a brownout detector on the board? A few analog components could do the trick. Compare 0.43 times the 5V0 line with the 1.8V (This puts the threshold at 4.2V). When this triggers for longer than a few microseconds, pull the reset line of the SOC AND turn on a LED for a few seconds.
• With the I2C pullups on SDA0 and SCL0 already on the 'pi, these pins are not easily used as GPIO anymore. Wouldn't it have been better to specify: "do those yourself if you want to do I2C"? On the other hand, changing it now is going to be a bigger nightmare....

Suggestions:

• Drop a couple of the LAN9512 status leds and subtitute GPIO leds. Just monitor dmesg for info on 10/100 and FDX.

Question:

• I read somewhere that there are two IO voltage banks. Is this true? (would it be possible to drop the IO voltage to the SD card to 1.8V or not.
• What is the nominal core voltage of the BCM 2835? I'm measuring about 1.3V, or is this changed with the load and CPU speed? (I graduated from the group that "invented" clock scaling and undervolting to save power. At least those guys claimed they invented that there. A few years down the road it was commonly known and done.
• I thought the Polyfuse would need to cool down (i.e. a few minutes) for it to reset. However people claim an hour or more. Do you know how this works? (Littlefuse hasn't answered their EMail yet).

Oddity:

Pete, I was just checking my 5V power, and found that with the Raspberry powered off, there was still about 1.3V on the 5V0 line. Removing cables one by one indicated this was coming from HDMI.

Turns out 5V0_HDMI is 2.7V. and the HDMI_SCL and HDMI_SDA are 3.something.

So apparently my IIYAMA monitor is partially powering the RASPI through the I2C datalines.... This leads to the question: Do the specs say something about WHO has to pull those lines up to 5V? (do the specs say which voltage?) I made the same mistake: I designed a GPIO breakout board and had designed pullup resistors on my board.

(and we can conclude that the whole raspberry pi draws less than 2 mA at 1.3V on the power supply.)