Missing labels on the schematics 4b

[RiccardoMoretti]

I'm looking at the schematics for the Pi 4b and whilst the information is there, you have no Idea what chip is being referred to, i see the other documents have the labels on,
Was this intentional or an omission ?

https://www.raspberrypi.org/documentati ... educed.pdf

[jamesh]

No idea, will check.

[RiccardoMoretti]

Thanks James,
just to clarify my initial post and to help you,

the previous docs had the component part number or its label
EG,

10103594-0001LF Power In
1-1734248-5 Display
XR77004 U8
A70-112-331-N126 J10 Ethernet

[James Adams]

These have been removed as they aren't generally useful - they aren't standard parts that are available in the channel and so you can't get hold of the parts or datasheets.

Also, we often change manufacturers (of connectors, say) so they can change over time.

Is there some specific info you need?

[RiccardoMoretti]

All good many thanks, I just wasn't sure of the power management chip (MxL7704) but I did find a picture of the Pi and I saw the part number from there, just would have been nice to see that on the schematic

[hippy]

All good many thanks, I just wasn't sure of the power management chip (MxL7704) but I did find a picture of the Pi and I saw the part number from there, just would have been nice to see that on the schematic

I agree. It may not be that useful for most people but is useful for some, but not that useful when the parts are proprietary and datasheets are not available.

Knowing what parts were originally used and what is actually fitted has been useful in identifying some issues which have arisen when parts have changed. This was the case when a replacement regulator made a Pi much more sensitive to small 5V drops than it previously was.

It also allows people to double-check that everything is how it should be, identify issues there may be, as might have been the case with the PoE HAT issue.

Knowing the part numbers is also useful for figuring out what the lowest voltage a Pi can run at is, though that's complicated by not having full circuit diagrams as well as not having datasheets.

Talking of 'wanting to know': Can anyone read the part number of the Boot Eeprom on a Pi 4 ?

I'm guessing it's the chip near the GPIO on the rear of the board given the solder pad jumpers -

https://raspi.tv/wp-content/uploads/201 ... _700-1.jpg

Thoygh it might be one of the two 8-legged chips to the lower left of the VIA USB chip -

https://raspi.tv/wp-content/uploads/201 ... 4B_700.jpg

I was thinking the Boot Eeprom could be used to hold Videocore bare metal code, so one can have an entirely embedded Pi system bootable without any SD Card or anything else connected. But to pursue that it would be useful to know its write-erase lifetime and for that one needs to know the part number. I haven't so far found a photo which I can read that on, and that's not even included on the schematics.


One thing which does intrigue me on the Pi 4B schematic is what "WB_VBAT" is used for ?

Could "WB" be Watch Battery ? Or Something Backup ?

[RiccardoMoretti]

hippy I agree with you,
It's exactly why I wanted the see the IC part number, so I could read the data sheet, which is available. having U8 on the schematic is way more useless to me!
https://www.maxlinear.com/ds/mxl7704.pdf

I can see there has been a revision to the mxl7704 IS from 03 to 04

WB could that refer to Wireless/Bluetooth ?

[chrisnz]

One thing which does intrigue me on the Pi 4B schematic is what "WB_VBAT" is used for ?

I too am curious as to what this net is for.

It's connected to 5V via a diode and supplies power to the power led.

Strangely, the LED is turned off by shorting it out - kind of weird.
Also of concern is the 470R R13, which appears to do nothing but sit there and burn 10mA for no apparent reason.
With GLOBAL_EN shorted to ground, turning off the PMIC, the LED and R13 appear to be the only things consuming power.
I get 22.2mA in power-off mode and 13.9mA when GLOBAL_EN is at 0V. The LED is consuming 3mA, due to its 1.8V drop and a 1K resistor. R13 must be taking the other 10mA.

What would be the harm in removing R13, or replacing it with a higher value? Is WB_VBAT connected to anything else? From the top and bottom layers it looks like it's not, but obviously I can't see the inner PCB layers and the foundation is shy on releasing more detailed schematics for some unknown reason.

Capture.PNG (13.86 KiB) Viewed 1312 times

WIN_20191018_23_15_50_Pro.jpg (129.39 KiB) Viewed 1312 times

[trejan]

Thoygh it might be one of the two 8-legged chips to the lower left of the VIA USB chip -

https://raspi.tv/wp-content/uploads/201 ... 4B_700.jpg

Boot EEPROM is a Winbond W25X40CLUXIG. It is the USON-8 closest to the headphone socket.

USB controller EEPROM is a Winbond W25X10CLUXIG. It is the USON-8 closest to the USB controller.

Manufacturer/Device ID and markings match up with the Winbond datasheets. Both parts are rated for > 100,000 erase/write cycles.

[hippy]

Is WB_VBAT connected to anything else? From the top and bottom layers it looks like it's not, but obviously I can't see the inner PCB layers

There are these, but not sure they will be a lot of use ...

https://www.raspberrypi.org/app/uploads ... tW8Qws.jpg

https://imgur.com/a/Y6mtR2E#A8ud1e6

[user12345]

It's connected to 5V via a diode and supplies power to the power led.

Strangely, the LED is turned off by shorting it out - kind of weird.
Also of concern is the 470R R13, which appears to do nothing but sit there and burn 10mA for no apparent reason.
With GLOBAL_EN shorted to ground, turning off the PMIC, the LED and R13 appear to be the only things consuming power.
I get 22.2mA in power-off mode and 13.9mA when GLOBAL_EN is at 0V. The LED is consuming 3mA, due to its 1.8V drop and a 1K resistor. R13 must be taking the other 10mA.

What would be the harm in removing R13, or replacing it with a higher value? Is WB_VBAT connected to anything else? From the top and bottom layers it looks like it's not, but obviously I can't see the inner PCB layers and the foundation is shy on releasing more detailed schematics for some unknown reason.

Seems for heating the PI if someone use it on the North or Southpol
I would guess its a error, cause noone of the guys who made this answers.
This shorting thing may cause you have a unchangeable Level(Power Chip gives high at undervoltage) and a PCh Fet/2 Nch Fets is/are more expensive than 1 Nch. People will wonder that turning off the LED dont save power...

R13 would make little sense if D2(Schottky <0.4V drop) would be a Zener Diode and the LED is switched, not shorted to add a little load,
for the case WB_VBAT is only 3V3 tollerant and the designer dont know that it is waste of energy to drive a red led(<1.5V) from 5V instead of 3V3.

Without the full Schematics the only save ways seem:
-Replace D2 against a 2.7V Zener Diode, remove R13. Load isnt required cause LED is shorted, WB_VBAT is always <3.3V.
or
-Attach D2 to 3V3 instead of 5V, remove R13.

Dont know why 10year old LEDs are used that needs mA for this brightness while LEDs today can reach that with µA.
Means replacing the LED and R17 would too possible.

On PI4 you have to use GLOBAL_EN Jumper to turn it off? On earlyer models i shutdown and the Powermeter shows 0W, the PI4 shows 1.3W on shutdown(Network LEDs are still on), more than a PI2 takes at operating(Idle)

[jamesh]

Checked with HW, the Pi4 is all working as designed - not seeing any faults here, so if you think something done in an odd fashion, it was done that way for a reason (technical, cost, manufacturing ease) . And reasons, on the whole, are proprietary information.

[James Adams]

R13: 5V bleeder to pull 5V down in sensible time when unplugged.
WB_VBAT: WiFi/BT supply, needs to be >3.6V for RF PA LDO to be in regulation.
"Strangely, the LED is turned off by shorting it out - kind of weird" - not really, it guarantees the LED is ON until actively driven OFF (which is only intended to be during an error state).

Note the board is designed to a cost - hence some of the engineering decisions. We generally expect these to be powered on all the time so saving a few mA when off is not a primary concern.

"Replace D2 against a 2.7V Zener Diode" - this won't help you'll still be drawing the same current you'll just burn the power in the diode instead.

"Dont know why 10year old LEDs are used that needs mA for this brightness while LEDs today can reach that with µA" - point me at a very cheap equivalent one please?

"On PI4 you have to use GLOBAL_EN Jumper to turn it off? On earlyer models i shutdown and the Powermeter shows 0W, the PI4 shows 1.3W on shutdown(Network LEDs are still on), more than a PI2 takes at operating(Idle)" - There is a boot ROM flag that will cause a shutdown to power down all rails (except 5V) - this interacts poorly with some HATs that don't like it if they have 5V but 3V3 goes away hence it's not the default.

[hippy]

R13: 5V bleeder to pull 5V down in sensible time when unplugged.
WB_VBAT: WiFi/BT supply, needs to be >3.6V for RF PA LDO to be in regulation.

Thanks for the clarifications, particularly as to what WB_VBAT is - With the datasheet seeming to indicate it was optimally functional between 3.2V and 4.8V I had presumed it to have been driven from the system-wide 3.3V supply. I had also overlooked that their power input is named "VBAT" which, with hindsight, is a pretty obvious clue.

[user12345]

R13: 5V bleeder to pull 5V down in sensible time when unplugged.
WB_VBAT: WiFi/BT supply, needs to be >3.6V for RF PA LDO to be in regulation.

Means WiFi/BT is powered through D2 from 5V and require at least 3.6V and require to be faster down than 5V main?
Cause there is no 3V6 Stepdown we can only replace the LED and resistor for Powersaving.
General i would assume, cause the new ARM has fewer nm, the PI4 require less power at same frequency.
If WiFi/BT is disabled, is WiFi/BT really powered down or is this not switchable for cost reasons?
If we would have the full schematics and know if there other Voltages on the wifi/bt we could know if this supply trace could be removed
for power saving.

"On PI4 you have to use GLOBAL_EN Jumper to turn it off? On earlyer models i shutdown and the Powermeter shows 0W, the PI4 shows 1.3W on shutdown(Network LEDs are still on), more than a PI2 takes at operating(Idle)" - There is a boot ROM flag that will cause a shutdown to power down all rails (except 5V) - this interacts poorly with some HATs that don't like it if they have 5V but 3V3 goes away hence it's not the default.

Yes, that could destroy devices if they not build for this. But how this was handled on earlyer PIs? Network Chip isnt switchable cause of cost reduction?