Here’s the final issue-list we had to be addressed with the OVMS v3 release candidate board:
Board layout to match enclosure (including mounting holes)
To be able to have a board + enclosure product.
CP2102 power from USB (with no connection to 3.3v on ESP side)
To fix sleep-mode power consumption issue with CP2102.
100nF for C21
To resolve auto-flash issue.
10K between pin 25 of the CP2102 and pin 35 of the ESP32
To protect against fake CP2102.
Change to resistor values for voltage divider on ESP32 ADC
To fix power consumption issue with that voltage divider network.
Tri-state buffer on MAX7317 MISO using 74AHC1G125
To fix MAX7317 MISO not co-existing with MCP2515 MISOs.
EGPIO P2 (MAX7317) 10K pull-up to 3.3v and connection to Rs of matching SN65
To fix sleep-mode power consumption issue of SN65.
MCP2515 #1 10K pull-up on RX0BF and connection to Rs of matching SN65
To fix sleep-mode power consumption issue of MCP2515 and SN65.
MCP2515 #2 10K pull-up on RX0BF and connection to Rs of matching SN65
To fix sleep-mode power consumption issue of MCP2515 and SN65.
#3 was the worst. That has been plaguing us for months, and got worse when we changed #2 to power the CP2102 from USB. We thought it might be a ‘fake’ component issue, but that turned out to be not so. In the end, it turned out to be a silicon issue on the ESP32. Frustrating in that Espressif knew about it all along, but it was not documented or discussed much. Anyway, a change from 1nF to 100nF capacitor works around the issue and will continue to work with newer silicon (when Espressif release it).
Apart from #3 and #6, these fixes are primarily to address power consumption. We want to get that as low as possible (particularly in sleep mode).
Without the modem, and with the above fixes, it seems that we can get down to around 23mA in normal operation, 19mA with CAN asleep, and 4.5mA with everything sleeping (all these @12v supply). By comparison, with modem, OVMS v2 is around 80mA @12v (and can’t sleep). We might be able to get the sleep figure a bit lower in software, but we think we are about as low as it is going to get. At such low currents, our switched mode power supply is not very efficient, and further work on this is really now hitting diminishing returns.
The modem should be able to be completely powered off, under control of the ESP32. It also has various low-power sleep modes. So, that will add to the consumption (as does wifi/bluetooth, if enabled), but only in use. Anyway, it is a separate board so any required fixes can be made directly without affecting the main board.
New layout is done, and release candidate board is being made now. It should be in my hands any day. Assuming that is ok, we’ll pull the trigger and make the first production run of developer boards.
Firmware is coming on, but we need more people working on it. Particularly for porting over vehicle support. So, #1 priority is getting those developer boards available.
Regards, Mark.
