mck1117 / proteus Goto Github PK

I believe that having silkscreen would improve usability.

Specifically silkscreen around connectors to facilitate troubleshooting & initial wiring.

With no grounds connected to the black 23pin connector and while powered on, the black 35 pin connector was disconnected and reconnected. When it was connected again the capacitor failed and the trace between the MR pin and capacitor was vaporized.

In an effort to control inductive loads, it would be effective to have freewheel diodes on the lowside outputs.

They're not easy to solder.

in light on humans freaking out about DBW for not good reason would proteus benefit from external to stm32 watchdog?

not DNP, or CAN won't work

BOM has diodes that are SOT323
Footprints are SOT23

Footprint isn't right, chip doesn't line up.

STM32F429 uses a different bootloader activation pattern than the F407, so not all board reliably make it to the loader when the PROG button is held.

Possible to bodge on v0.3 boards like this:

In favor of leaving it as 3.3v:

• fewer components, less board area
• potentially slightly more efficient

In favor of switching to 5v, and adding a separate 3.3v regulator:

• Potentially quieter 3.3v supply, if linear is used
• Prevents risk of 5v pull up from external shorts

We have ETB1_PWM and ETB2_PWM but DIR and DIR, DIS and DIS

schematics says "digital 7" but mapping says vr1? i think it wold be better to have those consistent

Several DRC and ERC errors reported. Many are "not" important, things KC wants a power flat to suppress the error things like that could be cleaned up.
-- Pin connected to.... should have power flag
-- PB2 could use nettie to suppress the message
-- Track to close to pad can be suppressed by adding a net name to the pad and making the via use that net name. AKA you really have a connected net.
-- components could be moved by like 5 mils to prevent courtyard overlap errors.

hopefully there are a number of unused pins to breakout on the board including a pair of UART pins

this would allow external Bluetooth if gnd +3v and +5v are also available on breakout

Or should I just add a 5x1 header for a sealed USB connector?

RC filter on vref+ would do good things, ST used a 47 ohm + 1.1uf on the f4 disco board.

that's confusing.

has JLC bom changed?
can mouser BOM be copied from 0.3 to reduce confusion to make 0.3a self-contained?

Some adapter boards might need SPI, so make it available.

It was pointed out that the 33k resistor may be a bit large, since that can put rise times in the 10s of us range, making us more susceptible to jitter and potentially losing short pulses altogether.

Noticed in mcu schematic that miso and mosi are swapped on the stm32 pins for SBC.
And SBC miso and sbc CSN are connected to each other on hierarchal root sheet

It's one of the harder ones to solder.

How much has to change to also support the lqfp144 STM32F407ZG?

What is the $10 part for USB connector with the sealed cap?

I believe it's not 5787834-1? https://github.com/mck1117/proteus/blob/master/export/v0.3/proteus_bom_mouser.csv mentions 5787834-1 in the same line with "if user wants USB in sealed conn" but I am confused what does 5787834-1 have to do with sealed?

0.2 has ibom

0.3 and 0.3a want iboms too!

Significant noise observed on the 3.3v rail, occuring every 100us (10khz) due to the fast ADC callback waking the cpu up, causing a ~70mvpp oscillation in the 3.3v power supply.

Possible solutions:

1. Disable sleep mode on the MCU so there isn't a corresponding jump in load when it wakes up
2. Split supplies, one running the 3.3v digital, and one running VDDA, VREF+, and the digital input buffer.
3. Investigate whether the TLS205 is actually any good, or whether using a different regulator would improve performance.
4. Better isolate VREF+ (and maybe VDDA) from VDD, potentially using an RC filter for VREF+.

How would Proteus handle Miata NB2 MAF 1k pull-down - where would it be located?

I recall there are some sensors with 12v output?

Would this case justify at least some prototyping area? I believe we can make PCB a bit larger if needed

That's the case @stefanst recommends with higher chance of water tightness

There are many many footprints used instead of via's. This means you cannot use the "delete extra footprints" option under "updated PCB from schematic" tool

The title block seems generic and the PCB does not reside within the title block boundaries. It seems odd to work off the side of the title block.

0.2 and 0.3 PDF schematics files are not searchable.

there are two ways to produce PDF export - kicad "PLOT" and kicad "print" + PDF printer. One way is searchable the other one is not

What needs to happen for:

• Reverse battery protection
• "ecu supply" protection (powering ECU thru high side driver)

Would help to say "white" above one of these 35 pin

Looks like we are getting closer to SD card read functionality please add SD card socket on the board

47219-2001 hinge type is really easy to solder

Knock sensors are high impedance - circa 1 megaohm.

Input impedance of the frontend looks like this:

Something in the 3-7k ohm range.

Which means that when you include the source impedance, the filter actually does this:

This is neither a flat passband, nor is -42dB of gain acceptable. That's down near the noise floor of the system, so only the loudest knocks would appear at all over the noise, and certainly not with any fidelity.

f767 in lqfp144
216mhz

Usability - mention LED colors on schematics