This board has an ATmega324pb which has three timers (Timer1, Timer3, Timer4) that have input capture (ICP) hardware all running from the same crystal. The main idea is to calibrate flow meters with a gravimetric method, so that means it needs to capture flow meter events (pulses) so they can be calibrated with a volume (corrected by weight). This method is also the preferred way to calibrate [provers]. ICP1 connects with the flow meter which produces pulses as a unit of volume flows (e.g., a magnet on a spinning turbine triggering a sensor). ICP3 is used to measure a start event and ICP4 a stop event. ICP3 and ICP4 inputs each have a one-shot pulse extender. Gravimetric proving is reasonably natural with a flow meter alone but may also benefit from ideas seen in flow labs. A diversion needs to switch the flow from the unmeasured side to the measurement side like a bistable switch. Usually, the measurement side fills a bucket (or ilk) that will be weighted. The one-shot that feds ICP3 on the ATmega324pb is also a circuit that starts the diversion instantly and gives the interrupt service routine (ISR) some time to latch the diversion. The one-shot pulse extender on ICP4 can override the diversion latch to send the flow away from the bucket. The ISR that ICP4 runs can then disable the latch that holds the diversion so it will stay off after the ICP4 one-shot ends. The result is that flow is immediately diverted from the instant START occurs to the instant STOP occurs. One helpful thing to do is slow the flow near the start and stop events, but that requires the magic of a computer (and SBC like an R-Pi) which needs to estimate when those events occur relative to the flow meter pulse count and thus require a somewhat iterative technique.
Bootloader options include optiboot and xboot. Serial bootloaders can not change the hardware fuse setting.
ATmega324pb application controler.
ATmega328pb managment controler.
Input power can range from 7 to 36V DC
Alternat input voltage (ALT_V) is sensed with manager ADC channel 1.
High side current sense of alternat input current (ALT_I) is sensed with manager ADC channel 0.
High side current sense of input current (PWR_I) is sensed with manager ADC channel 6.
Input voltage (PWR_V) is sensed with manager ADC channel 7.
Four analog (ADC channel 0..3) or digital connections with level conversion.
Three inputs for event capture: ICP1, ICP3, ICP4.
Two of the event capture inputs have a one shot pulse extender: ICP3, ICP4.
Event timers have a common crystal which eliminates correlation errors.
Event capture interface has a 17mA current source for pulse sensor (hall, VR, or a switch)
Event transition occures at about 6.5mA of current returned from the sensor to a 100 Ohm resistor on board.
Diversion control use ICP3 to START and ICP4 to STOP, and there ISR's to operate the CS_DIVERSION holding pin.
Diversion control is nearly instantaneous with a slight delay that has repeatable timing.
Control of a PMOS enables an alternate power supply (e.g. battery charging).
Control of a PMOS disables power to the 40PIN SBC header.
SPI between SBC and ATmega324pb with IOFF buffer so SBC can power off.
SMBus between SBC and ATmega328pb manag, SBC can power off without locking I2C.
I2C0 between ATmega324pb and ATmega328pb manager does not lockup on SBC power down.
I2C1 from ATmega324pb is for user and can be a master or slave.
Serial0 from ATmega324pb continues working with other RPUbus host when SBC is powered down.
Serial1 and Serial2 from ATmega324pb is for user.
Calibration of volumetric controlled rotation measured with input capture (ICP1) hardware.
Use a START (ICP3) and STOP (ICP4) sensor to capture displacer events while capturing flow meter pules.
Diversion and fast/slow flow control are used for gravimetric calibration methods.
One-shot pulse extenders on ICP3 and ICP4 for precise STAR and STOP events.
Use SPI only after ICP3 event is done, it will cut off CS_ICP3.
ADC channel 4..7 are not used at this time, the do have test points.
PB2 and PD7 are not used at this time, and also have test points.
^3 Done:
WIP:
Todo: Design, Layout (#=done), BOM, Review*, Order Boards, Assembly, Testing, Evaluation.
*during review the Design may change without changing the revision.
long term goal is to switch to the new AVR DB's
after development and testing of serial and I2C lib's for the DA's then start this update (DA's and DB's have same serial and I2C)
change m324pb to AVR128DB48 (DB has pins to drive HF crystal)
smoked manager with an extern pwr RPUusb (same problem will happen with extern pwr R-Pi)
change m328pb to AVR128DB32 (has dual power domains use primary for app and secondary for R-Pi)
^2 Done: Design, Layout (#=done), BOM,
WIP: Review*,
Todo: Order Boards, Assembly, Testing, Evaluation.
*during review the Design may change without changing the revision.
# Q105 and Q118 need to go into cut off (or hi-z) when the MOSFET gate is low.
# Add note on schem near *manager* ISP port "remove alternat power befor programing with ISP"
# Add test points on I2C0 nodes.
# Shutdown switch resistor needs moved so manager can overide switch while host is doing power UP.
This version may not be ordered and built, but the layout was done.
^1 Done: Design, Layout, BOM, Review*, Order Boards, Assembly,
WIP: Testing,
Todo: Evaluation.
*during review the Design may change without changing the revision.
Add to testing: calculate and save ref_intern_1v1 on manager
# add gap between RJ45 headers
# Power Protection
# Alternat power diode replaced with a P-CH MOSFET
# HOST_nCTS: on manager should move from PC0 to PD2
# keep I2C like ^0: TWI0 goes to the manager as I2C master, the other port TWI1 can be slave or master I2C.
# connect R1 (ALT_V divider to ADC4) directly to ALT input
# add second K38 under Q9 with its gate controled from ADC4 voltage, so 5..15V is needed to turn on ALT power.
# move MGR_STATUS LED to MGR_SCK.
# ALT_EN, PIPWR_EN: can go to the manager PB3, PB1
# connect PIPWR_EN to PB1 on bus manager
# connect ALT_EN to PB3 (MOSI) on bus manager (and do not ICSP program the manager with alt power applied).
# ALT_I, ALT_V, PWR_I, PWR_V: can go to the manager ADC 0, 1, 6,7
# connect ADC4..ADC7 to test points
# connect PB3 and PB1 to test points
# change pin numbers so PA0 (with ADC0) is pin 0 for both digital and analog functions
Debugging and fixing problems i.e. Schooling
Setup and methods used for Evaluation
The board is 0.063 thick, FR4, two layer, 1 oz copper with ENIG (gold) finish.
DIN rail
Check correct assembly and function with Testing
The BOM's are CVS files, import them into a spreadsheet program like Excel (or LibreOffice Calc), or use a text editor.
| Option | BOM's included |
|---|---|
| A. | BRD |
| J. | BRD SMD HDR |
| M. | BRD SMD HDR CAT5 |
| N. | BRD SMD HDR CAT5 POL |
| Y. | BRD SMD HDR CAT5 POL DIN PLUG |
The board is assembled with CHIPQUIK no-clean solder SMD291AX (RoHS non-compliant).
The SMD reflow is done in a Black & Decker Model NO. TO1303SB which has the heating elements controlled by a Solid State Relay and an RPUno loaded with this Reflow firmware.
The ATmega324pb has a good selection of IO. It has three hardware serial ports so one can be dedicated to the host, one to a scale, and one to a volumetric calibration device (prover). A volumetric prover will have two events that correspond to a precisely repeatable volume. Those events can be measured with ICP3 and ICP4. The method requires precisely diverting the flow of the volume into a measuring vessel between the events. The gravimetric weight of the diverted amount is then used to determine the volume. A flow meter can input pulses into ICP1 during the events and thus be calibrated as well. A flow meter can also be used to estimate when to slow the flow down so that the diversion precision can improve.
The diversion control is part hardware and part software; it is a work in progress at this time. The slow control will need flow meter pulses and values determined by a single board computer (e.g., an R-Pi); it is an idea that may turn into a work in progress.
How precise will the measurements be? Well, the idea is to sort out the method(s). I am not trying to get certified. That is for labs that want to do actual calibrations, though they may wish to ponder about this setup.