Configuration¶
Every field of max22200::Config and max22200::ChannelConfig, how a configuration is checked, and the optional
Kconfig layer. What the settings do to the load is on Drive modes.
Datasheet sections this page relies on
From Analog Devices 19-100531 1:
- The two parts and how each clears the FAULT register: General Description; Ordering Information, p. 1
- Oscillator and per-channel chopping frequency: Chopping Frequency (FREQM, FRQ_CFG_); Table 1, p. 16
- VDR duty-cycle codes: Voltage Drive Regulation (VDR); Table 3, p. 17
- CDR current codes, IFS from RREF and the half full-scale bit: Current Drive Regulation (CDR); Table 4; Full-Scale Current Setting; Half Full-Scale Setting, p. 18
- STATUS: Status Register Description (STATUS); Table 10; Register Details, STATUS (0x00), p. 26
- CFG_CHn: Configuration Register Description (CFG_CH_); Table 11; Register Details, CFG_CH, p. 27
- CFG_DPM: Detection of Plunger Movement Register Description (CFG_DPM); Figure 9, p. 29
- Register map and reset values: Register Map; Register Details, p. 35
All of them, with the driver API for each: Datasheet map.
Config{} holds the reset values of STATUS and every CFG_CHn (Register Details, p36, p38-39), for a part declared as
Max22200 (correct on both parts) whose RREF is not given: every channel CDR, low side, codes 0 (no current), 25 kHz,
every diagnostic off; COMER masked, the other faults signalled on FAULT; 100 kHz; every pair independent. The datasheet
gives no reset values for CFG_DPM (p40-41): dpm is 0, the value written at start-up.
max22200::Config ValveBoard() {
max22200::Config c{}; // the chip's reset state, MAX22200A
c.rref_ohm = 30000; // IFS = 500 mA
c.fault_mask.communication = false; // let COMER pull FAULT low as well
c.pairs[3] = max22200::PairMode::FullBridge; // channels 6 and 7: a latching valve
c.loaded_channels = 0xFF; // a load on every channel: diagnostics allowed
for (uint8_t ch = 0; ch < 6; ++ch) {
max22200::ChannelConfig& v = c.channels[ch];
v.hit = max22200::CurrentCode(400, c.rref_ohm, false).value; // checked: ok() in real code
v.hold = max22200::CurrentCode(120, c.rref_ohm, false).value;
v.hit_time = max22200::HitTimeCode(25000, c.clock, v.chop).value;
v.open_load_detect = true;
v.hit_current_check = true; // HHF if 400 mA is not reached within 25 ms
}
return c; // max22200::Validate(c) checks it without the chip
}
max22200::Error StartValves(MyPort& port) {
static max22200::MAX22200<MyPort> chip(port, ValveBoard()); // stored, no I/O
return chip.EnsureInitialized().error; // InvalidArgument if the config is out of range
}
Config¶
| Field | Type | Reset value | Register | Notes |
|---|---|---|---|---|
variant |
Variant |
Max22200 |
none | How a FAULT read clears; the default suits both parts, Max22200A only per the BOM (Hardware) |
rref_ohm |
uint32_t |
kNoRref (0) |
none | Not given by default: codes work, CurrentCode(), CurrentUa() and DpmCurrentCode() return InvalidArgument. 15000 to 100000 (15 kΩ: IFS = 1 A, KFS / RREF); never sent |
wake_timeout_us |
uint32_t |
100000 | none | How long the wake loop of EnsureInitialized() may take: 2500 (one tWU) to 10000000. tWU is 2.5 ms, but UVM clears only once VM is above UVLO (p22), so the wake also waits for the supply ramp; measure yours (bench step 1) |
loaded_channels |
uint8_t |
0 | none | The channels with a load fitted, bit n for channel n; never sent. Only these may turn on open_load_detect, plunger_detect or hit_current_check: on an unloaded channel they report the missing load as a fault and pull the shared FAULT pin (handbook KB-CPP-131) |
clock |
MasterClock |
Khz100 |
STATUS.FREQM | Global: every channel's chopping frequency and HIT time follow it |
fault_mask |
FaultMask |
COMER masked | STATUS[23:17] | A masked fault does not pull FAULT low; its flag is still set |
pairs[4] |
PairMode |
Independent |
STATUS.CMxy | Changes only with both channels of the pair off |
channels[8] |
ChannelConfig |
all zero | CFG_CH0 to CFG_CH7 | Below |
dpm |
DpmConfig |
none given (0 written) | CFG_DPM | start 0 to 127, debounce and threshold 0 to 15 |
ChannelConfig¶
| Field | Reset value | Register bits | Range and rules |
|---|---|---|---|
drive |
Cdr |
VDRnCDR (7) | Changes only with every channel off and TRIGA, TRIGB low |
side |
Low |
HSnLS (6) | The same; High needs Vdr |
hit |
0 | HIT[6:0] (22:16) | CDR 0 to 126 (CurrentCode()); VDR 0 to 100 (percent) |
hold |
0 | HOLD[6:0] (30:24) | The same; on the fly with SetHoldCode() |
hit_time |
0 | HIT_T[7:0] (15:8) | 0 none, 1 to 254 (HitTimeCode()), 255 forever (kHitTimeContinuous) |
chop |
Div4 |
FREQ_CFG (5:4) | fMAIN / 4, 3, 2, 1 |
half_scale |
false | HFS (31) | IFS halved; low side only |
trigger_pin |
false | TRGnSPI (23) | TRIGA (even channels) or TRIGB (odd) switches the channel instead of ONCH |
slew_rate_control |
false | SRC (3) | Low side; below 50 kHz only |
open_load_detect |
false | OL_EN (2) | Only on a channel in Config::loaded_channels |
plunger_detect |
false | DPM_EN (1) | Low side only; settings in Config::dpm; only on a loaded channel |
hit_current_check |
false | HHF_EN (0) | CDR only; only on a loaded channel |
EncodeChannel() and DecodeChannel() convert to and from the 32-bit register; EncodeStatus(), DecodeStatus(),
EncodeDpm(), DecodeDpm() and DecodeFaults() do the same for the others.
Validation¶
max22200::Validate(config) checks a configuration without touching the chip; EnsureInitialized() and
ApplyConfig() call it before any I/O, ConfigureChannel() calls ValidateChannel(). They return
Error::InvalidArgument for:
- an enum holding a value that is not an enumerator, or the reserved pair mode 3;
- an RREF that is neither
kNoRrefnor 15 kΩ to 100 kΩ (kNoRrefpasses: only the milliamp conversions need it); - a wake timeout outside 2.5 ms to 10 s;
- a CDR code above 126 (127 is "LS on", the coil at full VM with no regulation, Table 4) or a VDR code above 100;
- CDR on the high side (the chip ignores HSnLS in CDR), or SRC, HFS or DPM on the high side;
- SRC at 50 kHz or above ("SRC mode is only available ... for switching frequencies less than 50kHz");
- HHF in VDR (the HIT check is CDR only);
- a DPM field wider than its bits.
The runtime checks that need the chip's state return ChannelsActive instead: a drive-mode or side change while a
channel is on, a pair-mode change while one of the pair is on.
Kconfig¶
Under ESP-IDF, idf.py menuconfig shows MAX22200 solenoid driver when the driver is a component of the project
(Installation); the menu is _config/Kconfig. The options reach the code
only through inc/max22200_kconfig.hpp, never through a driver signature:
max22200::Error StartFromMenuconfig(MyPort& port) {
static max22200::MAX22200<MyPort> chip(port, max22200::KconfigDefaults());
return chip.EnsureInitialized().error;
}
| Option | Default | Config field |
|---|---|---|
CONFIG_MAX22200_VARIANT_PLAIN / CONFIG_MAX22200_VARIANT_A |
MAX22200 | variant |
CONFIG_MAX22200_CLOCK_80KHZ |
n (100 kHz) | clock |
CONFIG_MAX22200_RREF_OHM |
0 (not given) | rref_ohm |
The channel settings are not Kconfig options: set them in code. Without Kconfig (plain CMake, other platforms)
KconfigDefaults() returns Config{}. Both cases are host tests.
Next: Porting.