Working documentation
This page is generated from the project's working notes. It records a real retrofit in progress, corrections included. Anything marked 🔴 or ⚠️ is unresolved or carries a caveat - read those before acting.
The Mesa cards — every terminal, from the manuals¶
7 September 2026. Read directly from 7i77uman.pdf, 7i84uman.pdf and 7i76euman.pdf, all three on Drive. Every table below was cross-checked against the PDF text layer by word coordinate, because these are multi-column pin tables that interleave badly under plain text extraction.
This page exists to be printed. It is the terminal-by-terminal reference that turns "the X analog command goes to the 7I77U" into "land this wire on TB5 terminal 4."
🔴 Four things to check against the physical boards before wiring¶
✅ 7I77U Rev A: TB7 and TB8 silkscreen labels are SWAPPED |
The manual's own errata. ✅ DOES NOT APPLY TO THIS MACHINE - our 7I77U is Rev B (Grant, 1 Nov 2026). The tables on this page and the board print agree |
⚠️ 7I77U TB2 voltage rating disagrees with itself |
The TB2 terminal table says 8-32 V; the narrative and the specification table both say 5-28 V. Confirm before choosing a field supply. 24 V is inside both, so this does not bite us |
⚠️ 7I76EU calls its 5 V power connector both P4 and P5 |
Same connector, two names in one paragraph. Check the silkscreen |
🔴 The boards in THIS machine, by revision¶
Recorded 1 November 2026, on the bench, before they went into the Housing. Not recoverable once they are mounted and wired.
| Card | Revision | Known errata | Status |
|---|---|---|---|
| 7I77U | Rev B | TB7/TB8 silkscreen swap, Rev A only |
✅ Not affected. Board print and tables agree |
| 7I76EU | Rev A | None found | ⚠️ See below |
| 7I84U | Rev A | None found | ⚠️ Same caveat as the 7I76EU |
⚠️ On the two Rev A cards this is an absence of evidence, not a clean bill. Nothing in this repo, nothing in the manuals as captured, and nothing in a web search on 1 Nov 2026 turns up a revision issue for either the 7I76EU or the 7I84U - the swapped-silkscreen errata has only ever been a 7I77 Rev A problem. But mesanet.com is outside this session's allowed network domains, so Mesa's own current errata page could not be read.
✅ Cheap closure, and it beats any document. Mesa silkscreens every signal beside its terminal, so read the board:
| Card | Block | Terminals | Should print |
|---|---|---|---|
| 7I76EU | TB6 |
17-24 | outputs 0-7 |
| 7I76EU | TB5 |
17-24 | outputs 8-15 |
| 7I84U | TB3 |
1-16 / 17-24 | inputs 0-15 / outputs 0-7 |
| 7I84U | TB2 |
1-8 / 17-24 | inputs 16-23 / outputs 8-15 |
Definitively, once LinuxCNC is up, mesaflash --readhmid gives the pin assignment from the firmware itself, which outranks any printed label. Current mesaflash releases report TB1, TB2 and TB3 correctly; older ones did not, which is itself a source of confusion in the forum archives.
🔴 A 7I84U missing one field rail is NOT DETECTED AT ALL¶
Stronger than "will not start", and it is the symptom you meet first. Mesa's own forum guidance: both field supplies must be connected or the card is not detected.
⚠️ And the two cards take them the opposite way round:
| Card | Feeds | Feeds |
|---|---|---|
| 7I84U | VFIELDB -> TB2 |
VFIELDA -> TB3 |
| 7I77U | VFIELDA -> TB7 |
VFIELDB -> TB8 |
🔴 If the 7I84U does not appear at first power-up, check both field rails before suspecting the card, the SSerial cable or the firmware. With CN10 landed across both of its blocks, one missing rail presents as half the outputs being dead, which is a far more confusing symptom than the card simply being absent.
| ⚠️ 7I76EU jumper names conflict | IP address selection is W13/W14 in one section and W2/W3 in another; secondary flash is W7 in one place and W9 in another. Both look like manual typos. Confirm against the board before moving a jumper |
✅ Pin 1 is marked with a square pad on the 7I77U and 7I76EU. On all three cards, pin 1 of a 24-way field block is at the bottom edge; on the 8-way power blocks pin 1 is the bottom pin and pin 8 the top.
7I77U — analog servo interface¶
Connectors¶
| Type | Ways | For | |
|---|---|---|---|
P1 |
DB25F | 25 | Host interface to the 7I76EU |
TB1 |
screw | 2 | Aux 5 V for the analog / encoder / RS-422 section |
TB2 |
screw | 8 | Field power |
TB3 |
screw | 24 | 🔴 Encoders 0-2 |
TB4 |
screw | 24 | Encoders 3-5 |
TB5 |
screw | 24 | 🔴 Analog drive interface — command and enable |
TB6 |
screw | 8 | RS-422 |
TB7 |
screw | 24 | Field I/O — inputs 16-31, outputs 8-15 |
TB8 |
screw | 24 | Field I/O — inputs 0-15, outputs 0-7 |
🔴 TB5 — the analog drive interface¶
Four terminals per axis, and the enable sits alongside the command. This is the block the three servo drives land on.
| Pin | Signal | Pin | Signal | Pin | Signal | ||
|---|---|---|---|---|---|---|---|
| 1 | ENA0- |
9 | ENA2- |
17 | ENA4- |
||
| 2 | ENA0+ |
10 | ENA2+ |
18 | ENA4+ |
||
| 3 | GND |
11 | GND |
19 | GND |
||
| 4 | AOUT0 |
12 | AOUT2 |
20 | AOUT4 |
||
| 5 | ENA1- |
13 | ENA3- |
21 | ENA5- |
||
| 6 | ENA1+ |
14 | ENA3+ |
22 | AOUT5 region |
||
| 7 | GND |
15 | GND |
23 | ENA5+ |
||
| 8 | AOUT1 |
16 | AOUT3 |
24 | AOUT5 |
(Channels 4 and 5 follow the same 4-pin pattern; only 0-3 are used here.)
🔴 The enables are differential pairs (ENA+ / ENA-), not open-drain to ground. That is not the same shape as the SGDH's /S-ON, which is a sinking input against +24V-IN. How the two meet is a design question, not a wiring one — see servo-drive-interface.md.
⚠️ Analog wiring, from the manual: minimum load 2 kΩ; use shielded twisted pair with AOUTn to the drive's AIN+ and the adjacent GND to AIN-; tie the shield at the 7I77U end only; run the drive's common back to 7I77U power ground on a separate wire.
🔴 TB3 — encoders 0-2¶
Eight terminals per channel. +5V and GND are supplied by the card, so the encoder is powered from here.
| Pin | Ch 0 | Pin | Ch 1 | Pin | Ch 2 | ||
|---|---|---|---|---|---|---|---|
| 1 | QA0 |
9 | QA1 |
17 | QA2 |
||
| 2 | /QA0 |
10 | /QA1 |
18 | /QA2 |
||
| 3 | GND |
11 | GND |
19 | GND |
||
| 4 | QB0 |
12 | QB1 |
20 | QB2 |
||
| 5 | /QB0 |
13 | /QB1 |
21 | /QB2 |
||
| 6 | +5V |
14 | +5V |
22 | +5V |
||
| 7 | IDX0 |
15 | IDX1 |
23 | IDX2 |
||
| 8 | /IDX0 |
16 | /IDX1 |
24 | /IDX2 |
✅ Differential is the factory default on all six channels, which is what the SGDH's PG output is. Nothing to change.
✅ IDX is on the block whether or not it is used — so landing the index wire costs one terminal and enables HOME_USE_INDEX later (see the homing note in pre-strip-survey.md item 4).
TB4 carries channels 3-5 on the identical pattern.
TB2 — field power¶
| Pin | Signal | Feeds |
|---|---|---|
| 1, 2 | VFIELDA |
🔴 TB7 — its outputs, and its input thresholds |
| 3, 4 | VFIELDB |
🔴 TB8 — its outputs, and its input thresholds |
| 5 | VIN |
field I/O logic power |
| 6, 7 | NC |
|
| 8 | GROUND |
common to VIN and both VFIELD |
🔴 Both VFIELDA and VFIELDB must be supplied or the card reports a low-voltage fault and will not start. Even if only one I/O block is used.
TB8 and TB7 — field I/O¶
TB8 pin |
TB7 pin |
||
|---|---|---|---|
| 1-16 | INPUT0-INPUT15 |
1-16 | INPUT16-INPUT31 |
| 17-24 | OUTPUT0-OUTPUT7 |
17-24 | OUTPUT8-OUTPUT15 |
Inputs are sinking, ~11 kΩ, threshold 50 % of the field supply with 10 % hysteresis — on 24 V that is high above 14.4 V, low below 9.6 V. INPUT0-INPUT3 can also read analog, 8-bit.
Outputs are universal — sourcing, sinking or push-pull, selectable per output in software. 500 mA per output, 2.0 A per driver chip (each chip serves eight sequential outputs), short-circuit trip about 1.5 A. ✅ Built-in flyback clamps — no external diodes needed on inductive loads.
7I84U — isolated remote field I/O¶
| Type | Ways | For | |
|---|---|---|---|
TB1 |
screw | 8 | Power |
TB2 |
screw | 24 | Inputs 16-31, outputs 8-15 |
TB3 |
screw | 24 | Inputs 0-15, outputs 0-7 |
J1 |
RJ-45 | 8 | RS-422 to the 7I76EU |
TB1 — power¶
| Pin | Signal | Feeds |
|---|---|---|
| 1, 2 | VFIELDB |
🔴 TB2 |
| 3, 4 | VFIELDA |
🔴 TB3 |
| 5 | VIN |
logic power, 8-32 V |
| 6, 7, 8 | GROUND |
⚠️ Note the ordering is the opposite way round to the 7I77U — on the 7I84U, VFIELDB is on pins 1-2 and feeds TB2; on the 7I77U, VFIELDA is on pins 1-2 and feeds TB7. Easy to get wrong when wiring both cards in one session.
TB3 and TB2 — field I/O¶
Same layout as the 7I77U's blocks: pins 1-16 the sixteen inputs, pins 17-24 the eight outputs.
Inputs sinking, 22 kΩ (twice the 7I77U's), same 50 % threshold. Outputs 5-28 V, sourcing / sinking / push-pull per point, 500 mA.
Status LEDs — worth knowing before the first power-up¶
CR2 yellow, upper left |
VIN power |
CR5 red, lower left |
🔴 field I/O fault |
CR6 green, lower left |
field I/O activity |
CR7 yellow, bottom left |
RS-422 cable power |
✅ CR2 and CR7 must both be lit in normal operation.
7I76EU — the host card¶
| Type | Ways | For | |
|---|---|---|---|
| RJ-45 | 8 | Host Ethernet | |
P1, P3 |
IDC header | 26 | Expansion I/O |
TB1 |
screw | 8 | Field power |
TB2 |
screw | 24 | Step/dir 0-3 |
TB3 |
screw | 24 | Step/dir 4, encoder, RS-422, unregulated logic power in |
TB4 |
screw | 8 | Spindle |
TB5 |
screw | 24 | Inputs 16-31, outputs 8-15 |
TB6 |
screw | 24 | Inputs 0-15, outputs 0-7 |
P4 |
screw | 2 | 5 V logic power |
TB4 — the spindle connector¶
| Pin | Signal |
|---|---|
| 1 | SPINDLE- |
| 2 | SPINDLE OUT |
| 3 | SPINDLE+ |
| 4 | NC |
| 5 | SPINDLE ENA- / FWD- |
| 6 | SPINDLE ENA+ / FWD+ |
| 7 | SPINDLE DIR- / REV- |
| 8 | SPINDLE DIR+ / REV+ |
⚠️ The analog output here needs an external reference supply on SPINDLE+ / SPINDLE-, 5-20 V, SPINDLE+ always the more positive. Output current 5 mA. The ENA and DIR outputs are isolated, up to 100 V open circuit, 50 mA.
TB1 — field power¶
| Pin | Signal |
|---|---|
| 1-4 | VFIELD |
| 5 | VIN |
| 6, 7, 8 | GROUND |
⚠️ No VFIELDA/VFIELDB split on this card — one rail feeds both TB5 and TB6. Different from the other two.
TB5 and TB6 — field I/O¶
Same 16-in / 8-out layout. TB6 inputs 0-3 double as analog inputs; TB5 inputs 16-19 double as MPG quadrature. Inputs sinking, 10 kΩ. Outputs 500 mA.
Rules that apply to all three cards¶
🔴 VFIELD must connect directly to the DC source |
No switches, breakers or relay contacts in that circuit. A fuse is acceptable |
| 🔴 Field power ramp rate under 10 V/ms | Which is another way of saying the same thing — it cannot be switched with a mechanical contact |
| ✅ No flyback diodes needed | All three cards' output drivers have built-in clamps |
⚠️ Sinking outputs must not drive a load tied to a higher voltage than VFIELD |
It back-powers the field rail through the load |
| ✅ Frame ground | The 7I76EU has a 0.250" blade tab near the Ethernet jack. Earth it |
I/O capacity against what the machine needs¶
⚠️ The count is 44 inputs, not 41 — corrected 7 September¶
Grant: "We need to allow for 3 home sensors as inputs — the ones that come in on pin 6 of the D-sub connectors."
✅ Right. The 41/41 figure counts PLC addresses — IF.n and OF.n. The home switches BOX/BOY/BOZ have no PLC address at all; housing-connectors.md records them as "none — stops at the Housing". They arrive on CONN.X/Y/Z pin 6 and were never in the 41.
🔴 And there is a matching pair on the output side¶
Checking the axis D-sub table for anything else in the same gap turns up the servo enables. /S-ON on pin 13 of each axis connector (wires 71 / 85 / 99) also has no PLC address — it comes from the Tecnos axis card, not from an output module.
✅ But it costs nothing, because it lands on the 7I77U's TB5 ENA0-2± alongside the analog command, not on a field output block. The two gaps do not cancel: inputs go up by three, outputs do not move.
| Needs | Where it lands | |
|---|---|---|
PLC inputs IF.n |
41 | field input blocks |
| 🔴 Home switches, pin 6 | +3 | field input blocks |
| Field inputs, total | 44 | |
PLC outputs OF.n |
41 | field output blocks |
| Servo enables, pin 13 | +3 | 🔴 TB5, not a field output |
| Speed references, pin 8/15 ×3 + spindle | 4 | 🔴 TB5 AOUT0-3, analog |
| Encoder A/B/Z ± + common, ×3 | 3 channels | 🔴 TB3, not field I/O |
| Card | Inputs | Outputs |
|---|---|---|
| 7I76EU | 32 | 16 |
| 7I84U | 32 | 16 |
| 7I77U | 32 | 16 |
| Available | 96 | 48 |
| Needed | 44 | 41 |
✅ Inputs remain comfortable — 44 of 96. Outputs are the tight side: 41 of 48, seven spare.
✅ And the seven spare ARE freely placed — corrected 8 September. An earlier version of this said allocation was constrained by the 2.0 A per driver chip. Grant measured the module relays at 18.5 mA, so eight outputs on one chip draw 160 mA of a 2000 mA budget — eight per cent. Outputs can go wherever is tidiest.
Related¶
mesa-wiring-schedule.md— the signal-by-signal planmesa-hardware.md— what was ordered and whyservo-drive-interface.md— the SGDH side ofTB5andTB3io-map.md— every machine signal to its wire number