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.

Servo drive interface - CN1, the axis looms, and what the Tecnos actually sends¶

2 September 2026. Read directly from the wiring diagram sheets 7 (AZION. ASSE X), 8 (ASSE Y), 9 (ASSE Z) and 10 (ASSE 560), photographed from the full-resolution PDF by Grant. This supersedes everything earlier that was inferred from OCR or from conductor counts.

Sheet index correction¶

wiring-diagram-index.md could not pin the AZIONAMENTO ASSE sheets to page numbers. They are:

Page	Sheet
7	`AZION. ASSE X` - SGDH-15DE
8	`AZION. ASSE Y` - SGDH-10DE
9	`AZION. ASSE Z` - SGDH-10DE
10	`AZION. ASSE 560` - Panasonic MSS083A1XP 750 W
11	`AUX INSERITI` (auxiliaries on)
12	`CIRCUITO AUX`
15	drive-OK relays `KAX`/`KAY`/`KAZ` and brake relay `KAF`
16	`INPUT IF1`
33	I/O rack / module addressing (every module box cross-references `/33.n`)

WARNING: COLLEGAMENTI SERIALI is NOT page 11. An earlier reading of the OCR put it there; page 11 is AUX INSERITI. The OCR-derived page numbers are unreliable and only sheet titles should be trusted. The serial sheet has not yet been located - candidates are 13, 14, 22 and 34-36.

The command interface - identical on X, Y and Z¶

CN1 pin	Drawing label	Yaskawa signal	Wired to
5	`REFERENCE`	`V-REF` analog speed reference	`2X0.50 sch` screened pair
6	`0V`	`SG`	same pair
47	-	`+24V-IN`	24 V rail
40	`START`	`/S-ON` servo on	single wire to cabinet
31	`DRIVER OK`	`ALM+`	to cabinet
32	`DRIVER OK`	`ALM-`	to cabinet
42	`AVANTI`	`P-OT` forward overtravel	limit switch `-SQX+` / `-SQY+` / `-SQZ+`
43	`INDIETRO`	`N-OT` reverse overtravel	limit switch `-SQX-` / `-SQY-` / `-SQZ-`
44	`RESET`	`/ALM-RST`	`-SB15`, yellow pushbutton (sheet 5.3)
27, 28	-	`/TGON+`, `/TGON-`	14A and `-B` terminals

AVANTI / INDIETRO are not run commands. They are the overtravel inhibits. Earlier notes read the OCR as forward/reverse run; that was wrong.

Wire numbers, command side¶

Signal	X	Y	Z
`REFERENCE` / `0V` pair	69, 70	83, 84	97, 98
`START`	71	85	99
`DRIVER OK`	72	86	100
`P-OT` (`AVANTI`)	73	87	101
`N-OT` (`INDIETRO`)	74	88	102
`RESET`	75	89	103

Limit switches sit on terminal rail =BM+X43 (X) and =BM+X4 (Y, Z).

The encoder loom¶

8X0.22 sch - 8 core, 0.22 mm2, screened. Lands on CN1:

CN1 pin	Drawing label	Yaskawa signal
(blank box)	`+`	not a CN1 pin
1	`-`	`SG`
33	`A`	`PAO`
35	`B`	`PBO`
19	`O`	`PCO` index
34	`A-bar`	`/PAO`
36	`B-bar`	`/PBO`
20	`O-bar`	`/PCO`
50	(none)	cable shield

✅ Resolved 2 Sep 2026 from a higher-resolution read of sheets 7-9. The label row on the drawing is offset one box to the left of the pin-number row; reading it that way is electrically unambiguous and identical on all three axes. Pin 50 takes the screen.

Conductor colours named on the sheet: rosso, marrone, blu, giallo, grigio, verde, rosa, bianco - the eight DIN 47100 base colours (white, brown, green, yellow, grey, pink, blue, red).

Wire numbers: X 62-68, Y 76-82, Z 90-96.

Colour to pin, best reading from the page 7 zoom¶

Colour	Wire (X)	Terminal	CN1 pin	Signal
marrone	62	`-X 9`	1	`SG` (shown as `-`)
blu	63	`-X 4`	33	`PAO` (A)
giallo	64	`-X 3`	35	`PBO` (B)
grigio	65	`-X 5`	19	`PCO` (O)
verde	66	`-X 11`	34	`/PAO` (A-bar)
rosa	67	`-X 10`	36	`/PBO` (B-bar)
bianco	68	`-X 12`	20	`/PCO` (O-bar)
rosso	-	-	-	spare, not used
screen	-	-	50	cable shield

Seven of the eight conductors are used. Y uses wires 76-82, Z uses 90-96, same pattern.

✅ CONFIRMED 2 Sep 2026 by the SGM axis connector schedules. Every colour above is correct; the only refinement is that conductor 62/76/90 is red-brown rather than plain brown. Full pin-by-pin table in connector-schedules.md. The cable is SCH 8x0.25, not 8x0.22 as printed on sheet 7.

What this settles¶

The encoder feedback is present, differential, all three phases, from every drive back into the cabinet. Under Option B (analog velocity) Mesa takes it over at the cabinet end. Nothing needs adding at CN1, and the do-not-disturb rule can stand for the encoder pairs.

Earlier notes flip-flopped on this. The sheets settle it.

🔴 The motor encoders are serial - LinuxCNC can never read them directly¶

Sheets 29 and 30 draw the motor encoder connectors from the LATO SALDATURE (solder side). Only four conductors plus shield are used:

Encoder pin	Colour	`2CN` pin	Function
H	rosso	1	`PG5V`
G	blu	2	`PG0V`
C	verde	5	`PS`
D	nero	6	`/PS`
J	calza	shell	shield

PS//PS is Yaskawa's serial encoder channel. The FASE A/B/Z/A-bar/B-bar/Z-bar labels printed against the -ENC.X block are the connector's nominal pinout; only the serial pair is wired.

Consequence for the retrofit: the motor encoder speaks a proprietary serial protocol to the drive. LinuxCNC cannot read it directly under any scheme. Position feedback must come from the drive's divided PAO/PBO/PCO outputs, which the axis loom already carries. This removes any temptation to tap the motor encoder, and it is a further point in favour of Option B.

⚠️ Sheets 30 and 31 have their title blocks swapped. Sheet 30 is titled ENCODER ASSE Z but its header reads AXIS Y DRIVER CONNECTOR with -ENC.Y / -2CN.Y / =BM+X4. Sheet 31 is titled ENCODER ASSE Y but its header reads AXIS Z DRIVER CONNECTOR with -ENC.Z / -2CN.Z / =BM+X5. Sheet 30 is Y, sheet 31 is Z. All three are otherwise identical.

The carousel encoder is different - sheet 32¶

Unlike the servo motors, the carousel encoder is a full differential quadrature device read directly by the Tecnos. Connector -E.561 at =BM+X43 to -SIG2:

`-E.561` pin	Signal	Conductor	`-SIG2` pin
13	`+5Vdc`	nero	3
14	`GND`	viola	4
1	`FASE A`	grigio/rosa	1
2	`FASE A-bar`	blu/rosso	2
3	`FASE B`	bianco	7
4	`FASE B-bar`	marrone	8
5	`FASE Z`	grigio	9
6	`FASE Z-bar`	rosa	10
12	`/RX`	blu	11
11	`RX`	rosso	12
15	`SCH`	shield	17/18/20

Twelve conductors plus shield, in DIN 47100 colours.

🔴 CORRECTED 2 Sep 2026 (Grant). An earlier reading here said "the Tecnos reads carousel position directly". It does not. The device list settles it: -E.561 is part 025AMP0001, "Amphenol, Panasonic encoder", listed immediately after -M560 "Amphenol, Panasonic motor", and sheet 32's header reads CONNETTORE AZIONAMENTO ASSE 561 - drive connector. So -SIG2 is the Panasonic MSS083A1XP's own motor encoder input, closing the indexer's internal loop. It never goes near the Tecnos, which is exactly why it is not one of the four AXIS D-subs on the Monoscheda panel.

The controller's only carousel position sense is binary - COIN plus five P1OUT-P16OUT bits on IF.35-IF.40, through the ordinary I/O modules. The retrofit needs no encoder input for the ATC.

⚠️ The sheet's title block says ENCODER ASSE 560 while its header says CONNETTORE AZIONAMENTO ASSE 561. The Panasonic drive is ASSE 560.

✅ -SIG2 is not one of the four AXIS D-subs on the Monoscheda panel - sheet 35 shows those as X, Y, Z and S. Resolved above: it is on the Panasonic drive, not the Tecnos.

Everything lands on numbered terminals¶

Both cables run from CN1 to a cabinet terminal rail (-X, -Y, -Z) with numbered terminals, and every conductor carries a numbered ferrule. The pin-to-pin record should be built from wire numbers, not colours - more reliable than reading colours off a 24-year-old loom, and it matches the machine's own convention.

Three things not previously recorded anywhere¶

The Z axis has a motor brake, and the circuit is now fully traced.

rail 8A -> KAF contacts (9-10 and 5-6) -> wire 34 / 7 -> -FR.Z brake coil
                                               with -V2 (1N5402) flyback

-KAF's coil is on sheet 15, energised by wire 35 from the Z drive's CN1 pin 27 against rail 8A. Since 8A is gated by CNC OK (OF.1), the brake releases only when the drive asserts and the controller watchdog is healthy. The PLC has nothing to do with it, which is why it appears nowhere in MW310.EQU. It can be left exactly as built through the retrofit.

On X and Y, CN1 pins 27/28 instead tie into the -SB15 / 14A reset circuit. 2. X has a braking resistor, -RX, 75 ohm 1300 W, across B1/B2 on the X drive. 3. The Panasonic carousel is a standalone indexer, not a servo axis. Its I/O is now mapped exactly (see below). Page 10 shows MSS083A1XP with binary position select P1IN-P16IN, SVON, STB, EMG, an ALM output, COIN in-position, P1OUT-P16OUT position feedback, and its own home input ORGL fed from proximity switch -SQ560 via relay -KA560. It homes itself. No analog channel and no LinuxCNC axis is needed for the ATC.

The carousel I/O, mapped point by point¶

Inputs to the Tecnos - six for six against MW310.EQU:

Panasonic terminal	Wire	Sheet.col	Address	`MW310.EQU`
`COIN` (25)	E1	20.3	`IF.35`	`IF_POS_RAGG_CU_6` ✅
`P1OUT` (24)	E2	20.4	`IF.36`	BIT 1 ✅
`P2OUT` (23)	E3	20.5	`IF.37`	BIT 2 ✅
`P4OUT` (17)	E4	20.6	`IF.38`	BIT 3 ✅
`P8OUT` (15)	E5	20.7	`IF.39`	BIT 4 ✅
`P16OUT` (14)	E6	20.8	`IF.40`	BIT 5 ✅

Outputs from the Tecnos:

Wire	Sheet.col	Address	Panasonic terminal	`MW310.EQU` name
A1	28.1	`OF.41`	`EMG` (9)	`OF_ABIL_GIOSTRA_CU_6`
A2	28.2	`OF.42`	`SVON` (12)	`OF_POSIZ_GIOSTRA_CU_6`
A3	28.3	`OF.43`	`STB` (10)	`OF_BIT1_GIOSTRA_CU_6`
A4	28.4	`OF.44`	`P1IN` (35)	`OF_BIT2_...`
A5	28.5	`OF.45`	`P2IN` (34)	`OF_BIT3_...`
A6	28.6	`OF.46`	`P4IN` (33)	`OF_BIT4_...`
A7	28.7	`OF.47`	`P8IN` (32)	`OF_BIT5_...`
-	-	-	`P16IN` (31)	from a `-KA560` contact, wire 252

⚠️ The PLC symbol names do not line up with the hardware on the output side. What MW310.EQU calls BIT1 is physically the strobe, and every position bit shifts by one. P16IN is not driven by an output at all; it comes from -KA560, which is itself driven by the home proximity switch -SQ560. ORGL (13) comes from -KA562 on wire 251.

The drawing is authoritative. Do not build the ATC HAL logic from the symbol names.

Consequences for the retrofit¶

Item	Consequence
Overtravel	Hardwired drive to switch. Survives the Tecnos coming out. Keep it that way.
Alarm reset	Manual pushbutton `-SB15`. No software involvement
Servo enable	One wire per axis, `/S-ON` on pin 40. Mesa drives it, or better, keep it in the hardwired safety chain
Drive fault	✅ No new wiring needed (Grant, 2 Sep). `ALM` already drops `KAX`/`KAY`/`KAZ` in hardware, and the three drive alarms are already among the sources feeding the `IF.8` summary fault, which Mesa reads through the ribbon for free. Per-axis fault off the spare `MY4` poles is an optional diagnostics nicety, not a requirement
Analog command	One screened pair per axis on pins 5/6. Exactly what a 7I77 provides
Encoder	Three differential pairs per axis, already run. 7I77 encoder inputs
Z brake	New requirement. Not in the PLC map. Must be traced and reproduced
ATC	Binary I/O only, through UF6 and IF5. No encoder connection to the Housing at all - the Panasonic keeps its own
Watchdog	🔴 Rail `8A` is gated by `CNC OK` (`OF.1`). It powers the `KAX`/`KAY`/`KAZ` drive-OK relays and the `KAF` brake relay, so a dead controller drops all four. `OF.1` must be a genuine hardware watchdog from Mesa, not a software-set output

sgdh-drive-reference.md - CN1 pinout from the Yaskawa manual
servo-control-mode.md - the Pn000.1 decision
wiring-diagram-index.md - sheet index (page numbers corrected above)
mesa-hardware.md - the three-card chain

What Mesa needs from this loom, and where to get it¶

2 September 2026. Answering directly: does the Yaskawa-to-Tecnos loom carry everything a Mesa 7I77 needs for analog velocity control with encoder feedback, and on which pins at the Tecnos end?

Yes, bar one signal, and that one is better taken elsewhere.

The interception point¶

The AXIS 1 / AXIS 2 / AXIS 3 sockets on the Monoscheda front panel (X, Y, Z respectively; AXIS 4 is the spindle). The loom runs from that 15-pin D-sub straight to CN1 at the drive with no intermediate terminal strip, so this plug is the only clean interception point.

Mate with a standard DA-15 plug and hood - the original parts are Zaniboni 1402005 with 1413014 cover - and land on the 7I77's terminal blocks. The looms are never cut, the drives are never touched, and the Tecnos plugs straight back in.

Pin map for the Mesa build¶

Mesa 7I77 requirement	Pin	Conductor	Wire X	Wire Y	Wire Z	Drive CN1
Analog command +	8	BN	69	83	97	5 `V-REF`
Analog command −	15	WH	70	84	98	6 `SG`
Encoder A	4	BL	63	77	91	33 `PAO`
Encoder /A	11	GN	66	80	94	34 `/PAO`
Encoder B	3	YE	64	78	92	35 `PBO`
Encoder /B	10	PK	67	81	95	36 `/PBO`
Encoder Z (index)	5	GR	65	79	93	19 `PCO`
Encoder /Z	12	WH	68	82	96	20 `/PCO`
Encoder common	9	RD-BN	62	76	90	1 `SG`
Servo enable	13	- (0.5 mm2)	71	85	99	40 `/S-ON`
Home switch (bonus)	6	-	137	138	139	to `BOX`/`BOY`/`BOZ`
Drive 24 V input	7	-	`014`	`014`	`014`	47 `+24V-IN`

Pins 1, 2 and 14 unused.

⚠️ Keep the two grounds separate. Pin 9 (CN1-1) is the encoder common; pin 15 (CN1-6) is the analog reference ground. TEA separated them; so should the Mesa build - encoder common to the 7I77's encoder ground, analog common to its analog ground.

✅ The home switch arrives on the same plug (pin 6). Mesa gets its home input for free, on the same terminal block as everything else for that axis.

The one signal that is not on this connector¶

Drive fault, ALM+/ALM- (CN1 31/32). Wires 72 / 86 / 100 go straight from the drive to the KAX / KAY / KAZ relays in the cabinet, not through the axis loom.

✅ This does not need solving (Grant, 2 Sep). The fault already stops the machine without Mesa's involvement, and Mesa learns a drive faulted for free through the IF.8 summary input on the ribbon. What it does not learn is which drive. The spare MY4 poles below remain available if per-axis diagnostics are ever wanted; they are not part of the build.

That is fine and arguably better, because it keeps the drive-fault chain hardwired (see safety-and-enable-chain.md). To read fault status, Mesa takes a spare contact from those relays:

Relay	Contacts in use	Spare
`KAX` (Omron `MY4`, 4 pole)	9-5 to the fault chain, 10-6 to `KMX`	11-7 and 12-8
`KAY`	9-5, 10-6	11-7 and 12-8
`KAZ`	9-5, 10-6	11-7 and 12-8

Two spare changeover contacts per axis, already in the cabinet on RA78728 sockets. A pair of wires each to a Mesa input.

🔴 The enable is the wrong polarity for a 7I76EU output¶

The SGDH's sequence inputs are sinking: +24V-IN (CN1-47) feeds the internal opto commons and /S-ON (CN1-40) is pulled to 0 V to enable. That is what the Tecnos does today on wire 71 through pin 13.

Mesa's 7I76EU outputs are sourcing, so they cannot drive /S-ON directly.

Two clean fixes, to decide early because it changes which board the enables land on:

Put the three servo enables on the 7I84U, whose outputs are sourcing/sinking selectable. Costs nothing.
An interposing relay per axis. More robust, and it lets the enable sit in the hardware chain alongside OF.1 rather than depending on software.

Cable quality notes¶

Cable	Spec	Comment
Analog pair	`SCH 2x0.50`	Individually screened. Ideal for +/-10 V
Encoder	`SCH 8x0.25`	Single overall screen, not individually screened pairs. Seven conductors used, one spare

The encoder cable has run this machine for 24 years, so it is proven in service. But a single overall screen on differential encoder pairs is the known weak point if noise appears during commissioning, and there is one spare conductor in the cable if a signal needs re-routing.

Summary¶

Requirement	Present in the loom?
Analog velocity command, screened pair	✅ pins 8 and 15
Encoder A, /A, B, /B, Z, /Z, differential	✅ pins 4, 11, 3, 10, 5, 12
Encoder common, separate from analog common	✅ pin 9
Servo enable	✅ pin 13 - but needs a sinking output
Home switch	✅ pin 6, bonus
Drive fault	❌ not on the loom - take a spare `KAX`/`KAY`/`KAZ` contact instead