How to Reconstruct a Diagram From the Equipment

Why this matters

When nothing is published, or when what is published no longer matches what is installed, you can build the drawing yourself from the hardware. This is not a last resort, it is often better than the factory drawing, because a reconstructed diagram is a record of the machine in front of you including every field modification somebody made and never wrote down. The work takes an hour or two the first time. Skipping it means every future diagnosis on that machine starts from zero, and the shop pays that hour again on every call forever.

Step 1: Make it dead, prove it, then release stored energy

Open the disconnect, apply your own lock and tag, and prove the circuit dead: test the meter on a known live source, test the conductors you are about to touch, then test the meter on the known live source again. That sequence is NFPA 70E-2021, 120.5, and it exists because a meter that failed silently between the two tests is the one way a correct procedure still kills someone. The de-energize and lockout requirement for work on or near exposed energized parts is 29 CFR 1910.333(b)(2).

Dead is not the same as safe. Before your hands go in:

• Capacitors hold a charge after the supply is gone. Discharge them through an appropriate resistive load, then verify zero volts across the terminals with your meter rather than assuming the bleed resistor worked.
• Springs, counterweights, and raised components hold mechanical energy. Block or restrain them.
• Pressurized and refrigerant-bearing lines hold energy that does not care whether the power is off. Isolate and relieve where the work requires opening them.

Those three are stored-energy hazards under 29 CFR 1910.147, which is a different standard from the electrical one above and applies alongside it on the same machine.

If you skip this step you will still get a drawing. You will just get it with a meter lead on a live terminal in a cabinet you are also reaching into with the other hand.

Step 2: Photograph before you disturb anything

Wide shot of the open cabinet, then a straight-on shot of each device group, then close shots of every terminal block with the wire markers legible. Put something in frame for scale on at least one shot.

This costs two minutes and it is the only copy you will ever have of the original state. The moment you start pulling wires to ring them out, the as-found condition is gone, and half of what makes a reconstruction valuable is knowing what was there before you touched it. It is also the record that protects you when a customer later claims the machine worked differently before you arrived.

Step 3: Tag devices before you touch a wire

Give every device in the circuit a unique identifier and write it on the device with a marker or a label. Not the wires yet, the devices.

Use a scheme that says what the thing is: a prefix by function, then a number. Transformer, switch, operating control, limit, relay, contactor, motor. The scheme does not have to match any standard, it has to be unambiguous and consistent inside your drawing.

This ordering is deliberate. A wire's identity is which two devices it connects, so wire numbering is meaningless until the devices have names. Techs who start by numbering wires end up with a list of numbers and no map, and they redo the whole thing.

Step 4: Find the two rails and draw the power path first

Every control circuit has a source side and a return side, and every branch hangs between them. Find them and draw them as two vertical lines with everything else between.

Cross-trade, the rails are whatever the equivalent pair is: hot and neutral or two lines on an electrical circuit, supply and return on a hydronic loop, upstream and downstream of the shutoff on a gas train, high side and low side on a refrigeration circuit.

Draw the power path before the control path. Source, disconnecting means, overcurrent protection, switching device, load. That is the skeleton, it is usually four or five elements, and having it on paper gives every control element you find afterwards somewhere to attach.

Step 5: Ring out point to point and mark both ends

With the circuit dead and stored energy released, use a meter on continuity and work one conductor at a time. Put one lead on a known terminal, find the other end, and the instant you find it, write the same wire number on both ends with a marker or a wire tag.

Two rules make the difference between a reconstruction that closes and one you abandon halfway.

• Mark both ends immediately. A wire you identified but did not label is a wire you will identify again in twenty minutes.
• Record the element state as you go. When you ring across a switch or a limit, write down the resistance you measured and whether the device was open or closed at that moment, and note the condition it was in. Those readings are the raw data for the verification step, and going back for them means opening the cabinet again.

For non-electrical systems the equivalent is the same discipline with a different instrument: follow one pipe or one duct branch at a time, tag both ends, and record what valve or damper position it was found in.

Step 6: Redraw it in logical order, not physical order

You now have a physical map: which terminal connects to which terminal. That is a wiring diagram, and it is what you need to put a meter on a specific wire. It is close to useless for understanding what the machine does.

Redraw the same information in logical order, source rail at the left or top, return rail at the other side, each branch running through its switching elements in series to its load. That is a schematic, and it is what lets you say out loud what has to be true for a given load to energize. The library has a card on the distinction; the point here is that you produce both from the same ring-out, and the second drawing costs ten minutes once the first exists.

Step 7: Predict, then test the prediction

A reconstruction is not finished when it is drawn. It is finished when it has made a correct prediction.

Pick a load. Read your own schematic and state the full condition for it to energize: every series element that has to be closed, in order. Then verify that the machine agrees, by measurement rather than by hope. A drawing that has never been tested against the machine is a hypothesis.

Worked: the contradiction that found a device nobody drew

A packaged unit's control circuit, reconstructed after the panel door diagram was found to be for a different configuration. Six devices tagged in the control section, eleven conductors identified, and the series string as drawn ran: transformer secondary, service switch, operating control, high limit, low pressure switch, relay coil, back to transformer common. Four switching elements between the transformer and the coil.

De-energized, with the panel proved dead and the capacitor verified at zero volts, the tech ringed each element and recorded both the reading and the state:

• Service switch, closed, 0.2 ohms
• Operating control, closed at the time of test because the space was calling, 0.2 ohms
• High limit, closed at ambient, 0.3 ohms
• Low pressure switch, closed, 0.2 ohms

All four closed. The sum of those four readings is 0.9 ohms. Then he ringed the whole string end to end, transformer secondary terminal to relay coil terminal, and got an open circuit.

Those two results cannot both be true for the circuit as drawn. Four closed elements totalling 0.9 ohms in series cannot present as open. Either a reading was wrong or the drawing was missing something between two of the nodes.

He re-ringed segment by segment, terminal to terminal rather than element to element, and found the open in the run between the low pressure switch output and the relay coil, which his drawing showed as a single unbroken conductor. Following that conductor out of the cabinet, it entered a flexible whip and terminated at a flow switch mounted remotely on a pipe, several feet away, out of sight from the panel and absent from every drawing on the machine.

So the string had five switching elements, not four. The fifth had been added in the field, correctly wired and never documented, and it was open because the condition it senses was not satisfied.

Two things fell out of that. The immediate diagnosis was no longer a control problem at all, it was a flow problem, which is a different system and a different repair. And the reconstruction had done the job a factory drawing could never do, because the factory drawing would have shown four elements and been correct about the machine as built and wrong about the machine as installed.

The failure mode of stopping at the ring-out is a tech who trusts the four element readings, concludes the string is good, and condemns the relay or the board. He replaces it, the string is still open at the flow switch, and the new part changes nothing.

How to verify the drawing is right

1. Sum the series. End-to-end resistance through a string of closed elements should be close to the sum of the individual readings, allowing for lead and connection resistance. A large gap means an element you have not drawn.
2. Trace one branch backwards. Start at a load and read your way back to the source rail. If you cannot get there on the drawing, the drawing has a hole.
3. Count both ways. Count the devices on the drawing against the devices you physically tagged, and count the conductors on the drawing against the conductors in the panel. Both counts must match. A conductor with no home on the drawing is the most common signal of an undocumented modification.
4. Make one prediction and test it. Name a condition that should prevent a load from energizing, create that condition through the normal operating means, and confirm the load does not energize. Do not create it by defeating a safety device; a prediction proved by jumpering a limit has proved nothing except that you jumpered a limit.
5. Date it and note what you could not resolve. A drawing that says two conductors were traced to a junction that could not be opened is honest and useful. One that quietly omits them will send the next tech in a circle.

References

• NFPA 70E-2021, 120.5, verification of an electrically safe work condition including the test-before-and-after instrument check
• 29 CFR 1910.333(b)(2), de-energizing and lockout or tagging before work on or near exposed energized parts
• 29 CFR 1910.147, control of hazardous energy, for stored mechanical and pressure energy released before the work begins
• See related: The Difference Between a Schematic and a Wiring Diagram; How to Trace a Circuit on a Schematic; Discovering an Undocumented Modification Mid-Diagnosis