https://www.perplexity.ai/search/f2645053-8f0b-482b-ae31-6233a6595660
<img src="https://r2cdn.perplexity.ai/pplx-full-logo-primary-dark%402x.png"; style="height:64px;margin-right:32px"/> # I have an intel AGM T60. I think I shorted something. I found the attached documents about the model family. When I attach the 20V supply to the DC-in port (I think J24?), I get power that fluctuates always below 2V at its pins. Teach me what to check next. Your next step is to check the DC-in power path starting at connector J24. On the T60, 20V input from the adapter goes to J24, then immediately through fuse F2 and, if the fuse is good, continues to a group of MOSFETs for power gating and polarity protection like Q78, Q79, Q33, and others.[^1_1] ### Steps for Troubleshooting 1. **Measure Input at J24**: Confirm your power adapter is really outputting 20V with a multimeter before it ever meets the board.[^1_1] 2. **Check the Fuse (F2)**: - Set your meter to continuity mode and test across F2. If the fuse is open (no beep or infinite resistance), it must be replaced.[^1_1] - If it’s good (beep/small resistance), continue. 3. **Check MOSFETs (Q78, Q79, Q33)**: - These handle polarity protection and switching. If any are shorted, damaged, or not being correctly driven, voltage may never make it to the main power rail. - With the adapter plugged in, carefully measure voltage at the MOSFETs' drain and source (big pins)—in general, you should see 20V in, and 20V out (with sometimes a tiny drop or a switching waveform). If you see 20V entering but little to nothing leaving, a MOSFET is blown or not being turned on. If you see no voltage on both sides, the problem is upstream (fuse/open trace/socket issue).[^1_1] 4. **Inspect for Shorts**: On the low side (ground) of these components, or on the 20V rail after the first MOSFET, check for a short to ground. A dead short will load down the input to near 0V, as you described.[^1_1] 5. **Visual Inspection**: Look for damage, burned parts, or corrosion, especially near the input connector and big power transistors.[^1_1] ### Summary Table | Test Area | What You Should See | What to Do if Wrong | | :-- | :-- | :-- | | Adapter output | 20V DC | Replace adapter | | J24 input pins | 20V with adapter connected | Broken socket/trace | | Fuse F2 | Continuity beep, 20V across when powered | Replace if open | | Q78/Q79/Q33 | 20V in and out (each side of MOSFET) | Replace if shorted or not triggering | | 20V rail to GND | High resistance, not a direct short | Locate and repair short | If you have measurements at any step, share them—the next advice will be tailored based on what stage the voltage disappears or what anomalies you find.[^1_1] **Don’t continue feeding power to the board if you find a low resistance (short, a few ohms or less) between DC-in and ground, as this risks further damage**.[^1_1] <div align="center">⁂</div> [^1_1]: Schematic_IBM-T60.pdf --- # I’m not at the board right now, just planning. It definitely reads 20V from the supply when disconnected and under 2V when connected, so I don’t expect those MOSFET readings to look good. Do I just measure conductivity to ground over each component? How do I tell what component it is when they’re wired together (DOCK-PWR20)? I also have a .cst file which I guess includes the wire names, but it’s binary, I’m not sure if you can read it. You are on the right track — measuring resistance or continuity from each major power net (like DOCK-PWR20) to ground is a fundamental method for locating shorts when the board is unpowered.[^2_1] ### How to Check for Shorts 1. **Check Resistance to Ground**: - With the power **disconnected**, measure resistance (ohms mode) between the DOCK-PWR20 rail and ground. - On a good board, this should read high (hundreds of kΩ or open, sometimes tens of kΩ if batteries/capacitors are present). If you read something very low (like <10Ω), you likely have a shorted part somewhere along that net.[^2_1] 2. **Finding Components on Shared Nets**: - On the schematic, DOCK-PWR20 connects many parts: fuse, MOSFETs, diodes, sometimes nearby voltage regulators. - When measuring at any pin labeled as DOCK-PWR20, you’re electrically probing the whole net: each component tied to that net will read the same resistance to ground. - This means you **can’t** electrically distinguish which device is shorted by resistance alone — you are locating which net is shorted, not **which specific part**.[^2_1] - To isolate the bad part, follow the net on the schematic and board. Check for parts that commonly fail (ceramic capacitors, protection MOSFETs, diodes). Sometimes it’s necessary to desolder one leg at a time to ‘divide and conquer’ the short.[^2_1] 3. **Identifying Components**: - Use the schematic: DOCK-PWR20 net will have pointers to each component's value and position. - If you have boardview (.cst) files, they typically require proprietary software to view, but text netlists or schematic cross-references can also help. ### What to Do with a Short - Once you’ve found a low-resistance reading to ground, look up which parts are connected to that net. - Visually inspect each for damage. - If not obvious, selectively lift (desolder) one pin of likely suspects — start with ceramic capacitors, then check power FETs and diodes. - After each removal, repeat the resistance test. Once resistance jumps high, the just-removed part was likely the problem.[^2_1] ### Practical Notes - If any power FET or large multilayer capacitor on the net gets hot quickly with power applied, that’s a clue. - For common net names like DOCK-PWR20, the schematic will show how it branches to subcircuits, allowing you to plan methodical checks.[^2_1] You’re correct: the .cst file is a boardview layout for tracking component placement and net connectivity, but it’s not readable here. Stick to your schematic .pdf and text netlists for tracing.[^2_1] <div align="center">⁂</div> [^2_1]: Schematic_IBM-T60.pdf --- # I’m new to this and my hands shake. I should be aware of any points I might further damage the board, to be appropriately careful or practice or find better tools, first. Is desoldering pins an appropriate first diagnostic step once I see a net is shorted? Most of the parts near 20V are under a thin insulating plastic but the nets can generally be accessed outside it — should I remove this to inspect the components? I’m not certain how to replace it Desoldering is appropriate—but it is not the first thing you do when you discover a shorted net. It is a *last* resort after you confirm the short, visually inspect components, and do not see obvious damage. Beginners with shaky hands should practice on scrap boards, as repeated heating and scraping risks lifting traces or pads, especially on laptop motherboards.[^3_1][^3_2] ### Safety and Precautions - Do not start desoldering until you’re sure which component is suspect, or until you’ve safely ruled out all possibilities. - Each time you heat a joint or pry a part, you risk damaging the board, especially if you slip.[^3_1] - If you are not confident, consider seeking help from someone experienced or invest in a good quality soldering iron, a no-clean flux pen, and a desoldering pump or braid. ### Plastic Film/Insulator Handling - The thin insulating plastic is typically Kapton (polyimide) or sometimes vinyl/PVC; it prevents shorts from the heatsink or shields.[^3_3][^3_4] - You can gently peel it back or tastefully cut it to gain access—do not stretch or force it, as it can rip.[^3_3] - To replace: you can buy Kapton tape online and re-cover the area after the repair, ensuring all 20V components are insulated from potential shorts.[^3_4] - If you remove it, document (photo) the original arrangement for accurate restoration. ### Stepwise Diagnostic Approach - First, check resistance to ground on the 20V nets with the power off. - Visually examine all power components for burn marks or physical damage. - Only desolder if you cannot locate the fault with previous steps—start with likely suspects (tiny ceramic capacitors) rather than hard-to-replace ICs or MOSFETs.[^3_1] ### Summary Table | Step | Risk | Precaution | | :-- | :-- | :-- | | Resistance tests | None | Multimeter probes only | | Visual inspection | None | Good lighting/magnifier | | Lifting plastic | Low (if careful) | Pry gently, document | | Desoldering | Medium-high | Quality tip, flux, steady hands, practice beforehand | If you’re concerned about your dexterity, practice desoldering on unused scrap PCBs first, and don’t rush. Many laptop plastics and traces are delicate, but patience and good technique prevent damage.[^3_2][^3_4][^3_3][^3_1] <span style="display:none">[^3_10][^3_11][^3_12][^3_13][^3_14][^3_15][^3_16][^3_17][^3_18][^3_19][^3_20][^3_5][^3_6][^3_7][^3_8][^3_9]</span> <div align="center">⁂</div> [^3_1]: https://www.instructables.com/The-Ultimate-Guide-to-Desoldering/ [^3_2]: https://www.instructables.com/9-Different-Desoldering-Techniques/ [^3_3]: https://www.eevblog.com/forum/repair/what_s-this-black-material-sticked-over-my-laptop_s-cpu-chip-is-it-necessary/ [^3_4]: https://www.reddit.com/r/overclocking/comments/1duaf1m/help_i_dont_know_what_this_yellow_plastic_sticker/ [^3_5]: https://www.youtube.com/watch?v=bG7yW9FigJA [^3_6]: https://www.youtube.com/watch?v=Vou2xlJkuoU [^3_7]: https://www.reddit.com/r/soldering/comments/1nha5e3/my_first_time_soldering_and_im_having_a_hard_time/ [^3_8]: https://www.pcbbuy.com/news/How-to-Remove-PCB-Conformal-Coating-with-5-Tips.html [^3_9]: https://www.ifixit.com/Guide/How+To+Solder+and+Desolder+Connections/750 [^3_10]: https://forums.tomsguide.com/threads/laptops-are-thin-why-are-power-supplies-still-fat.411220/ [^3_11]: https://www.youtube.com/watch?v=414pbaLK0Os [^3_12]: https://www.reddit.com/r/soldering/comments/1eyqwju/how_do_i_desold_a_component_from_a_pc_motherboard/ [^3_13]: https://www.youtube.com/watch?v=D2zZN1UkOhw [^3_14]: https://www.candlepowerforums.com/threads/removing-pcb-protective-film.86485/ [^3_15]: https://www.facebook.com/groups/tec1z80/posts/2079344075907964/ [^3_16]: https://recoverit.wondershare.com/hard-drive-knowledge/small-cylinder-on-power-supply.html [^3_17]: https://www.reddit.com/r/techsupport/comments/7x5pe5/repairing_my_laptop_removed_some_black_film_what/ [^3_18]: https://www.youtube.com/watch?v=n9OLvsXalGQ [^3_19]: https://www.youtube.com/watch?v=Ki-qw2hIBv8\&vl=en [^3_20]: https://www.eevblog.com/forum/beginners/how-to-remove-solder-that-won_t-melt/ --- # Is there an even safer way to desolder — like hot air and lifting suction or something? If I have already left the board powered with this voltage drop, what is likely to have been further damaged? Hot air rework stations are a safer and more controlled way to desolder surface-mount parts compared to a basic soldering iron, especially for beginners or those with less steady hands. Hot air allows you to evenly heat all pins of a component, making it easier to lift with tweezers or a suction tool without putting mechanical stress on delicate pads. For even better safety:[^4_1][^4_2] - Practice on scrap boards first to get a feel for temperature, distance, and timing. - Use lower airflow and moderate temperatures (start around 330–350°C for laptop PCBs) to avoid lifting nearby components or overheating the board.[^4_2][^4_3] - Use aluminum foil or Kapton tape to protect nearby sensitive parts from excess heat.[^4_1] - Manipulate parts ONLY after the solder is fully molten; never pry or force, as pads on laptop boards are very thin and prone to lifting.[^4_1] Pairing hot air with a manual vacuum “solder sucker” pump helps for through-hole parts, but for SMD work, tweezers or a gentle suction pen are better suited.[^4_4][^4_5] ### Damage from Leaving Board Powered with a Short If the board was powered with a voltage drop (2V present instead of 20V) across a short, risk of further damage depends on what failed and for how long you left it powered. Typical concerns include:[^4_6] - MOSFETs or other power-path components overheating and failing more seriously. - Traces or vias exposed to excess current potentially burning open. - Downstream ICs (controllers, regulators) exposed to undervoltage or excess current may be damaged, especially if short persists for an extended period.[^4_7][^4_6] - Batteries or fuses in the power path may additionally fail. Usually, the root cause part (often a small capacitor, FET, or diode) takes the brunt of the damage first. Repeated powering during a hard short increases risk to nearby parts due to heat spread or electrical stress.[^4_7][^4_6] ### Recommendations - Practice hot air on scrap boards before trying on your laptop.[^4_2][^4_1] - Use tweezers or suction for SMD; don’t force parts off. - If not confident, consider having someone experienced do the initial component removal, as improper technique can make final repair much harder.[^4_2] - Once a board is found shorted, avoid repeated power application until you’ve located and removed the fault.[^4_6] Modern hot air stations, careful handling, and patience make SMD rework much less risky for beginners compared to older methods.[^4_5][^4_1][^4_2] <span style="display:none">[^4_10][^4_11][^4_12][^4_13][^4_14][^4_15][^4_16][^4_17][^4_18][^4_19][^4_20][^4_8][^4_9]</span> <div align="center">⁂</div> [^4_1]: https://www.youtube.com/watch?v=0as0NJpWafc [^4_2]: https://learn.sparkfun.com/tutorials/how-to-use-a-hot-air-rework-station/all [^4_3]: https://www.eevblog.com/forum/beginners/removing-smd-chip-with-hot-air-gun-what-is-a-safe-temperature-to-use/ [^4_4]: https://www.reddit.com/r/AskElectronics/comments/p8qqdd/looking_for_tools_to_aid_me_solderdesolder/ [^4_5]: https://www.eevblog.com/forum/chat/best-way-to-mass-de-solder-electronic-components-(through-hole-and-smd)/ [^4_6]: https://repair.wiki/w/Short_Circuits_-_Repair_Basics [^4_7]: https://www.reddit.com/r/AskElectronics/comments/mh62f6/finding_a_short_circuit_on_a_pc_motherboard/ [^4_8]: https://www.youtube.com/watch?v=H_XOsOlm6_w [^4_9]: https://www.reddit.com/r/soldering/comments/y9hrt7/hot_air_desoldering_throughhole_tutorial/ [^4_10]: https://www.instructables.com/The-Ultimate-Guide-to-Desoldering/ [^4_11]: https://www.diyaudio.com/community/threads/any-tricks-to-desoldering-using-a-cheapo-solder-sucker.167028/ [^4_12]: https://www.youtube.com/watch?v=Vou2xlJkuoU [^4_13]: https://www.reddit.com/r/soldering/comments/16g58cm/desoldering_with_a_solder_pump_do_you_remove_the/ [^4_14]: https://www.youtube.com/watch?v=gGSeSk8QWsE [^4_15]: https://www.youtube.com/watch?v=pZ2ADJGvVYo [^4_16]: https://www.reddit.com/r/buildapc/comments/1c5gtot/does_low_voltage_damage_pc_components/ [^4_17]: https://www.instructables.com/9-Different-Desoldering-Techniques/ [^4_18]: https://forums.tomshardware.com/threads/is-it-safe-to-use-some-parts-of-a-short-circuited-desktop.3733093/ [^4_19]: https://www.reddit.com/r/consolerepair/comments/rbw31t/my_desoldering_station_doesnt_suck_and_im_not/ [^4_20]: https://www.youtube.com/watch?v=0X2C2H_DgqA --- # I engaged another laptop and eventually learned when the power failed, it was always the same burnt shunt diode that needed to be replaced, and I could keep it alive forever through occasional big power anomalies by replacing this diode. Is this likely to have a similar failure mode that would be obvious to someone experienced … given my voltage drop measurement? Yes, it is quite likely that this failure mode has a close analogy on the T60 motherboard. Most laptops, including this model, place a shunt or clamp diode (usually a TVS or Schottky type) across the 20V DC-in rail near the input jack. Its job is to protect against voltage spikes or reversed adapter polarity: if something abnormal happens, the diode conducts heavily, protecting downstream circuitry but usually sacrificing itself in the process.[^5_1] Given your voltage drop symptom (adapter drops to under 2V when connected), an experienced technician would immediately suspect this input protection diode, a power MOSFET, or a shorted ceramic capacitor on that same high-current rail—just as you've seen in other laptops. On T60 schematics, components like D13, D106, or D45 (diodes marked at the input/power path) would be obvious first suspects; if the diode fails short, it will clamp the supply and cause exactly this kind of voltage sag.[^5_1] This is a very common scenario and is usually visually obvious: the diode might be cracked or show discoloration, though sometimes it fails silently. Experienced repairers will often check and temporarily desolder or clip one end of this input protection diode as a diagnostic step, then see if the short is cleared and the board powers up normally (and then replace it with a suitable part of the same specs).[^5_1] In summary, your symptom matches a classic "blown clamp diode" or similar passive-component input failure, and any experienced laptop repair technician would check those high-current input diodes first, based on your measurements.[^5_1] <div align="center">⁂</div> [^5_1]: Schematic_IBM-T60.pdf --- # You mentioned “clamp” — are you able to find RV7 in the .lst.txt file and explain the fields to me? Here is the entry for RV7 found in your .lst.txt board netlist: ``` RV7 NOASM UCLAMP3301D 1 3244 6568 3160 6568 84 3202 6568 L SOT2121601 H1.28 UCLAMP3301D-NOASM 73 ``` ### Meaning of the Fields - **RV7**: Reference Designator (this is the ID for finding the part in schematic and on the board) - **NOASM**: Not always assembled (sometimes on boards this means it is an optional or alternate part) - **UCLAMP3301D**: Part number/type — this is a Littelfuse uClamp3301D, which is a TVS clamp diode, made for ESD and surge protection - **1**: Quantity (possibly how many on the BOM, or single instance) - **3244 6568 3160 6568 84 3202 6568**: These are XY board coordinates for placing/locating the part on an assembly diagram or for automated pick-and-place machines - **L**: Layer — this typically refers to the side of the board (L usually means “Left”, but in board terms more often “Top” or “Bottom”, sometimes by convention) - **SOT2121601**: Package footprint (SOT-23 package, specifics here; "SOT2121601" is a manufacturer or designator code) - **H1.28**: Height in mm - **UCLAMP3301D-NOASM**: Alternative name (alternate part/assembly info) - **73**: Possibly sheet number or sorting code for design databases ### Application - The UCLAMP3301D is absolutely a clamp (shunt) protection device, meant to suppress voltage spikes and clamp the high-voltage net (like 20V) to a safe range, sacrificing itself if the spike is excessive. In a failure mode, it will usually short, causing exactly the "voltage drop" at input you described. An experienced engineer would check this part immediately on a board with your symptoms. <span style="display:none">[^6_1]</span> <div align="center">⁂</div> [^6_1]: T60-VGA-ATI.LST.txt --- # It looks like RV7 connects to a different DC-IN pin than the 20V source though, I think? I don’t understand quite why I would see a voltage drop on DOCK-PWR20 — shouldn’t F2 blow and prevent current from flowing? Could it be the power adapter that can’t provide current when loaded? Am I missing something? Your reasoning is solid, and the questions you raise reflect important diagnostic nuances. Here’s how the power entry works, why you’d see a voltage drop, and answers to your fuse and adapter concerns based on the T60 schematic:[^7_1] ### Power Path and Short Symptoms - The **input adapter (20V)** feeds the DC-in jack (J24), then goes through **fuse F2** before reaching the main power rail (DOCK-PWR20).[^7_1] - **RV7** is a surge clamp (TVS) to ground — its job is to "short" ANY excessive surge or reversed polarity to ground, but it connects before the fuse, *not always directly on the main DOCK-PWR20 rail*.[^7_1] - If there's a direct short *after* F2 (e.g., a failed ceramic cap or power MOSFET), you will see a voltage drop at the DC-in pins, but F2 often **does not blow immediately**. Fuses are slow: if the current is just above their rating but not a dead short, they may "hang on" for seconds or minutes, especially if you cycle power. - If there is still continuity (low resistance) across F2 and you read 2V at DOCK-PWR20 with a 20V adapter, there is a **significant load (short circuit) downstream**, pulling voltage down but not blowing the fuse outright.[^7_1] ### Why the Adapter May Sag - Most adapters have internal foldback or protection: instead of providing massive current into a dead short, they limit output. This causes the *output voltage to drop* well below nominal (e.g., your 2V reading), but doesn’t necessarily clear or "pop" F2.[^7_1] - It is possible, especially with modern adapters, that the current never exceeded F2’s trip point due to the adapter’s own protection, explaining why the fuse remains intact even during a short event. ### What You Might Be Missing - The most likely causes of voltage collapse without a blown fuse: - Shorted component on the DOCK-PWR20 net (such as a failed input capacitor, power FET, or a clamp diode — but not necessarily RV7, since that’s likely pre-fuse) - Adapter limiting current, not the fuse (this is very common) - If the short is upstream of F2 (rare, unless the socket itself is damaged), F2 would see enough current to blow ### Diagnostic Tips - Measure continuity across F2 to verify if it’s truly intact; sometimes a “blown” fuse will show only a few ohms continuity from carbonization. - Confirm voltage at each point: before F2, after F2, and directly on DOCK-PWR20. - If you see a drop to 2V with the supply connected but F2 is still fine, the **adapter is protecting itself from a hard load/short, not a gentle overcurrent**. ### Summary Table | Point | Voltage reading (ideal) | Voltage reading (short) | Diagnosis | | :-- | :-- | :-- | :-- | | Adapter unloaded | 20V | 20V | Adapter OK | | Before F2, loaded | 20V | 2V | Short after F2 or fuse blown/open | | After F2, loaded | 20V | 2V or 0V | If 2V, heavy short; if 0V, open fuse | | DOCK-PWR20, loaded | 20V | 2V or near 0V | Heavy short after F2 | So, you are not missing anything, but this is why a voltage drop can occur even with an unblown fuse, and why RV7 may not be to blame unless the short is truly at the entry (pre-fuse).[^7_1] <div align="center">⁂</div> [^7_1]: Schematic_IBM-T60.pdf --- # Some of your analysis is inaccurate — DOCK_PWR20 is before the fuse, DOCK_PWR20_F is after the fuse. RV7 connects ACDC_ID to GND. What is the format of the data you would need to assimilate that information automatically? Could I generate that format with openboardview? Is it simple enough to patch some code into openboardview that would spit it out? To automatically and accurately assimilate power path information (like what nets components connect to, before/after fuse, and their roles), the most useful machine-readable format is a **netlist** that includes: - Reference designator (e.g., F2, RV7) - Part type or value (e.g., "FUSE", "TVS Diode", resistance, voltage rating, etc.) - Pin numbers and the net each pin connects to (e.g., F2 pin 1: DOCK_PWR20, F2 pin 2: DOCK_PWR20_F) - Net names - Optionally: coordinates, package, and layer, for physical cross-reference A sample CSV or JSON representation might look like this: ```csv RefDes, Type, Value, Pin, Net F2, Fuse, 5A, 1, DOCK_PWR20 F2, Fuse, 5A, 2, DOCK_PWR20_F RV7, TVS, UCLAMP3301D, 1, ACDC_ID RV7, TVS, UCLAMP3301D, 2, GND ``` or in JSON: ```json [ {"refdes":"F2", "type":"Fuse", "value":"5A", "pins":{"1":"DOCK_PWR20", "2":"DOCK_PWR20_F"}}, {"refdes":"RV7", "type":"TVS", "value":"UCLAMP3301D", "pins":{"1":"ACDC_ID", "2":"GND"}} ] ``` ### Can You Get This from OpenBoardView? OpenBoardView by default displays physical layout and net connections, but it often works from proprietary `.brd` or `.bvr` formats. The OpenBoardView project and related tools don't natively export this kind of full pin-to-net mapping in a simple CSV/JSON format, but the underlying parsed data contains it. - **Yes, you can patch OpenBoardView**: The internal data structure of OpenBoardView includes component reference, pin numbers, and nets. Modifying it to export as CSV or JSON as above is entirely achievable and has even been requested on forums.[^8_1] - **Community precedents:** People in the repair community have written scripts/tools to export netlists or cross-reference tables from BoardView files for exactly your use-case.[^8_2][^8_1] ### How Simple? For a coder familiar with C++ or Python (depending on what version/branch you use), patching OpenBoardView to output a component-pin-net CSV would be a straightforward task. You'd walk the component list, print refdes and for each pin, print the net it connects to. There may also be standalone boardview parsing Python libraries available to do a similar job for batch formats.[^8_1][^8_2] In summary: - A flat CSV or JSON pin-net mapping enables automatic, accurate analysis. - You can generate this with a small patch to OpenBoardView or by exporting from some boardview processing scripts. - Community DIY repairers frequently perform such custom exports for their workflow.[^8_2][^8_1] <span style="display:none">[^8_10][^8_11][^8_12][^8_13][^8_14][^8_15][^8_16][^8_17][^8_18][^8_19][^8_20][^8_3][^8_4][^8_5][^8_6][^8_7][^8_8][^8_9]</span> <div align="center">⁂</div> [^8_1]: https://www.eevblog.com/forum/eda/seeking-pads-pcb-file-format-specification-to-write-decoder-for-openboardview/ [^8_2]: https://usoldering.com [^8_3]: https://forum.kicad.info/t/faq-how-to-export-to-openboardview/4349 [^8_4]: https://github.com/google/pcbdl/issues/26 [^8_5]: https://kibot.readthedocs.io/en/v1.8.2/configuration/outputs/boardview.html [^8_6]: https://www.quadcept.com/en/manual/schematic/post-107 [^8_7]: https://www.youtube.com/watch?v=g7ImUPXxgI8 [^8_8]: https://desktop.arcgis.com/en/arcmap/latest/extensions/schematics/automating-schematic-diagram-content-by-specifying-schematic-rules.htm [^8_9]: https://www.youtube.com/watch?v=6bZyT7AYTBU [^8_10]: https://camptechii.com/mastering-net-classes-in-altium-designer-a-comprehensive-guide-for-2024/ [^8_11]: https://openboardview.org [^8_12]: https://resources.altium.com/p/understanding-your-pcb-netlist [^8_13]: https://resources.altium.com/p/managing-net-connection-order-schematic-layout [^8_14]: https://github.com/OpenBoardView/OpenBoardView [^8_15]: https://github.com/OpenBoardView/OpenBoardView/issues/187 [^8_16]: https://pmc.ncbi.nlm.nih.gov/articles/PMC11103103/ [^8_17]: https://www.youtube.com/watch?v=rDL0Ya47oz0 [^8_18]: https://techwaveav.alboompro.com/post/261548-reverse-engineering-circuits-with-free-schematic-diagram-tools-is-it-practical [^8_19]: https://www.sciencedirect.com/science/article/pii/S2773126X25000243 [^8_20]: https://kibot.readthedocs.io/en/v1.6.5/configuration/outputs/boardview.html --- # Can you link me to the specification of a netlist format for a popular open source circuit simulator that could be used to simulate damaged circuits on guix if it imported a netlist I generated? The most widely used open source circuit simulator for scientific/educational purposes (and packaged for Guix) is **ngspice**, which uses the SPICE netlist format. This format is well-documented, human-readable, and can be created from a simple script or export tool. ### SPICE Netlist Format - **Official ngspice manual:** http://ngspice.sourceforge.net/docs/ngspice-manual.pdf - **Quick reference on SPICE netlist syntax:** https://web.stanford.edu/class/ee101a/supplements/SPICE_quick_reference.pdf[^9_1] - **Example and explanation of netlist format (edu):** https://ecircuitcenter.com/SpiceTopics/Netlists/Netlist.htm[^9_2] https://ngspice.sourceforge.io/docs/[^9_3][^9_4] You can use these resources to generate valid netlists for simulation. ### SPICE Netlist Structure Example Each SPICE netlist line generally looks like: ``` R1 1 2 1k ; a resistor named R1 between net/node 1 and 2, with 1kΩ C2 2 0 10u ; a capacitor C2 between node 2 and ground (node 0), 10µF D5 3 0 1N4148 ; a diode D5 between node 3 and ground, using the 1N4148 model ``` Dot commands like `.tran`, `.dc`, `.ac`, and `.end` specify simulation types and define the netlist boundaries. ### Compatibility - **ngspice** and **Xyce** are both free, open, SPICE-compatible simulators supported on Linux and Guix environments.[^9_5] - Many open-source schematic tools (like KiCad) and board parsers can export netlists in exactly this format. If you export your netlist as plain text following the SPICE format, you can directly simulate the effects of component damage (e.g., a shorted or open device) in ngspice on Guix.[^9_6][^9_7][^9_4] <span style="display:none">[^9_10][^9_11][^9_12][^9_13][^9_14][^9_15][^9_16][^9_17][^9_18][^9_19][^9_20][^9_8][^9_9]</span> <div align="center">⁂</div> [^9_1]: https://web.stanford.edu/class/ee133/handouts/general/spice_ref.pdf [^9_2]: https://www.ecircuitcenter.com/Basics.htm [^9_3]: https://ngspice.sourceforge.io [^9_4]: https://ngspice.sourceforge.io/ngspice-control-language-tutorial.html [^9_5]: https://xyce.sandia.gov [^9_6]: https://qucs-s-help.readthedocs.io/en/legacy/BasSim.html [^9_7]: https://vlsiarchitect.com/ng-spice-open-source-simulator-tool/ [^9_8]: https://www.altium.com/documentation/altium-designer/circuit-simulation/working-with-spice-netlist [^9_9]: https://help.simetrix.co.uk/8.4/simetrix/simulator_reference/topics/runningthesimulator_netlistformat.htm [^9_10]: http://ebook.pldworld.com/_semiconductors/ti/pspice%20for%20ti/doc/cap_ug/SPICE_netlist_format.html [^9_11]: https://simplis.com/documentation/simetrix/simulator_reference/topics/runningthesimulator_netlistformat.htm [^9_12]: https://www.youtube.com/watch?v=dZHyg8ravBE [^9_13]: https://product.tdk.com/system/files/dam/tvcl/an-nl06b001_en.pdf [^9_14]: https://ngspice.sourceforge.io/ngspice-tutorial.html [^9_15]: https://ra3xdh.github.io [^9_16]: http://ahkab.readthedocs.io/en/latest/help/Netlist-Syntax.html [^9_17]: https://openmodelica.org/forum/default-topic/1763-reading-a-ngspice-netlist.html [^9_18]: https://en.wikipedia.org/wiki/List_of_free_electronics_circuit_simulators [^9_19]: https://gab.wallawalla.edu/~larry.aamodt/engr434/eldosim.pdf [^9_20]: https://sourceforge.net/p/ngspice/discussion/133842/thread/d64d06e452/ --- # Is this format appropriate for preserving net and component names and part models in conversations with you, or would we need something with another field or such? The SPICE netlist format preserves basic connectivity (which component connects which nodes), reference designators, and model names, but it is not designed to retain descriptive **net names** (like DOCK_PWR20) or full component model information beyond what's specified in the model library or comments. By default, SPICE uses node numbers, not human-readable net names.[^10_1][^10_2] ### SPICE Format Details - **Reference designator and model:** Each component line typically uses the refdes (e.g., R1, C2) and model name (e.g., 1N4148 for diodes). - **Node numbers:** Connections are typically by node number, not by net name, unless comments are added. - **Part values:** Specified inline (resistance, capacitance, model for diode/transistor). - **Net names:** Usually lost unless you overlay manual comments or preprocess to translate names to nodes. ### For Conversations and Collaboration If your goal is: - **Simulation only**: SPICE is sufficient. - **Rich, traceable troubleshooting or schematic sharing (with humans or AI)**: Augment the netlist with a CSV/JSON table or inline comments that explicitly **map node numbers to net names** and components to part models and footprints. Example formats: ```csv RefDes,Type,Model,Pin,Net,Value,Footprint F2,Fuse,,1,DOCK_PWR20,5A,0603 F2,Fuse,,2,DOCK_PWR20_F,5A,0603 ``` Or in SPICE with comments: ``` * Node assignments: 1=DOCK_PWR20 2=DOCK_PWR20_F F2 1 2 5A ``` - For machine or AI communication, **CSV/JSON with fields for:** RefDes, Type, Net Names, Pin mapping, Model, Value, Footprint is best for preserving and referencing all relevant details. ### Summary - Use SPICE for simulation, but consider CSV/JSON (or a hybrid annotated netlist) to keep rich, human/AI-friendly context like net names and additional part info. - This approach ensures lossless, understandable collaboration across troubleshooting, simulation, and conversation threads.[^10_2][^10_1] <div align="center">⁂</div> [^10_1]: https://www.ecircuitcenter.com/Basics.htm [^10_2]: https://web.stanford.edu/class/ee133/handouts/general/spice_ref.pdf
![https://r2cdn.perplexity.ai/pplx-full-logo-primary-dark%402x.png"](https://r2cdn.perplexity.ai/pplx-full-logo-primary-dark%402x.png"){kind=link}