On day 22 we mapped the SMD pick-and-place line for SMEs. On day 23 we contrasted hand assembly vs machine productivity. On day 24 we covered MOQ, lot changes, and flexibility. Today we do not rewrite anatomy, throughput, or MOQ: we focus on quality, process consistency, SMD yield, and rework reduction on small and medium runs.
The usual shop-floor question is: “Why was last week’s lot clean and this week a festival of bridges?” The answer is almost never magic. It is first-article discipline, stencil/paste control, feeder polarity, a stable thermal profile , and a shared defect vocabulary. Useful anchors: pick and place hub · custom electronics · Downloads · shop · blog.
What yield (and consistency) mean in an SME SMD shop
Yield is the share of boards that pass the first time—without heroic rework—against the released lot. Process consistency means the same revision, frozen BOM, and setup behave the same on Tuesday as on Thursday. You do not need a global factory dashboard; you need not to depend on “the operator who remembers.”
- High yield with a loose process: luck from one run. It does not scale.
- Mediocre yield with a documented process: painful, but improvable (you know where to look).
- High yield with a consistent process: the real target for small and medium series.
Rework is not “alternative quality.” It is a hidden cost that eats the margin of a 50 / 100 / 200 board lot. If day 24 taught you to respect the changeover toll, today asks you to respect the toll of not improvising quality.
First article: the board that protects the rest of the lot
The first article is the board—or panel—you validate before releasing volume. It is not bureaucracy; it is insurance for the other 49 / 99 / 199. Without it, every systematic defect multiplies at pick-and-place speed.
| First-article check | What you validate | If you skip it |
|---|---|---|
| References vs BOM | Value and footprint match the frozen revision | You run a whole lot with a “close enough” resistor |
| Polarity / pin 1 | Diodes, LEDs, ICs, connectors oriented | Massive rework or silent scrap |
| Paste on critical pads | Reasonable volume; no obvious bridges on fine-pitch | Blame reflow when the problem started at the stencil |
| Post-reflow solder | Wetting, opens, tombstoning at extremes | “We’ll catch it at functional”—expensive and late |
| Release signature | Someone names “OK for lot” | Nobody owns it; the defect belongs to everyone and no one |
On small lots first article “hurts more” because “there are so few boards.” That is exactly why it matters: you already paid the setup fixed cost; multiplying a systematic error is paying that setup twice.
A practical SMD defect taxonomy (no failure lab required)
You do not need a full IPC dictionary to start. You do need shared names on the floor so “looks weird” stops being the official diagnosis.
| Defect | Typical signals | First places to look |
|---|---|---|
| Bridging | Solder joining neighboring pads; shorts on fine-pitch | Too much paste; stencil aperture; pad spacing; print pressure |
| Tombstoning | Passive standing on one pad | Paste or heat imbalance between pads; pad design; aggressive profile |
| Opens / cold solder | Pad without wetting; intermittent continuity | Too little paste; oxidation; short profile; high placement |
| Wrong part | Value or footprint that does not match the BOM | Misloaded feeder; lot mix-up; no clear-down between runs |
| Reversed polarity | Diode/LED/IC backwards | Tape load; program; skipped first article |
| Shift / rotation | Part off-pad or rotated | Vision/offsets; nozzle; paste that “drags” |
Consistency wins when the team names the defect before “touching up with solder and hoping.” Heroic rework hides symptoms; taxonomy points at process cause.
Stencil and paste: where yield is born (or dies)
Much of what you later fight in reflow or functional testing was born at print. Minimum discipline:
- Stencil tied to revision: if Gerber changed, the previous stencil is suspect until proven otherwise (echo of day 24).
- Cleaning and wear: a clogged aperture or tired foil changes paste volume without the placer being at fault.
- Repeatable print parameters: pressure, speed, separation—documented, not “by eye for this shift.”
- Paste inspection on critical pads: fine-pitch, QFN, BGA-lite: look before placing, not only after the oven.
If your line includes the ZB3040H stencil printer in the Zheng Bang ecosystem ElectroTAS represents in Argentina, the point is not the model name: it is using it with the same discipline you demand from pick and place. Inconsistent paste + perfect placement still yields mediocre results.
Feeder polarity: the ghost of the reversed part
Pick and place is excellent at repeating a load error. That is why polarity and clear-down between lots are quality controls, not “warehouse details”:
| Control | How it protects yield |
|---|---|
| Pin-1 / cathode mark at load | Avoids reversing a whole column of diodes or an entire IC |
| Clear-down checklist between designs | Stops the leftover part from the previous lot (mix-up) |
| BOM check on first article | Catches wrong value before volume |
| Named owner for feeder load | Without an owner, the error is anonymous and repeats |
On day 24 , clear-down showed up as a changeover toll. Today it is also a yieldtoll: one misloaded feeder can collapse the percentage for the whole lot.
Thermal profile consistency: the oven is not “set 250 and go”
Reflow (in the published ecosystem, ZB5040HL is the oven reference on ElectroTAS hub/Downloads) needs a stable profile for your paste + parts + board thermal mass. Thermal consistency is not an improvised max temperature.
- Same paste, same documented profile for that revision—not “the other product’s profile because it looked similar.”
- Thermal mass: large copper planes and heavy connectors change behavior; first article validates the oven too.
- Do not blame reflow first: massive bridging: check stencil/paste; systematic tombstoning: look at pad balance and profile together.
- Paste or supplier changes: treat as a mini process change; revalidate, do not assume.
No invented curves or times: the operating message is repeatability. A profile that “worked once” and nobody wrote down is not a process; it is an anecdote.
AOI and visual checks: where to look without theater
AOI (Automated Optical Inspection) and visual review do not replace first article or paste discipline. They amplify detection when volume or pitch justify it. Practical SME framing:
| Level | When it is enough | Honest limit |
|---|---|---|
| Critical visual post-reflow | Small lots; known critical pads; first article + sampling | Human fatigue; cannot see under every package |
| AOI / process vision | More volume or finer pitch; you want detection consistency | You must teach defect vs false positive |
| Functional / ICT later | Closes the electrical loop | If you arrive here with systematic bridging, you already overpaid |
Zheng Bang models ElectroTAS publishes (ZB3245T · ZB3245TS with 2 cameras · ZB3245TSS with 4 cameras) use vision in placement; that is not the same as post-solder AOI, but it reinforces the idea: eyes (human or camera) at the right moment, not only at the end when the lot is already done.
- 1. Freeze revision, BOM, and stencil
- Without a freeze there is no measurable yield—only a prototype dressed as a series.
- 2. First article with a signature
- References, polarity, critical paste, solder. Someone says “OK for lot.”
- 3. Feeder load with an owner
- Polarity, value, clear-down of the previous lot. Write who loaded.
- 4. Paste/stencil in repeatable mode
- Parameters and cleaning; not “a bit more pressure because today felt light.”
- 5. Thermal profile documented for that paste/board
- Revalidate if paste, mass, or revision changed. Do not copy the neighbor’s profile.
- 6. Name defects; do not only “touch up”
- Bridge, tombstone, open, wrong part, polarity: name + likely cause before heroic solder.
Two ElectroTAS doors for quality and yield
Same series framing, applied to consistency:
| Door | When to look | What it means for yield / rework | Link |
|---|---|---|---|
| A. Zheng Bang equipment | You want your own SMD capacity and internal process ownership | Pick and place + stencil + reflow; official Argentina representative; certified gear; ElectroTAS warranty and manufacturer-backed support. You build yield with discipline | Pick and place hub |
| B. Custom development + production | You need small/medium series on a running process | You use line and first-article / control practices as a service (Planning · Design · Prototype · Production); yield is discussed as a deliverable, not a surprise | Custom electronics |
Contact: info@electrotas.com · desarrollos@electrotas.com · +54 (011) 3221-4474 · WhatsApp +54 9 11 2245-2000 · Martínez, San Isidro, Buenos Aires.
Zheng Bang family: brief context (not fine selection)
Confirmed features only from hub / Downloads. No CPH or prices. Fine T vs TS vs TSS choice stays for a later post in the series.
| Model | Profile | Notes for consistency (qualitative) |
|---|---|---|
| ZB3245T | Entry / prototypes and small runs | Useful to leave tweezers; yield still depends on first article and stencil |
| ZB3245TS | Mid | 27 feeders; 2 cameras; HIWIN; servo — more vision and mix help repeat placement; they do not replace correct polarity |
| ZB3245TSS | More feeders / more cameras | 54 feeders; 4 cameras — more mix capacity; load and clear-down discipline weighs more, not less |
| ZB3040H | Stencil | Where much bridging/opens is decided; print discipline |
| ZB5040HL | Reflow | Stable documented profile; not improvised “max mode” |
Manuals and support material: Downloads · support post manuals, drivers, and firmware.
Yield and rework myths that burn lots
| Myth | Operating reality |
|---|---|
| “If the machine places well, yield takes care of itself” | Paste, polarity, and profile also place (badly) at scale |
| “On a 50-board lot you can skip first article” | On 50, a systematic error can wipe almost the whole lot |
| “Solder touch-up is quality” | It is containment. Quality is not needing a hero every run |
| “We’ll catch the defect at functional” | Late functional = expensive scrap + muddy diagnosis |
| “AOI fixes loose processes” | AOI detects; it does not invent a clean stencil or a frozen BOM |
| “Yesterday’s profile is close enough” | Without a document there is no consistency—only fragile memory |
Signals your “quality” is really chronic rework
- The same bridge or tombstone appears run after run and nobody touches stencil or profile.
- There is no first-article signature; “we started because we were in a hurry.”
- Feeders load without clear-down and “mystery” wrong parts appear.
- The oven is tuned by ear every shift.
- There is no defect log—only a soldering iron that is always hot.
- You celebrate a clean run without knowing what you did differently (so you cannot repeat it).
What this post is not
- Not a deep dive on lead time or hidden costs from prototype to short series (that is the day-26 teaser).
- Not fine Zheng Bang model choice T vs TS vs TSS by volume (day 27).
- Not the e2e path spec to board + assembly (day 28) nor firmware + PCB + assembly (day 29).
- Not the SEO hub / product map (day 30).
- Does not rewrite day 22anatomy, day 23 productivity, or day 24 MOQ—only cross-links them.
- Does not invent prices, CPH, reflow curves, or magic yield percentages.
Frequently asked questions
1. What is SMD yield in an SME (without giant-factory jargon)?
How many boards in the lot pass first time without depending on heroic rework. It is built with process consistency (first article, stencil, polarity, profile), not luck from one run.
2. Is first article worth it if I only run 50 boards?
Yes. On 50, a systematic defect can wipe almost the whole lot. First article protects the setup you already paid when loading feeders, program, and stencil.
3. Are bridges always the reflow oven’s fault?
No. Look at paste/stencil first (volume, aperture, pressure). The oven can worsen things, but many bridges were born at print.
4. Does AOI replace visual inspection and first article?
No. AOI (or process vision) amplifies detection. First article and paste/polarity discipline remain the base. Without criteria, AOI mostly creates false positives.
5. How does this relate to day-24 MOQ?
MOQ and changeover define the setup toll. Yield defines whether that setup produces good boards. A 100-board lot with 30% rework was not “flexible”—it was expensive.
6. Does ElectroTAS only sell machines, or also produce?
Two doors: A Zheng Bang equipment ( pick and place hub) with official Argentina representation, certification, ElectroTAS warranty, and manufacturer-backed support; B custom electronics and made-to-order production (typical service MOQ 50/100/200).
7. Where do I download manuals and equipment materials?
At Downloads (ZB3245T / TS / TSS / ZB5040HL / ZB3040H). Also the post on manuals, drivers, and firmware.
8. What comes on day 26?
Lead time and hidden costs from prototype to short series—no URL yet. Today we close quality/consistency/yield; next, the calendar and what the quote does not show.
Close with consistency and look to day 26
Less rework is not a slogan: it is a signed first article, repeatable stencil/paste, owned feeders, a documented thermal profile, and named defects. Process consistency beats the heroic soldering iron on small and medium series. If you are building your own capacity, start from the pick and place hub and Downloads. If you prefer made-to-order production, see custom electronics.
Next in the series (day 26): lead time and hidden costs from prototype to short series —no link yet. Meanwhile, revisit the day-22 line map, the day-23 tweezers/machine contrast, and day-24 MOQ: quality is the layer that makes those three worth the effort.
ElectroTAS · Martínez, San Isidro, Argentina · info@electrotas.com · desarrollos@electrotas.com · +54 (011) 3221-4474 · WhatsApp +54 9 11 2245-2000 · shop · blog.
On day 22 we mapped the SMD pick-and-place line for SMEs. On day 23 we contrasted hand assembly vs machine productivity. On day 24 we covered MOQ, lot changes, and flexibility. Today we do not rewrite anatomy, throughput, or MOQ: we focus on quality, process consistency, SMD yield, and rework reduction on small and medium runs.
The usual shop-floor question is: “Why was last week’s lot clean and this week a festival of bridges?” The answer is almost never magic. It is first-article discipline, stencil/paste control, feeder polarity, a stable thermal profile , and a shared defect vocabulary. Useful anchors: pick and place hub · custom electronics · Downloads · shop · blog.
What yield (and consistency) mean in an SME SMD shop
Yield is the share of boards that pass the first time—without heroic rework—against the released lot. Process consistency means the same revision, frozen BOM, and setup behave the same on Tuesday as on Thursday. You do not need a global factory dashboard; you need not to depend on “the operator who remembers.”
- High yield with a loose process: luck from one run. It does not scale.
- Mediocre yield with a documented process: painful, but improvable (you know where to look).
- High yield with a consistent process: the real target for small and medium series.
Rework is not “alternative quality.” It is a hidden cost that eats the margin of a 50 / 100 / 200 board lot. If day 24 taught you to respect the changeover toll, today asks you to respect the toll of not improvising quality.
First article: the board that protects the rest of the lot
The first article is the board—or panel—you validate before releasing volume. It is not bureaucracy; it is insurance for the other 49 / 99 / 199. Without it, every systematic defect multiplies at pick-and-place speed.
| First-article check | What you validate | If you skip it |
|---|---|---|
| References vs BOM | Value and footprint match the frozen revision | You run a whole lot with a “close enough” resistor |
| Polarity / pin 1 | Diodes, LEDs, ICs, connectors oriented | Massive rework or silent scrap |
| Paste on critical pads | Reasonable volume; no obvious bridges on fine-pitch | Blame reflow when the problem started at the stencil |
| Post-reflow solder | Wetting, opens, tombstoning at extremes | “We’ll catch it at functional”—expensive and late |
| Release signature | Someone names “OK for lot” | Nobody owns it; the defect belongs to everyone and no one |
On small lots first article “hurts more” because “there are so few boards.” That is exactly why it matters: you already paid the setup fixed cost; multiplying a systematic error is paying that setup twice.
A practical SMD defect taxonomy (no failure lab required)
You do not need a full IPC dictionary to start. You do need shared names on the floor so “looks weird” stops being the official diagnosis.
| Defect | Typical signals | First places to look |
|---|---|---|
| Bridging | Solder joining neighboring pads; shorts on fine-pitch | Too much paste; stencil aperture; pad spacing; print pressure |
| Tombstoning | Passive standing on one pad | Paste or heat imbalance between pads; pad design; aggressive profile |
| Opens / cold solder | Pad without wetting; intermittent continuity | Too little paste; oxidation; short profile; high placement |
| Wrong part | Value or footprint that does not match the BOM | Misloaded feeder; lot mix-up; no clear-down between runs |
| Reversed polarity | Diode/LED/IC backwards | Tape load; program; skipped first article |
| Shift / rotation | Part off-pad or rotated | Vision/offsets; nozzle; paste that “drags” |
Consistency wins when the team names the defect before “touching up with solder and hoping.” Heroic rework hides symptoms; taxonomy points at process cause.
Stencil and paste: where yield is born (or dies)
Much of what you later fight in reflow or functional testing was born at print. Minimum discipline:
- Stencil tied to revision: if Gerber changed, the previous stencil is suspect until proven otherwise (echo of day 24).
- Cleaning and wear: a clogged aperture or tired foil changes paste volume without the placer being at fault.
- Repeatable print parameters: pressure, speed, separation—documented, not “by eye for this shift.”
- Paste inspection on critical pads: fine-pitch, QFN, BGA-lite: look before placing, not only after the oven.
If your line includes the ZB3040H stencil printer in the Zheng Bang ecosystem ElectroTAS represents in Argentina, the point is not the model name: it is using it with the same discipline you demand from pick and place. Inconsistent paste + perfect placement still yields mediocre results.
Feeder polarity: the ghost of the reversed part
Pick and place is excellent at repeating a load error. That is why polarity and clear-down between lots are quality controls, not “warehouse details”:
| Control | How it protects yield |
|---|---|
| Pin-1 / cathode mark at load | Avoids reversing a whole column of diodes or an entire IC |
| Clear-down checklist between designs | Stops the leftover part from the previous lot (mix-up) |
| BOM check on first article | Catches wrong value before volume |
| Named owner for feeder load | Without an owner, the error is anonymous and repeats |
On day 24 , clear-down showed up as a changeover toll. Today it is also a yieldtoll: one misloaded feeder can collapse the percentage for the whole lot.
Thermal profile consistency: the oven is not “set 250 and go”
Reflow (in the published ecosystem, ZB5040HL is the oven reference on ElectroTAS hub/Downloads) needs a stable profile for your paste + parts + board thermal mass. Thermal consistency is not an improvised max temperature.
- Same paste, same documented profile for that revision—not “the other product’s profile because it looked similar.”
- Thermal mass: large copper planes and heavy connectors change behavior; first article validates the oven too.
- Do not blame reflow first: massive bridging: check stencil/paste; systematic tombstoning: look at pad balance and profile together.
- Paste or supplier changes: treat as a mini process change; revalidate, do not assume.
No invented curves or times: the operating message is repeatability. A profile that “worked once” and nobody wrote down is not a process; it is an anecdote.
AOI and visual checks: where to look without theater
AOI (Automated Optical Inspection) and visual review do not replace first article or paste discipline. They amplify detection when volume or pitch justify it. Practical SME framing:
| Level | When it is enough | Honest limit |
|---|---|---|
| Critical visual post-reflow | Small lots; known critical pads; first article + sampling | Human fatigue; cannot see under every package |
| AOI / process vision | More volume or finer pitch; you want detection consistency | You must teach defect vs false positive |
| Functional / ICT later | Closes the electrical loop | If you arrive here with systematic bridging, you already overpaid |
Zheng Bang models ElectroTAS publishes (ZB3245T · ZB3245TS with 2 cameras · ZB3245TSS with 4 cameras) use vision in placement; that is not the same as post-solder AOI, but it reinforces the idea: eyes (human or camera) at the right moment, not only at the end when the lot is already done.
- 1. Freeze revision, BOM, and stencil
- Without a freeze there is no measurable yield—only a prototype dressed as a series.
- 2. First article with a signature
- References, polarity, critical paste, solder. Someone says “OK for lot.”
- 3. Feeder load with an owner
- Polarity, value, clear-down of the previous lot. Write who loaded.
- 4. Paste/stencil in repeatable mode
- Parameters and cleaning; not “a bit more pressure because today felt light.”
- 5. Thermal profile documented for that paste/board
- Revalidate if paste, mass, or revision changed. Do not copy the neighbor’s profile.
- 6. Name defects; do not only “touch up”
- Bridge, tombstone, open, wrong part, polarity: name + likely cause before heroic solder.
Two ElectroTAS doors for quality and yield
Same series framing, applied to consistency:
| Door | When to look | What it means for yield / rework | Link |
|---|---|---|---|
| A. Zheng Bang equipment | You want your own SMD capacity and internal process ownership | Pick and place + stencil + reflow; official Argentina representative; certified gear; ElectroTAS warranty and manufacturer-backed support. You build yield with discipline | Pick and place hub |
| B. Custom development + production | You need small/medium series on a running process | You use line and first-article / control practices as a service (Planning · Design · Prototype · Production); yield is discussed as a deliverable, not a surprise | Custom electronics |
Contact: info@electrotas.com · desarrollos@electrotas.com · +54 (011) 3221-4474 · WhatsApp +54 9 11 2245-2000 · Martínez, San Isidro, Buenos Aires.
Zheng Bang family: brief context (not fine selection)
Confirmed features only from hub / Downloads. No CPH or prices. Fine T vs TS vs TSS choice stays for a later post in the series.
| Model | Profile | Notes for consistency (qualitative) |
|---|---|---|
| ZB3245T | Entry / prototypes and small runs | Useful to leave tweezers; yield still depends on first article and stencil |
| ZB3245TS | Mid | 27 feeders; 2 cameras; HIWIN; servo — more vision and mix help repeat placement; they do not replace correct polarity |
| ZB3245TSS | More feeders / more cameras | 54 feeders; 4 cameras — more mix capacity; load and clear-down discipline weighs more, not less |
| ZB3040H | Stencil | Where much bridging/opens is decided; print discipline |
| ZB5040HL | Reflow | Stable documented profile; not improvised “max mode” |
Manuals and support material: Downloads · support post manuals, drivers, and firmware.
Yield and rework myths that burn lots
| Myth | Operating reality |
|---|---|
| “If the machine places well, yield takes care of itself” | Paste, polarity, and profile also place (badly) at scale |
| “On a 50-board lot you can skip first article” | On 50, a systematic error can wipe almost the whole lot |
| “Solder touch-up is quality” | It is containment. Quality is not needing a hero every run |
| “We’ll catch the defect at functional” | Late functional = expensive scrap + muddy diagnosis |
| “AOI fixes loose processes” | AOI detects; it does not invent a clean stencil or a frozen BOM |
| “Yesterday’s profile is close enough” | Without a document there is no consistency—only fragile memory |
Signals your “quality” is really chronic rework
- The same bridge or tombstone appears run after run and nobody touches stencil or profile.
- There is no first-article signature; “we started because we were in a hurry.”
- Feeders load without clear-down and “mystery” wrong parts appear.
- The oven is tuned by ear every shift.
- There is no defect log—only a soldering iron that is always hot.
- You celebrate a clean run without knowing what you did differently (so you cannot repeat it).
What this post is not
- Not a deep dive on lead time or hidden costs from prototype to short series (that is the day-26 teaser).
- Not fine Zheng Bang model choice T vs TS vs TSS by volume (day 27).
- Not the e2e path spec to board + assembly (day 28) nor firmware + PCB + assembly (day 29).
- Not the SEO hub / product map (day 30).
- Does not rewrite day 22anatomy, day 23 productivity, or day 24 MOQ—only cross-links them.
- Does not invent prices, CPH, reflow curves, or magic yield percentages.
Frequently asked questions
1. What is SMD yield in an SME (without giant-factory jargon)?
How many boards in the lot pass first time without depending on heroic rework. It is built with process consistency (first article, stencil, polarity, profile), not luck from one run.
2. Is first article worth it if I only run 50 boards?
Yes. On 50, a systematic defect can wipe almost the whole lot. First article protects the setup you already paid when loading feeders, program, and stencil.
3. Are bridges always the reflow oven’s fault?
No. Look at paste/stencil first (volume, aperture, pressure). The oven can worsen things, but many bridges were born at print.
4. Does AOI replace visual inspection and first article?
No. AOI (or process vision) amplifies detection. First article and paste/polarity discipline remain the base. Without criteria, AOI mostly creates false positives.
5. How does this relate to day-24 MOQ?
MOQ and changeover define the setup toll. Yield defines whether that setup produces good boards. A 100-board lot with 30% rework was not “flexible”—it was expensive.
6. Does ElectroTAS only sell machines, or also produce?
Two doors: A Zheng Bang equipment ( pick and place hub) with official Argentina representation, certification, ElectroTAS warranty, and manufacturer-backed support; B custom electronics and made-to-order production (typical service MOQ 50/100/200).
7. Where do I download manuals and equipment materials?
At Downloads (ZB3245T / TS / TSS / ZB5040HL / ZB3040H). Also the post on manuals, drivers, and firmware.
8. What comes on day 26?
Lead time and hidden costs from prototype to short series—no URL yet. Today we close quality/consistency/yield; next, the calendar and what the quote does not show.
Close with consistency and look to day 26
Less rework is not a slogan: it is a signed first article, repeatable stencil/paste, owned feeders, a documented thermal profile, and named defects. Process consistency beats the heroic soldering iron on small and medium series. If you are building your own capacity, start from the pick and place hub and Downloads. If you prefer made-to-order production, see custom electronics.
Next in the series (day 26): lead time and hidden costs from prototype to short series —no link yet. Meanwhile, revisit the day-22 line map, the day-23 tweezers/machine contrast, and day-24 MOQ: quality is the layer that makes those three worth the effort.
ElectroTAS · Martínez, San Isidro, Argentina · info@electrotas.com · desarrollos@electrotas.com · +54 (011) 3221-4474 · WhatsApp +54 9 11 2245-2000 · shop · blog.
Start writing here...