200922422 九、發明說明 【發明所屬之技術領域】 本發明係關於焊球印刷裝置，特別是關於將焊球印刷 於基板上之焊球印刷裝置。 【先前技術】 已知的印刷法，爲了以1 8 0〜1 5 0 V m的節距（p i t c h )來形成焊球（直徑80〜1 〇〇 # m )，是用公知之高精度 網版印刷裝置印刷無鉛焊料後，進行熔焊而形成焊球。網 版印刷裝置例如是具備：基板搬入輸送器、基板搬出輸送 器、具備昇降機構之印刷台、開口部具有轉印圖案之遮罩 、刮刀、具備刮刀昇降機構及水平方向移動機構之刮刀裝 置、用來控制該等機構之控制裝置。 在從基板搬入輸送器將基板搬入裝置內部後，將基板 暫時定位固定於印刷台，接著將基板和具有開口部（對應 於電路圖案）的遮罩（網版）雙方的標記用攝影機辨識， 修正雙方的偏差量以將基板對準於網版後，讓印刷台上昇 以使基板接觸網版，在網版和基板接觸的狀態下藉由刮刀 將無鉛焊料等的糊劑充塡於網版的開口部，接著讓印刷台 下降，使基板和網版分離（網版分離）以將糊劑轉印於基 板上，然後將基板從裝置搬出，藉此完成印刷。 另一種方法是焊球轉移法，係將焊球轉移到以高精度 穿設有微細孔之治具上而使其以既定節距整齊排列，然後 直接移載到基板上，接著進行熔焊。 -5- 200922422 又專利文獻1揭示出，讓遮罩擺動或振動以將焊球充 塡於既定開口的方法、藉由刷子的倂進運動進行充塡後實 施加熱之方法等等。又專利文獻2揭示出，將焊球裝載於 托盤上，用管子吸附來再度充塡於電極墊之方法。 〔專利文獻1〕日本特開2000-49 1 8 3號公報 〔專利文獻2〕日本特開2003 -3 09 1 3 9號公報 【發明內容】 使用無鉛焊料之印刷法，其設備成本低且能一次形成 大量的凸塊（bump )，因此具有高產能、低製造成本的優 點。然而，由於印刷法難以確保轉印體積的均一性，因此 必須進行平坦化處理（flattering )，以將熔焊後的焊料凸 塊擠壓而使高度變得平坦化，因此存在著步驟數變多 '設 備成本增加的問題。又隨著元件的高密度化，當進展到節 距1 5 0〜1 2 0 # m之微細化的情形，印刷良率低而造成生產 性變差。 另一方面，焊球轉移法雖能確保焊球的分級精度而形 成具有穩定高度的凸塊，但必須藉由高精度的焊球吸附治 具並用機器人來批式充塡焊球，在微細化的情形，存在著 產距時間（tact )增加、治具設備價格高昂所造成之凸塊 形成成本增加等的問題。 此外，專利文獻1之讓網版擺動或振動以將焊球充塡 於既定開口的方法，隨著焊球粒徑變小，會發生粒子間的 凡得瓦力所造成之密合現象、靜電所造成之吸附現象’而 -6 - 200922422 存在著無法充塡於遮罩開口部的問題。又同樣地，利用刮 刀和刷子的倂進運動等來進行充塡時，也存在著同樣的問 題。 又在專利文獻2的方法，即使可以進行修補（repair )，但殘存助焊劑量變少的可能性極大，在批式溶焊時若 焊料之附著性（wetting )變差的情形，即使焊球熔融，仍 可能發生附著不良而造成電極墊部之焊接不完全。 本發明的目的是爲了提供一種焊球充塡用印刷裝置及 凸塊形成方法，在形成超微細節距的凸塊時，能像印刷法 般批式形成大量的凸塊，且能像焊球轉移法般形成穩定高 度的凸塊，又能進行低成本、高速、高效率的印刷暨充塡 而具有高生產性。 爲了達成上述目的，本發明之焊球印刷裝置，係具備 :在基板的電極墊上印刷助焊劑之助焊劑印刷部、對印刷 前述助焊劑後的電極上供應焊球之焊球充塡暨印刷部、檢 查印刷焊球後之基板狀態且對應於不良狀態來進行修補之 檢查暨修補部，而構成焊球印刷裝置；其特徵在於： 前述助焊劑印刷部係具備：對應於複數個電極墊的位 置而設置開口之網版、用來裝載並固定基板之載台、爲了 進行基板和網版的對準而對定位標記進行拍攝之定位用攝 影機、用來觀測印刷後的網版開口部的狀況之檢查用攝影 機 '用來將網版開口部的堵塞或網版背面之助焊劑附著污 染予以清掃除去之清掃手段、將檢查用攝影機所拍攝的影 像和預先記錄的基準模型的畫面做比較以判斷印刷不良之 200922422 判定手段'根據判定手段的判定結果來決定是將基板送往 下個步驟或從生產線排除，在排除的情形會發生網版的清 掃指示之印刷不良發生手段。 又在設置於焊球充塡暨印刷部的印刷頭設置充塡單元 ，該充塡單元係具備：由框體、蓋、篩狀體所構成之焊球 盒、以和篩狀體隔著間隔的方式設於焊球盒的下部之狹縫 狀體；透過支承構件在蓋設置加振手段，該加振手段係用 來對篩狀體施加振動而使設於篩狀體的開口大小形成可變 以將焊球供應至狹縫狀體。 依據本發明，針對焊球充塡不良的重大原因之助焊劑 印刷不良，在前頭步驟就進行提早處理，因此能提昇生產 性。 又依據本發明，可提昇焊球充塡效率以縮短產距時間 並進行高充塡率之焊球充塡暨印刷，因此能提昇生產性。 又從助焊劑印刷〜焊球充塡〜檢查、修補爲止’可提 昇各裝置的運轉效率，可縮短產距時間，因此能以批式、 低成本、高速的方式穩定形成大量的焊料凸塊高度精度佳 之焊料凸塊。又裝置之構造簡單，而能降低設備成本。 【實施方式】 以下參照圖式來說明本發明的印刷裝置及凸塊形成方 法之較佳實施形態。 第1圖係顯示助焊劑印刷部及焊球充塡暨印刷部之印 刷步驟的槪要。第1 ( a )圖係顯示助焊劑印刷步驟’第1 200922422 (b )圖係顯示焊球充塡暨印刷的狀態。 第1(a)圖，係在設有開口部（對應於預先5又疋匕基 板21之電極墊22的形狀）的網版20上’裝載助焊劑’ 藉由移動刮刀3來將既定量的助焊劑2 3印刷於基板21的 電極墊22上。 在本實施例，網版2 0是助焊劑印刷用的網版’爲了 保障高精度的圖案位置精度，係使用加成法所製成之金屬 網版。刮刀3，是使用方刮刀、劍刮刀、平刮刀中之任一 者。按照助焊劑2 3的黏度、觸變性來設定刮刀間隙、印 壓、刮刀速度，以進行印刷動作。 若助焊劑23的印刷量過少，在充塡焊球24時’焊球 無法附著在電極墊22上。又在焊球印刷後之後步驟的熔 焊時，會構成焊料附著不良的主要原因，無法形成形狀美 觀的焊料凸塊，也會構成焊料凸塊高度不良或焊料連接強 度不足的主要原因。 又若助焊劑23的量過多，在焊球充塡、印刷時，在 用來將焊球24供應到電極墊22上之網版20上所設置之 開口部等’可能會附著助焊劑2 3。若在網版開口部附著助 焊劑23，焊球24會附著於網版的開口，而造成無法轉印 在電極墊22上的問題。因此助焊劑印刷，就焊球充塡品 質而言乃影響最大的步驟。 接著，如第】（b)圖所示，在具備充塡單元60 (參 照第7圖）之焊球充塡暨印刷部，將焊球24充塡、印刷 於印刷助焊劑23後的基板2丨的電極墊22上。用來將焊 200922422 球24充塡於電極墊22之網版20b，爲了保障高精度 案位置精度’係使用加成法所製作出之金屬網版° 該焊球充塡用的網版2 0 b的材質，係使用可和裝載 板的磁鐵載台（印刷台1 〇 )進行磁性吸引之磁性體材料 以使基板2 1和網版20的間隙成爲〇 ’如此可避免焊球 鑽入基板2 1和網版20之間而發生剩餘焊球不良。 再者，在網版2 0 b的背面（和基板21接觸側）， 有樹脂製或金屬製之微小支柱20a，以在和剛印刷好助 劑23的基板21密合時，避免助焊劑23滲透到網版開 部周圍。藉此構成助焊劑2 3滲透液之釋放部。又，從 在起將網版20b和支柱20a的組合稱爲焊球印刷用的網 20 ° 爲了高精度地將焊球24供應至既定位置的電極墊 ，在基板2 1的四個角落設有定位標記（未圖示）。對 於設在基板2 1側之定位標記，在網版20側也設有定位 記。藉由CCD攝影機1 5 (參照第4圖）來視覺辨識該 定位標記，以網版20側的定位標記和基板2 1側的定位 記一致的方式實施高精度的對準。在本實施例的對準， 使裝載基板2 1之印刷台1 〇沿水平方向移動來進行。 當對準完成後，縮小基板21和網版2 0的間隔，使 版20接觸基板21後，讓充塡單元60動作，將焊球24 網版20的開口部供應至印刷助焊劑23後的基板21面 的電極墊22。在焊球供應用的充塡單元60 (參照第7 )的下部側’設有狹縫狀體6 3，藉由使充塡單元6 0進 圖 基 24 設 焊 □ 現 版 22 ffM 標 等 標 係 網 從 上 圖 行 -10- 200922422 擺動、前進動作，使焊球24滾動、振動，以將焊均 塡至網版開口部。 第2圖係顯示焊球印刷裝置的一實施例。本圖 ，係將助焊劑印刷部、焊球充塡暨印刷部、檢查暨 都設成一體的裝置。各部位也能以單獨裝置的方式 。在本裝置，首先在助焊劑印刷部（網版印刷方式 助焊劑23印刷於基板上的各電極墊22°藉由搬送 (從助焊劑印刷部側看爲搬出輸送器’從焊球充塡 部側看爲搬入輸送器）’在焊球充塡暨印刷部將焊 至電極墊（隔著助焊劑）。 助焊劑印刷部和焊球充塡暨印刷部之主要不同 刷頭部，助焊劑印刷部是採用刮刀構造’焊球充塡 部是由用來供應焊球之充塡單元所構成。檢查暨修 印刷頭部，是採用分配器型之吸引供應頭構造。在 修補部，由於不須使用網版，故並未設置網版安裝 框支座等。 第3圖係顯示本實施例的凸塊形成之流程圖。 搬入（S T E P 1 )後，在電極墊上印刷既定量的助 STEP2 )。接著，檢查印刷助焊劑後之網版的開口 STEP3 )。檢查結果爲NG時，將基板搬出至NG 存部，用版下清掃裝置45自動實施清掃（STEP4 ) ，視需要來補充供應助焊劑。 N G基板，係和不實施焊球印刷以後的步驟之 號一起在後步驟的輸送器上待機並排出至生產線外 $ 24充 的裝置 修補部 來構成 )，將 輸送器 暨印刷 球供應 點在印 暨印刷 補部之 檢查暨 用的版 在基板 焊劑（ 狀況（ 基板儲 。然後 NG訊 。也能 -11 - 200922422 使用生產線上之N G基板儲存器等，用基板匣一倂排出。 N G基板在生產線外的步驟洗淨後，可再度使用於助焊劑 印刷。 接著，實施焊球充塡、印刷（STEP5 )。在焊球充塡 、印刷後，在進行網版分離後，從網版上方檢查網版開口 內之焊球充塡狀況（STEP6 )。檢查結果存在充塡不足部 位的情形，在進行網版分離前再度實施焊球充塡、印刷動 作（STEP7 )。藉此，可提昇焊球充塡率。 若STEP6結果OK，則實施網版分離（STEP8 )。接 著，在焊球充塡後，用檢查暨修補裝置來檢查充塡狀況（ STEP9 )。充塡狀況檢查NG的情形，在供應助焊劑後， 對NG點之電極墊部再度供應焊球（STEP 1 0 )。充塡狀況 檢查爲OK的情形，用熔焊裝置將焊球再度熔融，藉此完 成焊料凸塊。 第4圖係顯示本發明的網版印刷裝置（主要爲助焊劑 印刷部）的槪略構造。第4 ( a )圖係從網版印刷裝置的正 面觀察的構造，第4 ( b )圖係顯示系統構造圖。又第5 ( a )( b )圖係網版印刷裝置的說明圖。 本體框1上設有未圖示的版框支座，在版框支座安裝 遮罩（將開口部設有印刷圖案之網版20張設於版框20c ( 參照第6圖）而構成）。在本圖，在網版2 0的上方，配 置著設有刮刀3的印刷頭2。 在助焊劑印刷部的情形，是在印刷頭2裝設聚胺酯製 的刮刀3。在焊球充塡暨印刷部的情形，在印刷頭2，是 -12- 200922422 取代刮刀3而裝設狹縫狀體63等所構成之充塡單元60。 印刷頭2，係藉由印刷頭移動機構6來沿水平方向移動’ 藉由印刷頭昇降機構4來上下移動。在將刮刀3置換成充 塡單元60時，充塡單元60是藉由印刷頭昇降機構4進行 上下移動。 在網版20的下方，以和網版20相對向的方式’設有 印刷台1 〇 (用來裝載並保持作爲印刷對象物之基板2 1 ) 。該印刷台1 〇係具備：使基板21沿水平方向（XY 0方 向）移動以和網版20進行對準之ΧΥ β載台1 1 ;從搬入 輸送器25接收基板21，且讓基板21接近或接觸網版20 面之載台昇降機構12。 設置在印刷台1 〇上面的基板接收輸送器26，可在印 刷台10上接收藉由基板搬入輸送器25搬入之基板21，又 能在印刷終了時將基板2 1排出至基扳搬出輸送器27。 網版印刷裝置係具備將網版20和基板2 1自動對準的 功能。亦即，藉由CCD攝影機15來拍攝分別設於網版20 和基板2 1上的對準用標記，經影像處理求出位置偏差量 ，以校正該偏差量的方式驅動ΧΥ 6»載台1 1而進行對準。 此外’在印刷機本體框的內部設置印刷機控制部3 0， 其具備：網版分離控制部3 9、各部的驅動控制部等所構成 之印刷控制部3 6 '用來處理來自c C D攝影機1 5的影像訊 號之影像輸入部3 7。又在印刷機的外側設置：用來改寫控 制用資料或變更印刷條件等之資料輸入部5 〇、用來監視印 刷狀況等和所取得的辨識標記之顯示部40。 -13- 200922422 印刷機控制部3 0係具備用來控制充塡單元6 0之印刷 控制部3 6，依照所生產的凸塊節距、焊球粒徑及所使用的 金屬遮罩種類，可簡單地選定適當的充塡印刷模式。 又具備：按照輸入影像來計算相關値之相關値計算部 3 1、根據所取得的影像和資料庫3 8的資料來求出形狀之 形狀推定部32、求出位置座標位置之位置座標運算部3 3 、尺寸計算部34 ;根據CCD攝影機15所拍攝的資料，按 照設於基板2 1和網版2 0之位置辨識標記來求出位置偏差 量。根據ΧΥ 0台控制部的指令來驅動XY e台11以進行 對準。 接著以焊球充塡暨印刷部爲例來說明印刷裝置的動作 。形成有焊料凸塊的基板21，藉由基板搬入輸送器25供 應給基板接收輸送器26。當基板2 1搬送至印刷台1 0的位 置時，印刷台1 〇會上昇，而從基板接收輸送器2 6將基板 21交接給印刷台1 〇上。交接給印刷台1 〇之基板21，係 固定於印刷台1 〇的既定位置。將基板2 1固定後，將CCD 攝影機1 5移動至預先設定之基板標記位置。其狀況顯示 於第5 ( a )圖。 接著，CCD攝影機15拍攝設於基板21及網版20之 位置辨識用標記（未圖示），傳送至印刷機控制部3 0。在 印刷機控制部內之影像輸入部3 7，根據影像資料來求出網 版20和基板21的位置偏差量，印刷機控制部3 0是根據 其結果來作動ΧΥ Θ控制部3 5以移動印刷台1 〇 ,藉此修 正基板2 1的位置而使其對準網版2 0。 -14 - 200922422 對準動作完成後的狀況顯示於第5 ( b )圖。首先， C CD攝影機1 5退避既定量至不會和印刷台1 〇干涉的位置 。當CCD攝影機15退避完成後，印刷台1〇會上昇，而 使基板2 1和遮罩20接觸。在該狀態下，讓印刷頭昇降機 構4動作，讓刮刀（圖中是顯示刮刀3，但在焊球充塡步 驟是成爲充塡單元60前端之狹縫狀體63)接觸網版面。 接著’邊將狹縫狀體63施加振動而使其擺動，邊旋轉驅 動印刷頭驅動用的馬達2g而使狹縫狀體63在網版面上水 平移動’藉此將焊球2 4從狹縫狀體6 3的開口透過設於網 版面的開口來充塡於基板2 1的電極墊22部。 印刷頭2，係沿水平方向移動一定距離的衝程後上昇 。接著，印刷台1 0下降，網版20和基板2 1分離，充塡 於網版2 0的開口部之焊球2 4轉印於基板21。然後，印刷 焊球24後之基板2 1經由基板搬出輸送器27送到下個步 驟。 又如前述般，在基板21和網版20上，是以相對向且 在同一部位的方式設置2個以上辨識對準用標記。藉由具 有上下方向2視野之特殊的CCD攝影機1 5，從下方辨識 網版2 0的標記，從上方辨識基板21的標記，讀取設於既 定部位的標記之位置座標，運算並修正基板21相對於網 版20之偏差量，藉此將基板21對準網版20。 第6圖係顯示印刷助焊劑後的網版之開口狀態。第6 (a )圖係顯示網版全體的形狀，第6 ( b )圖係顯示設置 1個電極群後之開口狀況，第6 ( C )係顯示印刷助焊劑後 -15- 200922422 之開口部的狀況。藉由設定成適當的網版間隙（網版和基 板的間隔）、印壓（刮刀在網版上的擠壓力）、刮刀速度 ，能將助焊劑2 3充分地充塡於網版2 0的開口 2 0 k ,在通 過刮刀3的同時使基板21和網版20分離，藉此能確實地 將助焊劑23轉印至基板21之電極墊22部。又網版20是 固定於版框2 0 c。 受到網版印刷用的助焊劑2 3的黏度、觸變性、網版 20的開口 20k 口徑微細等的影響，印刷後之網版20的開 口 2 0k的狀況，在正常的印刷狀態下，在開口內並非完全 沒有助焊劑23的存在，而是會形成薄薄的皮膜。 基於助焊劑23之滲透、飛散、乾燥等的原因，網版 20的開口 20k可能會堵塞，或造成網版分離或轉印性變差 ，這時印刷結果變得不均一。此印刷狀態，即使不用確認 基板2 1，只要確認印刷用的網版2 0即可判斷是否合格。 圖6 ( c )的（1 )是顯示網版開口部正常的狀態，（2 )是 顯示發生局部堵塞的狀態，（3)是顯示全部都發生堵塞 的狀態。基板側之轉印量較多的部分，網版開口側之助焊 劑殘留量少；相反地，基板側之轉印量較少的部分，網版 開口側之助焊劑殘留量多。亦即基板2 1上的印刷狀態之 相反狀態，可從網版20側觀察。 網版2 0的開口狀態之合格判定係如下述般進行。用 CCD攝影機1 5拍攝網版20的開口狀態’將拍攝的影像經 由影像輸入部3 7送到印刷機控制部3 0。接著’將預先儲 存於資料庫3 8之網版2 0的開口狀態的基準模型影像和上 -16- 200922422 述取得的網版20的開口狀態的影像做比較，在尺寸 部34進行「正常（OK )」或「不良（NG )」的判定 定結果，「正常（〇Κ )」代表網版開口部正常的狀 「不良（NG )」代表網版開口部發生局部堵塞的狀 或全部都堵塞的狀態。 在印刷助焊劑後，判定爲不良（NG )之網版20 口狀態顯示於第6 ( c )圖之（2 )及（3 ) 。（ 2 )代 刷不均一而產生斑點狀的圖案。利用黑白攝影機之圖 配，即可簡單地進行判定。 另一方面，（3 )之NG的情形，未印刷到基板2 殘留在網版20開口 20k的助焊劑23量很多。因此， 劑殘留的程度依據色濃度的不同即可判定，所以藉由: 處理之濃淡灰階模型進行比較即可簡單地判定。或是 用彩色攝影機進行色差比較等來進行判定也可以。 又，爲了用定位用的C CD攝影機15來確認網）¾ 開口部的狀況，係將光線從網版20的下部向上照射 配置於網版20上方的CCD攝影機進行確認，藉此取 定的影像。也能將光線從網版20上方向下照射。 CCD攝影機15在上下都具有鏡頭（攝影部），在當 攝定位標記的定位用攝影機使用時，是運用向上和向 鏡頭；在當作觀測印刷後的網版20開口部的狀況之 用攝影機使用時，是運用上部的鏡頭。 關於檢查網版20的狀態後之檢查結果’若網版 部的堵塞、或發生附著污染等而從尺寸計算部34發出200922422 IX. Description of the Invention [Technical Field] The present invention relates to a solder ball printing apparatus, and more particularly to a solder ball printing apparatus for printing a solder ball on a substrate. [Prior Art] In the known printing method, in order to form a solder ball (diameter 80 to 1 〇〇 #m) at a pitch of 180 to 150 V, a known high-precision screen is used. After the printing device prints the lead-free solder, it is welded to form a solder ball. The screen printing apparatus includes, for example, a substrate loading conveyor, a substrate carrying conveyor, a printing table including an elevating mechanism, a mask having a transfer pattern in the opening, a doctor blade, a blade device including a blade lifting mechanism and a horizontal moving mechanism, Control devices used to control these mechanisms. After loading the substrate into the device from the substrate, the substrate is temporarily positioned and fixed to the printing table, and then the mark of both the substrate and the mask (screen) having the opening (corresponding to the circuit pattern) is recognized by the camera, and corrected. After the substrate is aligned with the screen, the printing table is raised to bring the substrate into contact with the screen, and the paste of the lead-free solder or the like is filled on the screen by the doctor blade while the screen and the substrate are in contact with each other. The opening is then lowered, the substrate and the screen are separated (screen separated) to transfer the paste onto the substrate, and then the substrate is carried out of the apparatus, thereby completing the printing. Another method is the solder ball transfer method, in which the solder balls are transferred to a jig having fine holes with high precision and arranged at a predetermined pitch, and then directly transferred onto the substrate, followed by fusion welding. Further, Patent Document 1 discloses a method of causing a mask to be oscillated or vibrated to fill a solder ball to a predetermined opening, a method of performing heating by charging the brush, and the like. Further, Patent Document 2 discloses a method in which a solder ball is mounted on a tray and adsorbed by a tube to be refilled to the electrode pad. [Patent Document 1] Japanese Laid-Open Patent Publication No. 2000-49 No. Hei. No. 2003-3 09 1 3 9 [Invention] The use of a lead-free solder printing method is low in equipment cost and can be A large number of bumps are formed at a time, so that it has the advantages of high productivity and low manufacturing cost. However, since it is difficult to ensure the uniformity of the transfer volume by the printing method, flattening must be performed to press the solder bump after the fusion to flatten the height, so that the number of steps is increased. 'The problem of increased equipment costs. Further, as the density of the element is increased, when the pitch is reduced to a pitch of 1 500 to 1 2 0 # m, the printing yield is low and the productivity is deteriorated. On the other hand, the solder ball transfer method can ensure the bump accuracy of the solder ball to form a bump with a stable height. However, it is necessary to use a high-precision solder ball adsorption jig and use a robot to batch-fill the solder ball to make it fine. In the case of the increase in the production time (tact) and the increase in the cost of the bumps caused by the high cost of the jig equipment. In addition, in Patent Document 1, the screen is oscillated or vibrated to fill the solder ball with a predetermined opening. As the particle size of the solder ball becomes smaller, the adhesion caused by the van der Waals force between the particles and the static electricity may occur. The resulting adsorption phenomenon 'and -6 - 200922422 There is a problem that it cannot be filled in the opening of the mask. In the same manner, the same problem occurs when the charging is performed by the squeezing motion of the doctor blade or the brush. Further, in the method of Patent Document 2, even if repair can be performed, there is a high possibility that the amount of residual flux is small, and if the soldering property of the solder is deteriorated during batch soldering, even if the solder ball is melted The adhesion may still occur and the welding of the electrode pad portion may be incomplete. SUMMARY OF THE INVENTION The object of the present invention is to provide a solder ball charging device and a bump forming method. When forming a bump of ultra-fine pitch, a large number of bumps can be formed in a batch manner like a printing method, and can be like a solder ball. The transfer method forms a bump of a stable height, and is capable of low-cost, high-speed, high-efficiency printing and charging, and has high productivity. In order to achieve the above object, a solder ball printing apparatus according to the present invention includes a flux printing portion for printing a flux on an electrode pad of a substrate, and a solder ball filling and soldering portion for supplying a solder ball to an electrode after printing the flux. And the inspection and repairing portion for inspecting the state of the substrate after the solder ball is printed, and the repairing and repairing portion corresponding to the defective state constitutes a solder ball printing device; wherein the flux printing portion is provided with a position corresponding to the plurality of electrode pads And a screen for opening the screen, a stage for loading and fixing the substrate, a positioning camera for photographing the positioning mark for alignment of the substrate and the screen, and a condition for observing the opening of the screen after printing. The inspection camera' is used to compare the screen of the screen opening or the flux adhesion on the back of the screen to remove the cleaning means, compare the image captured by the inspection camera with the screen of the pre-recorded reference model to determine the printing. Defective 200922422 The means of judgment 'determines according to the judgment result of the judgment means is to send the substrate to the next step or from production Exclusion, the exclusion of bad situations occur sweep the screen clear indication of the means of printing occurs. Further, a charging unit is provided in the printing head provided in the solder ball filling and printing unit, and the charging unit includes a solder ball box including a frame body, a cover, and a mesh body, and is spaced apart from the sieve body. The method is provided in a slit-like body at a lower portion of the solder ball box; and a vibration absorbing means is disposed on the cover through the support member, wherein the vibration absorbing means is configured to apply vibration to the sieve body to form an opening size of the sieve body Change to supply the solder balls to the slit-like body. According to the present invention, the flux is poorly printed for the major cause of defective solder ball soldering, and the premature processing is performed in the first step, so that productivity can be improved. According to the present invention, the solder ball filling efficiency can be improved to shorten the production time and perform high-filling solder ball charging and printing, thereby improving productivity. From flux printing to solder ball filling ~ inspection and repair, the operating efficiency of each device can be improved, and the production time can be shortened. Therefore, a large number of solder bump heights can be stably formed in a batch, low cost, and high speed manner. Solder bumps with good precision. Moreover, the structure of the device is simple, and the equipment cost can be reduced. [Embodiment] Hereinafter, preferred embodiments of the printing apparatus and the bump forming method of the present invention will be described with reference to the drawings. Fig. 1 shows a summary of the printing steps of the flux printing portion and the solder ball filling and printing portion. The first (a) diagram shows the flux printing step '1st 200922422 (b) shows the state of solder ball filling and printing. Fig. 1(a) shows a 'loading flux' on the screen 20 provided with an opening (corresponding to the shape of the electrode pad 22 of the substrate 2 in advance), by moving the blade 3 to a predetermined amount. The flux 2 3 is printed on the electrode pad 22 of the substrate 21. In the present embodiment, the screen 20 is a screen for flux printing. In order to ensure high-precision pattern position accuracy, a metal screen made by an additive method is used. The scraper 3 is any one of a square scraper, a sword scraper, and a flat scraper. The blade gap, the press, and the blade speed are set in accordance with the viscosity and thixotropy of the flux 23 to perform a printing operation. If the amount of printing of the flux 23 is too small, the solder ball cannot adhere to the electrode pad 22 when the solder ball 24 is filled. Further, during the welding after the solder ball printing, the solder adhesion failure is caused, and the solder bumps having a beautiful shape cannot be formed, which also causes a defect in solder bump height or insufficient solder connection strength. Further, if the amount of the flux 23 is too large, the flux may be attached to the opening portion provided on the screen 20 for supplying the solder ball 24 to the electrode pad 22 when the solder ball is filled and printed. . If the flux 23 is attached to the opening of the screen, the solder ball 24 adheres to the opening of the screen, causing a problem that the electrode pad 22 cannot be transferred. Therefore, flux printing is the most influential step in terms of solder ball quality. Next, as shown in the figure (b), the solder ball 24 is filled and printed on the substrate 2 after the printing flux 23 is provided in the solder ball filling and printing unit of the charging unit 60 (see FIG. 7). On the electrode pad 22 of the crucible. Used to weld the 200922422 ball 24 to the screen 20b of the electrode pad 22, in order to ensure the positional accuracy of the high-precision case, the metal mesh version produced by the additive method is used. The material of b is a magnetic material that can be magnetically attracted to the magnet stage (printing table 1 装载) of the loading plate so that the gap between the substrate 21 and the screen 20 becomes 〇', so that the solder ball can be prevented from being drilled into the substrate 2 The remaining solder ball is bad between 1 and the screen 20. Further, on the back surface of the screen 20b (on the side in contact with the substrate 21), there are micro pillars 20a made of resin or metal to avoid the flux 23 when it is adhered to the substrate 21 just after the auxiliary 23 is printed. Penetrate around the opening of the screen. Thereby, the release portion of the flux 2 3 permeate is formed. In addition, the combination of the screen 20b and the support 20a is referred to as a net 20° for solder ball printing, and an electrode pad for supplying the solder ball 24 to a predetermined position with high precision is provided at four corners of the substrate 2 1 . Positioning mark (not shown). For the positioning mark provided on the side of the substrate 21, a positioning mark is also provided on the side of the screen 20. The positioning mark is visually recognized by the CCD camera 15 (refer to Fig. 4), and high-precision alignment is performed in such a manner that the positioning mark on the screen 20 side coincides with the positioning on the substrate 2 1 side. In the alignment of this embodiment, the printing table 1 of the substrate 2 1 is moved in the horizontal direction. After the alignment is completed, the interval between the substrate 21 and the screen 20 is reduced, and after the plate 20 contacts the substrate 21, the charging unit 60 is operated to supply the opening of the solder ball 24 screen 20 to the printing flux 23. The electrode pad 22 on the surface of the substrate 21. The slit-shaped body 6 3 is provided on the lower side of the charging unit 60 (see the seventh) for solder ball supply, and the charging unit 60 is provided with the welding unit □ From the above figure, the line -10-200922422 swings and advances, so that the solder balls 24 roll and vibrate to evenly weld the solder to the screen opening. Figure 2 shows an embodiment of a solder ball printing apparatus. This figure is a device that integrates the flux printing department, the solder ball filling and printing department, and the inspection and cum. Each part can also be in the form of a separate device. In the present device, first, in the flux printing portion (the screen printing method flux 23 is printed on each of the electrode pads 22 on the substrate, the solder ball is filled by the transfer (from the side of the flux printing portion as the carry-out conveyor) The side view is the loading conveyor.) 'The solder ball is filled and the printing part will be soldered to the electrode pad (via flux). The flux printing part and the solder ball filling and printing department are mainly different from the brush head, flux printing The part is made of a doctor blade. The solder ball filling part is composed of a charging unit for supplying solder balls. The inspection and repairing printing head is a dispenser type suction supply head structure. Since the screen is used, the screen mounting frame holder or the like is not provided. Fig. 3 is a flow chart showing the formation of the bumps of this embodiment. After loading (STEP 1), a certain amount of assist STEP2 is printed on the electrode pads. Next, the opening of the screen after printing the flux is inspected STEP3). When the inspection result is NG, the substrate is carried out to the NG storage portion, and the cleaning is automatically performed by the under-cleaning device 45 (STEP4), and the flux is replenished as needed. The NG substrate is formed by waiting for the step after the solder ball printing is performed on the conveyor in the subsequent step and discharging it to the device repairing portion of the $24 charging device outside the production line. The conveyor and the printing ball supply point are printed. The inspection of the cum printing and the version of the substrate used in the soldering agent (the condition (the substrate is stored. Then NG News. -11 - 200922422 using the NG substrate storage on the production line, etc., using the substrate 匣 倂 。. NG substrate in the production line After the outer step is washed, it can be used again for flux printing. Next, solder ball charging and printing (STEP5) is performed. After the solder ball is filled and printed, after the screen separation, the net is checked from the top of the screen. The solder ball fill condition in the opening of the plate (STEP6). If there is a shortage of the inspection result, the solder ball charging and printing operation (STEP7) is performed again before the screen separation. If the result of STEP6 is OK, the screen separation (STEP8) is carried out. Then, after the solder ball is filled, the inspection and repair device is used to check the filling condition (STEP9). After the flux is supplied, the solder ball is again supplied to the electrode pad portion of the NG point (STEP 1 0). When the filling condition is checked as OK, the solder ball is melted again by the welding device, thereby completing the solder bump. Fig. 4 is a schematic view showing a schematic structure of a screen printing apparatus (mainly a flux printing unit) of the present invention. Fig. 4(a) is a structure viewed from the front of the screen printing apparatus, and Fig. 4(b) is a diagram Figure 5 (a) (b) is an explanatory diagram of the screen printing device. The main body frame 1 is provided with a frame support (not shown), and a mask is installed on the frame support (opening 20 screens having a printed pattern are provided on the frame 20c (see Fig. 6). In the figure, a print head 2 provided with a doctor blade 3 is disposed above the screen 20. In the case of the flux printing unit, the squeegee 3 made of polyurethane is attached to the printing head 2. In the case of the solder ball filling and printing unit, the printing head 2 is -12-200922422 instead of the doctor blade 3, and the slit-shaped body is installed. a charging unit 60 constituted by 63 or the like. The printing head 2 is moved in the horizontal direction by the printing head moving mechanism 6' The print head elevating mechanism 4 moves up and down. When the doctor blade 3 is replaced with the charging unit 60, the charging unit 60 is moved up and down by the printing head elevating mechanism 4. Below the screen 20, the net plate 20 is opposed to the screen 20 The method of 'direction' is provided with a printing table 1 (a substrate 2 1 for loading and holding a printing object). The printing table 1 is provided with a substrate 21 moving in the horizontal direction (XY 0 direction) and a screen. 20 Alignment ΧΥ The β stage 1 1 is received; the substrate 21 is received from the carry-in conveyor 25, and the substrate 21 is brought close to or in contact with the stage elevating mechanism 12 of the screen 20 surface. The substrate receiving conveyor 26 disposed on the printing table 1 can receive the substrate 21 carried by the substrate loading conveyor 25 on the printing table 10, and can discharge the substrate 2 1 to the base loading and unloading conveyor at the end of printing. 27. The screen printing apparatus has a function of automatically aligning the screen 20 and the substrate 2 1 . That is, the alignment mark provided on the screen 20 and the substrate 21 is photographed by the CCD camera 15, and the positional deviation amount is obtained by image processing, and the ΧΥ 6»stage 1 1 is driven to correct the deviation amount. And the alignment is performed. Further, a printer control unit 30 is provided inside the printer main body frame, and includes a screen separation control unit 39, a drive control unit for each unit, and the like, and a print control unit 36 for processing the CD camera. The image input unit 3 7 of the image signal of 1 5 . Further, on the outside of the printing machine, a data input unit 5 for rewriting the control data or changing the printing conditions, and a display unit 40 for monitoring the printing status and the acquired identification mark are provided. -13- 200922422 The printing machine control unit 30 is provided with a printing control unit 3 for controlling the charging unit 60, in accordance with the bump pitch, the size of the solder ball, and the type of metal mask used. Simply select the appropriate fill print mode. Further, the correlation calculation unit 31 for calculating the correlation 按照 according to the input image, the shape estimation unit 32 for obtaining the shape based on the acquired image and the data of the data library 38, and the position coordinate calculation unit for obtaining the position coordinate position 3 3 and size calculation unit 34. Based on the data captured by the CCD camera 15, the positional deviation amount is obtained in accordance with the position identification marks provided on the substrate 21 and the screen 20. The XY stage 11 is driven to perform alignment in accordance with an instruction from the 控制 0 control unit. Next, the operation of the printing apparatus will be described by taking the solder ball charging and printing department as an example. The substrate 21 on which the solder bumps are formed is supplied to the substrate receiving conveyor 26 by the substrate loading conveyor 25. When the substrate 2 1 is transported to the position of the printing table 10, the printing table 1 上升 rises, and the substrate receiving conveyor 26 transfers the substrate 21 to the printing table 1 . The substrate 21 delivered to the printing table 1 is fixed at a predetermined position on the printing table 1 . After the substrate 21 is fixed, the CCD camera 15 is moved to a predetermined substrate mark position. The status is shown in Figure 5 (a). Next, the CCD camera 15 captures a position identification mark (not shown) provided on the substrate 21 and the screen 20, and transmits it to the printer control unit 30. The image input unit 307 in the printer control unit obtains the positional deviation amount between the screen 20 and the substrate 21 based on the image data, and the printer control unit 30 operates the Θ control unit 35 to move the printing based on the result. The stage 1 is thereby corrected for the position of the substrate 21 to be aligned with the screen 20. -14 - 200922422 The status after the alignment is completed is shown in Figure 5 (b). First, the C CD camera 15 retreats to a position that does not interfere with the printing table 1 既. When the retraction of the CCD camera 15 is completed, the printing table 1 is raised, and the substrate 21 and the mask 20 are brought into contact. In this state, the print head elevator mechanism 4 is operated, and the doctor blade (the blade 3 is shown in the figure, but the slit-shaped body 63 which is the tip end of the charging unit 60 in the solder ball filling step) contacts the screen surface. Then, while the slit-shaped body 63 is vibrated and vibrated, the motor 2g for driving the print head is rotationally driven to move the slit-like body 63 horizontally on the screen surface, thereby taking the solder ball 24 from the slit. The opening of the body 63 is filled in the electrode pad 22 of the substrate 21 through an opening provided in the screen surface. The print head 2 is raised after a stroke that moves a certain distance in the horizontal direction. Then, the printing table 10 is lowered, the screen 20 is separated from the substrate 2, and the solder balls 24 which are filled in the opening of the screen 20 are transferred onto the substrate 21. Then, the substrate 2 1 after the solder ball 24 is printed is sent to the next step via the substrate carry-out conveyor 27. Further, as described above, in the substrate 21 and the screen 20, two or more alignment alignment marks are provided so as to face each other at the same position. The mark of the screen 20 is recognized from below by a special CCD camera 15 having a field of view of 2 in the vertical direction, the mark of the substrate 21 is recognized from above, and the position coordinates of the mark provided at the predetermined portion are read, and the substrate 21 is calculated and corrected. The substrate 21 is aligned with the screen 20 with respect to the amount of deviation of the screen 20. Fig. 6 is a view showing the opening state of the screen after printing the flux. Figure 6 (a) shows the shape of the entire screen, Figure 6 (b) shows the opening condition after setting one electrode group, and Section 6 (C) shows the opening of the -15-200922422 after printing the flux. The situation. By setting the appropriate screen gap (interval between the screen and the substrate), printing (squeezing force of the blade on the screen), and the speed of the blade, the flux 2 3 can be fully charged to the screen 2 0 The opening 20 k is separated from the screen 20 by the doctor blade 3, whereby the flux 23 can be surely transferred to the electrode pad 22 of the substrate 21. The screen version 20 is fixed to the frame 2 0 c. The viscosity of the flux 2 3 for screen printing, the thixotropy, and the opening 20k of the screen 20 are fine, and the opening 20 0 of the screen 20 after printing is in the normal printing state. The presence of the flux 23 is not completely present, but a thin film is formed. The opening 20k of the screen 20 may be clogged due to penetration, scattering, drying, etc. of the flux 23, or the screen separation or transferability may be deteriorated, and the printing result may become uneven. In this printing state, even if it is not necessary to confirm the substrate 2 1, it is judged whether or not it is acceptable by checking the screen 20 for printing. (1) of Fig. 6(c) shows a state in which the opening of the screen is normal, (2) shows a state in which partial clogging occurs, and (3) shows a state in which clogging occurs. In the portion where the amount of transfer on the substrate side is large, the amount of flux remaining on the screen opening side is small. Conversely, the portion having a small amount of transfer on the substrate side has a large amount of flux remaining on the opening side of the screen. That is, the opposite state of the printed state on the substrate 2 1 can be observed from the screen 20 side. The qualification determination of the opening state of the screen 20 is performed as follows. The aperture state of the screen 20 is taken by the CCD camera 15, and the captured image is sent to the printer control unit 30 via the video input unit 37. Then, the reference model image stored in the open state of the screen 20 in the database 38 is compared with the image of the open state of the screen 20 obtained in the above -16-200922422, and the size portion 34 is "normal" ( "OK" or "bad" (NG) determines the result. "Normal (〇Κ)" means that the screen opening is normal. "Bad (NG)" means that the screen opening is partially blocked or all blocked. status. After printing the flux, the screen state of the screen which is judged to be defective (NG) is shown in (2) and (3) of Fig. 6(c). (2) The pattern is not uniform and produces a speckled pattern. With the black and white camera's map, you can easily make a decision. On the other hand, in the case of NG of (3), the amount of the flux 23 which is not printed on the substrate 2 and remains in the opening 20k of the screen 20 is large. Therefore, the degree of residue of the agent can be determined based on the difference in color density, so that it can be easily determined by comparing the processed light gray scale model. Alternatively, it may be determined by color difference comparison using a color camera. Further, in order to confirm the condition of the opening portion of the screen by the C CD camera 15 for positioning, light is emitted from the lower portion of the screen 20 to the CCD camera disposed above the screen 20 to confirm the image. . It is also possible to illuminate the light downward from above the screen 20. The CCD camera 15 has a lens (photographing unit) on the upper and lower sides, and is used for the camera for positioning the positioning mark, and is used for the camera for observing the position of the screen 20 after printing. When using the upper lens. The result of the inspection after checking the state of the screen 20 is issued from the size calculating unit 34 if the screen portion is clogged or adhesion contamination occurs.
計算 。判 態， 育旨， 的開 表印 案匹 1而 助焊 影像 ，使 i 20 ，用 得穩 由於 作拍 下的 檢查 開口 NG -17- 200922422 訊號的情形，根據印刷機控制部3 0的指令而用印刷裝置 內的版下清掃裝置45 (參照第5圖）實施自動清掃，並可 按照需要來供應補充助焊劑23。又NG基板，是和不實施 焊球印刷以後的步驟之N G訊號一起，依據印刷機控制部 3 0的指令在後步驟的輸送器上待機，以排出生產線外。也 能使用生產線上的N G基板儲存器等，用基板匣一倂排出 。N G基板在生產線外的步驟貫施洗淨後，可再度使用於 助焊劑印刷。 第7圖係顯示焊球印刷頭（充塡單元6 0 )的構造。充 塡單元6 0係具備：在由框體6 1和蓋6 4和篩狀體6 2所形 成的空間收納焊球24之焊球盒、與篩狀體62隔著間隔之 狹縫狀體63。篩狀體62，係對應於供應對象之焊球24的 直徑而具有網狀開口（或連續長方形的狹縫等）之極薄金 屬板。在篩狀體6 2的下部，配置狹縫狀體6 3，且讓狹縫 狀體63和網版20形成面接觸。 藉由未圖示之印刷頭昇降機構4，可微調狹縫狀體6 3 對網版20的接觸程度（間隙）。狹縫狀體63是使用磁性 體材料，係對應於供應對象之焊球2 4的直徑及網版2 0的 開口尺寸而具有網狀的開口（或連續長方形的狹縫等）之 極薄金屬板。 第8圖係顯示將篩狀體62 (設於焊球收納部之焊球盒 )在水平方向施加振動之水平振動機構。在蓋64的上部 設置支承構件70，該支承構件70在與焊球盒側面平行的 位置裝設加振手段65。依據此構造，從焊球盒側面側藉由 -18- 200922422 加振手段65施加振動，能對篩狀體62施加振動。藉由使 篩狀體62振動，會使設於篩狀體62之狹縫狀開口擴大成 比焊球24直徑更大。 藉此，收納於焊球盒的焊球24，可從篩狀體62的狹 縫部落至狹縫狀體63上。落至狹縫狀體63上的焊球24 的量、亦即焊球24的供應量，可改變加振手段65的加振 能量來做調整。 本圖所示的加振手段65，係使用氣旋式振動器，以數 位控制來微調壓縮氣體壓力，可控制其振動數。也能藉由 改變壓縮氣體流量來改變振動數。又篩狀體62及焊球盒 ，係藉由加振手段6 5對收納於焊球盒之焊球2 4賦予振動 ，而抵銷作用於焊球24間之凡得瓦力所產生之吸引力， 藉此讓焊球24分散。爲了避免因焊球24材料或生產環境 的溫度濕度的影響而使焊球供應量發生變化，可考慮生產 效率來調整前述分散效果。 第9圖係顯示充塡單元60的水平擺動機構。狹縫狀 體63是由磁性材料所形成。藉由使用磁性材料，利用內 設磁鐵之載台（印刷台10)的磁力，可將狹縫狀體63吸 附於磁性材料所形成的網版2 0。如第9圖所示，水平擺動 機構是採用以下構造。在支承構件7〇的上部設置線性導 件67，並設有具備線性導軌（用來導引前述線性導件67 移動）之充塡單元支承構件71。在充塡單元支承構件71 設置驅動用馬達6 8，又在該驅動用馬達的軸裝設偏芯凸輪 66，利用偏芯凸輪66之旋轉而使支承構件70在左右方向 -19- 200922422 移動。 亦即’水平方向之水平擺動機構，是藉由驅動用馬 6 8讓偏芯凸輪旋轉’而在任意的衝程量對狹縫狀體6 3 予㉟動動作’因此在狹縫狀體6 3和網版2 0之間能以不 生間隙的方式確實地進行焊球24之滾動。又利用狹縫 體63的開口尺寸’能使焊球24確實地補滿狹縫狀體 的開口並進行高效率的充塡動作。網版20的擺動動作 周期速度’可控制驅動用馬達6 8的速度來任意地改變 可考慮生產線的均衡來設定焊球24的充塡產距時間。 ，可對應於焊球24材料的種類、網版24的開口、環境 件來調整周期速度，藉此來控制充塡率。 第1 〇圖係顯示在充塡頭設置篦狀體的構造。藉由 塡單元60將焊球24供應至基板21上後，當網版20和 板2 1面分離時’亦即進行網版分離而將焊球轉印至基 上時’若在網版20的版面上有焊球24殘留時，焊球 會通過有網版20的開口而落至基板21上導致過剩焊球 良的原因。因此，本實施例是在充塡單元60的進行方 ，以和焊球盒隔著間隔的方式設置篦狀體6 9 (高度和狹 狀體63大致相同）。篦狀體69的前端極薄且硏磨成平 精度極高的狀態，以和網版20形成密合狀態，而避免 球20溢出至充塡單元60的外部。 又，篦狀體69是使用磁性體材料，和狹縫狀體63 樣的是利用磁力吸附於網版20，如此可避免焊球20溢 至充塡單元6 0的外部。此外，篦狀體6 9也能設置於焊 達 賦 發 狀 63 之 ， 又 條 充 基 板 24 不 向 縫 坦 焊 同 出 球 -20- 200922422 盒的整個外周部。 第11圖係顯示在充塡單元部設置氣簾的構造。藉由 篦狀體69，能使網版20的版面上幾乎沒有焊球的殘留。 然而’起因於網版20的版面之微小移位，可能仍會有焊 球殘留。於是’在本實施例，爲了徹底解決過剩焊球造成 的不良影響’係設置有氣簾。亦即，在支承昇降機構4 ( 用來構成印刷頭2之上下移動馬達）的馬達支承構件設置 空氣噴出口 75’以在充搞卓兀的周圍形成氣簾。由未圖示 的壓縮空氣供應源將壓縮空氣供應至該噴出口 75。 利用該氣簾’在充塡單元朝基板端面方向移動時，藉 由壓縮空氣溢出的焊球朝充塡單元動作方向側推擠，而使 版面上不會發生焊球殘留。 第1 2圖係用來說明焊球印刷後的網版充塡狀態的檢 查。第12(a) (b)圖是和第6圖相同，在此省略其詳細 說明。 第12 ( c )圖的（1 )〜（3 )顯示焊球充塡、印刷後 之網版20上的焊球充塡狀態。（1 )是代表，在網版2〇 開口全部都充塡有焊球24的狀態。（2 )是代表焊球充塡 不完全的狀態。（3 )是代表，充塡時複數個焊球24彼此 吸附成雙焊球的狀態，或在網版的版面上殘留過剩焊球的 狀態。 在前述（2 )( 3 )的狀態下進行網版分離時，即使將 基板送到後步驟仍會生產出不合格品。於是，在實施網版 分離之前，藉由檢查網版20的版面上的充塡狀況，再用 -21 - 200922422 充塡單元60重試充塡印刷動作，如此可將不合格品修正 爲良品。該檢查’可經由和良品模型比較之圖案匹配來進 行判定。在焊球充塡印刷後，用安裝於印刷頭側之線性感 測攝影機以區域爲單位進行批式辨識。判定爲NG時，再 度實施焊球充塡印刷。判定爲合格時，實施網版分離動作 後將基板2 1排往後步驟。 第1 3圖係用來說明焊球充塡後在檢查修補部之修補 作業。第14圖係用來說明焊球充塡後之充塡不良狀況。 如第14圖所示，焊球充塡不良是包括··無焊球、雙焊球 、位置偏移焊球、變形焊球、過剩焊球等的不良模式。 在檢查修補部，首先，當焊球充塡印刷完成後，用 CCD攝影機來確認基板上的充塡狀況。若檢測出不良，求 出不良部位的位置座標。在發生雙焊球、位置偏移焊球、 變形焊球、過剩焊球等不良的情形，將真空吸附用嘴8 6 移動至焊球的位置，進行真空吸附而移動至不良焊球廢棄 站，並具備藉由解除真空而使焊球落下並廢棄之廢棄箱。 又在檢測出因焊球24的供應不足而未供應焊球之電 極墊部的情形，將收納於焊球收納部84之正常焊球24用 分配器87吸附，將吸附著焊球24之分配器87移動至助 焊劑供應部85內儲存的助焊劑23·，將焊球24浸漬於助焊 劑23中，藉此對焊球24添加助焊劑23。將吸附著焊球 24 (添加助焊劑2 3後）之分配器8 7移動至基板的缺陷部 ，將焊球供應至缺陷部而完成修補作業。 又，在前述的檢查，除了變形焊球 '位置偏移焊球等 -22- 200922422 的不良焊球以外，都能利用上述修補作業來修復缺陷。 第15圖係顯示檢查修補裝置的槪略構造。在本圖， 檢查修補部是1個獨立的裝置。 搬入側輸送器88上的檢查對象基板82沿空白箭頭方 向搬送至檢查部輸送器90上。在檢查部輸送器90的上部 設有門型框80，在門型框80的搬入側輸送器88側，沿基 板搬送方向（空白箭頭方向）的垂直方向設置線性感測器 8 1。藉由該線性感測器8 1來檢測基板2 1的電極墊22上 所印刷的焊球24狀態。 在用來支承門型框8 0的一方足側設置有：用來收納 正常焊球之焊球收納部84、助焊劑供應部85。在另一方 足側設置廢棄箱。在門型框部設有：用來吸引除去不良焊 球之真空吸附嘴86 (可藉由線性馬達進行左右移動）、用 來修補基板上的缺陷之分配器8 7。真空吸附嘴8 6及分配 器87，可沿斜線箭頭方向移動。 檢查部輸送器90可沿空白箭頭方向往復移動。可對 應於基板的缺陷位置而使分配器或真空吸附嘴的位置對準 缺陷位置。檢查修補完成後的基板，被搬送輸送器8 9搬 出而送往熔焊裝置。依據上述構造，能用第14圖所說明 的動作來進行檢查修補。 依據上述構造所提供之印刷裝置，可將焊球正確地供 應至基板的電極墊部，且儘量防止不良品的發生。 【圖式簡單說明】 -23- 200922422 第1(a) ( b )圖係顯示助焊劑印刷及焊球充塡印刷 步驟的槪要。 第2圖係顯示焊球印刷之凸塊形成裝置的一例。 第3圖係本實施形態的凸塊形成的流程圖。 第4 ( a )( b )圖係顯示網版印刷裝置的槪略構造。 第5 ( a )( b )圖係網版印刷裝置的動作說明圖。 第6 ( a ) ( b ) ( c )圖係顯示網版印刷後的網版的開 口狀態。 第7圖係顯示焊球印刷頭的構造。 第8圖係顯示焊球收納部篩狀體的水平振動機構。 第9圖係顯示焊球印刷頭的水平擺動機構。 第1 〇圖係焊球印刷頭用篦狀體的說明圖。 第1 1圖係焊球印刷頭用氣簾之說明圖。 第12(a) ( b ) ( c )圖焊球印刷後的網版狀態例的 說明圖。 第1 3圖係焊球的修補之說明圖。 第1 4圖係焊球印刷不良的狀況之說明圖。 第15圖係檢查修補裝置的槪要說明圖。 【主要元件符號說明】 1 :印刷機 2 :印刷頭 3 :刮刀 1 0 :印刷台 -24- 200922422 1 1 : X Υ 0 台 1 5 _·攝影機 20 :網版 2 1 :基板 3 0 :印刷機控制部 3 4 :尺寸計算部 45 :清掃裝置 60 :充塡單元 63 :狹縫狀體 65 :加振手段 69 ’·篦狀體 -25Calculation. Judgment, the purpose of the education, the opening of the printed case 1 and the welding of the image, so that i 20, used to stabilize the inspection of the opening NG -17- 200922422 signal, according to the instructions of the printing machine control unit 30 On the other hand, automatic cleaning is performed by the under-plate cleaning device 45 (see Fig. 5) in the printing apparatus, and the supplementary flux 23 can be supplied as needed. Further, the NG substrate, together with the N G signal of the step after the solder ball printing is not performed, is placed on the conveyor in the subsequent step in accordance with the instruction of the printing machine control unit 30 to discharge the outside of the production line. It is also possible to use a substrate such as an N G substrate storage device on the production line to discharge it. The N G substrate can be reused for flux printing after being cleaned by the steps outside the production line. Fig. 7 shows the construction of the solder ball print head (filling unit 60). The charging unit 60 includes a solder ball box that houses the solder balls 24 in a space formed by the frame body 61, the cover 64, and the mesh body 62, and a slit-shaped body that is spaced apart from the sieve body 62. 63. The sieve body 62 is an extremely thin metal plate having a mesh opening (or a continuous rectangular slit or the like) corresponding to the diameter of the solder ball 24 to be supplied. In the lower portion of the sieve body 64, the slit-like body 63 is disposed, and the slit-like body 63 and the screen 20 are brought into surface contact. The degree of contact (gap) of the slit-like body 6 3 with respect to the screen 20 can be finely adjusted by the print head elevating mechanism 4 (not shown). The slit-shaped body 63 is a very thin metal having a mesh-like opening (or a continuous rectangular slit, etc.) in accordance with the diameter of the solder ball 24 and the opening size of the screen 20 using a magnetic material. board. Fig. 8 is a view showing a horizontal vibration mechanism for applying vibration to the horizontal direction of the sieve body 62 (the solder ball case provided in the solder ball housing portion). A support member 70 is provided on the upper portion of the cover 64, and the support member 70 is provided with a vibration absorbing means 65 at a position parallel to the side surface of the solder ball cartridge. According to this configuration, vibration is applied to the sieve body 62 by applying vibration from the side surface of the solder ball cartridge by the -18-200922422 vibration absorbing means 65. By vibrating the sieve body 62, the slit-like opening provided in the sieve body 62 is enlarged to be larger than the diameter of the solder ball 24. Thereby, the solder balls 24 accommodated in the solder ball box can be separated from the slits of the sieve body 62 to the slit-like body 63. The amount of the solder balls 24 falling on the slit-like body 63, that is, the supply amount of the solder balls 24, can be adjusted by changing the vibration energy of the vibration-increasing means 65. The vibration absorbing means 65 shown in the figure uses a cyclone vibrator to finely adjust the pressure of the compressed gas by digital control, and the number of vibrations can be controlled. It is also possible to change the number of vibrations by changing the flow rate of the compressed gas. Further, the mesh body 62 and the solder ball box are vibrated by the vibration means 65 for the solder balls 24 accommodated in the solder ball box, thereby offsetting the attraction of the van der Waals force acting on the solder balls 24. Force, thereby allowing the solder balls 24 to disperse. In order to avoid a change in the supply amount of the solder ball due to the influence of the temperature and humidity of the solder ball 24 material or the production environment, the above-described dispersion effect can be adjusted in consideration of the production efficiency. Fig. 9 shows the horizontal swinging mechanism of the charging unit 60. The slit body 63 is formed of a magnetic material. By using a magnetic material, the slit-like body 63 can be adsorbed to the screen 20 formed of a magnetic material by the magnetic force of the stage (printing table 10) on which the magnet is placed. As shown in Fig. 9, the horizontal swing mechanism adopts the following configuration. A linear guide 67 is provided at an upper portion of the support member 7''', and a charging unit support member 71 having a linear guide for guiding the movement of the linear guide 67 is provided. The driving unit 6 is provided in the charging unit supporting member 71, and the eccentric cam 66 is attached to the shaft of the driving motor, and the supporting member 70 is moved in the left-right direction -19-200922422 by the rotation of the eccentric cam 66. That is, the horizontal swinging mechanism in the horizontal direction is to rotate the eccentric cam by the driving horse 6 8 and to move the slit-shaped body 6 3 to an arbitrary stroke amount. Therefore, the slit-shaped body 6 3 The rolling of the solder balls 24 can be surely performed without gaps between the screen and the screen 20. Further, by using the opening size ' of the slit body 63, the solder ball 24 can surely fill the opening of the slit-like body and perform an efficient charging operation. The swinging motion of the screen 20 can be arbitrarily changed by controlling the speed of the driving motor 6.8. The charging time of the solder ball 24 can be set in consideration of the equalization of the production line. The charging rate can be controlled by adjusting the cycle speed in accordance with the type of the material of the solder ball 24, the opening of the screen 24, and the environmental member. The first figure shows the structure in which the scorpion is placed in the filling head. After the solder ball 24 is supplied onto the substrate 21 by the germanium unit 60, when the screen 20 and the surface of the board 21 are separated, that is, when the screen is separated and the solder ball is transferred onto the substrate, the screen 20 is When the solder ball 24 remains on the layout, the solder ball will fall onto the substrate 21 through the opening of the screen 20, resulting in a good solder ball. Therefore, in the present embodiment, the weir-like body 6 9 (the height is substantially the same as the narrow body 63) is provided on the side of the charging unit 60 so as to be spaced apart from the solder ball case. The front end of the scorpion 69 is extremely thin and honed to a state of extremely high flatness to form a close contact with the screen 20, and to prevent the ball 20 from overflowing to the outside of the charging unit 60. Further, the ridge body 69 is made of a magnetic material, and is adsorbed to the screen 20 by a magnetic force like the slit-like body 63, so that the solder ball 20 can be prevented from overflowing to the outside of the charging unit 60. In addition, the ridges 69 can also be disposed in the weld flaring 63, and the slabs 24 are not welded to the entire outer circumference of the -20-200922422 box. Fig. 11 shows a structure in which an air curtain is provided in the charging unit. With the ridges 69, there is almost no residue of solder balls on the screen surface of the screen 20. However, due to the slight shift in the layout of the screen 20, there may still be solder ball remaining. Thus, in the present embodiment, an air curtain is provided in order to completely solve the adverse effect caused by the excess solder ball. That is, the motor supporting member supporting the elevating mechanism 4 (which is used to constitute the upper and lower moving motors of the printing head 2) is provided with an air ejection port 75' to form an air curtain around the charging. Compressed air is supplied to the discharge port 75 by a compressed air supply source not shown. When the charging device is moved toward the substrate end surface by the air curtain, the solder ball overflowing by the compressed air is pushed toward the charging unit side of the charging unit, so that the solder ball remains without being deposited on the plate surface. Fig. 1 2 is a view for checking the state of the screen after the solder ball is printed. The figure 12(a)(b) is the same as Fig. 6, and a detailed description thereof is omitted here. (1) to (3) of Fig. 12 (c) show the state in which the solder balls are filled and the solder balls on the screen 20 after printing are filled. (1) is a state in which all of the openings of the screen 2 are filled with the solder balls 24. (2) is a state in which the solder ball is not fully charged. (3) is a state in which a plurality of solder balls 24 are adsorbed into a double solder ball at the time of filling, or a state in which excess solder balls remain on the screen surface of the screen. When the screen separation is carried out in the state of the above (2) (3), a defective product is produced even if the substrate is sent to the subsequent step. Therefore, before the screen separation is performed, by checking the filling condition on the layout of the screen 20, the charging operation is retried by the charging unit 60 with the -21 - 200922422, so that the defective product can be corrected as a good product. This check ' can be determined by pattern matching compared to the good model. After the solder balls are filled and printed, batch identification is performed in units of areas using a line-sensing camera mounted on the side of the print head. When it is judged as NG, the solder ball is printed again. When it is judged to be acceptable, the substrate 2 1 is discharged to the subsequent step after the screen separation operation is performed. Fig. 3 is a diagram for explaining the repair work of the inspection portion after the solder ball is filled. Figure 14 is used to illustrate the poor condition of the solder ball after it has been filled. As shown in Fig. 14, the defective solder ball is a defective mode including no solder balls, double solder balls, positionally offset solder balls, deformed solder balls, and excess solder balls. In the inspection repairing section, first, after the solder ball is filled and printed, the CCD camera is used to confirm the filling state on the substrate. If a defect is detected, find the position coordinate of the defective part. In the case where a double solder ball, a positional offset solder ball, a deformed solder ball, or an excessive solder ball is defective, the vacuum suction nozzle 8 6 is moved to the position of the solder ball, and vacuum suction is performed to move to the defective solder ball disposal station. There is also a waste box that allows the solder balls to fall and be discarded by releasing the vacuum. Further, when the electrode pad portion of the solder ball is not supplied due to insufficient supply of the solder ball 24, the normal solder ball 24 accommodated in the solder ball housing portion 84 is adsorbed by the dispenser 87, and the distribution of the solder ball 24 is adsorbed. The device 87 moves to the flux 23· stored in the flux supply unit 85, and immerses the solder ball 24 in the flux 23, thereby adding the flux 23 to the solder ball 24. The dispenser 8 7 which adsorbs the solder ball 24 (after the addition of the flux 2 3) is moved to the defective portion of the substrate, and the solder ball is supplied to the defective portion to complete the repair work. Further, in the above-described inspection, in addition to the defective solder balls of the deformed solder ball 'position offset solder balls -22-200922422, the above repair work can be used to repair the defects. Fig. 15 shows a schematic structure of the inspection repairing device. In this figure, the inspection repair section is a separate device. The inspection target substrate 82 on the carry-in side conveyor 88 is conveyed to the inspection portion conveyor 90 in the direction of the blank arrow. A door frame 80 is provided on the upper portion of the inspection unit conveyor 90, and a line sensor 8 1 is disposed in the vertical direction of the substrate conveyance direction (the direction of the blank arrow) on the side of the carry-in conveyor 88 of the door frame 80. The state of the solder ball 24 printed on the electrode pad 22 of the substrate 2 1 is detected by the line sensor 81. On the one side of the foot for supporting the door frame 80, a solder ball accommodating portion 84 for accommodating a normal solder ball and a flux supply portion 85 are provided. Set the waste bin on the other side of the foot. The door frame portion is provided with a vacuum suction nozzle 86 (which can be moved left and right by a linear motor) for sucking and removing defective solder balls, and a distributor 87 for repairing defects on the substrate. The vacuum suction nozzle 86 and the distributor 87 are movable in the direction of the oblique arrow. The inspection portion conveyor 90 is reciprocally movable in the direction of the blank arrow. The position of the dispenser or vacuum nozzle can be aligned to the defect location corresponding to the defect location of the substrate. The substrate after the repair is completed is carried out by the transport conveyor 8 9 and sent to the welding device. According to the above configuration, the inspection and repair can be performed by the operation described in Fig. 14. According to the printing apparatus provided by the above configuration, the solder ball can be correctly supplied to the electrode pad portion of the substrate, and the occurrence of defective products can be prevented as much as possible. [Simple description of the diagram] -23- 200922422 The 1st (a) (b) diagram shows the main steps of flux printing and solder ball filling. Fig. 2 is a view showing an example of a bump forming device for solder ball printing. Fig. 3 is a flow chart showing the formation of bumps in the present embodiment. The fourth (a)(b) diagram shows the schematic structure of the screen printing apparatus. The fifth (a)(b) diagram is an operation explanatory diagram of the screen printing apparatus. The 6th ( a ) ( b ) ( c ) diagram shows the open state of the screen after screen printing. Figure 7 shows the construction of the solder ball print head. Fig. 8 is a view showing a horizontal vibration mechanism of a mesh body of a solder ball housing portion. Figure 9 shows the horizontal swing mechanism of the solder ball print head. The first drawing is an explanatory view of the beak for the solder ball print head. Fig. 1 is an explanatory view of an air curtain for a solder ball print head. Section 12(a) (b) (c) An illustration of the example of the screen state after the solder ball is printed. Figure 13 is an explanatory diagram of the repair of the solder ball. Fig. 14 is an explanatory view showing a state in which solder ball printing is defective. Figure 15 is a schematic diagram of the inspection of the repairing device. [Description of main component symbols] 1 : Printing machine 2 : Print head 3 : Scraper 1 0 : Printing station - 24 - 200922422 1 1 : X Υ 0 Table 1 5 _· Camera 20 : Screen 2 1 : Substrate 3 0 : Printing Machine control unit 34: size calculation unit 45: cleaning device 60: charging unit 63: slit-like body 65: vibration-inducing means 69 '·篦--25