195 改性尼龙供应商审核该看什么
"吴工，下周我们去你们厂做个供应商审核，麻烦准备一下。"
我回第一句通常是："好。我这边有份清单，先发给你对着过一遍，省得来回折腾。"
电话那头一般会愣一下。因为他们准备的那张表——营业执照、体系证书、产能、主要客户——这些东西当然有，打印一下十分钟。
但真正决定你后两年会不会出事的，基本不在那张表上。
我把话说得再白一点：审核这件事，最容易做的部分是把所有证书收齐；最难的部分，是判断一个工厂在没有外部压力的时候会不会自觉。
这篇讲的，就是怎么用一天时间，把第二件事看出来。
先说一个前提：这套方法适合改性塑料厂，尤其是玻纤增强、阻燃、配色这类有配方自由度的料。如果是买标准化树脂，看的东西另说。
供应商审核看什么？表面看设备与检测，往里看是配方与批次管控，最里面看的是这支团队对改性尼龙工况的理解深度。设备可以买，理解力买不来。
一、审核最常见的失败：变成了参观
多数审核之所以流于形式，是因为它变成了一场参观。
参观的问题在于：你能看到的，都是对方愿意让你看的。干净的大厅、擦亮的机台、挂满墙的看板。两个工厂之间，这块区域的差别不大。
真正有差别的区域，通常在三个地方：配料间、检测室、留样柜。 其中最后那个，甚至经常不在参观路线上。
一条动线，五站半
建议这样安排审核顺序（这个顺序很重要，先前台后再后台，容易被"刚打扫过"误导）：
| 次序 | 站点 | 要回答的问题 |
|---|
| 一 | 资料核对 | 证书上的范围与实际产品是否一致 |
| 二 | 配料间 | 配方是怎么称出来的，精度靠什么 |
| 三 | 挤出产线 | 玻纤加入方式、回掺怎么管 |
| 四 | 检测室 | 自己能测什么，多久测一次 |
| 五 | 留样与追溯 | 一批货能不能倒追到基料批 |
| 半站 | 变更台账 | 配方改过没有，有没有通知过谁 |
最后那个"半站"是重头戏，后面单独讲。
二、配料间：看秤，不看秤的数量
配料是把几十公斤的树脂、玻纤、阻燃剂、色粉、助剂按配方称到一起的地方。改性料的稳定性，一半从这里开始。
三个值得蹲下来看的细节
其一，玻纤和色粉怎么进配方。
比较讲究的做法是失重式计量秤，玻纤单独一路侧喂，加入量按主机产量自动跟随。差一档的做法是把玻纤按比例预混后进主喂料斗——粒度不同、流动性不同，预混容易分层。
其二，秤的校准记录。
不要问"秤准不准"（没人会说不准），要看校准记录卡挂在哪个位置、上次校准是什么时候。
一张积了灰的记录卡，和一张上周刚填的记录卡，说明的不是一个东西。
其三，剩料的去向。
每批配料难免有剩余，这些剩料怎么处理？是登记回库、按批次参与下一批，还是随手倒进下一个配方里？
第三个答案是致命的——它意味着你的"专属配方"里可能混进了别人的料。
一句行话判定
问一句："你们的玻纤含量实际控制精度是多少？"
答得上 ±1.5% 以内并掏得出数据，说明配料是真在管；答不上或者含糊到"差不多就行"，基本可以判断配料环节是粗放型的。
三、挤出产线：看螺杆位置、真空口和回掺
挤出这一段，看三件事够用。
一看螺杆组合与侧喂位置
玻纤从哪个位置加入，决定了它在螺杆里挨多少剪切。加得早，分散好但玻纤被剪短；加得晚，玻纤留得长但分散差。
这两条路各有适用场景，关键是：这家工厂是不是清楚自己这个料号用的哪一种，为什么这么定。
能讲清楚取舍逻辑的，是配方工程师在管；讲不清楚的，多半是照着一个老方子在跑。
二看真空脱挥
抽真空口有没有在开、真空表读数是多少、口模有没有定期清理残留——这三项影响的是低分子物残留和气味，做汽车内饰件的对此特别敏感。
三看回掺
这是产线上最值得深问的一项。
每批开机有头料、收尾有尾料、换配方有过渡料。这些东西能不能回掺、回掺比例上限是多少、会不会进到客户的料里？
行内比较体面的规则：同牌号同批次的水口可以按规定比例回掺，而且要在记录里写明；不同牌号之间的过渡料一律降级处理，不得进正式订单。
问出这句话就够了：那么三个牌号切换的时候，中间的过渡料去哪了？
答"降级卖给做非受力件的客户"——这是有管理的工厂。 答不上来的，或者答"都用了"的，要认真想一想。
四、检测室：看设备清单，更看使用痕迹
检测室是最容易"看起来差不多"的地方，因为设备大家都买得起。
一组基础设备清单
做改性尼龙的工厂，下面这些东西属于及格线：
熔体流动速率仪：看流动性的基本盘
万能材料试验机：拉伸、弯曲
冲击试验机：简支梁或悬臂梁
灼热丝与漏电起痕试验仪：电气类客户的需求
水平垂直燃烧箱：阻燃等级
马弗炉：玻纤含量（灼烧法）与灰分
干燥箱与水分测定仪：含水率
注塑打样机：自己打标准样条
但这还不够。 真正区分水平的是这几项：能不能自己做热老化与湿热老化、有没有老化箱在跑、有没有一部老化箱同时挂着不同批次的留样。
看使用痕迹的三招
看设备上的运行记录本——一台三个月没动过的箱。
看试验机的夹具磨损——天天在用的机器，夹具接触面必有压痕。
看样条框里的存条——存条堆到什么程度、标注清不清楚，这一眼基本能看出检测频次。
一句总结：设备代表能力上限，使用痕迹代表实际水平。这两者之间的差距，往往是审核里最值钱的信息。
五、留样柜：这半小时最值得花
如果一天里只能认真看一个地方，我建议看留样。
为什么要看留样
因为留样是唯一一个没有当下利益驱动、纯粹为未来准备的动作。
给客户做的检测报告，有商业压力；设备的采购，有面子因素；只有留样这件事，没人看，也没人催。 一个工厂愿不愿意在没人看的地方花成本，本身就是结论。
具体看什么
看标签：有没有批号、生产日期、牌号、客户名。手写还是打印不重要，关键是信息全不全。
看留样期限：问一句"留多久"。答三个月以内的，追溯能力有限；能留一年以上的，说明这套体系是为追溯建的，不是为应付检查建的。
看密度与整齐度：这一项很主观，但我一直很信。柜子里塞得满满当当、按月份码齐的，和不规则地堆了半柜的，是两种管理状态。
看是常温柜还是恒温柜：有条件的工厂会把留样放在恒温恒湿柜里，这样一年后做对比才准。
一个可以现场做的动作
随手挑一个三个月前的留样批号，让对方当场把这个批号的完整记录调出来：
用了哪一批基料、哪一批玻纤、配料记录是谁签的字、当时出的是哪几项检测数据、发给哪个客户。
能在十分钟内调齐这五项的工厂，追溯体系是活的。 调不齐的，前面说再多也不作数。
六、变更台账：整个审核里最值钱的一问
最后回到开篇说的那"半站"。
这一问怎么提
不要问"你们会不会偷偷改配方"（没人承认）。建议这样提：
"过去两年，有没有做过配方调整或者主要原料换供应商的情况？当时是怎么通知客户的？"
这个问题回答的难度，恰恰在于它逼对方承认"我们改过东西"。
三种回答的含金量
最高一档：立刻拿出一份记录，上面写着某年某月因某个原料断供切换了某家上游货源，做了哪些验证，通知了哪些客户。有这种记录的工厂，把变更当成一件要留痕的事。
中间一档：口头承认改过，但拿不出书面。"当时打过电话说过。" 这类属于管理意识有，体系没有。
最低一档：一口否认，两年一次没改过。这一档要警惕。改性料的配方在两年里完全不动，在原料价格剧烈波动的年份几乎不可能。
为什么要这么较真
因为隐性切换是最难被发现、代价又最高的风险。
表面上牌号没变、性能数据还在范围内，但换了阻燃剂供应商、换了偶联剂品牌、换了玻纤厂的浸润剂体系——这些改动在常规检测里常常看不出来，要等半年后、客户端才现形。
这一问的价值，就在于它让对方知道：你是会追问的那一类客户。
这一句话本身，就是一种约束力。
七、一张可以在一天内跑完的行程与打分表
把上面这些拧成一天：
| 时段 | 内容 | 打分（一至五） |
|---|
| 上午前半 | 资料核对、体系文件 | |
| 上午后半 | 配料间 + 挤出产线 | |
| 午饭后 | 检测室 | |
| 下午前半 | 留样柜 + 现场调档实操 | |
| 下午后半 | 变更台账 + 技术访谈 | |
| 收尾 | 反馈与整改事项确认 | |
打分建议按二比三比三比二分到四块：配料与产线三成、留样追溯三成、变更管理两成、其余两成。
分数不是最重要的。 真正有用的是最后那二十分钟的反馈——把看到的三两处具体差距讲出来，看对方的反应：
立刻掏本子记、当场叫技术负责人过来的，可以进下一轮。 全程解释、"我们行业都这样"的，你自己判断。
八、去不了现场怎么办：五份能要到的文件
有些情况确实跑不成一趟——距离远、疫情后的差旅预算收紧、或者只是先做个初筛。
这时候不要退而求其次去看宣传册。可以换个思路：要数据，不要材料。
下面这五份东西，任何一家像样的工厂都拿得出，而且是很难临时造假的那种：
| 要什么 | 能看出什么 |
|---|
| 过去六个月同一牌号的出货检验数据 | 批次波动的真实幅度，比任何承诺都实在 |
| 一次延期或客诉的完整处理记录 | 出事时的反应速度与闭环方式 |
| 最小起订量与当前排产周期 | 真实产能裕度，也是议价空间所在 |
| 配方冻结政策（书面） | 多长时间内不得擅自变更，白纸黑字那种 |
| 是否有第二产地或备用产线 | 极端情况下的保供能力 |
第三项最容易被忽略，其实很说明问题。 一家排产周期常年在一周以内、最小起订量很低的工厂，通常意味着产能有余；反过来，开口就要三十天交期、起订一吨的，多半是产能真的紧，这时候压价压不下来，且插单基本没戏。
拿到数据后怎么看
六个月的出货数据摊开做三件事：看极差、看趋势、看有没有突然的跳变。
极差小说明过程受控；趋势单调往下要警惕（可能是偷偷降本）；突然的跳变——比如某一批之后强度整体抬了百分之十——那就值得追问：那个月动了什么？
这种问法比"你们质量稳不稳"有效得多。因为它问的是事实，不是态度。
审核结束后的跟踪比审核本身重要：把改性尼龙供应商的关键参数变更报备写进协议，审核才真正落地。
一句收拢
把判断写成表，把表发给改性尼龙供应商对答案，比电话里来回问省一半时间——这一篇就是那张表的底稿。
结语
我把话说得朴素一点：审核不是去找完美工厂，是去找"知道自己哪里不完美、并且愿意说出来"的工厂。
设备可以买、证书可以办、车间可以临时打扫。只有配料精度、批次追溯、变更自觉这三样，是花钱也买不来、只能靠时间养出来的东西。
我把这套行李箱版的审核清单——五站、二十多个具体观察点、含一份现场打分表——整理成了一页纸：
前两天接了个电话，是位不认识的采购打来的。开口就问：你们厂能不能让我看留样柜？
我说能。问留样柜的人，通常都在这行里吃过亏。真正内行的审核，从来不问你有多少台机器。
195 What to look at in a supplier audit for modified nylon
"Engineer Wu, next week we’re coming to your factory for a supplier audit, please prepare accordingly."
My usual first response is: "Okay. I have a checklist here; I’ll send it to you first so you can go through it, to avoid back-and-forth."
The person on the phone usually pauses for a moment. Because the form they’ve prepared—business license, system certificates, capacity, main customers—of course they have these, a ten-minute printout.
But what really determines whether you’ll have problems in the next two years is basically not on that form.
To put it more plainly: in audits, the easiest part is collecting all the certificates; the hardest part is judging whether a factory will act conscientiously without external pressure.
This article explains how to see that second part in one day.
First, a premise: this method suits modified plastic factories, especially materials like glass fiber reinforced, flame retardant, and color-compounded types that allow formulation flexibility. If you’re buying standardized resins, the things to check are different.
What do you look at in a supplier audit? On the surface, equipment and testing; inside, formulation and batch control; at the core, the depth of the team’s understanding of modified nylon processing. Equipment can be bought, understanding cannot.
1. The Most Common Audit Failure: Turning Into a Tour
Most audits become a mere formality because they turn into a tour.
The problem with a tour is: what you see is only what the other party wants you to see. Clean lobbies, polished machines, walls full of display boards. Between two factories, this area shows little difference.
The areas that truly make a difference are usually in three places: the ingredient room, the testing lab, and the sample storage cabinet. The last one, in particular, is often not even included in the tour route.
One Route, Five and a Half Stops
It is recommended to arrange the audit sequence like this (this sequence is very important; starting at the front and then going to the back can easily be misleading due to 'just cleaned' appearances):
| Order | Site | Questions to Answer |
|---|
| One | Data Verification | Whether the scope on the certificate matches the actual product |
| Two | Ingredient Room | How the formula is measured, what ensures the accuracy |
| Three | Extrusion Production Line | Method of adding fiberglass, how recycling is managed |
| Four | Testing Room | What can be tested in-house, how often it is tested |
| Five | Sample Retention and Traceability | Whether a batch can be traced back to the raw material batch |
| Half-Station | Change Ledger | Whether the formula has been modified, and who has been notified |
The last "Half-Station" is the main focus, which will be discussed separately later.
2. Ingredient Room: Observe the scales, not the quantity on the scales
The ingredient room is where tens of kilograms of resin, fiberglass, flame retardants, colorants, and additives are weighed together according to the formula. The stability of modified material starts here, for half of it.
Three Details Worth Squatting Down to Observe
First, how fiberglass and color powders enter the formulation.
The more meticulous approach is a loss-in-weight scale, with fiberglass fed separately on one line, and the amount added automatically following the main machine's output. A less precise method is to premix the fiberglass in proportion before putting it into the main feed hopper—different particle sizes and flowabilities make premixing prone to layering.
Second, the calibration records of the scale.
Don't ask "Is the scale accurate" (no one will admit it's not); look at which position the calibration record card is hanging and when it was last calibrated.
A dusty record card and one filled out just last week do not indicate the same thing.
Third, the destination of leftover materials.
There are inevitably leftovers from each batch of ingredients. How are these leftovers handled? Are they recorded and returned to inventory, used in the next batch by lot, or casually dumped into the next formulation?
The third approach is fatal—it means someone else's material could end up mixed into your "exclusive formulation".
Judging with a Phrase of Industry Lingo
Ask a question: "What is the actual control accuracy of your glass fiber content?"
If they can answer within ±1.5% and provide data, it shows that the dosing is truly under control; if they cannot answer or their answer is vague like 'roughly okay,' it can basically be judged that the dosing process is handled carelessly.
3. Extrusion Production Line: Look at Screw Position, Vacuum Port, and Reblending
For extrusion, focusing on these three things is sufficient.
First, check screw combination and side feeding position
Where the glass fiber is added determines how much shear it experiences in the screw. If added early, it disperses well but the fibers get shortened; if added late, fibers stay long but dispersion is poor.
Both methods have their applicable scenarios. The key question is: does this factory know which method to use for this material and why it was chosen?
Those who can clearly explain the trade-off logic are the formulation engineers in control; those who cannot are mostly just following an old recipe.
Second, check vacuum degassing
Whether the vacuum port is on, the vacuum gauge reading, and whether the die is regularly cleaned of residue — these three aspects affect low-molecular residue and odor, which is especially important for automotive interior parts.
Three Checks on Reblending
This is one of the most important things to investigate on the production line.
For each batch start, there is head material; for batch end, there is tail material; when changing formulas, there is transitional material. Can these materials be reblended? What is the maximum reblending ratio? Will it end up in the customer's materials?
The relatively proper rule in the industry: sprues of the same grade and batch can be reblended according to the specified ratio, and it must be recorded; transitional materials between different grades must be downgraded and must not go into official orders.
Just asking this is enough: so when switching between three grades, where did the transitional material go?
Answer: "Downgraded and sold to customers making non-load-bearing parts" — this is a factory with proper management. If one cannot answer, or answers "all was used," they need to seriously reconsider.
Four, Testing Room: Check the Equipment List, But Pay More Attention to Usage Marks
The testing room is the place that most easily looks "similar," because everyone can afford the equipment.
A Basic Equipment List
For factories making modified nylon, the following items are the baseline:
Melt Flow Indexer: basic test for flowability
Universal Testing Machine: tensile, bending
Impact Tester: simply supported or cantilever beam
Hot Wire and Tracking Resistance Tester: required by electrical clients
Horizontal and Vertical Burn Boxes: flame retardant levels
Muffle Furnace: glass fiber content (ashing method) and ash content
Drying Oven and Moisture Meter: water content
Injection Molding Sample Machine: to make standard test bars
But this is still not enough. What really differentiates skill levels are these: whether they can perform thermal and humidity aging on their own, whether aging ovens are running, and whether one aging oven has samples from different batches hanging simultaneously.
Three Tricks to Check Usage Marks
Look at the operation log on the equipment—a box that hasn’t been touched for three months.
Check the wear on the testing machine's fixtures—a machine used every day will definitely have indentations on the contact surfaces of the fixtures.
Look at the stored samples in the sample frame—how piled up the samples are, and how clearly they are labeled; at a glance, you can basically tell the frequency of inspections.
One summary sentence: Equipment represents the upper limit of capability, while usage marks represent actual level. The gap between these two is often the most valuable information in an audit.
5. Sample Cabinet: The Half Hour Most Worth Spending
If you can only seriously look at one place in a day, I recommend checking the samples.
Why Look at Samples
Because samples are the only action that is not driven by current interests, purely prepared for the future.
Testing reports for clients carry commercial pressure; equipment purchases involve face considerations; only sample retention is not seen by anyone, nor pressured. Whether a factory is willing to invest in areas unseen by others is itself a conclusion.
What to Specifically Look At
Check the labels: Are there batch numbers, production dates, brand numbers, customer names? Whether they are handwritten or printed is not important; the key is whether the information is complete.
Check the sample retention period: Ask one question, "How long is it kept?" If the answer is less than three months, the traceability is limited; if it can be kept for more than a year, it indicates that this system is built for traceability, not just to pass inspections.
Check density and neatness: This is very subjective, but I always trust it. A cabinet packed full and organized by month is different in management from a half-cabinet piled irregularly.
Check whether it is a room-temperature cabinet or a constant-temperature cabinet: Factories with the means will place samples in constant temperature and humidity cabinets, so that comparisons after a year are accurate.
An action you can do on-site
Randomly pick a sample batch from three months ago and have the other party pull up the complete record for this batch number on the spot:
Which batch of base material and which batch of fiberglass were used, who signed the formulation record, which tests were conducted at the time, and which customer it was sent to.
A factory that can gather these five items within ten minutes has an active traceability system. If it cannot, everything said before does not count.
6. Change Ledger: The most valuable question in the entire audit
Finally, return to the "half station" mentioned at the beginning.
How to ask this question
Do not ask "Will you secretly change the formula?" (no one will admit it). It is recommended to ask like this:
"In the past two years, has there been any formula adjustment or change of main raw material supplier? How was this communicated to the customers at the time?"
The difficulty in answering this question lies precisely in forcing the other party to admit "we have changed something".
The value of three types of answers
Top tier: Immediately provide a record showing that in a certain year and month, due to a supply shortage of a certain raw material, a certain upstream source was switched, what verifications were done, and which customers were notified. Factories with such records treat changes as something that must be documented.
Middle tier: Verbally admit changes, but cannot provide written proof. "At the time we called and mentioned it." This reflects management awareness but lack of system.
Lowest tier: Deny outright, saying nothing was changed in two years. Be cautious with this. Formulas of modified materials remaining completely unchanged in two years is almost impossible in years with significant raw material price fluctuations.
Why take this seriously
Because hidden changes are the most difficult to detect and the most costly risk.
On the surface, the grade hasn't changed, and performance data is still within range, but changing the flame retardant supplier, switching the coupling agent brand, or changing the fiberglass factory's sizing system—these changes often cannot be detected in routine tests and only appear six months later with the client.
The value of this question lies in letting the other party know: you are the type of customer who will follow up.
This sentence itself is a form of constraint.
7. An itinerary and scoring sheet that can be completed in one day
Compress the above tasks into one day:
| Time Slot | Content | Score (1 to 5) |
|---|
| Morning first half | Document verification, system files | |
| Late morning | Ingredient room Extrusion production line | |
| After lunch | Testing room | |
| Early afternoon | Sample storage cabinet On-site file adjustment practice | |
| Late afternoon | Change log Technical interview | |
| Wrap-up | Feedback and confirmation of corrective actions | |
The scoring recommendation is to allocate four parts according to 2:3:3:2: three parts for ingredients and production line, three parts for sample traceability, two parts for change management, and two parts for others.
The score is not the most important. What is truly useful is the last twenty minutes of feedback — point out two or three specific differences you noticed and observe the other party’s reaction:
Immediately take out your notebook and write it down, call over the technical person in charge on the spot, this can enter the next round. If they explain everything in detail, saying 'our industry is like this,' you make your own judgment.
8. What to do if you cannot go on-site: five documents you can request
In some cases, it's really not feasible to make a trip — the distance is far, post-pandemic travel budgets are tight, or you just want to do an initial screening first.
At this time, don’t settle for just looking at brochures. You can think differently: ask for data, not materials.
The following five items, any decent factory can provide, and they are very hard to fake on short notice:
| What to Ask For | What It Reveals |
|---|
| Shipment inspection data of the same product grade for the past six months | The real batch fluctuation, which is more reliable than any promise |
| Complete records of a delay or customer complaint handling | Response speed and closure method when issues occur |
| Minimum order quantity and current production scheduling cycle | Actual production capacity surplus, which is also room for negotiation |
| Formulation freeze policy (in writing) | How long changes are not allowed without authorization, the kind that is written in black and white |
| Whether there is a second production site or backup production line | Supply guarantee capability under extreme circumstances |
The third item is the most easily overlooked, and it actually tells a lot. A factory with a production cycle of less than a week year-round and a very low minimum order quantity usually means there is excess capacity. Conversely, if their lead time is 30 days and the minimum order is one ton, it often indicates that capacity is tight. In such cases, negotiating lower prices is difficult, and squeezing in additional orders is basically impossible.
How to Analyze the Data Once Received
Spread out the six months of shipment data to do three things: look at the range, check the trend, and see if there are sudden jumps.
A small range indicates controlled processes; a monotonically downward trend should be watched (it might suggest covert cost-cutting); sudden jumps — for example, if the strength as a whole increased by 10% after a certain batch — is worth questioning: what changed that month?
This way of questioning is much more effective than asking "Is your quality stable?" Because it asks for facts, not attitudes.
Follow-up after the audit is more important than the audit itself: putting the key parameter changes of modified nylon suppliers into the agreement is when the audit truly takes effect.
A Summary Statement
Write your judgments into a table, send the table to the modified nylon supplier to confirm answers, and you can save half the time compared to asking back and forth over the phone — this article is the draft of that table.
Conclusion
Let me put it simply: an audit is not to find a perfect factory, but to find a factory that "knows where it is imperfect and is willing to admit it."
Equipment can be bought, certificates can be obtained, workshops can be temporarily cleaned. Only the precision in formulation, batch traceability, and conscientious reporting of changes are things that money cannot buy and can only be cultivated over time.
I have organized this suitcase-version audit checklist — five stations, more than twenty specific observation points, including an on-site scoring sheet — into a one-page document:
A few days ago, I got a call from a procurement person I didn’t know. They opened by asking: Can your factory let me see the sample cabinet?
I said yes. People who ask about the sample cabinet usually have had setbacks in this field. Real professional audits never ask how many machines you have.
