I wonder if the accuracy is granted by the process itself,by testing samples,or automatic test one by one; the last would be amazing in case of very cheap and very small resistors(or other passive components).The same question could be extended to semiconductors,despite test on wafer is probably easier, by mechanical point of view,what about the sealed component? Thanks and forgive improper use of english language. Diego
Are small SMD resistors factory-tested one by one?
Started by ●September 20, 2020
Reply by ●September 20, 20202020-09-20
On Sun, 20 Sep 2020 19:28:06 +0200, blisca <blisca@tiscali.it> wrote:>I wonder if the accuracy is granted by the process itself,by testing samples,or automatic test one by one; >the last would be amazing in case of very cheap and very small resistors(or other passive components).The same question could be extended to semiconductors,despite test on wafer is probably easier, by mechanical point of view,what about the sealed component? >Thanks and forgive improper use of english language. >DiegoResistors are generally laser trimmed. Look at some under a microscope. Trimming requires that every resistor be measured. We only buy 1% resistors, and they are probably all trimmed. Apparently some sloppier parts, like 10%, aren't trimmed. Heck, a resistor that fails one bin can just go into the next one. But we never see bad resistors. Whatever the process, it sure works. An automatic test machine could measure resistors as fast as they can be processed. A millisecond is plenty of time to measure resistance. Youtube knows all. https://www.youtube.com/watch?v=wshRwO0MCSU -- John Larkin Highland Technology, Inc Science teaches us to doubt. Claude Bernard
Reply by ●September 22, 20202020-09-22
On Sunday, September 20, 2020 at 2:02:40 PM UTC-4, jla...@highlandsniptechnology.com wrote:> On Sun, 20 Sep 2020 19:28:06 +0200, blisca <bli...@tiscali.it> wrote: > > >I wonder if the accuracy is granted by the process itself,by testing samples,or automatic test one by one; > >the last would be amazing in case of very cheap and very small resistors(or other passive components).The same question could be extended to semiconductors,despite test on wafer is probably easier, by mechanical point of view,what about the sealed component? > >Thanks and forgive improper use of english language. > >Diego > Resistors are generally laser trimmed. Look at some under a > microscope. Trimming requires that every resistor be measured. > > We only buy 1% resistors, and they are probably all trimmed. > Apparently some sloppier parts, like 10%, aren't trimmed. Heck, a > resistor that fails one bin can just go into the next one. > > But we never see bad resistors. Whatever the process, it sure works. > > An automatic test machine could measure resistors as fast as they can > be processed. A millisecond is plenty of time to measure resistance. > > Youtube knows all. > > https://www.youtube.com/watch?v=wshRwO0MCSUI just got another 700+, 7" reels of resistors, capacitors and transistors. Most appear to be full reels. It is going to take a while to sort everything out.
Reply by ●September 22, 20202020-09-22
On Tue, 22 Sep 2020 07:32:48 -0700 (PDT), Michael Terrell <terrell.michael.a@gmail.com> wrote:>On Sunday, September 20, 2020 at 2:02:40 PM UTC-4, jla...@highlandsniptechnology.com wrote: >> On Sun, 20 Sep 2020 19:28:06 +0200, blisca <bli...@tiscali.it> wrote: >> >> >I wonder if the accuracy is granted by the process itself,by testing samples,or automatic test one by one; >> >the last would be amazing in case of very cheap and very small resistors(or other passive components).The same question could be extended to semiconductors,despite test on wafer is probably easier, by mechanical point of view,what about the sealed component? >> >Thanks and forgive improper use of english language. >> >Diego >> Resistors are generally laser trimmed. Look at some under a >> microscope. Trimming requires that every resistor be measured. >> >> We only buy 1% resistors, and they are probably all trimmed. >> Apparently some sloppier parts, like 10%, aren't trimmed. Heck, a >> resistor that fails one bin can just go into the next one. >> >> But we never see bad resistors. Whatever the process, it sure works. >> >> An automatic test machine could measure resistors as fast as they can >> be processed. A millisecond is plenty of time to measure resistance. >> >> Youtube knows all. >> >> https://www.youtube.com/watch?v=wshRwO0MCSU > > > I just got another 700+, 7" reels of resistors, capacitors and transistors. Most appear to be full reels. It is going to take a while to sort everything out.I, and the rest of my engineers, have been forbidden from taking parts out of the stock room. We have to send an email, and one of the production people bring us the parts. I don't like to wait, so we've bought a lot of various R and C and L sample kits for engineering, and I have a private stock of ICs and discretes. And there's always nights and weekends. I have a career-long record of successfully burgling stockrooms. -- John Larkin Highland Technology, Inc Science teaches us to doubt. Claude Bernard
Reply by ●September 22, 20202020-09-22
On Tuesday, September 22, 2020 at 11:07:23 AM UTC-4, John Larkin wrote:> On Tue, 22 Sep 2020 07:32:48 -0700 (PDT), Michael Terrell wrote: > > >On Sunday, September 20, 2020 at 2:02:40 PM UTC-4, John Larkin wrote: > >> On Sun, 20 Sep 2020 19:28:06 +0200, blisca wrote: > >> > >> >I wonder if the accuracy is granted by the process itself,by testing samples,or automatic test one by one; > >> >the last would be amazing in case of very cheap and very small resistors(or other passive components).The same question could be extended to semiconductors,despite test on wafer is probably easier, by mechanical point of view,what about the sealed component? > >> >Thanks and forgive improper use of English language. > >> >Diego > >> > >> Resistors are generally laser trimmed. Look at some under a > >> microscope. Trimming requires that every resistor be measured. > >> > >> We only buy 1% resistors, and they are probably all trimmed. > >> Apparently some sloppier parts, like 10%, aren't trimmed. Heck, a > >> resistor that fails one bin can just go into the next one. > >> > >> But we never see bad resistors. Whatever the process, it sure works. > >> > >> An automatic test machine could measure resistors as fast as they can > >> be processed. A millisecond is plenty of time to measure resistance. > >> > >> Youtube knows all. > >> > >> https://www.youtube.com/watch?v=wshRwO0MCSU > > > > > > I just got another 700+, 7" reels of resistors, capacitors and transistors. Most appear to be full reels. It is going to take a while to sort everything out. > > > I, and the rest of my engineers, have been forbidden from taking parts > out of the stock room. We have to send an email, and one of the > production people bring us the parts. I don't like to wait, so we've > bought a lot of various R and C and L sample kits for engineering, and > I have a private stock of ICs and discretes. > > And there's always nights and weekends. > > I have a career-long record of successfully burgling stockrooms.Me, too. The first time was in the Army. They wouldn't give me what I needed to keep a TV station on the air. In a three week period, I used 1400 line items, at a total of a little over 2100 parts. The station was still in good shape almost two years later when it was decommissioned and dismantled. I've talked to people who were there after I left for home. I have a room full of discretes. There were a lot of ROHS COG SMD capacitors in the last purchase. Most are under 1000pF. A real boon for people who play with RF. I have a local electroncs store that sells components to the public that I can trade parts with on a reel by reel basis. I also have dozens of 'Taiyo Yuden SMD/SMT Capacitor Kits' in small footprints.
Reply by ●September 22, 20202020-09-22
On Tue, 22 Sep 2020 09:59:34 -0700 (PDT), Michael Terrell <terrell.michael.a@gmail.com> wrote:>On Tuesday, September 22, 2020 at 11:07:23 AM UTC-4, John Larkin wrote: >> On Tue, 22 Sep 2020 07:32:48 -0700 (PDT), Michael Terrell wrote: >> >> >On Sunday, September 20, 2020 at 2:02:40 PM UTC-4, John Larkin wrote: >> >> On Sun, 20 Sep 2020 19:28:06 +0200, blisca wrote: >> >> >> >> >I wonder if the accuracy is granted by the process itself,by testing samples,or automatic test one by one; >> >> >the last would be amazing in case of very cheap and very small resistors(or other passive components).The same question could be extended to semiconductors,despite test on wafer is probably easier, by mechanical point of view,what about the sealed component? >> >> >Thanks and forgive improper use of English language. >> >> >Diego >> >> >> >> Resistors are generally laser trimmed. Look at some under a >> >> microscope. Trimming requires that every resistor be measured. >> >> >> >> We only buy 1% resistors, and they are probably all trimmed. >> >> Apparently some sloppier parts, like 10%, aren't trimmed. Heck, a >> >> resistor that fails one bin can just go into the next one. >> >> >> >> But we never see bad resistors. Whatever the process, it sure works. >> >> >> >> An automatic test machine could measure resistors as fast as they can >> >> be processed. A millisecond is plenty of time to measure resistance. >> >> >> >> Youtube knows all. >> >> >> >> https://www.youtube.com/watch?v=wshRwO0MCSU >> > >> > >> > I just got another 700+, 7" reels of resistors, capacitors and transistors. Most appear to be full reels. It is going to take a while to sort everything out. >> > >> I, and the rest of my engineers, have been forbidden from taking parts >> out of the stock room. We have to send an email, and one of the >> production people bring us the parts. I don't like to wait, so we've >> bought a lot of various R and C and L sample kits for engineering, and >> I have a private stock of ICs and discretes. >> >> And there's always nights and weekends. >> >> I have a career-long record of successfully burgling stockrooms. > >Me, too. The first time was in the Army. They wouldn't give me what I needed to keep a TV station on the air. In a three week period, I used 1400 line items, at a total of a little over 2100 parts. The station was still in good shape almost two years later when it was decommissioned and dismantled. I've talked to people who were there after I left for home. > >I have a room full of discretes. There were a lot of ROHS COG SMD capacitors in the last purchase. Most are under 1000pF. A real boon for people who play with RF. I have a local electroncs store that sells components to the public that I can trade parts with on a reel by reel basis. I also have dozens of 'Taiyo Yuden SMD/SMT Capacitor Kits' in small footprints.I never worked for a company so big that I couldn't walk to the stockroom and production floor and machine shop. I've known engineers who worked in a "office building" where there they didn't even have a lab. And I know outfits that send everything, even prototypes, out to a contract manufacturer. I could tell a funny story.
Reply by ●September 22, 20202020-09-22
On Tuesday, September 22, 2020 at 3:31:23 PM UTC-4, John Larkin wrote:> > On Tue, 22 Sep 2020 09:59:34 -0700 (PDT), Michael Terrell wrote: > > >On Tuesday, September 22, 2020 at 11:07:23 AM UTC-4, John Larkin wrote:> >> I have a career-long record of successfully burgling stockrooms. > > > >Me, too. The first time was in the Army. They wouldn't give me what I needed to keep a TV station on the air. In a three week period, I used 1400 line items, at a total of a little over 2100 parts. The station was still in good shape almost two years later when it was decommissioned and dismantled. I've talked to people who were there after I left for home. > > > >I have a room full of discretes. There were a lot of ROHS COG SMD capacitors in the last purchase. Most are under 1000pF. A real boon for people who play with RF. I have a local electroncs store that sells components to the public that I can trade parts with on a reel by reel basis. I also have dozens of 'Taiyo Yuden SMD/SMT Capacitor Kits' in small footprints. > I never worked for a company so big that I couldn't walk to the > stockroom and production floor and machine shop. I've known engineers > who worked in a "office building" where there they didn't even have a > lab. And I know outfits that send everything, even prototypes, out to > a contract manufacturer. > > I could tell a funny story.Tell it. I don't have a full machine shop, but I do have some nice tools for building prototypes. I havee a Cameron precision drill press, a floor model drill press and a small milling machine. My latest addition was a small lathe. They are in my 1200 Sq foot detached garage.
Reply by ●September 22, 20202020-09-22
On 9/20/2020 2:02 PM, jlarkin@highlandsniptechnology.com wrote:> On Sun, 20 Sep 2020 19:28:06 +0200, blisca <blisca@tiscali.it> wrote: > >> I wonder if the accuracy is granted by the process itself,by testing samples,or automatic test one by one; >> the last would be amazing in case of very cheap and very small resistors(or other passive components).The same question could be extended to semiconductors,despite test on wafer is probably easier, by mechanical point of view,what about the sealed component? >> Thanks and forgive improper use of english language. >> Diego > > Resistors are generally laser trimmed. Look at some under a > microscope. Trimming requires that every resistor be measured. > > We only buy 1% resistors, and they are probably all trimmed. > Apparently some sloppier parts, like 10%, aren't trimmed. Heck, a > resistor that fails one bin can just go into the next one. > > But we never see bad resistors. Whatever the process, it sure works. > > An automatic test machine could measure resistors as fast as they can > be processed. A millisecond is plenty of time to measure resistance. > > Youtube knows all. > > https://www.youtube.com/watch?v=wshRwO0MCSU >For the best quality you have to do both I believe. Testing each units rejects duds at the individual level, and statistical sampling to reject bias or errors in your trimming and measurement apparatus at the process level.
Reply by ●September 22, 20202020-09-22
On 9/20/2020 1:28 PM, blisca wrote:> I wonder if the accuracy is granted by the process itself,by testing > samples,or automatic test one by one; > the last would be amazing in case of very cheap and very small > resistors(or other passive components).The same question could be > extended to semiconductors,despite test on wafer is probably easier, by > mechanical point of view,what about the sealed component? > Thanks and forgive improper use of english language. > DiegoStatistical sampling can ensure whole batches are very close to perfect, like with prescription drug manufacturers. Testing pills is destructive, though, so obviously they don't test each one in that biz they do it statistically and set up the manufacturing process with hard limits so if it's going to be off it's going to be off-low. To ensure there are exactly 0 out-of-spec resistors in a batch you have to test each one there's no way around it and for precision resistors that's what they do, they're easier to test as individuals than pills. But I believe for the best quality they also do batch sampling because there can be errors in your overall process that just testing individuals may not detect. No "precision" manufacturing pr mechanical process stays precision forever, there's always some drift, and what happens if your manufacturing apparatus and test apparatus drift the same way, just sampling individuals won't detect that.
Reply by ●September 22, 20202020-09-22
