Back in the early 70's I read an article that talked of the importance of checking each resistor to verify it before using in any circuit that requires accuracy. Today, one would expect the manufacturing of the components to be higher quality then it was 45 years ago. But I guess that old advice still holds true today.
"But I guess that old advice still holds true today." Well today things are made in the hundreds of thousands, if not multi-millions. SO yeah, there is a chance that some of those will be defective.
Benjamin Esposti I understand what you are saying, but I disagree with you. It’s the manufactures responsibility to make sure their product is as advertised. It doesn’t matter if it’s quad trillion pieces that are produced. Otherwise the problem would get worse and worse.
I recently purchased some resistors that were an order of magnitude off from spec. Correct bands, correct packaging, wrong resistor. Could have blow something! I think I'm going to check all components upon receiving them in the future. Looks like it's not an impossibility even today.
@@BenjaminEsposti -- But, given how good mass production is these days, the percentage of defective resistors should be remarkably small, perhaps in the hundred-thousandths.
Obviously this is an old post but this is precisely why the practice if removing the leading zero from values is a dangerous one. Ive fought this problem for years with purchasing agents, inventory clerks, engineers. Etc. That tiny little decimal point vanishes far too easily. But that big fat Zero doesn't.
After watching a bunch of these videos from almost a decade ago, its just dawned on me how amazing the video quality was in these videos, considering it was 2010.... Even in 2019 this is great
Well I'm sure Digi-Key will hear about this. Hah! Their buyers thought they knew how to avoid being snookered by Chinese alleyway manufacturers of fake components. Reminds me of Radio Shack's replacement semi's in the '70's. They were Texas Instruments culls. Just like then, and as it has always been, the QC department is right on your work bench, Dave
Dave Thanks for the excellent tutorial. I've been in the electronics manufacturing industry for 25 years and learned that you ALWAYS test ALL your product specs during test and cal. That's been learned the hard way just as you have experienced. It does increase the cost slightly though.
I think this video shows that you have to verify everything yourself. Once you take ownership of those parts you are ultimately responsible for them and the results they produce in the equipment they get installed in. If you're going to sell a product you should also test your product to make sure it meets the specs you are guaranteeing to you customers. It's more costly not only in recalls and refunds but in reputation in the long run to sell dodgy gear.
Jud Clark It isn't practical to verify the detailed specs of every component supplied. Not if you are doing more than a one off prototype. You pay a premium to have them tested before supply.
@@MrSwanley - Why do you think, nomal manufacturers use bed of nails? Of course you have to test your product and materials. Your customers do not care what is practical for you. They care what is possible and that they get what you advertised.
Digikey has always been solid for me. No component problems here In fact I just placed an order for some supplies that were supposed to ship UPS 2 day and they accidentally were keyed in on their end as UPS Ground. I needed one of the items fast (small BUD enclosure) and they shipped it to me free overnight at no charge. I can't recommend them enough but I understand how something like this would be irritating.
It's a risk that happens when Digi-key breaks down a manufacturer's packaging and puts a lot of resistors into their own packaging. They likely filled that packet from the wrong spool of components. The best you can do to work around a mistake like this is to only buy full spools that were straight from the mfg.
@yanava Not in this case. The resistor is isolated with a mechanical switch. And the resistor was probed directly so traces have no effect. Capacitance does not matter for DC resistance measurements.
I would absolutely love any one of those really bang on meters, Dave’s got about twenty of the things just kicking about in his lab! The man’s got some kit!
Depending on how they are made they my be static sensitive...?! We had a class on static at one of our Western Union computer schools. They related a story of how a particular lot of super precision resistors made by a German company were all being measured slightly off when tested on the bench. Believe it was NASA, not sure. This was back in the late 1960's, before the era of anti static procedures & packaging, and ever shrinking conductor spacing. When they were removed from the paper envelope the Germans had packed them in, they picked up a static charge that was enough to arc across between two conductors, dragging a bit of molten metal from one to the other, shorting out a turn. That began the development of anti-static procedures, or so we were told.
If he wanted 1.5% error components he would have spec'ed it out for that and he would have bought components to match those tolerances. A 10 ohm 0.1% SMD part costs $101.18 for a batch of 1000, whereas a 10 ohm 1% SMD part costs $12.59 for a batch of 1000 from digikey.
@EEVblog On Eckert building the ENIAC:"Whole boxes of resistors were brought in and their tolerances checked one by one.... The best resistors were kept to one side for the most critical parts of the machine wile others were used in less sensitive parts. On other occasions whole boxes...would be rejected and shipped back to the manufacturer." ---Computer, a History of the Information Machine, 1996 There's nothing new under the sun.
I remember doing 4-wire resistance measurement on my introduction to metrology lab, we got 0.1 ohm reference resistor (one of these old, crusty big ones in big cylinder-shaped boxes) and we got 0.2 on 2-wire and 0.1 on 4-wire, that was a good eye-opener for someone trusting all you can get on the multimeter display (I was an IT student and this lab was kinda funny because most of students didn't even know an Ohm's law...) :)
I'm on Information Science, we have one electronics course, and we have to make a few measurements like resistance of a bulb (with voltage and ampere meter). Maybe there were more complicated too, but I haven't take that course, maybe next year, but my friend have already did. And not optional. And we had a PLC and other control stuff too, RS232, current loop, ladder programming, and who knows what. Studends didn't care, teacher didn't care, so I brought my arduino and did something less useless. 90% of class is like my grandmother to computers. But not beacuse they don't know what a phone charger does (220V- >5V), we tried to make a remote door plan (like we need door close sensor, door open sensor, something in the way sensor, remote controller, etc, if this does that, blablabla). The possible exam questions was sent out, and the project was make a traffic light, with the built in leds of the PLC dev kit, or with the "IDE" simulator INSTED OF making a small pneumatic assembly line piece that they told us. That wouldn't be more complicated but it would be interesting for the "noobs". Because of this nearly everybody borrowed the project file and sent in the same.
@CampKohler No, no incoming component inspection apart from labeling, nothing unusual there, very common procedure. You build up confidence in your system and supply chain, but sometimes it lets you down like this. The antistatic mat is just that, anti-static, or what's called "static dissipative", it is not conductive.
Digikey has offered me a job on at least 10 occasions. They want me to work as a sales engineer. This is where lots of American EEs wind up after the age of 50. If their call center was not in Thief River Falls Minnesota, I might consider it. The average winter temperature there is -3 F. PASS! I build boutique guitars, guitar amps, bass guitar amps, & other audio amplification & signal processing products. I usually do test every component that goes into my stuff. If I need 3 of something, I buy 5 of them. If I need one, I buy 3. I select the best part, and keep the rest for repairs. I stay so busy with repairs my building has kind of stopped for the time being. I'd be hyper pissed if those parts were not within 1/5th of a percent. I keep my crap decade boxes at +/-3%. I keep my calibration standards at 0.05% maximum. I use a 4 lead Kelvin clip leads to test the tiny resistances, and use a $5000 LCR as my meter. It is accurate down to micro-Ohms, and reads up into 100s or GigOhms. I feel for Dave having paid postage from the USA, and getting a bum batch of parts.
I know this comment is several years old, but I want to mention that I often ask Digikey sales and tech support engineers how the weather is in Thief River Falls, only to find that they actually live in some other state, and often have separate 'day jobs' and just log into Digikey to work a few hours at night or on the weekend and pick up some extra cash.
DigiKey has send the wrong part twice during various orders (~1% screw-up rate) and both times they squared me away but seemed like it was business as usual over the phone. I am trying different vendors such as Mouser to see if I get better results. This video and my prior experiences have strengthened my resolve to do parametric testing for each lot of parts received and to perform small initial builds that are to be tested extensively.
I had a weird experience with my last order. Got some tiny SMT varactors on cut-tape, and the tape only had EVERY OTHER spot filled!! So, my quantity was 50% short! At least they have good customer service when there's a problem.
@EEVblog Dave, I agree that in this case it's ok. But it would be good advice to tell people that measuring in circuit would cause problems. If you did this in a previous blog, then it's ok. Capacitance could affect the measurement because as long as you keep the meter current source, parasitic capacitance will provide voltage across (integration of applied current). It's quite marginal, but it's there. Great job on this blog, keep it up mate!
I got caught out with some resistors from Farnell that weren't even the right value ! The bag was marked right but the resistors were wrong. Farnell didn't want to know, they just said they would remove the wrong resistors from stock. Had to remove a few resistors and put in the right value.
Having once identified numerous problems on a stack of newly manufactured micro-controller boards to be due to a bag of 5000 1N4148 diodes with the cathode band printed on the anode end, I never trust anything going into a new build. Now that we are in the depths of manufacturing, labor, and shipping shortages and difficulties it becomes ever more important to verify EVERYTHING.
An old question: How many 5% resistors will be found in a lot of 20% resistors? None. All the 5% units are sold as 5%, the 10% are sold as 10% and the rest are sold as 20%. Nowadays probably any units that are further out are returned to the worker that produced them, who is shot by the Chinese army ("Quality control that works!).
Digi-Key is dead to me. I have BOMs on their server to order parts for production. Two out of the last three orders had errors in Digi-Reeled components. In one case, a different three-terminal device was substituted for a SOT-23-pkg transistor, and in the more recent one, they reeled the wrong resistors (0603s vs 0805s). It's unfortunate--we've used and liked DK for years, and their BOM/reeling services were a great fit when we brought some manufacturing back from China. But these errors...
@ubuntututorials I don't know. This series of resistors are at worst 1% tolerance, so the manufacturing tolerances would be pretty tight. Something must have gone horribly wrong to get several percent out. It's more likely there ha been a mixup in the labeling or something. Will have to wait and see.
@EEVblog How could it be static dissipative without being conductive? Is it radioactive (ala polonium brushes for film, etc.)? All the mats that I am familiar with are conductive in the multi-megohms/square range and are tested with ohmmeters that measure in that range. There are even go/no-go circuits that monitor the conductivity during use and give off an alarm if it no longer conducts, which is neat.
What did Digikey say and/or how did they respond to this? Also it's funny that you mention using two multimeters for measuring with four leads. I did that at work once when I had to check out a 1.5 watt 1 2/3 Ohm resistor (amp meter shunt 30 Amps 50 milivolts) and a power supply that can only go up to 5 Amps.
I would just like to add to the explanation of the 4 wire measurement. One key benefit of the 4 wire (kelvin) measurement is the special 4 wire clips actually make 4 separate points of contact. in contrast to the statement that the wires join at the very tip of the probe. The difference is that the contact resistance is negated by making four separate points of contact. This can be significant for high precision measurements. In the two wire delivering the constant current source, the contact resistance does not matter. It is a current source after all. This leaves the sensing wires circuit, which has a very low current due to the high input impedance of the meter. So the voltage drop due to the contact resistance of this circuit is V=IR .... = almost nil current times what ever the resistance equals almost nil voltage drop and high accuracy. If the proper 4 contact points are not used then the significant current of the current source times whatever the contact resistance is equal to a non negligible voltage drop. A second comment for everyone out there that does not have a 4 wire ohm meter is that just use a separate current source and a standard volt meter and you can make 4 wire kelvin measurements. Additionally depending on what you are measuring, you can use a higher current to improve the signal to noise ratio and get better accuracy as long as the current is not enough to heat the parts under test.
You are more educated than me in electrical theory and practice, and I love your videos. Thank you for taking the time to share with everyone and for explaining as you go. I have learned a lot from you. Is there a provision on your microcurrent board for calibration? I didn't see it in the schematic, but could possibly mitigate problems with components out of spec.
@QBMan My second incident in almost as many days. I had actually already shot another blog about another problem with dodgy Maxim chips, but the Digikey one got bumped up the list.
Digikey has gotten much worse in the past few years about sending completely wrong parts - the label will indicate the correct part, but the bag will contain a completely different part. Not just a different value - but a different part! In my experience, they screw up about 20% of the time. I place about 50 orders per year.
I recently (2019) purchased a component that was an order of magnitude out. Banding was correct, packaging was correct, part number was correct, but the thing was not what all that said. Always check your components prior to installation.
I know how you feel Dave. I had gotten a Samsung 52inch DLP TV from my wifes sister because it had just stoped working. No surprise it was the voltage cap on the balast baord so I ordered a new one from digikey. 450v 47uf had it replaced with a 400v 47uf cap from an abandon switch mode supply and that worked for 2.5 months. I put the new cap in from Digikey and it didn't last a week. Had a new cap overnighted to me and it lasted 3 days! So I said screw it and just got a used board from eBay. =p
Very nice tutorial on 4-terminal measurements. Thank you. Sorry about you getting bent on the component values. Even if they reimburse you, there are still compromised boards out in the wild now.
I sat through a sales presentation where the rep referred to them as Kevin measurements. Nobody said a word. But you could see an occasional twitch in the corners of the engineers eyes.
Dave, where do you get your boards fabricated and how much does it cost per board. I cant find any place that doesnt fabricate small boards for less than $120 except for Seeed Studio which goes for 35 which is a good deal.
I bought some components from Digi-key and they were way out of spec. Not only that but one of the bags was empty, it had the label on it, but no components in the bag. and seeing they were expensive devices I was not amused.
So Dave, how do you get around this problem?? I am building a BLDC position controller and I will be getting it sent off to get made, my first run will be 1000 and I don't want this problem to happen to me. Do I order one of each component and test all of them first then get it built with the same items??
@CampKohler Antistatic mats have surface resistances of at least hundreds of millions of ohms per inch, specifically so that it will not interfere with stuff like a component being tested or a PCB with live traces in contact with the mat or whatever.
Love your videos, Dave. Just a tip for probing those tiny 0805's (and 0603's, its especially handy for checking LED orientation) - leave it in the tape and, using sharp probes, probe the component through the tape.
I've had instances where I ordered a small part and Digikey sent me some eval board instead. They constantly send out wrong parts. I only had one order in the past three years where they sent out everything correctly, and that order only had one item.
@DJMC5ive Well, that's what "going green" is really all about. I saved about $80 in antistatic packaging material on my last production shipment by reusing what one supplier had sent their parts in. Recycling, going green. Whatever. It's "green" (USA dollars are green) in my pocket.
+KpilotRCHelis What if you want to know how much current a watch use? or how much current a microcontroller use in sleep mode? It may not look much, but if you have a tiny battery then you need a tiny current drain to get a long battery life.
The only problem I've had with ordering from Digikey is when I once received some transistors when I actually ordered microcontrollers. I called them and they had someone check the bin in the warehouse and--sure enough--they were stocked wrong. Digikey sent me the right parts and let me keep the transistors. My only real complaint is that the lady in customer service was really cranky about the whole affair.
sounds like someone was caught charging premium prices for an inferior product and it bit em in the ass...they were relying on the fact that most people wont verify to the extent that dave did....
Are they in fact thick film ones? Try warming it up with a soldering iron and check how the resistance varies? Thick film have at least 100ppm temp co, thin film should be considerably better.
I had a major recall incident with ECS crystals ECS-160-20-3X-TR from Mouser. These crystal would suddenly stop oscillating. Real bad news because while it did the quadcopters' motors would spin up full throttle even though the motors were in a "disarmed" mode. Intermittent, hard to diagnose. Assumed a reflow problem for a long time but turned out to be bad parts. Replaced with an ABM7 crystal and not an issue since. I recalled 220 units, which cost me a bundle for a small biz. @flyingeinstein
That's a good point. I'm wondering if it's possible that these resistors were spot on immediately after manufacture, but drifted over time due to some obscure reason?
You realize measuring components in circuit is a bit lame right? There are resistances in contact, trails resistance, capacitance and all sorts of stuff. You did it correctly by measuring the resistor out of the package.
Just so you know Digikey employs as bunch of rural clueless Minnesota housewives who don't know a resistor from a capacitor. For anyone who's seen the movie Fargo that's pretty much the Digikey workforce. Among the major online component distributors Digikey usually comes in dead last despite having one of the better website search engines. I once ordered 200 surface mount resistors as cut tape and Digikey shipped 20 bags with 10 resistors in each bag which is useless to any pick and place SMT machine on the planet. When I called to ask for 200 resistors on a single tape none of three people I talked to could understand why what they shipped was a problem. Digikey=Fail. I have had far fewer problems with Mouser, Avnet, Arrow, Future, etc. Digikey is the Walmart of components and, like Walmart, their employees mostly don't have a clue. Buyer beware. They might as well be some Chinese supplier on eBay.
Er... Did you call them /before/ placing your order? From their site: "Digi-Key tries to ship in one strip when possible however, if you require your order to be one continuous strip, this should be requested at the time you place your order." If you didn't, that's on you, not them. It's understandable why that'd be frustrating if you didn't know it beforehand, but you can't make assumptions and blame others when they turn out to be wrong. Another assumption you're making here is that the other distributors don't do the same thing. Mouser did that to me with some ceramic caps, and they don't even directly say they do it on their site. It's only said indirectly by specifying their custom reels will be continuous, with the unsaid implication that normal cut tape orders might be discontinuous.
You can always order your parts on a Digi-Reel; it's a continuous strip with as many parts as you'd like, and comes on a reel for the pick and place machine.
You probably dont give a shit but if you guys are stoned like me atm you can stream pretty much all of the latest series on InstaFlixxer. Been binge watching with my girlfriend recently xD
I'm already triggered by the ".1%" notation. Is it that hard to print the extra zero? What if the ink is low and the dot gets lost in the print? "01%" is quite obvious and you at least question it, rather than "1%" which looks perfectly fine.
Do you realize how many part numbers Digikey carries? You can't fault them for a single manufacturer's batch of 10 ohm resistors that were out of spec.
Mallom Bum They are suppliers, not manufacturers. Therefore, NO you cannot fault them for a few units out of millions being out of spec. If you think so, you should pack up a multimeter and head over there and beg for a job testing each and everything that goes out the door.
***** I can't blame digikey for getting it wrong some times, and it's not at all reasonable for them to test every single 0402 resistor on a reel of 10,000 :D At the same time, when they get parts wrong they are pretty good about setting it right. I've had them ship me some obvious mistakes like several thousands of dollars worth of eeproms when I had ordered $100 worth of barometers - they were very polite that time because they wanted their parts back unharmed :D The worst error I've had from them was getting 5.76k ohm resistors labeled as 57.6k ohm - luckily the power supply circuit wasn't damaged but was clearly non functional so it was easy to diagnose - I've luckily never run in to such a subtle error as this one.
So, I assume the lesson is - do spot-checks on batches of supplies? I mean, not that genius a thought, but I guess sometimes you really do need to get burned once to avoid this happening in the future
Back in the early 70's I read an article that talked of the importance of checking each resistor to verify it before using in any circuit that requires accuracy. Today, one would expect the manufacturing of the components to be higher quality then it was 45 years ago. But I guess that old advice still holds true today.
"But I guess that old advice still holds true today."
Well today things are made in the hundreds of thousands, if not multi-millions. SO yeah, there is a chance that some of those will be defective.
Benjamin Esposti I understand what you are saying, but I disagree with you. It’s the manufactures responsibility to make sure their product is as advertised. It doesn’t matter if it’s quad trillion pieces that are produced. Otherwise the problem would get worse and worse.
I recently purchased some resistors that were an order of magnitude off from spec. Correct bands, correct packaging, wrong resistor. Could have blow something! I think I'm going to check all components upon receiving them in the future. Looks like it's not an impossibility even today.
@@BenjaminEsposti -- But, given how good mass production is these days, the percentage of defective resistors should be remarkably small, perhaps in the hundred-thousandths.
I sill do it for prototypes. But for manufacturing just cross your fingers.
Obviously this is an old post but this is precisely why the practice if removing the leading zero from values is a dangerous one. Ive fought this problem for years with purchasing agents, inventory clerks, engineers. Etc. That tiny little decimal point vanishes far too easily. But that big fat Zero doesn't.
George Carlin joke: "Everybody worries about all hell breaking loose. What if just a little piece broke loose and nobody noticed?"
After watching a bunch of these videos from almost a decade ago, its just dawned on me how amazing the video quality was in these videos, considering it was 2010.... Even in 2019 this is great
Yep, but his lighting was atrocious in old videos. Still 10x better than most vids from the 2010 era tho.
Well I'm sure Digi-Key will hear about this. Hah! Their buyers thought they knew how to avoid being snookered by Chinese alleyway manufacturers of fake components. Reminds me of Radio Shack's replacement semi's in the '70's. They were Texas Instruments culls. Just like then, and as it has always been, the QC department is right on your work bench, Dave
Dave
Thanks for the excellent tutorial.
I've been in the electronics manufacturing industry for 25 years and learned that you ALWAYS test ALL your product specs during test and cal. That's been learned the hard way just as you have experienced. It does increase the cost slightly though.
The level of double triple checking you go through is outstanding, great stuff no mucking around at EEV!
I think this video shows that you have to verify everything yourself. Once you take ownership of those parts you are ultimately responsible for them and the results they produce in the equipment they get installed in. If you're going to sell a product you should also test your product to make sure it meets the specs you are guaranteeing to you customers. It's more costly not only in recalls and refunds but in reputation in the long run to sell dodgy gear.
Jud Clark It isn't practical to verify the detailed specs of every component supplied. Not if you are doing more than a one off prototype. You pay a premium to have them tested before supply.
@@MrSwanley - Why do you think, nomal manufacturers use bed of nails? Of course you have to test your product and materials.
Your customers do not care what is practical for you. They care what is possible and that they get what you advertised.
Digikey has always been solid for me. No component problems here In fact I just placed an order for some supplies that were supposed to ship UPS 2 day and they accidentally were keyed in on their end as UPS Ground. I needed one of the items fast (small BUD enclosure) and they shipped it to me free overnight at no charge.
I can't recommend them enough but I understand how something like this would be irritating.
Are there crickets in your lab? Sounds like you have bugs.
+Picobyte 'Strailya!
It's a risk that happens when Digi-key breaks down a manufacturer's packaging and puts a lot of resistors into their own packaging. They likely filled that packet from the wrong spool of components. The best you can do to work around a mistake like this is to only buy full spools that were straight from the mfg.
@yanava Not in this case. The resistor is isolated with a mechanical switch. And the resistor was probed directly so traces have no effect. Capacitance does not matter for DC resistance measurements.
The customer is always the final QA/QC inspector.
I would absolutely love any one of those really bang on meters, Dave’s got about twenty of the things just kicking about in his lab! The man’s got some kit!
Depending on how they are made they my be static sensitive...?! We had a class on static at one of our Western Union computer schools. They related a story of how a particular lot of super precision resistors made by a German company were all being measured slightly off when tested on the bench. Believe it was NASA, not sure. This was back in the late 1960's, before the era of anti static procedures & packaging, and ever shrinking conductor spacing. When they were removed from the paper envelope the Germans had packed them in, they picked up a static charge that was enough to arc across between two conductors, dragging a bit of molten metal from one to the other, shorting out a turn. That began the development of anti-static procedures, or so we were told.
If he wanted 1.5% error components he would have spec'ed it out for that and he would have bought components to match those tolerances. A 10 ohm 0.1% SMD part costs $101.18 for a batch of 1000, whereas a 10 ohm 1% SMD part costs $12.59 for a batch of 1000 from digikey.
@EEVblog On Eckert building the ENIAC:"Whole boxes of resistors were brought in and their tolerances checked one by one.... The best resistors were kept to one side for the most critical parts of the machine wile others were used in less sensitive parts. On other occasions whole boxes...would be rejected and shipped back to the manufacturer." ---Computer, a History of the Information Machine, 1996
There's nothing new under the sun.
There is no excuse for any company with that sh_tty of a mass production failure rate to still be in business.
the emperor of chin could behead the maker of out of tolerance arrowheads
I remember doing 4-wire resistance measurement on my introduction to metrology lab, we got 0.1 ohm reference resistor (one of these old, crusty big ones in big cylinder-shaped boxes) and we got 0.2 on 2-wire and 0.1 on 4-wire, that was a good eye-opener for someone trusting all you can get on the multimeter display (I was an IT student and this lab was kinda funny because most of students didn't even know an Ohm's law...) :)
kroplaaaa just remember, multimeters do not measure resistance, they are actually measuring time in the D-A conversion.
D-A conversion? Typical multimeter is double integrating ADC and it actually counts impulses during which capacitor in the integrator discharges ;)
Since when did IT studies include metrology, or was that an optional subject you could take for your grade point count ?
I'm on Information Science, we have one electronics course, and we have to make a few measurements like resistance of a bulb (with voltage and ampere meter). Maybe there were more complicated too, but I haven't take that course, maybe next year, but my friend have already did. And not optional.
And we had a PLC and other control stuff too, RS232, current loop, ladder programming, and who knows what. Studends didn't care, teacher didn't care, so I brought my arduino and did something less useless. 90% of class is like my grandmother to computers. But not beacuse they don't know what a phone charger does (220V- >5V), we tried to make a remote door plan (like we need door close sensor, door open sensor, something in the way sensor, remote controller, etc, if this does that, blablabla). The possible exam questions was sent out, and the project was make a traffic light, with the built in leds of the PLC dev kit, or with the "IDE" simulator INSTED OF making a small pneumatic assembly line piece that they told us. That wouldn't be more complicated but it would be interesting for the "noobs". Because of this nearly everybody borrowed the project file and sent in the same.
Those wires for the 4W measurement looked kinda thin. I'm guessing you would get even better accuracy with low resistivity cables?
@CampKohler No, no incoming component inspection apart from labeling, nothing unusual there, very common procedure. You build up confidence in your system and supply chain, but sometimes it lets you down like this.
The antistatic mat is just that, anti-static, or what's called "static dissipative", it is not conductive.
So Dave, what was the outcome of this? Did Digikey say anything useful? How about Bournes?
Ha! I've fallen victim to the "Comment on YT, make a G+ post" trap.
Digikey has offered me a job on at least 10 occasions. They want me to work as a sales engineer. This is where lots of American EEs wind up after the age of 50. If their call center was not in Thief River Falls Minnesota, I might consider it. The average winter temperature there is -3 F. PASS! I build boutique guitars, guitar amps, bass guitar amps, & other audio amplification & signal processing products. I usually do test every component that goes into my stuff. If I need 3 of something, I buy 5 of them. If I need one, I buy 3. I select the best part, and keep the rest for repairs. I stay so busy with repairs my building has kind of stopped for the time being. I'd be hyper pissed if those parts were not within 1/5th of a percent. I keep my crap decade boxes at +/-3%. I keep my calibration standards at 0.05% maximum. I use a 4 lead Kelvin clip leads to test the tiny resistances, and use a $5000 LCR as my meter. It is accurate down to micro-Ohms, and reads up into 100s or GigOhms. I feel for Dave having paid postage from the USA, and getting a bum batch of parts.
I know this comment is several years old, but I want to mention that I often ask Digikey sales and tech support engineers how the weather is in Thief River Falls, only to find that they actually live in some other state, and often have separate 'day jobs' and just log into Digikey to work a few hours at night or on the weekend and pick up some extra cash.
DigiKey has send the wrong part twice during various orders (~1% screw-up rate) and both times they squared me away but seemed like it was business as usual over the phone. I am trying different vendors such as Mouser to see if I get better results. This video and my prior experiences have strengthened my resolve to do parametric testing for each lot of parts received and to perform small initial builds that are to be tested extensively.
I had a weird experience with my last order. Got some tiny SMT varactors on cut-tape, and the tape only had EVERY OTHER spot filled!! So, my quantity was 50% short! At least they have good customer service when there's a problem.
Digikey cannot check every part they sell. The fault is clearly with the point of origin.
They should check the product with at least the same dedication as Dave, right ?
And he didn't catch it either.....
the least DigiKey can do is to provide you free components for a year. Excellent investigation, thanks for sharing !
99,9 problems but a resistor aint one!
Finni M. You mean - 99.9 problems and a resistor ain't zero point one
Finni M. Hit me!
+Roastysprouts *SMACK* Happy?
+Finni M. The internet approves.
@EEVblog Dave, I agree that in this case it's ok. But it would be good advice to tell people that measuring in circuit would cause problems. If you did this in a previous blog, then it's ok. Capacitance could affect the measurement because as long as you keep the meter current source, parasitic capacitance will provide voltage across (integration of applied current). It's quite marginal, but it's there. Great job on this blog, keep it up mate!
I got caught out with some resistors from Farnell that weren't even the right value ! The bag was marked right but the resistors were wrong. Farnell didn't want to know, they just said they would remove the wrong resistors from stock. Had to remove a few resistors and put in the right value.
Having once identified numerous problems on a stack of newly manufactured micro-controller boards to be due to a bag of 5000 1N4148 diodes with the cathode band printed on the anode end, I never trust anything going into a new build. Now that we are in the depths of manufacturing, labor, and shipping shortages and difficulties it becomes ever more important to verify EVERYTHING.
An old question: How many 5% resistors will be found in a lot of 20% resistors? None. All the 5% units are sold as 5%, the 10% are sold as 10% and the rest are sold as 20%. Nowadays probably any units that are further out are returned to the worker that produced them, who is shot by the Chinese army ("Quality control that works!).
Digi-Key is dead to me. I have BOMs on their server to order parts for production. Two out of the last three orders had errors in Digi-Reeled components. In one case, a different three-terminal device was substituted for a SOT-23-pkg transistor, and in the more recent one, they reeled the wrong resistors (0603s vs 0805s).
It's unfortunate--we've used and liked DK for years, and their BOM/reeling services were a great fit when we brought some manufacturing back from China. But these errors...
@ubuntututorials I don't know. This series of resistors are at worst 1% tolerance, so the manufacturing tolerances would be pretty tight. Something must have gone horribly wrong to get several percent out. It's more likely there ha been a mixup in the labeling or something. Will have to wait and see.
@EEVblog How could it be static dissipative without being conductive? Is it radioactive (ala polonium brushes for film, etc.)? All the mats that I am familiar with are conductive in the multi-megohms/square range and are tested with ohmmeters that measure in that range. There are even go/no-go circuits that monitor the conductivity during use and give off an alarm if it no longer conducts, which is neat.
What did Digikey say and/or how did they respond to this? Also it's funny that you mention using two multimeters for measuring with four leads. I did that at work once when I had to check out a 1.5 watt 1 2/3 Ohm resistor (amp meter shunt 30 Amps 50 milivolts) and a power supply that can only go up to 5 Amps.
I would just like to add to the explanation of the 4 wire measurement. One key benefit of the 4 wire (kelvin) measurement is the special 4 wire clips actually make 4 separate points of contact. in contrast to the statement that the wires join at the very tip of the probe. The difference is that the contact resistance is negated by making four separate points of contact. This can be significant for high precision measurements. In the two wire delivering the constant current source, the contact resistance does not matter. It is a current source after all. This leaves the sensing wires circuit, which has a very low current due to the high input impedance of the meter. So the voltage drop due to the contact resistance of this circuit is V=IR .... = almost nil current times what ever the resistance equals almost nil voltage drop and high accuracy. If the proper 4 contact points are not used then the significant current of the current source times whatever the contact resistance is equal to a non negligible voltage drop. A second comment for everyone out there that does not have a 4 wire ohm meter is that just use a separate current source and a standard volt meter and you can make 4 wire kelvin measurements. Additionally depending on what you are measuring, you can use a higher current to improve the signal to noise ratio and get better accuracy as long as the current is not enough to heat the parts under test.
The best part of the video is when he shows how to do 4-Wire resistor measurements with two multimeters.
Ordered NPN audio power output devices, but tested as PNP and one as a double diode even though they were labeled correctly.
You are more educated than me in electrical theory and practice, and I love your videos. Thank you for taking the time to share with everyone and for explaining as you go. I have learned a lot from you. Is there a provision on your microcurrent board for calibration? I didn't see it in the schematic, but could possibly mitigate problems with components out of spec.
@nealsprouse Yes, sorry for the delay. Will ship when I have the new resistors and have tested them.
Fab intro animation on that one
Bournes - I remember paying through the nose for the right pots to restore my dad's old guitar - things failed after a bit of light soldering
If the value were being affected by other components in the circuit, it'd be reading low, not high.
@QBMan My second incident in almost as many days. I had actually already shot another blog about another problem with dodgy Maxim chips, but the Digikey one got bumped up the list.
Digikey has gotten much worse in the past few years about sending completely wrong parts - the label will indicate the correct part, but the bag will contain a completely different part. Not just a different value - but a different part! In my experience, they screw up about 20% of the time. I place about 50 orders per year.
did you put your curent mesuring instrument before or after the resistor? i haven't heard you say that information.
@EEVblog No problem Dave. I'm glad you found the problem. Keep up the good work!
Where can I get one of those res ref boxes ???????
I recently (2019) purchased a component that was an order of magnitude out. Banding was correct, packaging was correct, part number was correct, but the thing was not what all that said. Always check your components prior to installation.
I know how you feel Dave. I had gotten a Samsung 52inch DLP TV from my wifes sister because it had just stoped working. No surprise it was the voltage cap on the balast baord so I ordered a new one from digikey. 450v 47uf had it replaced with a 400v 47uf cap from an abandon switch mode supply and that worked for 2.5 months. I put the new cap in from Digikey and it didn't last a week. Had a new cap overnighted to me and it lasted 3 days! So I said screw it and just got a used board from eBay. =p
@joe72205 They are crickets. I get lots of local wildlife noises around my house.
Very nice tutorial on 4-terminal measurements. Thank you. Sorry about you getting bent on the component values. Even if they reimburse you, there are still compromised boards out in the wild now.
I sat through a sales presentation where the rep referred to them as Kevin measurements. Nobody said a word. But you could see an occasional twitch in the corners of the engineers eyes.
Dave, where do you get your boards fabricated and how much does it cost per board. I cant find any place that doesnt fabricate small boards for less than $120 except for Seeed Studio which goes for 35 which is a good deal.
+Nicholas Lenzo Dangerous prototypes
+Nicholas Lenzo I recently found OSHPark. They do 5$ for sq.in. of board for three boards.
nice
I bought some components from Digi-key and they were way out of spec. Not only that but one of the bags was empty, it had the label on it, but no components in the bag. and seeing they were expensive devices I was not amused.
RS components did that to me with expensive conmponents for a time-sensitive project. Such a pain!
Why were they doing this?
Did you spec invisible parts?
@@mcuembedded I had some pcb mount lithium coin cells on order, Farnell delivered them loose in a plastic bag, totally discharged.
So Dave, how do you get around this problem?? I am building a BLDC position controller and I will be getting it sent off to get made, my first run will be 1000 and I don't want this problem to happen to me. Do I order one of each component and test all of them first then get it built with the same items??
No... just test a few of the ones you get and make the call to the supplier if they are out of spec.
Dave,
You said it over and over...
They are not .1% resistors...LOL.
@CampKohler Antistatic mats have surface resistances of at least hundreds of millions of ohms per inch, specifically so that it will not interfere with stuff like a component being tested or a PCB with live traces in contact with the mat or whatever.
this is a lesson about incoming quality inspection.
@TheRoknemec Yep, notified.
Love your videos, Dave.
Just a tip for probing those tiny 0805's (and 0603's, its especially handy for checking LED orientation) - leave it in the tape and, using sharp probes, probe the component through the tape.
@abraxalito They are thin film 25ppm.
@billysgeo I've got one of those, had it for a while now, but like you haven't had time to really play with it yet.
I've had instances where I ordered a small part and Digikey sent me some eval board instead. They constantly send out wrong parts. I only had one order in the past three years where they sent out everything correctly, and that order only had one item.
I got the Same question as Lionel what happeneD??? I wish you could like a video multiple times lol.
@DJMC5ive Well, that's what "going green" is really all about. I saved about $80 in antistatic packaging material on my last production shipment by reusing what one supplier had sent their parts in. Recycling, going green. Whatever. It's "green" (USA dollars are green) in my pocket.
can someone explain why you'd need a microcurrent supply? What is there that needs such ridiculously accurate and *small* amounts of current?
Calibration and verification.
+KpilotRCHelis What if you want to know how much current a watch use? or how much current a microcontroller use in sleep mode? It may not look much, but if you have a tiny battery then you need a tiny current drain to get a long battery life.
Audio signals. Scientific instrument signals. Lots of stuff.
Nano amp current measurements for photon imaging bias settings, among other things.
@EEVblog With the 4-wire you showed 0.8%, so maybe you got some 1% ones? Or was that only one resistor?
The only problem I've had with ordering from Digikey is when I once received some transistors when I actually ordered microcontrollers. I called them and they had someone check the bin in the warehouse and--sure enough--they were stocked wrong. Digikey sent me the right parts and let me keep the transistors.
My only real complaint is that the lady in customer service was really cranky about the whole affair.
sounds like someone was caught charging premium prices for an inferior product and it bit em in the ass...they were relying on the fact that most people wont verify to the extent that dave did....
Are they in fact thick film ones? Try warming it up with a soldering iron and check how the resistance varies? Thick film have at least 100ppm temp co, thin film should be considerably better.
What's the possibility that the resistors are actually faulty, i.e. manufactured wrong?
Dave, where i can order these fine quality products (uCurrent, etc)? @ the forum?
maybe i was drunk when you put em for sale?
@qwaqwa1960 I have no doubt Digikey will come good on replacing them ASAP.
Best electronics videos ever! #NSA/CIA telepathy.
kinda funny to see someone saying he is not happy while similing
Id love to see what he's like if they get his order wrong at the drive thru. lol
insylem, no you wouldn't LOL
@neutron7 Maybe, only Digikey can tell us that. We'll see what they have to say (if anything other than sorry)
Sucks! My last digikey screwup was receiving resistors instead of gate drive transformers that I ordered. Easy to catch that.
I had a major recall incident with ECS crystals ECS-160-20-3X-TR from Mouser. These crystal would suddenly stop oscillating. Real bad news because while it did the quadcopters' motors would spin up full throttle even though the motors were in a "disarmed" mode. Intermittent, hard to diagnose. Assumed a reflow problem for a long time but turned out to be bad parts. Replaced with an ABM7 crystal and not an issue since. I recalled 220 units, which cost me a bundle for a small biz. @flyingeinstein
So what happened in the end, did you get any compensation for this horror from DigiKey?
Those china batteries can be dodgy, as they frequently don't have the best seal. Air gets in over time and makes them go high resistance.
That's a good point. I'm wondering if it's possible that these resistors were spot on immediately after manufacture, but drifted over time due to some obscure reason?
You realize measuring components in circuit is a bit lame right? There are resistances in contact, trails resistance, capacitance and all sorts of stuff. You did it correctly by measuring the resistor out of the package.
What is Digi-Key? and what is TEC1 - 12706?
Man that microscope is cool, but sadly expensive. I'll just stick with a dvd player lens on my iphone case for looking at components.
How many multimeters do you have?? I'm currently accepting donations in my workshop
Anything do do with the zero missing before the decimal point. Someone may have read them as 1.0% tolerance. (P/N matched exactly though...!)
@vpapanik It happens all the time unfortunately. I'd be happy if they just let me know what happened!
what kind of DMM is that? You said it but I couldnt understand you.
Just so you know Digikey employs as bunch of rural clueless Minnesota housewives who don't know a resistor from a capacitor. For anyone who's seen the movie Fargo that's pretty much the Digikey workforce. Among the major online component distributors Digikey usually comes in dead last despite having one of the better website search engines. I once ordered 200 surface mount resistors as cut tape and Digikey shipped 20 bags with 10 resistors in each bag which is useless to any pick and place SMT machine on the planet. When I called to ask for 200 resistors on a single tape none of three people I talked to could understand why what they shipped was a problem. Digikey=Fail. I have had far fewer problems with Mouser, Avnet, Arrow, Future, etc. Digikey is the Walmart of components and, like Walmart, their employees mostly don't have a clue. Buyer beware. They might as well be some Chinese supplier on eBay.
Er... Did you call them /before/ placing your order? From their site: "Digi-Key tries to ship in one strip when possible however, if you require your order to be one continuous strip, this should be requested at the time you place your order." If you didn't, that's on you, not them.
It's understandable why that'd be frustrating if you didn't know it beforehand, but you can't make assumptions and blame others when they turn out to be wrong.
Another assumption you're making here is that the other distributors don't do the same thing. Mouser did that to me with some ceramic caps, and they don't even directly say they do it on their site. It's only said indirectly by specifying their custom reels will be continuous, with the unsaid implication that normal cut tape orders might be discontinuous.
You can always order your parts on a Digi-Reel; it's a continuous strip with as many parts as you'd like, and comes on a reel for the pick and place machine.
love the tandy calculator!
9 years later I’m here to comment that it’s a rebadged Casio from my youth
Darn those refurbished resistors.
You probably dont give a shit but if you guys are stoned like me atm you can stream pretty much all of the latest series on InstaFlixxer. Been binge watching with my girlfriend recently xD
@Steven Noah yup, have been using InstaFlixxer for years myself =)
@Steven Noah yea, I have been using InstaFlixxer for since december myself :)
Where can I get a uCurrent device?
@DJMC5ive I love real catalogs, too - but I'm easily distracted by them (especially the full color Allied), so this isn't a clear loss for me.
Once you notice those crickets in the background, you can't un-notice them. UGH!!!
I'm already triggered by the ".1%" notation. Is it that hard to print the extra zero? What if the ink is low and the dot gets lost in the print? "01%" is quite obvious and you at least question it, rather than "1%" which looks perfectly fine.
I guess this explains why the uCurrent I ordered has not shipped yet.
Incoming inspection??
Good lesson!
Digi Key! You had one job....
Hmmm, no response from Dave or Digikey??? So for the past 4 years people shouldn't trust digikey parts?
Do you realize how many part numbers Digikey carries? You can't fault them for a single manufacturer's batch of 10 ohm resistors that were out of spec.
Mallom Bum They are suppliers, not manufacturers. Therefore, NO you cannot fault them for a few units out of millions being out of spec. If you think so, you should pack up a multimeter and head over there and beg for a job testing each and everything that goes out the door.
Mallom Bum Oh I see, I misunderstood you the first time.
***** I can't blame digikey for getting it wrong some times, and it's not at all reasonable for them to test every single 0402 resistor on a reel of 10,000 :D At the same time, when they get parts wrong they are pretty good about setting it right. I've had them ship me some obvious mistakes like several thousands of dollars worth of eeproms when I had ordered $100 worth of barometers - they were very polite that time because they wanted their parts back unharmed :D
The worst error I've had from them was getting 5.76k ohm resistors labeled as 57.6k ohm - luckily the power supply circuit wasn't damaged but was clearly non functional so it was easy to diagnose - I've luckily never run in to such a subtle error as this one.
+RN1441 Why didn't you just keep the EEPROMS and sell them on ebay/use them in projects?
I take it you sent this video to your supplier?
So, I assume the lesson is - do spot-checks on batches of supplies? I mean, not that genius a thought, but I guess sometimes you really do need to get burned once to avoid this happening in the future