I do design and development for the company I work for in electronics. We get all of our pcbs assembled by a high quality LCD manufacturer in China..they make LCD's for many companies like Newhaven. This company has been fantastic to work with. We have swapped out some of our parts with chinese manufactured equivalents..not all but some of them. We used to use a TI temperature chip now we use the chinese version and in 50,000 we have had maybe 1 failure. We also use a capacitive touch chip for Silicon Labs but they were getting expensive and we got a EOL notice so they sources a chinese made similar version. They sent me samples and I had to do a slight redesign of the pcb as it was a little different but it works great and no failures..These chinese manufacturers have been around for quie a long time. In these times where it is hard to get supply for manufacturing We are selective of course but chinese companies do make very good chips
Beware as well of those weasel words in the datasheet, where they tell you that the parameters are guaranteed, but are not tested in production. Often only those tested to the military grade will actually be tested fully, but the rest of the parts off the line, destined for the industrial or commercial market, will either be untested for all parameters, or only tested by sampling, or just be given a wide binning for those parameters, so all will pass. You can see that in the binning, where that bin likely was only the military devices, tested over full operating range, but your commercial part, tested only at 25C, and only during the initial trim operation where they zener zap the offset to fit into a bin, likely has, for the stable process they have in the fab, all parts sitting in the middle bin, or just on the one side, but they will use the full temperature range data to tell you just how many will be outliers, mostly due to how many junctions they had to blow to get them into spec. That on die trim is the biggest part of the offset change with temperature, because the resistors no longer track perfectly when you reach the trim extremes. Yes that other device is likely to be very variable, even made with the same dies and plant, because they likely are making batches at different foundries, with slightly different outcomes each time, as they do not follow exactly the same processing steps, times and process paths for each batch, so you end up with some differences in the final product. Making dies is very much process dependent, change the fab, and even on supposedly identical lines, with supposedly identical processing, there will be enough subtle differences to make the first batch or five at the new plant a dice roll, till they learn just how much they have to change to get the product the same.
Agree. And I usually leave sufficient (2x-5x) headroom in design to overcome inconsistency. But as a hobbyist, inconsistency could be a source of fun. Just like over-clocking CPUs, you can try to find ICs that work well far beyond specs. I once identified a few TA7642 ("single chip AM radio" for MW,
That was a nice discussion of interpreting the details of data sheets. Thanks! You may be running into a language or cultural barrier. First, the English data sheets may be pro forma. Especially for a startup, it might not make business sense to invest resources in a web site and data sheets that aren't for their primary market. They might even expect that you will translate the Chinese data sheet if you are serious about using the part. Second, Chinese business culture emphasizes relationships more than US culture does. It may be expected that requesting and receiving additional data is part of building that relationship. It's a little like the US companies that require contact info to get a data sheet so their salesperson can follow up, but to a greater degree. You may have to go farther down the sales path and get qualified as a customer to get the data. Finally, a lot of Chinese data sheets are marked "confidential", perhaps because of the rampant cloning. It may be that the graphs are in a secret version of the data sheet, even if a US company would publish them openly. This too would require building a relationship.
I spent months in China designing and manufacturing products. I understand the culture and their standards are very different when it comes to engineering products Maybe I will tell those stories someday
Also, I wouldn't read too heavily into the "foreign" part. It's cultural again. In this case "foreign" is probably a point of prestige. In US culture you sometimes see "foreign" used as a mark of prestige for cars and wines. Same kind of thing...
There is this Chinese company called Xinluda which makes clones of ICs from Analog, TI and others. Interestingly enough their datasheets look like exact copy of the big brand equivalent. See XD07/XL07Z for instance - 1:1 copy of Analog's OP07, just the branding was replaced. It is similiar for other chips as well. Makes me wonder - did they steal the IP or are their ICs performing equally as well as the original ones.
yes they can, but seldom do. As an EE myself, I concur with your assessment - however you were too kind. I once mentioned this type behavior to another engineer I know from China, he just laughed as if I was silly. There is a huge gap in basic integrity. How it happens I don't know, but it strains my ability to be accepting of the culture. - in the case you outlined it would be interesting to actually purchase the part and test it.
Yes, this is not the first time I have seen this. It is a stochastic way of looking at the quality of mass-produced products. The product goes through a first sieve (A) and the elements that pass this first examination are subjected to a second, stricter sieve (B). This is the case for table tennis balls. The number of stars is supposed to indicate the number of quality controls, in increasing order of rigour, to which these balls have been subjected. A ball with three stars (C) is worse than a ball with five stars (E). They were all made on the same production line, but by taking advantage of variance, normally viewed as a negative, the five-star balls are more symmetrical, balanced, uniform, etc.
So one Cosine data sheet says 2Mhz and the other says 1.5Mhz, which is it? I suspect that whatever they have it's probably copy of the mask for an Analog Devices part.
The first data sheet was for the COS77 which IMSAI Guy couldn't find on the website, so he randomly clicked a link and eventually ended with a datasheet for the COS358. That is clearly not the same part as the COS77, so it's hardly surprising that it has a different GBWP.
Try 3Peak, also sold on LCSC. They seem to have temperature data and lots of graphs in their datasheets. I know nothing about the company, they were just the first Chinese company besides Cosine to show up on an LCSC search.
Where I work we sometimes have issues getting not-so-old replacement parts because a lot of times the components are no longer made or the manufacturer couldn't be bothered to come up with a different solution after the initial warranty period is past. By the way, I have designed and 3D printed some holders for the 42mm x 11mm magnifier lens that you had a video about... I see that you are holding yours bare. Could I send you one?
China makes some exceptionally good parts. Take apart any Chinese radar system, ICBM , avionics package, military communication systems etc and the parts will meet or exceed specs. They just don't send them over here. And hopefully not any time soon.
Sure they have excellent components, but I'm sure many of them have patent infringement or obtained by espionage. So they don't publish too many details to the west.
@@vladnickul @Shazam @Shazam Not "because they are making their own", but because how they are able to make "their own". And this isn't only my opinion.
@@IMSAIGuy I called them, out of curiosity. A sales rep picked up immediately, who spoke tiny bit of English so I used Chinese. She told me COS77 is a trim-down version of COS1177, tailored for their biggest customer, and will be discontinued this year. (guess why their price is lower). COS1177 is the recommended replacement now (more expensive). I need to provide company name, reg code and email to get full datasheet and free samples so I didn't continue. Looks like she only serves large customers since she asked me how many reels I use on a monthly basis....😂😂😂
China does not fake stuff, small Chinese vendors do the same way thousands of people in USA do. If you want to buy Chinese parts go to LCSC. If you are expecting the same level of support, especially in English, you are going to be disappointed. Most of them cater to the China market.
@@galileo_rs If you have watched my recent videos, I show an LM741 that has a very low Vos that is as good as the Op-07, sometimes statistics gives you a good one. The problem is statistics give you really bad ones too unless you actually test the parts and reject the bad ones. The OP-07 is designed, manufactured (trimmed) , and then tested to meet a Max specification. It is also designed for performance over a wide temperature range. You can't just get a part and if it works design it into a product. Engineers need to understand how to read a datasheet and buy the parts that will accomplish the task.
@@galileo_rs How can you check whether the components meet the advertised specifications if none that matter for your design are published? Even if you measure the characteristics of a sample, there is no guarantee that the next batch will meet those specifications, as the manufacturer makes no written promises. Instead of uploading a PDF, guaranteeing that these parts will likely fit your design, you now have to commit resources to a test facility, maintain measuring instruments and pay for extra labour, or send a legal team to draft a contract they can't get around on a technicality. All because the supplier is taking shortcuts. And what happens if you don't stumble upon a component that works for you right away? Do you start this random exploration again with another unknown manufacturer? Don't forget that electronic products are assemblies of many components. They all need to meet your specifications. No serious engineer would waste time sorting through all these companies with dubious or approximate data sheets, unless you yourself are a manufacturer of low-quality, low-volume products destined to end up quickly in the bin. And don't get the impression that these components are discounted because of some prejudice against foreign manufacturers. The vast majority of engineers are quite willing to order from anywhere in the world. This is why standards like ISO 9000 were established by engineers: to increase the options available internationally. Designers, even those on the other side of the street from those companies with incomplete data sheets, will not order these parts either. Beyond the ideological divisions, both real and exaggerated, some cost-saving measures are proving too costly for the good of those who fall for them.
I do design and development for the company I work for in electronics. We get all of our pcbs assembled by a high quality LCD manufacturer in China..they make LCD's for many companies like Newhaven. This company has been fantastic to work with. We have swapped out some of our parts with chinese manufactured equivalents..not all but some of them. We used to use a TI temperature chip now we use the chinese version and in 50,000 we have had maybe 1 failure. We also use a capacitive touch chip for Silicon Labs but they were getting expensive and we got a EOL notice so they sources a chinese made similar version. They sent me samples and I had to do a slight redesign of the pcb as it was a little different but it works great and no failures..These chinese manufacturers have been around for quie a long time. In these times where it is hard to get supply for manufacturing We are selective of course but chinese companies do make very good chips
BTW I love your channel and all the things you do...I am a RF guy originally and love especially your RF videos
Beware as well of those weasel words in the datasheet, where they tell you that the parameters are guaranteed, but are not tested in production. Often only those tested to the military grade will actually be tested fully, but the rest of the parts off the line, destined for the industrial or commercial market, will either be untested for all parameters, or only tested by sampling, or just be given a wide binning for those parameters, so all will pass.
You can see that in the binning, where that bin likely was only the military devices, tested over full operating range, but your commercial part, tested only at 25C, and only during the initial trim operation where they zener zap the offset to fit into a bin, likely has, for the stable process they have in the fab, all parts sitting in the middle bin, or just on the one side, but they will use the full temperature range data to tell you just how many will be outliers, mostly due to how many junctions they had to blow to get them into spec. That on die trim is the biggest part of the offset change with temperature, because the resistors no longer track perfectly when you reach the trim extremes.
Yes that other device is likely to be very variable, even made with the same dies and plant, because they likely are making batches at different foundries, with slightly different outcomes each time, as they do not follow exactly the same processing steps, times and process paths for each batch, so you end up with some differences in the final product. Making dies is very much process dependent, change the fab, and even on supposedly identical lines, with supposedly identical processing, there will be enough subtle differences to make the first batch or five at the new plant a dice roll, till they learn just how much they have to change to get the product the same.
Agree. And I usually leave sufficient (2x-5x) headroom in design to overcome inconsistency.
But as a hobbyist, inconsistency could be a source of fun. Just like over-clocking CPUs, you can try to find ICs that work well far beyond specs. I once identified a few TA7642 ("single chip AM radio" for MW,
That was a nice discussion of interpreting the details of data sheets. Thanks!
You may be running into a language or cultural barrier. First, the English data sheets may be pro forma. Especially for a startup, it might not make business sense to invest resources in a web site and data sheets that aren't for their primary market. They might even expect that you will translate the Chinese data sheet if you are serious about using the part.
Second, Chinese business culture emphasizes relationships more than US culture does. It may be expected that requesting and receiving additional data is part of building that relationship. It's a little like the US companies that require contact info to get a data sheet so their salesperson can follow up, but to a greater degree. You may have to go farther down the sales path and get qualified as a customer to get the data.
Finally, a lot of Chinese data sheets are marked "confidential", perhaps because of the rampant cloning. It may be that the graphs are in a secret version of the data sheet, even if a US company would publish them openly. This too would require building a relationship.
I spent months in China designing and manufacturing products. I understand the culture and their standards are very different when it comes to engineering products Maybe I will tell those stories someday
Also, I wouldn't read too heavily into the "foreign" part. It's cultural again. In this case "foreign" is probably a point of prestige. In US culture you sometimes see "foreign" used as a mark of prestige for cars and wines. Same kind of thing...
There is this Chinese company called Xinluda which makes clones of ICs from Analog, TI and others. Interestingly enough their datasheets look like exact copy of the big brand equivalent. See XD07/XL07Z for instance - 1:1 copy of Analog's OP07, just the branding was replaced. It is similiar for other chips as well. Makes me wonder - did they steal the IP or are their ICs performing equally as well as the original ones.
yes they can, but seldom do. As an EE myself, I concur with your assessment
- however you were too kind.
I once mentioned this type behavior to another engineer I know from China, he just laughed
as if I was silly. There is a huge gap in basic integrity. How it happens I don't know, but it strains my ability to be accepting of the culture.
- in the case you outlined it would be interesting to actually purchase the part and test it.
Interesting how grade "B" is the better part vs. "A"...you'd almost think it was the other way around.
I was wondering the same thing
The B parts try harder
Yes, this is not the first time I have seen this. It is a stochastic way of looking at the quality of mass-produced products. The product goes through a first sieve (A) and the elements that pass this first examination are subjected to a second, stricter sieve (B).
This is the case for table tennis balls. The number of stars is supposed to indicate the number of quality controls, in increasing order of rigour, to which these balls have been subjected. A ball with three stars (C) is worse than a ball with five stars (E). They were all made on the same production line, but by taking advantage of variance, normally viewed as a negative, the five-star balls are more symmetrical, balanced, uniform, etc.
Maybe it's so they can add letters for higher-grade parts later?
The same is true for maple syrup - grade A is the weak sauce w/ no color or flavor. Grade B is the dark rich stuff that tastes like maple.
So one Cosine data sheet says 2Mhz and the other says 1.5Mhz, which is it? I suspect that whatever they have it's probably copy of the mask for an Analog Devices part.
The first data sheet was for the COS77 which IMSAI Guy couldn't find on the website, so he randomly clicked a link and eventually ended with a datasheet for the COS358. That is clearly not the same part as the COS77, so it's hardly surprising that it has a different GBWP.
Try 3Peak, also sold on LCSC. They seem to have temperature data and lots of graphs in their datasheets. I know nothing about the company, they were just the first Chinese company besides Cosine to show up on an LCSC search.
Where I work we sometimes have issues getting not-so-old replacement parts because a lot of times the components are no longer made or the manufacturer couldn't be bothered to come up with a different solution after the initial warranty period is past. By the way, I have designed and 3D printed some holders for the 42mm x 11mm magnifier lens that you had a video about... I see that you are holding yours bare. Could I send you one?
mine is in a translucent plastic piece I made on the lathe. thanks for the offer though
China makes some exceptionally good parts. Take apart any Chinese radar system, ICBM , avionics package, military communication systems etc and the parts will meet or exceed specs. They just don't send them over here. And hopefully not any time soon.
LOL
Sure they have excellent components, but I'm sure many of them have patent infringement or obtained by espionage.
So they don't publish too many details to the west.
@@paulcohen1555 You think that they are banned from buying USA and EU tech because they are making their own?:))
@@vladnickul @Shazam @Shazam Not "because they are making their own", but because how they are able to make "their own". And this isn't only my opinion.
Have you tried to call them?
😂😂😂
@@IMSAIGuy I called them, out of curiosity. A sales rep picked up immediately, who spoke tiny bit of English so I used Chinese. She told me COS77 is a trim-down version of COS1177, tailored for their biggest customer, and will be discontinued this year. (guess why their price is lower). COS1177 is the recommended replacement now (more expensive). I need to provide company name, reg code and email to get full datasheet and free samples so I didn't continue. Looks like she only serves large customers since she asked me how many reels I use on a monthly basis....😂😂😂
China does not fake stuff, small Chinese vendors do the same way thousands of people in USA do. If you want to buy Chinese parts go to LCSC. If you are expecting the same level of support, especially in English, you are going to be disappointed. Most of them cater to the China market.
The cosine part is on LCSC. That is where I got the datasheet.
@@IMSAIGuy Order some of the cheapest parts there and see if those meet the specs?
@@galileo_rs If you have watched my recent videos, I show an LM741 that has a very low Vos that is as good as the Op-07, sometimes statistics gives you a good one. The problem is statistics give you really bad ones too unless you actually test the parts and reject the bad ones. The OP-07 is designed, manufactured (trimmed) , and then tested to meet a Max specification. It is also designed for performance over a wide temperature range. You can't just get a part and if it works design it into a product. Engineers need to understand how to read a datasheet and buy the parts that will accomplish the task.
@@IMSAIGuy Yes I have seen that but this is more like: test the claims in datasheet. Just an idea for a video ...
@@galileo_rs How can you check whether the components meet the advertised specifications if none that matter for your design are published?
Even if you measure the characteristics of a sample, there is no guarantee that the next batch will meet those specifications, as the manufacturer makes no written promises. Instead of uploading a PDF, guaranteeing that these parts will likely fit your design, you now have to commit resources to a test facility, maintain measuring instruments and pay for extra labour, or send a legal team to draft a contract they can't get around on a technicality. All because the supplier is taking shortcuts.
And what happens if you don't stumble upon a component that works for you right away? Do you start this random exploration again with another unknown manufacturer? Don't forget that electronic products are assemblies of many components. They all need to meet your specifications.
No serious engineer would waste time sorting through all these companies with dubious or approximate data sheets, unless you yourself are a manufacturer of low-quality, low-volume products destined to end up quickly in the bin.
And don't get the impression that these components are discounted because of some prejudice against foreign manufacturers. The vast majority of engineers are quite willing to order from anywhere in the world. This is why standards like ISO 9000 were established by engineers: to increase the options available internationally. Designers, even those on the other side of the street from those companies with incomplete data sheets, will not order these parts either. Beyond the ideological divisions, both real and exaggerated, some cost-saving measures are proving too costly for the good of those who fall for them.
Rambling on too much!
I'm a ramblin' man