Now, what would counterfeiters buy, rub the numbers off and resell as OPA627 on eBay?
I think something cheap and pin-compatible so that most that will just plug it into their circuits will find it works and since it's an OPA627 it must sound better, right?
Here's a set of measurements on a TI LF411CP. At first glance, it doesn't look that different. What stands out in the first image is the undershoot after the falling edge. But the closeups of the slopes reveal this opamp isn't up there with the genuine OPA627.
But would this opamp sound different from the real deal? Somehow, I doubt it.
As an aside, I found another LF411 in my stash, a National Semi LF411CN. This one drew less current than the TI item (2.4 mA vs 2.8 mA) and also didn't nearly have as much undershoot after the falling edge.
I think something cheap and pin-compatible so that most that will just plug it into their circuits will find it works and since it's an OPA627 it must sound better, right?
Here's a set of measurements on a TI LF411CP. At first glance, it doesn't look that different. What stands out in the first image is the undershoot after the falling edge. But the closeups of the slopes reveal this opamp isn't up there with the genuine OPA627.
But would this opamp sound different from the real deal? Somehow, I doubt it.
As an aside, I found another LF411 in my stash, a National Semi LF411CN. This one drew less current than the TI item (2.4 mA vs 2.8 mA) and also didn't nearly have as much undershoot after the falling edge.
Attachments
Time to rejuvinate this thread once again.
Beginning last year I commented that I hadn't seen any recent datecodes on the OPA627BP for a very long time.
We kept using this opamp at the rate of 600-1000 per year, but not once did I see a datecode later than June 2011, until now. In the end, even some OPA627BPs from as long ago as 2008 (!) found their way to our factory.
As you can see in the image, we finally got a recent batch of July 2016. That's actually quite a bit more recent than I would normally expect. Normally, I expect to see a code of at least half a year old, so at two to three months old (when we received them), they're unusually young. (And no, these are not from 2006, at that time the encircled R would still have been present next to the BB logo.)
All this leads me to believe TI/BB produced a large batch and supplied from it until it ran out.
If this should be true, then any datecode between June 2011 and July(ish) 2016 should be regarded as suspicious.
As an aside, the same product also uses an Analog Devices AD797, and its datecode always progressed normally throughout the years. The OPA627AU (the SMD version) also progressed as normal. That made this "gap" in the production of the OPA627BP all the more remarkable.
Beginning last year I commented that I hadn't seen any recent datecodes on the OPA627BP for a very long time.
We kept using this opamp at the rate of 600-1000 per year, but not once did I see a datecode later than June 2011, until now. In the end, even some OPA627BPs from as long ago as 2008 (!) found their way to our factory.
As you can see in the image, we finally got a recent batch of July 2016. That's actually quite a bit more recent than I would normally expect. Normally, I expect to see a code of at least half a year old, so at two to three months old (when we received them), they're unusually young. (And no, these are not from 2006, at that time the encircled R would still have been present next to the BB logo.)
All this leads me to believe TI/BB produced a large batch and supplied from it until it ran out.
If this should be true, then any datecode between June 2011 and July(ish) 2016 should be regarded as suspicious.
As an aside, the same product also uses an Analog Devices AD797, and its datecode always progressed normally throughout the years. The OPA627AU (the SMD version) also progressed as normal. That made this "gap" in the production of the OPA627BP all the more remarkable.
Attachments
Last edited:
I've seen this happen. A decision is made to move some manufacturing steps from one place to another and product is built ahead to cover any short-term hiccups in the manufacturing flow. Sometime people over-do it. It's also possible if you buy from a distributor, that they may have taken on more stock in anticipation of purchases from other customers that didn't materialize. Unlike cars, op-amps don't roll stock (and have accompanying year-end sales) with every model year.
I considered this, but a move in location almost always end up with slight differences in how the components look. And this is not the case, as you can see in the photo in my previous post, the lot trace code still ends in "B", which indicates the production site hasn't changed.
A huge stockpile at one of our suppliers could be a reason for this lack of intermediate datecodes, however, our ERP software revealed that we now also make use of other suppliers, which reduces the probabilty of this. No, I still think that TI/BB may have been the ones with the huge stock.
Coincidentally, Analog Devices have moved production of the AD797ANZ. Those made last year (the newest I have are #1534) had a sharp looking casing and the No. 1 circle, not unlike the OPA627. With the last batch (#1611) my eye was caught by a slightly different look, more rounded edges and no No. 1 circle. A look at the bottom revealed that the "old" ones came from Malaysia, the new ones are made in the Philippines. So that is a confirmed move, however AFAIK, it didn't affect availability.
A huge stockpile at one of our suppliers could be a reason for this lack of intermediate datecodes, however, our ERP software revealed that we now also make use of other suppliers, which reduces the probabilty of this. No, I still think that TI/BB may have been the ones with the huge stock.
Coincidentally, Analog Devices have moved production of the AD797ANZ. Those made last year (the newest I have are #1534) had a sharp looking casing and the No. 1 circle, not unlike the OPA627. With the last batch (#1611) my eye was caught by a slightly different look, more rounded edges and no No. 1 circle. A look at the bottom revealed that the "old" ones came from Malaysia, the new ones are made in the Philippines. So that is a confirmed move, however AFAIK, it didn't affect availability.
TI may have changed the wafer fab but not the assembly site (where the wafers are diced, molded into the final package and tested). In that case the old parts would not look any different than the new parts.
Semiconductor manufacturing is not an 'under one roof' operation. The details of the external appearance of the final component depend on the assembly site and has little or nothing to do with what's inside (the wafer).
Another note: "Country of origin" (i.e., your example of Malaysia vs Philippines) for semiconductor devices is where they are assembled, tested, and packed for shipment to warehouses, distributors, and customers. That is no indicator of where the wafers -- the important part -- are made. I.E., Analog Devices may have changed their assembly factory to a different country, but the wafers likely still come from the wafer fab that they've been coming from.
Semiconductor manufacturing is not an 'under one roof' operation. The details of the external appearance of the final component depend on the assembly site and has little or nothing to do with what's inside (the wafer).
Another note: "Country of origin" (i.e., your example of Malaysia vs Philippines) for semiconductor devices is where they are assembled, tested, and packed for shipment to warehouses, distributors, and customers. That is no indicator of where the wafers -- the important part -- are made. I.E., Analog Devices may have changed their assembly factory to a different country, but the wafers likely still come from the wafer fab that they've been coming from.
Last edited:
So, like the country of origin, does the date code tell when the chip was packaged, and NOT when the wafer was made?
This is peripheral to identifying genuine parts, though it may in some was be pertinent. I read about a manufacturer who discontinued a part, but sent wafers to Rochester Electronics which made and sold the "obsolete" part for years until the wafers ran out.
And of course being obsolete now, they're still desirable to some people for repair or whatever, but not in the volume to put them back into production, making them a "perfect" target for counterfeit parts.
This is peripheral to identifying genuine parts, though it may in some was be pertinent. I read about a manufacturer who discontinued a part, but sent wafers to Rochester Electronics which made and sold the "obsolete" part for years until the wafers ran out.
And of course being obsolete now, they're still desirable to some people for repair or whatever, but not in the volume to put them back into production, making them a "perfect" target for counterfeit parts.
I have no idea if the datecode represents the date of manufacture of the important bit (the wafer) or the date of assembly of the finished part.
I could speculate until I'm blue in the face and still not know the answer, but in the end it doesn't really matter, it's what the lot trace code part on the IC is for.
As an example: at work I must assign a datecode to some batches of pcbs produced for one of our customers. These are a mix of locally and remotely produced products, and almost always I have no idea of when the manufacture actually started, so I simply assign the date of when I start testing to the whole batch.
And that is regardless of whether we received the whole lot at once (rare) or in parts (as usual), often with several weeks between first and last shipments.
If push comes to shove, since our ERP software does lot tracking as a matter of course, we can always trace everything back regardless of what date I arbitrarily assigned. I guess it's no different for the plant that assembles the supplied wafers into a final product. In the end, I think the date part of a lot trace code is mainly there as a little bit of help to know where to start looking in your huge pile of data.
I could speculate until I'm blue in the face and still not know the answer, but in the end it doesn't really matter, it's what the lot trace code part on the IC is for.
As an example: at work I must assign a datecode to some batches of pcbs produced for one of our customers. These are a mix of locally and remotely produced products, and almost always I have no idea of when the manufacture actually started, so I simply assign the date of when I start testing to the whole batch.
And that is regardless of whether we received the whole lot at once (rare) or in parts (as usual), often with several weeks between first and last shipments.
If push comes to shove, since our ERP software does lot tracking as a matter of course, we can always trace everything back regardless of what date I arbitrarily assigned. I guess it's no different for the plant that assembles the supplied wafers into a final product. In the end, I think the date part of a lot trace code is mainly there as a little bit of help to know where to start looking in your huge pile of data.
Last edited:
Date code for almost all semiconductors refers to the packaging/testing process, usually the week that the assembly started. So, yes, you could conceivably start the assembly process this week using 10-year-old wafers. Internally, companies' systems track more information -- like what wafers were used in which lot of assembled parts, but that isn't visible. Many companies use a simple YYWW (year/week) date code. National Semi. used to have a more complex code that could give you some wafer info (not that it would mean much to a customer). TI has a more complex code though it only relates to the assembly and test process.
Caveat: as surface-mount packages get smaller, there is less information on the part as there's no room for it. Often on tiny parts, you get only a few characters, which given a lookup tool on the manufacturer's site, may enable you to determine what the part is, but nothing else.
Caveat: as surface-mount packages get smaller, there is less information on the part as there's no room for it. Often on tiny parts, you get only a few characters, which given a lookup tool on the manufacturer's site, may enable you to determine what the part is, but nothing else.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.