It is commercial, and it is a board which is assembled on the manual P&P machine because it is only used in small quantities.

Since this is a manual process manufacturing needs to have a way to tell the differences and to pay special attention to the orientation of the chip. QC also needs to have the same training. And perhaps the placement operation needs to be something that is QC verified BEFORE the PCB runs through the oven. Then a second check may be wise. However, if confidence is high then a sample should be taken by some metrics, and perhaps the test department (if you have one) should have a way of Low Current Testing to make sure the chip is properly oriented.

This may be not addressing what you wish to know or discover but in the industry I spent about 40 years in - that of electronics manufacturing.

 

Since this is a manual process manufacturing needs to have a way to tell the differences and to pay special attention to the orientation of the chip. QC also needs to have the same training. And perhaps the placement operation needs to be something that is QC verified BEFORE the PCB runs through the oven. Then a second check may be wise. However, if confidence is high then a sample should be taken by some metrics, and perhaps the test department (if you have one) should have a way of Low Current Testing to make sure the chip is properly oriented.

This may be not addressing what you wish to know or discover but in the industry I spent about 40 years in - that of electronics manufacturing.

My solution is a bit more drastic. No more MB6S bridges. Next time I order PCBs I'll design it out. Generic components which are not interchangeable are not generic components I want to use.
It can be four S1G diodes, they are already on the BOM. Slightly longer assembly time, but fewer tubes.

 

The answer seems pretty simple to me, but I'm a pragmatic guy.....

Label the pins as AC, +, & –

Since the boards are being hand assembled, it's the easy solution. If pick&place is being used, whomever is programming the p&p can easily verify the settings. Or ask for a verification drawing before approving assembly. What? Your assembler don't offer that? Well, too bad for you. The absolutely cheapest assembler I know (JLC) does.

 

I suppose you check every resistor you use in case the manufacturer has decided to use a different colour code to the rest of us.

Actually I do check every resistor I use, but not for the reason stated.

I check every resistor to be sure I am getting what I ordered and there aren't any in the lot that are out of spec or defective.

Also because of my poor vision I need to verify each one.

 

I am a tad confused.
I see two pics, one is a TO-269AA and the other is a 751EP.
So I guess it matters which one you buy?

I agree, different packages that are close... but
Onsemi refers to package as SOIC4 W and this matches other uses.
And you might have missed this note on the Onsemi data sheet:


And they are not the same part numbers (see Digikey)
Vishay: MB6S-E3
OnSemi: MB6S

Definitely confusing but they are not identical as detailed on the datasheet.
No such think as "generic". When generating second source lists, must consult and review datasheets for all electrical/mechanical details.

 

I suppose you check every resistor you use in case the manufacturer has decided to use a different colour code to the rest of us.

Exactly, I've been burned by the same type of 'clone' parts or slightly changed technical procedures. Some seem to forget that humans are creatures of pattern and habit, not computers or machines. If we expect it to be the same, then a casual look with the similar (like the first few digits or colors) will confirm that belief created from previous action patterns.

 

I agree, different packages that are close... but
Onsemi refers to package as SOIC4 W and this matches other uses.
And you might have missed this note on the Onsemi data sheet:
View attachment 341152

Unless OTHERWISE specified. . . .
but the vast majority of specifications DO agree with TO-269, so they are not OTHERWISE specified.
The minimum length of the overall length of the overmould differs from TO-269 by 50μm, but the allowable flash is 130μm.
Neither TO-269 nor SOIC4 specifies the functions of various pins. There is no excuse for whoever made the copy (I don't know who made the first MB6S) to have used a different pinout, then called it an MB6S.

 

Unless OTHERWISE specified. . . .
but the vast majority of specifications DO agree with TO-269, so they are not OTHERWISE specified.
The minimum length of the overall length of the overmould differs from TO-269 by 50μm, but the allowable flash is 130μm.
Neither TO-269 nor SOIC4 specifies the functions of various pins. There is no excuse for whoever made the copy (I don't know who made the first MB6S) to have used a different pinout, then called it an MB6S.

Specifications (i.e. numbers) that are same/similar to TO-269 does not imply compatibility.
You are misinterpeting what the "otherwise" means. In this case the phrase " Unless OTHERWISE Specified" indicates a precedence in case of conflicting specs on the datasheet with the Notes taking lower precedence. So IF (and it doesn't) it says somewhere else on the datasheet that this part SPECIFICALLY meets TO-269 spec, than the Note would be overridden. You can't look at a bunch of mechanical numbers/spec and infer that it meets TO-269 unless the datasheet owner specifically says it does.

 

I already suggested duplicating the marks on the PCB so that any part can be used. My suggestion was ignored. (deleted)

 

"Unless otherwise specified, this package does not conform to JEDEC TO269AA."

So the pinout can be anything they choose.

 

The answer seems pretty simple to me, but I'm a pragmatic guy.....

Label the pins as AC, +, & –

Can the AC, + & - be easily read on the chip itself? After a while the chips start to look the same and once again the operator is mounting chips backwards. That's the human element that needs to be eliminated.

 

It appears the half moon option on 751EP is confusing in the norms of moon designations, but the middle placement should be a red flag.

Also, they all appear to have the DC pins on the beveled side.

Post #1 is a bit confusing. Did the buddy have 751EP tracks and then got 269AA and installed 269AA, or did buddy just install 269AA on 269AA tracks the wrong way?

 

I suppose you check every resistor you use in case the manufacturer has decided to use a different colour code to the rest of us.

Sorry, I still don't see the very problem here.

The color code problem you are suggesting as an analogy is different than what you are describing originally in this thread, at least the way it sounds to me.
1. You have a part number RXG10k and the color code is brown, black, orange, and it's a 10k 1/4 watt 5 percent resistor.
2. You have a part number TMX10k and the color code is Red, Red, Red, and it's a 10k 1/4 watt 5 percent resistor.
If you look at the data sheet it will clearly state 10k 1/4watt 5 percent for either of these.

I do realize that red red red would be very unusual though, but what you seem to be suggesting is that they made two resistors with part number RXG10k and one had color code brown black orange and the other had color code red red red. That's not the case here.

I do not think I would ever think that two bridge rectifiers just because they come in the same package, would have the same pinout. In most cases they show (+, -, ~, ~) for plus, minus, AC, AC, right on the package, but if not then I would be forced to look at the data sheet. I don't see how it could be any other way, but perhaps you can elaborate your idea on what has happened here.

Side note: I finally got to use the tilde (~) for something as that closely matches the symbols they use for AC and AC terminals on bridge rectifiers

 

I do not think I would ever think that two bridge rectifiers just because they come in the same package, would have the same pinout. In most cases they show (+, -, ~, ~) for plus, minus, AC, AC, right on the package, but if not then I would be forced to look at the data sheet.

But if you go look at all those rectifier packages, the DC pins do appear to be on the beveled side of the package.

If you rotate the 751 180deg, it matches the 269AA, matches on bevel, and matches pinout.

I guess for the 751 (or that specific vendor 751) the half moon option indicates DC pins, but actual location of + and - could vary I assume, but in this case the 751 as shown is just rotated 180 from the 269, case wise.

 

Also, they all appear to have the DC pins on the beveled side.

As far as indicating pins 1 on all IC's, there's either a dot ( ° ) by pin 1, the half moon on the pin 1 end or the bevel on the pin 1 side. Wherever any of those marks are placed, you start pin 1 going in a counter clockwise direction. At least that's what I think. Unless I got the bevel thing wrong.

If anything, I would suspect that a deviation from the standardization would lead me to think I have a counterfeit chip.

[edit] found this:
https://www.tech-sparks.com/how-to-identify-pin-1-on-different-chip/
[end edit]

 

you start pin 1 going in a counter clockwise direction.

Yes, but...
An end moon, and bevel, seem to be std indicators. Half moon in middle of side is not standard from what I can see, and is only an option on that onsemi item. The bevel indicators match 100%. The only odd thing is the half moon option, and in this case the indicator may not be pin order, but rather just indicator of the DC side of the package. Rotate it 180deg and all is the same as 269.

 

Yes, but...
An end moon, and bevel, seem to be std indicators. Half moon in middle of side is not standard from what I can see, and is only an option on that onsemi item. The bevel indicators match 100%. The only odd thing is the half moon option, and in this case the indicator may not be pin order, but rather just indicator of the DC side of the package. Rotate it 180deg and all is the same as 269.

The OnSemi parts we used never had the half-moon between pins 1 and 2. They all had a chamfer down the pin 1 side. Whilst the Vishay parts had the normal half-moon at the top, both are now completely standard ways of indicating pin 1. As the package dimensions are identical, it never occurred to us that the pinouts might not be identical.
Perhaps if they had half-moons at the side they might have stood out as being weird, prompting us to visit the datasheet.
We have used 7805s from many manufacturers for years, in slightly differing TO220 packages, but have never come across one where the pinout wasn't IN-GND-OUT. Why would anyone expect the MB6S to be different?

 

But if you go look at all those rectifier packages, the DC pins do appear to be on the beveled side of the package.

If you rotate the 751 180deg, it matches the 269AA, matches on bevel, and matches pinout.

I guess for the 751 (or that specific vendor 751) the half moon option indicates DC pins, but actual location of + and - could vary I assume, but in this case the 751 as shown is just rotated 180 from the 269, case wise.

Well, the bottom line is that they are different part numbers, and that's the only way you knew they were different as well. So you did use the data sheet after all or else you could not have known that.

 

Long gone are the days where simple semiconductor packages were uniform and identical, and if one purchased a part number from ANY vendor around the world, they would be interchangeable, no buts, ands or ifs.

I have also been burned by minuscule variants, and nowadays I make sure to include all the letters and numbers for a part number which I have seen some to be almost 20 characters long, and also specify the approved manufacturer(s).

 

I have a bin full of 222s that have a different pinout then one might expect.

Of course, they also have a specific part number.