How to interpret this package drawing?

What's stunning is that the nominal lead pitch is not specified. So if you guess or average to get the pitch, any error stacks up over all 5 pins! So get a sample and measure it.

And the compensation/stability guidelines suck, as is standard for LDOs. The "Stability" paragraph is just qualitative blather.

I wouldn't use this part unless I absolutely had to.

John

That's what really caught my eye. It's ridiculous - I've never seen such a bad drawing. I got something hacked together that should work though, by averaging all the values.

I found another company's drawing for TO-263-5s and they specified the pitch as 1.7mm, which is essentially the average of the min and max values (1.702mm).

I'm saving a decent amount of money going with this part though, so the extra hassle is worth it, I hope. I don't need a particularly smooth supply - I just need an adjustable vreg capable of taking a 5V signal down to 3-3.5 V (adjustable) with ~200mv or less ripple, all at about 2-3A. (constant load)

-Michael

datasheet:

This looks like a design that was in inches that has been converted to metric. Try converting the values to inches and see if a sensable sounding number lands right near the middle of the span.

eg: 0.4" = 10.16mm. This is near the center of 10.05 and 10.668 so there is a good chance the design started with a 0.4" dimension. I would expect the real units to land most often just about at that point.

I'm short on time for details, but that is a standard 5 pin D-PAK or D-PAK2 that's been thinned out to reduce the height requirements - the pads should be the same. The other is a standard SOIC-8.

datasheet:

Long time ago, when still useful in design, and constantly using metric and 'colonial' measuring units we agreed to re-dimention the inch for easier conversion between the decimal- binary fragments of partial measurements. 1 inch = 25.4 mm { standard definition of inch } we defined as 1 inch = 25.6 mm and it made the conversions more practical.

HTH

Stanislaw

It's also the same as IRF's drawing for the TO-262 (1.700 mm nominal), as you can see from this package drawing:

It's laid out with proper GD&T tolerancing-- the numbers in boxes are "basic" dimensions that are theoretically exact, and the position of each lead is defined off the 'A' datum in the maximum material condition so you can easily figure worst case conditions.

It would be tempting to guess that the nominal dimension is the same as the SDIP pitch of 1.778mm nominal (0.070"), but it'snot.

Best regards, Spehro Pefhany

datasheet:

In the early days of IC design, when mask patterns were cut from rubylith, an inch was defined as 25 mm.

John

One nice thing about the GD&T (eg. ASME Y14.5M or ISO) method of tolerancing (as on the other IRF drawing I linked to) compared to coordinate tolerancing is that there is no build-up of tolerances. The theoretically exact "basic dimension" is given and the position of each pin is specified to lie within a certain tolerance zone. To get the same effect with coordinate tolerancing you'd either have to put a stupidly tight tolerance on the pitch, dimension every pin, or add some kind of note to the drawing.

Best regards, Spehro Pefhany

Late at night, by candle light, John Fields penned this immortal opus:

a - b Let's write that as ------- + b which we then can write as 2

a - b + 2b a + b

---------- = ------- 2 2

Sure looks like an arithmetic average to me.

- YD.