Eagle library editor

A deviation is sort of an "economy classs ECO". It requires less signatures, usually puts more responsibility on the engineering department not to screw up, and generally has a limited lifetime. Think of it like the two-month extension Congress signed. It didn't fix the problem but did avoid a meltdown of sorts.

Example: Part X has become unobtanium. Panic sets in. Engineering finds out that part Y can be used but requires a major design spin that can't be scheduled in until early summer. The VP of Sales is about to go ballistic. So Engineering figures out how to do this with part Y on the existing board but it is a labor-intensive and thus expensive rework. Management decides to go this route and Engineering to prepares a deviation. This deviation is not a hack but makes sure of a clean business process. For example, it has to come with test data and "shovel-ready" (I guess that's the new word of 2011 ...) rework instructions. I've written many of those.

Usually just about anything but the kitchen sink is released via ECO so this means that there is now a justified deviation from an existing ECO. Can't be overdone though, if a company ran on lots of deviations for a long time that would raise some red flags during audits.

In medical electronics and aerospace we can't :-)

for

The silk screen has a "whiteboard style" field for the assy rev level. Usually EE rev levels with X1, X2 and so on get written on there with a fine point Sharpie, or onto a sticker. If you want to sell some of the EE boards you could qualify them, use cleaning alcohol to remove X3 and write A on there.

home

design is

defined by

separate

You said "raw board". Now I am not sure what that means, but obviously raw does not mean bare :-)

become

We number ECOs (we're up to ECO832 or so now, after 25 years) and log them, and we know which ECOs are applicable to any assembly. How are "deviations" identified and controlled?

In my company, an ECO needs exactly one signature.

My sales are about 50% aerospace; we don't certify anything and have never had a government inspector on premises. But very little of our stuff actually flies; it's mostly ground test gear.

John

for

So anybody with a sharpie can change the rev level! That could get interesting. Maybe I can take my 2008 Audi and Sharpie it to be a

Our first etch is rev A, and we assume that we will sell it.

Assuming multiple prototypes ensures that there will be multiple prototypes.

John

Only 832 ECOs after 25 years with you rather massive and diverse product portfolio is a remarkably low number.

Deviations are controlled via the doc center which (usually) must be managed by QC. Just like ECOs are, same system. Their impact penetrates into the MRP system. This is usually also where sunset requirements are logged.

In medical devices that would cause the goons to come out and put padlocks on the doors. No joke, it happened to a competitor of my first employer. The minimum is the signature from quality control. They don't care so much about whether or not the biz side of things might get screwed up but they won't likely accept the EE decision maker and the QC function to reside inside the same human body :-)

^^^^^^^^^^^^^^^^^^^^^

That's the reason. If it does fly and it ain't fully RTCA/DO-160 compliant and STC'ed that could lead to serious trouble. Well, at least in the US.

for

that

That constitutes fraud, and in the med world would probably lead to an extended stay at Club Fed in case something bad happens because of it.

So lets assume you build and sell rev A and then later find out that a part on there is not up to snuff and leads to problems that can't be corrected. Or a part becomes unobtanium. Your EEs find another form-fit-function compatible part and this gets ECO'ed in but you still have 50 bare boards in the stock room. What do you do then?

to

numbers for

that

changes.

Build them with the new part, obviously. It doesn't take an ECO to add a new part/vendor as an acceptable part for an inventory item. We don't need full traceability.

If rev A has a problem that can be fixed with an ECO, we do it. If it can't, obviously we have to trash the bare boards, or even the assembled or sold boards. What else could anyone do?

We have occasionally made adapters...

ftp://jjlarkin.lmi.net/Break2.jpg

ftp://jjlarkin.lmi.net/OnBoard.jpg

Manufacturing doesn't like kluges, so we have a background process of revving any boards that aren't clean.

John

Really? NASA bought a modified Microdyne 700 series Ku band telemetry reciever for the ISS. The modifications were to replace the

part

become

of

I hope you believe this, i have seen specialized military software progress from AFXU through DBXK and eventually EFXY. Not all possible letter combinations were used in distributed versions to be sure, but a heck of a lot more were used internally.

whose

it.

become

They're fundamentally different, though. You can call every piece of paper you have and "ECO", if you want.

So you can do anything. Isn't that the same as no process? ;-)

become

Or deviations had all the same (e-) signatures as an ECO. The purpose was different.

A deviation also has a defined end-of-life.

An ECO can't be "backed out", without another ECO. A deviation can, and it reverts to the base ECO.

Don't do that. ;-)

become

Same way, with a deviation number.

Why bother?

Aerospace life sustaining

to

numbers for

that

changes.

Even a different value? How do you tell them apart?

I thought you said your assembly and bare board ECOs were in step?

ECO controlled, one assumes?

home

design is

defined by

ECOed

separate

BOMs,

Board with printed wires, as it comes from, raw, the board house. ...a component. ;-)

to

numbers for

that

changes.

Oooh, that would almost be a capital offense in medical electronics :-)

If the part wasn't in the ECO for the old board there needs to be a deviation or (if this is permanent) another ECO.

But then you must bump the rev level.

Neat! But don't does this on a medical device without an ECO.

With some aircraft electronic (not mine or my client's) you'd be surprised. Lots of rework. All properly ECO'ed I am sure but still ugly. I have a NAV/COM radio that almost made me sick when I opened it.

NASA might be exempt from RTCA standards requirements, I've never directly worked for them (so far).

way.

It can. I guess when one engine on a 767 I was a passenger on quit over the ocean (it has only two engines) there was a thorough investigation of why it did. And one could say that the 2nd engine had now become life-sustaining :-)

But seriously, the ECO rules have nothing to do with life-sustaining. Med gear is held to the same standards no matter what it is used for. Because consequences can be dire. For example, when one of the patient interfaces that I designed quits at a Murphy moment that's almost the same as having both headlights fail in the dead of the night, at 80mph.

home

design is

defined by

ECOed

separate

BOMs,

Ok, but why on earth must that have the same rev level as the schematic?