Spartan 3 documentation confusing...

Who writes this stuff? I can speak from personal experience: When I arrived at Xilinx in 1988 I shuddered at the poor organization of the very first Data Book. And I volunteered to create a new one from scratch. And that's what I did, for 5 years in a row. I did not perform the measurements, but I either wrote or edited every single word and every comma, semicolon and hyphen.( I am hot on hyphens...) The advantage of a start-up is that the company can hire overqualified people to do mundane jobs, and do them very well. With over 30 years of engineering experience, I just rolled up my sleeves and dug into it. (The stock options were a nice motivator...) I disagree with the statement that engineers cannot write, and that the job has to be delegated to tech writers. For that's what you get in a more mature company, where the really savvy engineers are up to their ears in design, if they have not yet been promoted to a managerial position. The tech writers know their grammar and spelling, and the Chicago Book of Style, but they usually have never been working engineers. So you get good grammar, but not necessarily good sense. Sometimes not even good grammar... Excellent documentation is rarely elevated to a top priority. That is a sad fact. No need to convince me or Austin or Steve Knapp about the errors of that reasoning. We know and we try our best to help out

Peter Alfke

Reply to
Peter Alfke
Loading thread data ...

It has been my experience that tech writers provide a lot more than just a grammer check. They understand how to present information in a clear, concise and complete manner; all things that we are discussing as being missing from the Xilinx documentation. Tech writers understand documentation the way you understand your chips. They know the process of taking raw input and working it into a document by writing, reviewing, finding the holes and plugging them. Tech writers find and fix the problems so your customers don't have to.

That is true here as well. I am constantly being told that the document does not need to be complete, we just need something to meet our schedule. But just like the design, I am responsible for the completeness and accuracy of the documentation. My last ICD was good enough that our customer decided to use it as the governing document for the interface and his ICD removed the related technical info and references our document.

Reply to
rickman

Rick,

I have just checked the Spartan 3 changes, and I am pleased to see what I think you need.

formatting link

page 56, Table 32 has the ranges of pullup currents, and their equivalent ranges in ohms.

page 97, Table 68 through page 100 Table 69 has the definitions of the pins, with a statement of when the pullups are active, and when it matters.

page 109, Table 74 reinforces the definition of the pins, and the pullups.

Do you agree?

My only complaints are that 'pullup' and 'pull-up' are both used, so searching for "strength" needed 'pull-up' and searching for "definition" needed 'pullup.' And W is used instead of omega (for ohms) in four places (describing pullup for DONE, etc.).

I hope this will allow everyone to go back to work, now that they have the pull-up (pullup) resistors fully specified for 1.14 to 3.45 volts (as active devices, their strength varies with supply voltage).

If anyone at your place of work is still unhappy/uncomfortable with your design choices, I would be happy to remind them (in writing) that the data sheet (product specification, combined with the encrypted spice models, IBIS models, and speeds files) are the controlling documents, and that what may be suggested by the hotline, or by a FAE (or even by me!) may not be the most optimal solution (anything that is not in writing and approved by Xilinx may have been subject to mis-interpretation, or may be in error), and that good engineering judgement combined with observance of the data sheet and other specifications I have listed above is what actually matters (in a practical and legal sense).

Austin

Reply to
Austin Lesea

At least it was written in the one language, rather than being translated. The worst documentation I came across was for an ASIC made in Germany. The original German datasheet was very poor, and had been translated into English by someone who seemed to have little knowledge of either German or English, and no knowledge at all of the electronics involved. One entire chapter (the one containing all the register settings - kind of important) was translated to the single sentence "original German unclear".

Reply to
David Brown

I, for one, am very pleased to see the changes. The push to take care of the nitty gritty details that really do matter is sincerely appreciated. I saw great stuff with the Spartan3E design notes in the data sheet and now the clarification of trivial (i.e. rudamentary) information here in the Spartan3.

Thanks from me to the Xilinx folks that for pushed the changes.

- John Handwork

Aust> Rick,

Reply to
John_H

Something is wrong. I still get Ver 2.0. I am pretty sure I was able to see ver 2.1 last night from home. But I am getting the old version still here at work.

Aust> Rick,

Reply to
rickman

Perhaps your pdf got cached at work? A fresh load comes up with 2.1 on all four modules here.

Reply to
John_H

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.