Junction Dot Conspiracy!

Junctions must be marked with 3D balls. Junction balls.

1 Like

So I’d like to know what rules did apply on these:
cypress-an95089

Sorry, forgot to mention the source:
First one is from Cypress AN95089, second one from the data sheet of the PD70100

Nick

1 Like

These should have never existed in this form…

1 Like

From context: the first one must have a non marked 4 wire connection as the circuit would not work otherwise.
The second one goes double sure as it marks connected junctions with a dot and wire crossings with that nice jump over mark.

1 Like

Well, not always! Look at the left side! Between R_ref and “Internal Reference” and between R_class and I_class.
Plus no clear dots. Have a look at Linear’s schematics. They are clear like no other.

linear-lt6106fa

Nick

This is because they most likely used some normal graphics program to make this drawing (it looks a bit childish with all these colours so i would be wary of it anyways)


I found in general that there are some manufacturers who have very good documentation (TI, LT, …) and others that are publishing mixed quality documentation (infineon, on semi, …)

2 Likes

image

1 Like

Anyone wondering about the two left hand side crossings (???) would preferably be better of to stay clear of electronics in the first place. :sunglasses:

1 Like

Sure, I should stay far away from electronics because admitteldly I know too little, but that’s not the point with that schematic… It’s intolerable to have those 4-way wire crossings without explanation while there are both dots and “hoops” in other places. It makes me believe this shcematic was drawn with a generic diagram editor, not with a schematic editor. The one who published it should stay clear of electronics. Or change to KiCad for drawing, at least, and avoid 4-way junctions and use dots in every junction…

Agreed that it can be ‘tricky’ at times to properly read a chip manufacture’s internal functional representation of any given chip. Those are just functional ‘summaries’ of a chip.

You could look at it as an executive summary. If you want to know more, or exactly what happens inside it’s time to really dig in much deeper.

Having said that it is well acceptable/tolerable since both left hand side crossings constitute implicitly in their functionality no connections. In this sense there is no need to explicitly point out that fact. Too many connection information can clog up a schematic. It’s meant to be a summary (block diagram) after all.

One could also argue that those yellowish rectangles are resistors while they clearly stand for pins. Better still one could argue that the right hand side NC pin should have a dot to the fine line boxed around the block schematic. No need for any of the two above. Besides page 12 of the datasheet clearly states who is who of the pins.

On a side note, the PD70100 schematic extract is a misrepresentation in nickm’s above post. Refer to page 4, figure 2 of the datasheet. Those pins are anything but colored in, which could change the junction dot discussion for that case in point quite a bit.

@jos you are not entirely wrong, but we all started somewhere. @eelik I have to agree mostly with you here.

Any inconstancy requires extra time to decipher.

If just one of those dots gets scrubbed off, it makes the schematic inconsistent. If a second dot gets rubbed off it makes even more inconsistent.

Any inconstancy requires extra time to decipher.

I started fiddling around with computers and electronics at age ~13 with Radio Shack documents. It was a bit of a shock to end up at age 18 working with documents that were significantly different. I’ve been there, done that, and was issued a T-shirt to wear. I get the reluctance to change.

Much blood, sweat, tears, time, and money has been spent for me to form my opinion on Junction Dots; anyone reading my posts on this topic gets that expensively created opinion for free.

And, finally, it seems to me to be such a silly topic, I do not understand why the forum threads seem to get so emotionally heated. In the end, it is just a dot, or it is not.

I am not arguing against avoiding 4 wire joints!

It tells you that a dot got rubbed of. Which tells you that the printout got damaged in some way. Which then tells you that you might want to take it with a grain of salt. Again it adds information. In this case information about the state of the printout.

If it was an inconsistency at design time yes. Otherwise i would argue it is not extra time but a well deserved reduction in trust in the document.


In your scheme i would simply compare the junction dots to some error detection code. It is redundant information that can help in detecting the fact that the data you have at hand got damaged.

1 Like

How many times do you think the radar schematic is printed out for one of these?

The document is always going to be somewhat questionable with all the circumstances.

@Rene_Poschl You can have all the dots you want. In my opinion, you might change your mind after you have had to take them into the field.

So you tell me it is better not to know that a schematic is damaged?

Remember the schematic stays readable without dots. As they are only redundant information.

Also i was in the field. Not military but very old industrial machinery. Oil is quite efficient in destroying paper documentation so i have seen my fair share of damaged documents. I was very lucky (at one of my jobs) that the documents had a lot of redundant info in there which made it possible to reconstruct of the required information. Reduced the amount of work i had to do with reverse engineering. Still took us quite some time to find out what that thing was supposed to do. (That particular job was replacing the “brain” of that old thing.)

1 Like

From now on “bikeshedding” will be called “junction dotting”.

5 Likes

Sums it up

3 Likes

@Rene_Poschl

I was working on a real world answer when the thread went sideways.

If you care for honest argument, I’l try to keep my side influential.

By reading your original post which started this thread I can’t tell how serious or to-the-point this thread should have been.

1 Like

Try to…

…no more.

I have to push back on the idea that major government industries know best, therefore we should all follow what they do. By blindly following all the little rules you get perfect results every time… not. Major government industries are renowned for going spectacularly over budget and schedule. Even when they are on schedule, there is a large price premium.

I have worked for several major government contractors, and often their processes and procedures are absolutely terrible. It does not surprise me there are rules based on obsolete methods from 40 years ago. I am often saying “why on earth are we doing like this?” and the answer is someone 30 years ago wrote a document saying that was the best way, and no ever questioned it since.

Without cost plus, these companies wouldn’t last 5 seconds in a commercial environment. My instinct when a major government industry says “everyone should do it this way” is do the opposite.

1 Like