Discussion: Rules For Drawing Readable Schematics

[Tony Stewart]

Is Fig1 supposed to be an example of a bad schematic? Q1 has no emitter/collector distinction .

It could be improved in many ways besides BJT, missing details, minimal line paths, but the concept of fill factor, font,/ trace size, symbols dots, redundancy reduction makes it very easy on the eyes. But then this is only a conceptual schematic unlike professional standards.

 

[NorthGuy]

It could be improved in many ways besides BJT, missing details, minimal line paths, but the concept of fill factor, font,/ trace size, symbols dots, redundancy reduction makes it very easy on the eyes. But then this is only a conceptual schematic unlike professional standards.

They probably paid at least few hundered dollars for that drawing

 

[Tony Stewart]

They probably paid at least few hundred dollars for that drawing

I think it could be done is about 10 minutes.

 

[Tony Stewart]

Few years ago, I've got a virus through Adobe Acrobat Reader which automatically opens PDFs. Had to rebuild the computer. Took me few days. I don't use Adobe any more. I now save PDFs and watch them with off-line reader. I have quite a collection

I never use Adobe but it is very secure with the proper setup. I use Foxit or Sumatra with Firewall blocked .

 

[NorthGuy]

I think it could be done is about 10 minutes.

May be I should hire you to draw schematics. How much would you request for such a drawing?

 

[Tony Stewart]

My draftsman would take half an hour to turn my paper napkin block diagrams like my example 2 and turn them into example 1. Manfred was a great artist whom I worked with at two different R&D companies.

 

[MrAl]

Hi,

A draftsman at a company i once worked for used to draw lettering that looked like he used a stencil. I actually saw him do it by hand too, and almost couldnt believe it. Back then we used 2 feet by 4 feet paper drawings (like blueprints) so all the drawing was done by hand (also took a large cabinet with huge drawers to hold the drawings flat).
He had a little trick but i cant remember what it was now...like holding a ruler under the place where the lettering would be going, to keep the line straight...but cant remember that far back now. But there was no other guidance for drawing the lettering. Using those special tipped ink pens made just for drafting.

 

[DerStrom8]

Hi all,

I have split this discussion off into its own thread. I think it is better placed here than in the original sticky.

Regards,
Matt

 

[DerStrom8]

I think it's time we come up with a standard, to which we can point people who have questions regarding schematics, styling, or who come to us with sloppy ones. I have locked the original sticky but will make changes to it if several people agree. So let's start brainstorming and come up with some ideas that others can "vote" on to add. I'll start with the following:

1) Avoid angled wires in schematics
2) Schematics should be presented in PDF format or a high-quality PNG. No posting program-specific files (i.e. ".sch") simply to show a schematic.
3) Chips should be depicted using pin positions that best prevent crossing wires. If you can match the physical pinout without causing many of the wires to cross, then by all means do it.

Those are the three points I've picked up so far from the thread. I am hoping to get a collection of rules that most everyone agrees on, which I can then add to the sticky.

You may also put forward arguments for why proposed rules should not be implemented, but let's try to keep it on-topic and avoid side-discussions.

Regards,
Matt

 

[MrAl]

Hi,

I found a flaw in the 'corrected' drawing in post #1. This is a minor point, but a point nonetheless.

If you look at Item #3, it shows a transistor circuit not draw well at all. If you look at the 'corrected' drawing it is better, but not quite up to par. Notice the dots on the base connections and the collector connections. We could have draw these two with two less dots, by combining the connection points for the upper base resistor and output collector terminal connection.

This kind of thing should also be avoided. So i guess the rule here is, "Minimize the number of dots for connections when possible".

When i saw that thread again (after seeing the thread had been moved) i saw that drawing first before the one before it, and thought that was going to be an example of using too many dots in a schematic drawing but lo and behold it was the 'corrected' one (har har)
Not bad, but not as professional as it could be.

 

[DerStrom8]

Hi,

I found a flaw in the 'corrected' drawing in post #1. This is a minor point, but a point nonetheless.

If you look at Item #3, it shows a transistor circuit not draw well at all. If you look at the 'corrected' drawing it is better, but not quite up to par. Notice the dots on the base connections and the collector connections. We could have draw these two with two less dots, by combining the connection points for the upper base resistor and output collector terminal connection.

This kind of thing should also be avoided. So i guess the rule here is, "Minimize the number of dots for connections when possible".

When i saw that thread again (after seeing the thread had been moved) i saw that drawing first before the one before it, and thought that was going to be an example of using too many dots in a schematic drawing but lo and behold it was the 'corrected' one (har har)
Not bad, but not as professional as it could be.

My concern with that would be at 4-way junctions. We should try to avoid 4-way junctions whenever possible, and while only one dot would be needed, for clarity sake we should jog one of the wires over and add a second dot. Below is an example:

The above should generally be avoided, as it may be difficult to see the dot and one might assume the wires do not connect. Or, perhaps, they are simply crossing over one another and someone assumes the above is true (they are connected). Instead, this type of 4-way junction should be drawn as follows:

This eliminates all doubt that the top wire and bottom wire are connected to the horizontal one. However, it has two dots rather than one.

I understand where you're coming from, but I feel it should be reworded. Otherwise it could be more confusing.

 

[alec_t]

3) Chips should be depicted using pin positions that best prevent crossing wires.

I nearly always create schematics with LTspice. The chip symbols in the standard libraries don't usually have such conveniently-positioned pins. It would be a major effort to convert the symbols and a major inconvenience to have to use alternative drawing software.

 

[MrAl]

My concern with that would be at 4-way junctions. We should try to avoid 4-way junctions whenever possible, and while only one dot would be needed, for clarity sake we should jog one of the wires over and add a second dot. Below is an example:

View attachment 93018
The above should generally be avoided, as it may be difficult to see the dot and one might assume the wires do not connect. Or, perhaps, they are simply crossing over one another and someone assumes the above is true (they are connected). Instead, this type of 4-way junction should be drawn as follows:

View attachment 93019

This eliminates all doubt that the top wire and bottom wire are connected to the horizontal one. However, it has two dots rather than one.

I understand where you're coming from, but I feel it should be reworded. Otherwise it could be more confusing.

Hi there,

Well, i have to disagree on the basis that if we have dots that are legible then the drawing should be clear. Jogging the dots just makes the connections seem to be required to be near a certain component, which brings up another related point.

Sometimes it is very necessary to show a connection not only as a connection is normally shown on a schematic, but in a way that mimics it's physical placement as well. In this case 'jogging' the dots would be necessary to show that the leads are connecting very close to a given object. For example, a sense resistor that needs the two sense wires shown near the resistor. If they are far, it does not convey the message as clearly when reading the schematic. Yes a secondary note is in order, but the placement is important too. An extreme case is where we need something like a 'star' ground, where all the leads must go to a common point, which is a physical point not just a connecting point. For example on a modern buck regulator circuit.

We could go on and on about this stuff, but i see where this is going already. It's going in the direction of personal preferences. You like this, i like that, someone else likes something totally different...it's very hard to nail this stuff down in part because it is a drawing and a drawing is artwork and artwork is not totally objective, but feel free to keep going with it and see where it ends up

 

[NorthGuy]

I agreee about 4-way connections. They're quite common, and I don't remember ever missing a dot. The suggested alternative looks messier.

 

[DerStrom8]

Jogging the dots just makes the connections seem to be required to be near a certain component

Not quite sure I'm following...?

I agreee about 4-way connections. They're quite common, and I don't remember ever missing a dot. The suggested alternative looks messier.

I came across an old Protel drawing at work where the dots had faded, and the other engineers and I were struggling to determine which were connected and which weren't. It would have been far easier if they had been jogged, whether the connection dots were visible or not.

 

[KeepItSimpleStupid]

Another point that I don't think he made, in schematics wires should ALWAYS go either left, right, up, or down. When people draw schematics with wires going off at angles it makes it very sloppy and difficult to follow. The only exception I can think of is in a bridge rectifier schematic, but even then I would prefer that it be redrawn to be either vertical or horizontal.

There was one particular schematic that I had to use that used a unique way of connecting stuff across the page. The lines were vertical and horizontal, but some had angled ends with arrows. You could then take a ruler and connect to the signal across the page. Very cool.

Of course, schematics should be populated with cartoons like the Tek schematics had. https://mindjaunts.blogspot.com/2006/01/cartoons-in-tektronix-schematics.html

 

[NorthGuy]

I came across an old Protel drawing at work where the dots had faded, and the other engineers and I were struggling to determine which were connected and which weren't. It would have been far easier if they had been jogged, whether the connection dots were visible or not.

Dots cannot fade in the schematics posted to the forum!

 

[MrAl]

Hi again,

Not quite sure I'm following...?

I thought i made it clear with the couple of examples but let me expound this a little more.

One example is a drawing for a buck regulator like maybe the LM2576 if i remember right. That shows 'heavy' lines for the important lines that all connect together, and i think they all come to the center of a 'star' connection. This example helps to show the true artistic nature of schematics (ie not a science).

For another example, say we have a sense resistor to sense current, say it is to measure the current on the output of our newly designed power supply. The sense resistor is physically connected in series with the output; a necessity. We intend to measure the voltage across this resistor with an AD converter in a uC chip, something we all have done at one time or another. The best way to run the sense wires in real life (not the heavier current wires) to the resistor is to connect them right to the terminals of the resistor, not say on the output of the regulating transistor and one somewhere else, even though the heavier wiring is of a heavy gauge. Our intent with the wiring is to keep the voltage drops anywhere else from affecting the reading and help prevent noise pickup, we just want the voltage drop across the resistor, and we might even twist the wires together. So for this reason the wires are connected directly from the resistor leads to the circuit board near the uC chip or to leads that run to the uC chip. So we want to show that this is important on the schematic, so we draw lines right to the resistor rather than somewhere else that also connects to that resistor. We have four lines now coming from the resistor, two for one terminal and two for the other. One lead on each terminal is heavy for the current itself, and the other lead is thin for the sense lead and has little current flow. But to show this on the drawing we run the lines right to the resistor, and that means we ignore other rules and run the lines to the resistor showing a dot where they connect. If we twist the wires then we want to show that too, as a twisted pair often with lines that criss cross multiple times on the drawing to clearly show those twisted lines.

I came across an old Protel drawing at work where the dots had faded, and the other engineers and I were struggling to determine which were connected and which weren't. It would have been far easier if they had been jogged, whether the connection dots were visible or not.

Well, take the number of drawings you could not read and divide by the number of drawings that you could read plus the number you could not read, and then multiply by 100. That will give you the percentage of drawings that you could not read over your lifetime so far:
PercentCouldNotRead=NumberCouldNotRead/(NumberCouldRead+NumberCouldNotRead)

I think that will put the problem is a more balanced light, but please read the third point next...

This also brings up a third point about the connections. A convention we always used back in the day was to draw a small "half circle" to sort of "jump" over the line when there was no connection. This would mean that any connection that was not really a connection would cross the other non connected line with a half circle rather than just a straight line crossing another straight line.

I hope i have made this clear but i could provide a drawing if that would help.

 

[DerStrom8]

Dots cannot fade in the schematics posted to the forum!

They can if they're scanned, or hand-drawn!

(I've seen both here)

I thought i made it clear with the couple of examples but let me expound this a little more.

I can understand what you mean when referring to shunt resistors--Connecting the wires directly on each side in the schematic makes sense, because it's clear to see that it is a shunt resistor. But what does that have to do with a 4-way junction? Shifting one of the wires over a fraction of an inch shouldn't be confusing at all.

Well, take the number of drawings you could not read and divide by the number of drawings that you could read plus the number you could not read, and then multiply by 100. That will give you the percentage of drawings that you could not read over your lifetime so far:
PercentCouldNotRead=NumberCouldNotRead/(NumberCouldRead+NumberCouldNotRead)

It takes just one bad schematic to ruin your day, and maybe your next batch of boards

I think that will put the problem is a more balanced light, but please read the third point next...

This also brings up a third point about the connections. A convention we always used back in the day was to draw a small "half circle" to sort of "jump" over the line when there was no connection. This would mean that any connection that was not really a connection would cross the other non connected line with a half circle rather than just a straight line crossing another straight line.

An excellent point! I still prefer this practice and use it in my schematics. Some people do not, however, which is why it is important to work out a "standard" for the forum.

Regards,
Matt

 

[MaxHeadRoom78]

. A convention we always used back in the day was to draw a small "half circle" to sort of "jump" over the line when there was no connection. This would mean that any connection that was not really a connection would cross the other non connected line with a half circle rather than just a straight line crossing another straight line.

.

I have never been in favour of this method, in this style termination dots are not used at all, but if a line crosses or terminates at a line without the half circle then it is deemed connected.
I much prefer the dot termination and straight line crossing.
Especially using drawings where each one uses one of these alternate methods.
Max.