To Vias or not to Vias

Hi,

Do Vias have a disadvantage ? My placement of components on the board is forcing me to put quite some amount of Vias. Is this bad ?

Thanks and regards Rainer

Someone posted a study saying vias were worse than the dreaded 90 degree corners for signal reflection/degradation. I didn’t read the thing thoroughly but if you are doing high speed signals it might be best to avoid them on those lines if possible.

I don’t remember the paper quantifying vias though. He was testing 90 angles and the part I read was just a throw in line without backup.

Which router are you using? I did a simple board but the shove router eliminated all the vias I had used prior to discovering it.

Hi,

I am routing manually. Will try the different routers and no I do not have high speed signals. Nothing radio involved. Just getting voltage signals from Photoresistors or …transistors and using them in order to send signals to switches. Very low speed.

Rainer

It all depends on your application. There is always a tradeoff. Yes vias should be avoided. But other factors might mean that a via is still the best option in this case.

As far as i understand it, using a via in a signal line forces the signal return current to switch layers as well. (in high frequency applications. >=10kHz)
At least in 4 layer boards where you have the inner layers as supply layers.
This layer switching induces a voltage fluctuation into the supply planes which generates an EM wave. (That wave is later reflected by the board edge.)

For two layer pcbs the danger is that you create a slot beneath a fast signal where its return current might normally want to flow. Such a slot creates quite an effective antenna.
Signals don’t take the bath of least resistance but the path of least impedance. Which means if you have a power plane, the return current will want to take the path directly beneath your signal trace. (not the shortest distance as you might expect.)

This post on the contextual electronics forum is a good read for emi. (It is a bit simplified and i’m not 100% certain everything is truly correct. But the basics seem to be there.)

Hi,

No big secrecy in what I am doing and the market is not for millions of pieces :slight_smile:

I am measuring Sunlight with 2 LED, maybe 10 times a second, and send the voltages to a ADS1115 and get a differential voltage value which is then interpreted with an Arduino UNO as a plus or minus value and that values closes one or the other switch.

The cables from the LED to the Arduino are maybe 60cm long at maximum.

If you want to see the eeschema I can post it here.

Thats it.

Rainer

Hi,
I did take a look at my layout and I think I also have a problem of how I positioned the components. I will work on that again.

I will also exchange some single components to a Quad Component.
Rainer

Well, after a while of tinkering and making some changes I was able to get down from 23 to only o1 Via. I think that is a major improvement.

Thanks for the comments which helped me to rethink the whole enchilada.

regards Rainer

1 Like

Glad to hear this went well for you. KiCAD is a fantastic tool and the learning curve isn’t too steep for what it does.

Hi Hermit, Hi Everybody,

I tried the OpenGL but honestly the routing is so nervous and it was always jumping around and so I finished routing manually. Today afternoon I will try the other one.

I would like to thank everybody for the help with the Input you give which makes me rethink what I am doing and rethink and expereiment with new stuff.

I am now in my third or fourth week with this Project and still do not see the silverstripe at the horizon :slight_smile: but on the other side it is very interesting to develop a Hardware which will have a use in the Amateur Astronomy area.

:slight_smile:

Shove can do some weird things. I wonder if the grid setting affected it. You might play with that once you do angled chip. What kind of hardware? I had a friend making a microstepping controller for the Apple interface and the ATM list has been on my server for quite a while. A few other good lists out there.

I normally don’t use shove. I set up interactive drag and route in walk around. (I move everything out of the way via the interactive drag if necessary. I feel i have more control this way.)

Hi Hermit,

I am making a Solar guider for the Solar imagers who want to make time lapses. I already made one based on pure analog components. No microcontroller involved.

regards Rainer

According to Saturn PCB toolkit a via infers slightly over 1 nH inductance, has a step response of some 25 pS and a resonant frequency of ~6800 MHz. Capacitance 0.4 pF.
I would not worry about signal quality in non-RF applications. Even with MCUs at high clock frequencies (mine has 80 MHz): what is the frequency the traces really carry? Some 5 MHz for a SPI to a memory or so?

As for DC: 1.3 mOhms for a 10mil via for a 20° rise @ 25° ambient. So consider that for tracks carrying high current. And I mean HIGH, not 1-2 A :slight_smile:

1 Like

The article below adresses some via reliability issues.

  • Via’s might crack if the PCB suffers big temperature differnces.
  • Vias might explode if tented on both sides.
  • Good overall read.

https://www.eeweb.com/blog/eeweb/pcb-reliability-via-design1

3 Likes

You probably didn’t read the article?
My post was a very brief overview of linked article.
It does not say that vias are unreliable, it is an overview of some ways in which vias can fail (Which is exactly what OP asked for) and it also goes into measures to prevent these failures.
Excellent read.

1 Like

The first sentence of that article is:

And unfortunately lots of designers do not consider long term reliability very much.
The marketing department often decides it is a good feature when stuff breaks down after 5 year or so.
You are not right when you say:[quote=“Andy_P, post:16, topic:7266”]
but if it’s a choice between using vias and making a board, and not using vias and not making a board,[/quote]
The choice is between not giving it a moment of thought and designing PCB’s which might break down just after the warrenty expires or designing a robust product which will last 20 years.
For an engineer the difference might be so simple as moving a few vias away from a heat sink or power resistor. (or one of the other suggestrions in that article).

How many PCB’s have you seen in which low temp elco’s are glued to heat sinks? Is that ignorance, negliance, or do they do that on purpose because of marketing reasons?

Interesting reading and yes interesting Exchange of opinions.

In my former life I worked for a Company which made things " As good as possible " and now this Company was bought by another one which made things " As good as necessary " …

So, nowadays and I think especially in electronics the life time of a board is by far higher then the Use time of the electronics …

My boards are for Hobby persons and well maybe in 10 years I have a better idea of how to do it or there are other ways to do what my Hardware will do… Who knows …

The times since the really upcoming of all this short lived electronics has inmensly changed the way we consume …

OK that is 21st Centrury anyway … I am from the Generation of children who played in the dirt and now live surrounded by short term electronic Hardware …

So the decision is for the single Person what they expect from what they invent …

Anyhow. Thanks for the help so far to everybody :wink: Yesterday I did send the Gerber files to China, they did not have a single question about it. When I get them I will present them here :slight_smile:

There are some guidelines in that article that will be good to add to KiPadCheck. Look for them in the near future by watching that thread!