Errors in my Schematic?

Thread Starter

urb-nurd

Joined Jul 9, 2014
269
I am submitting this schematic with a piece of school work in the near future.

I was wanting to have a second set of eyes make sure there is nothing erroneous in the layout if possible.

The circuit is a data logger for recording direction and acceleration of a car.

Here it is
schematic.PNG

Thanks folks!
 

jpanhalt

Joined Jan 18, 2008
7,925
Looks pretty good.
You have a few junctions that aren't dotted:

upload_2015-4-9_16-3-21.png upload_2015-4-9_16-4-29.png upload_2015-4-9_16-5-55.png
And this has an unnecessary bend:
upload_2015-4-9_16-5-21.png

You do not show a voltage for Vdd to the MCU. Of course, the devil is in the details. Does your program work? What are you outputting?

John
 

WBahn

Joined Mar 31, 2012
24,703
The Vdd value is shown at the output of the regulator. It is good practice to only show this value in one place (or, more specifically, to only have it defined in one place -- if you can define it in one place and have that value be used to show the value in multiple places, that's wonderful -- but few schematic capture packages allow that).
 

WBahn

Joined Mar 31, 2012
24,703
that pic has internal osc upto. 8mhz, no need for the Xtal.
Unless you want an accurate 8Mhz clock.

Having said that, they are calibrated at the factory so they are much better than the internal oscillators of old. I don't know what their specs are.
 

jpanhalt

Joined Jan 18, 2008
7,925
The Vdd value is shown at the output of the regulator. It is good practice to only show this value in one place (or, more specifically, to only have it defined in one place -- if you can define it in one place and have that value be used to show the value in multiple places, that's wonderful -- but few schematic capture packages allow that).
From Eagle 7.2.0

upload_2015-4-9_18-6-30.png

I will send Cadsoft-USA your advice. I think Cadsoft will ignore it. It has been that way for more than 10 years.

John
 

WBahn

Joined Mar 31, 2012
24,703
I'm sure they will ignore it. That doesn't change anything regarding what I said. Software vendors routinely follow bad practices and are hesitant to correct them because of the significant maintenance effort involved in doing so. Just because Vendor X has been doing it that way for Y years doesn't mean that it is good practice.

It is bad practice to manually label something in multiple places. What happens when you adapt the design and change the logic from 5V to 3.3V? Now you have to go through all of the schematics making sure that you change all of the labels. What are the odds that at least one label isn't going to get changed? If you have a CAD package that supports defining global attributes, then by all means do so, but if it doesn't, then you are setting yourself up for lots of problems if you define something in umpteen different places.

If you encode the value as part of the global label, such as naming the node Vdd5 instead of just Vdd or putting the voltage value as a label in the symbol, then when you adapt the schematics you can change the global label definition from Vdd5 to Vdd3 or change the label in the symbol and all of the schematic renderings will be updated. That is 100% in agreement with what I said.
 

jpanhalt

Joined Jan 18, 2008
7,925
First, in Eagle, one does not "manually" label the supply. It just so happens that the device chosen for 3.3V and 5 V have the same appearance on the schematic, except for the label (name).

You are assuming the supply to the MCU is also 3.3 V. Can you cite an authority, such as a standard, for the practice you suggest?



John
 

WBahn

Joined Mar 31, 2012
24,703
The output of the regulator is connected to a symbol labeled Vdd.

The supply pin of the MCU is connected to an identical symbol labeled Vdd.

If two identical symbols that are both labeled the same thing are not the same node, then that is really poor.
 

jpanhalt

Joined Jan 18, 2008
7,925
The output of the regulator is connected to a symbol labeled Vdd.

The supply pin of the MCU is connected to an identical symbol labeled Vdd.

If two identical symbols that are both labeled the same thing are not the same node, then that is really poor.
But can you cite a standard that supports your conclusion?

John
 

WBahn

Joined Mar 31, 2012
24,703
No, I can't. I can probably dig up several places that emphasize that over-defined systems and documentation represent poor practice, precisely for the reason I gave (as well as several others).

Can you cite a standard that supports YOUR position that a specific voltage has to be shown everywhere that Vdd appears?
 

jpanhalt

Joined Jan 18, 2008
7,925
Without some standard, you might have been more accurate calling it your opinion.

I do think that the TS might be advised to label each supply as either just Vdd or with the voltage. I prefer the voltage, as that is unambiguous and is what Eagle default uses. Eagle will catch you in ERC, if you use Vdd for two different voltages.

John
 

kubeek

Joined Sep 20, 2005
5,607
As far as I remeber from past experience with eagle, whatever you name that voltage connection is the name of the net it is connected to. Easy and simple, especially whe you have more than one supply rail, and of course nothing prevents you from using different symbols for different supply levels.
 

WBahn

Joined Mar 31, 2012
24,703
Without some standard, you might have been more accurate calling it your opinion.
Just where did I say that it was required or that it was per some standard? I said that it is "good practice" to only define things in one place, and it is.

Did you label it as YOUR opinion when you critiqued his schematic for not showing the specific voltage value for Vdd at the MCU, despite the voltage associated with the node Vdd being documented at the regulator?
 

Thread Starter

urb-nurd

Joined Jul 9, 2014
269
That's a good question.

My guess is that it is a 5V to 3.3V voltage divider (which he almost certainly does NOT want).
I copied those resistors from an example schematic, i didn't really look into the purpose as it was the last thing i did before giong to bed for the night.

Heres where i copied it from.
sdcard.jpeg

I didn't know why they were present, but added them just before shut down after seeing them on a few schematics.

Ill read through the comments in a moment and reply properly.

mornings are torture.
 
Last edited:

Thread Starter

urb-nurd

Joined Jul 9, 2014
269
It just occurred to me that the resistors could be to allows a MCU with a 5V supply to interface the 3.3V sd card. Which is what WBahn was saying i think.
If the MCU is powered with 3.3v (mine is to allow one regulator to power all components for ease) then i guess the resistors are un needed as the High voltage on the pins is 3.3V due to the supply to the UC being 3.3V .

Just a guess
 

Thread Starter

urb-nurd

Joined Jul 9, 2014
269
The SD card is connected to port pins RC2 through RC5 and is operated in SPI mode. A card holder is used to physically make connections to card pins. The voltage at output pins of the microcontroller is too high and can damage the input circuitry of the SD card. A pair of potential divider resistors (using 2.2K and 3.3K resistors) is used to lower the microcontroller output voltages to a level acceptable by the SD card inputs. The SD card is powered using a 3.3V regulated supply, obtained using a MC33269DT-3.3.


Given i am using a 3.3V supply to my MCU, then i can omit the resistors.

I have also updated the schematic to remove the erroneous connections, though i am still unsure the concensus of my labelling of VDD and voltages haha.
 

Thread Starter

urb-nurd

Joined Jul 9, 2014
269
Although, i think i will look into changing my supply regulator for a switching one.
That way i can power the device with one 3.7V cell.
The gps module i have chosen requires at least 4V, so i will need to be using both 5 and 3.3V for this it seems.
I guess i will change the output of my supply circuit to 5V and then use the resistors to supply 3.3V to the SD card.

if i look at the power requirements of the device, maybe i can get away with a 1.5V cell and a boost converter.

We shall see
 
Last edited:
Top