Check my PCB layout before I send it to PCBfabExpress

Started by Dr Ron, May 17, 2006, 02:14:57 PM

Previous topic - Next topic

Dr Ron

I have finished laying out my PCB (using Osmond software) and am now ready (I think) to send it to PCBfabExpress.com
It's a two-layer board, so I can't really make a prototype with Press and Peel Blue, which has me a bit concerned since it's my first PCB.

If you are familiar with Osmond, can you tell me if it's as simple as printing it out to Gerber?
(It passed the "Check Layers" test and all the paths are connected.)

You can view the schematics and PCB layout online:
http://www.naturdoctor.com/Looper/Looper_Schematics.pdf
http://www.naturdoctor.com/Looper/Looper_PCB1.pdf
http://www.naturdoctor.com/Looper/Looper_PCB2.pdf

FYI - this is a looper with 4 x 15 second loops.

What I'm looking for is simple things ... for instance, I went through and curved most of the paths.
There are some right angles, however,

Thanks!!

(PS - Is it not a good idea to post schematics and layouts?)


Dr Ron

Quote from: TELEFUNKON on May 17, 2006, 04:35:32 PM
oouuuff !

Yeah ... it is more complicated than I expected when I started.

Here's an easy question:
One layout book said that curved paths are "old school" and that 45˚ angles are more professional.
Osmond has a handy "curve path" tool, which I used a lot ... it that OK?

And another one:
Should I increase the size of the ground and power traces, or are they fine?

TELEFUNKON

haven`t seen groundplanes, only traces. where are they?

Dr Ron

Quote from: TELEFUNKON on May 17, 2006, 04:46:56 PM
haven`t seen groundplanes, only traces. where are they?

That's because I'm not really sure that I need a groundplane with only two layers.
One side is connections, and the other is ground, 9V and 5V.

Are you recommending that I increase the thickness of the power and ground traces?

Jaicen_solo

#5
That looks like a really nice design. Is that logic circuit addressing the chip from the apps note??
I was studying that a while back but it's a bit beyond me. How did you force the chip to record only 15s per address?
Also, will it go directly from record to playback without pressing reset? Have you actually breadboarded this beast??
Again, nice work!  :icon_eek:

EDIT: I forgot to mention that a ground plane is a very good idea with this chip, as it's notorious for being quite noisy. I haven't checked the layout yet, but have a few comments.
It's quite important to have decoupling caps on the both the digital and analogue power pins right next to the chip. Also important to keep the digital power seperate from the analogue as much as possible to avoid bleedthrough. Personally I don't have a problem with the old school curves in your layout, except where signal lines are running near power lines. You'll want them to cross at right angles wherever possible, again, to minimise noise. If you're having the PCB's manufactured, thin traces aren't usually that big a deal, though I make mine as large as is practical normally. I'd be interested in getting one of these PCB's if they work, pm me if you're having a few done.

Dr Ron

>> That looks like a really nice design. Is that logic circuit addressing the chip from the apps note??

I combined several application notes, and then designed the logic to do what I wanted based on what the chip needed.
It's a complicated chip, and it took a lot of thought to get things right.

>> I was studying that a while back but it's a bit beyond me. How did you force the chip to record only 15s per address?

The chip reads the memory address when CE is pulled low, which occurs at the end of a message.
Everything became a lot simpler when I realized that only the top 2 address bits needed to change.

>> Also, will it go directly from record to playback without pressing reset?

I'm not sure about that.

>> Have you actually breadboarded this beast??

I think I should do that before getting a PCB done.

>> Again, nice work!  icon_eek

EDIT: I forgot to mention that a ground plane is a very good idea with this chip, as it's notorious for being quite noisy. I haven't checked the layout yet, but have a few comments.
It's quite important to have decoupling caps on the both the digital and analogue power pins right next to the chip. Also important to keep the digital power seperate from the analogue as much as possible to avoid bleedthrough. Personally I don't have a problem with the old school curves in your layout, except where signal lines are running near power lines. You'll want them to cross at right angles wherever possible, again, to minimise noise. If you're having the PCB's manufactured, thin traces aren't usually that big a deal, though I make mine as large as is practical normally. I'd be interested in getting one of these PCB's if they work, pm me if you're having a few done.

Jaicen_solo

I think you should definitely breadboard this design before building it as I have a few concerns.
Primarily, according to the info from Windbond, there is an audible click at the beginning of playback when any address except 0 is used, due to the way that the sound is stored. I don't know how big a deal this is, but it's certainly something check before paying out for an expensive PCB.
I think the design as it stands will also not go from playback to record without resetting first, which is not a big deal just another button click. If you search for my looper thread, theres a neat little circuit to get around that.

If my understanding of the way you've implemented this is correct, you have the CE pin wired to the EOM pin. So every time the looper reaches an EOM marker, it will reset the pin, and begin playback depending on the address selected by the logic circuits??

Dr Ron

>> I think you should definitely breadboard this design before building it as I have a few concerns.

I'm now realizing how important it is to breadboard a prototype!

>> Primarily, according to the info from Windbond, there is an audible click at the beginning of playback when any address except 0 is used, due to the way that the sound is stored. I don't know how big a deal this is, but it's certainly something check before paying out for an expensive PCB.

At one point, I thought I'd read that this wasn't a problem ... but maybe not.

>> I think the design as it stands will also not go from playback to record without resetting first, which is not a big deal just another button click. If you search for my looper thread, theres a neat little circuit to get around that.

Thanks!

>> If my understanding of the way you've implemented this is correct, you have the CE pin wired to the EOM pin. So every time the looper reaches an EOM marker, it will reset the pin, and begin playback depending on the address selected by the logic circuits??

Yes.
In this way, you could record several ways: one 60 second, a 15 and a 45 second, 30 and 30, etc.
An audible click on playback will definitely ruin the whole thing, however!

Dr Ron

About the audible click on looping:
I designed the circuit around the ISD paper: "Circuit Examples of ISD1000A and ISD2500 Products", Figure 4: Chip Enable Initiated Looping, which states: "Since the original message has continued to play during this pulse, the message is smoothly restarted without a break."

As I understand this, a click will occur if we start with something other than loop 1.
The circuit is designed so that the when you press a loop button, the current loop will finish before the new loop is played.

I am reminded again of the importance of breadboarding a prototype.

Jaicen_solo

I'd say breadboarding is absolutely essential! I'm guilty of the same thing with many of my own layouts.
It seems there's not a great deal of interest in this thread, perhaps it's looking too complex I don't know. I'm excited by it, except for one absolutely obvious omission that i've only just noticed!
Looking at the analogue portion of your input and output circuits, there's no mixer for the circuit in and circuit output signals. As it stands, you're only going to be able to hear anything when you have the pedal in play mode. When you click it into record, you won't be able to hear what you're playing, nor will you be able to play over what's being played back.
Sorry to rain on your parade a little  :icon_redface:

snap

What happened to this project? Did it work out well?