Shop OBEX P1 Docs P2 Docs Learn Events
Circuit Overlay Modules - Page 3 — Parallax Forums

Circuit Overlay Modules

13

Comments

  • idbruceidbruce Posts: 6,197
    edited 2020-09-04 22:38
    Publison
    Placement looks good, but I was having a hard time with traces landing on center of pads.

    You will have to be a little more specific. :) Are you refering to a layout created from the program or templates?
    To allow snapping on a 0.10" grid, except for final connection to the I/O pin of the Propeller, I will have to increase the board size for the overlays. I am starting to look into this now :)
    I will be testing out new board dimensions and header centers with the following parameters:

    NEW BOARD DIMENSIONS
    2.3 X 1.6

    NEW HEADER CENTERS
    JP2 - X 0.9 Y 2.15
    JP1 - X 0.115 Y 1.15
    JP3 - X 0.9 Y 0.15
  • Oh.... You were joking :)
  • Looks like 2-layer with vias going to the ground plan on the bottom?
  • idbruceidbruce Posts: 6,197
    edited 2020-09-04 23:05
    Yep.... It is going to be a 2-layer board, with approximately 8-13 vias, with the majority of them represented by the blue line in the picture below. However the gound plane is going to be on top.

    I also nudged a couple items. I think it is better now.

    board.jpg

    416 x 494 - 47K
  • If the ground plane is on top, why the vias below the chip?
  • There is a rectangular copper pour on the bottom layer, which aids in the heat dissapation from the stepper driver IC.
  • It is called a Power Pad IC. The attachment will tell you about these ICs.
  • Then the ground layer is on the bottom, right?
  • Publison

    I already told you, the ground layer is on top :) There is an open perimeter around the rectangular copper pour on the bottom layer, so it is isolated from the rest of the bottom layer.
  • idbruceidbruce Posts: 6,197
    edited 2020-09-04 23:58
    The power pad on the IC must be soldered to ground, because the pad is attached to ground on the IC. If the IC is to be mounted to the top layer, then the top layer must be ground, or there must be an isolated rectangular copper pour under the IC which goes to ground.

    EDIT: Two layers of copper, interconnected by vias, dissapates heat better than one layer.
  • Of course, it is no secret that EAGLE is my PCB editor of choice :) , however there are many things about EAGLE that I still have to learn. For example, I am currently working on designing a board and I wanted to add some additional items to the silkscreen, as well as hiding and viewing certain aspects of the board, so I needed to learn more about the layers of the board.

    During my quest for knowledge, I came across this handy Autodesk reference, which explains the purpose of every layer in EAGLE, so I thought I would share this information here.

    https://autodesk.com/products/eagle/blog/every-layer-explained-autodesk-eagle/
  • idbruceidbruce Posts: 6,197
    edited 2020-09-05 13:03
    Okay.... I had to do a bit more nudging, alter a couple of packages, and add a few things to the board layout that I thought was pertinent information. Here is a picture of the current board layout:

    board.jpg
    463 x 570 - 69K
  • Peter JakackiPeter Jakacki Posts: 10,193
    edited 2020-09-05 15:08
    If your smd components are on top then use that layer for the signal traces and keep the bottom layer for the large ground and heatsink, otherwise you always need vias to route signals.

    For instance, the P2 is mounted on the top in the P2D2 and I have the pad with thermal vias connecting to the main ground and heat sink under the pcb.
  • idbruceidbruce Posts: 6,197
    edited 2020-09-05 16:06
    Peter
    otherwise you always need vias to route signals.

    :)

    You and I both know that I would not have a clue, unless I had some serious guidance :)

    In this particular instance, I am using the Gerber files for the DRV8825 Evaluation Board as my guide. In the image below, you will see three panels, a board rendering, a top layer, and a bottom layer. In the board rendering, you will see two red rectangles, which encapsulate various vias, and I believe they are an absolute necessity, considering the board design. The vias in the left rectangle are necessary to get the motor outputs past the sensing resistors, from the IC, and of course, some of the vias in the right rectangle are necessary for similar reasons, or should I say, at least in my case. I am certain that someone with your skills could do it differently, but as for me, I need a guide, such as the Gerber files.

    3Panel.jpg

    EDIT: On second thought, I don't need the vias in the left rectangle :) or at least I don't think I will :)
    739 x 282 - 50K
  • I now noticed that both the top and bottom are ground planes.... My bad :)
  • idbruceidbruce Posts: 6,197
    edited 2020-09-05 17:08
    Alright, here is the latest news on the templates.

    The JP1 header, which houses all of the I/O pins, was so close to the board that the DRC was spitting out errors for each outlying pin. Instead of fussing around with the DRC and making alterations, I decided to increase the board size once again, on the X axis, by an additional 0.050 (in), which would allow me to bump over the JP1 header as well as all of the other components, along the X axis, in a (+) positive direction. I chose 0.050 (in) because I did not want a big, honking overhang, at the I/O pin header. This decision had the consequence of making the JP2 and JP3 headers center on a 0.05 grid, instead of a 0.10 grid. Sorry guys :(

    So once again......

    NEW BOARD DIMENSIONS
    2.3 X 1.65

    NEW HEADER CENTERS
    JP2 - X 0.95 Y 2.15
    JP1 - X 0.165 Y 1.15
    JP3 - X 0.95 Y 0.15
  • idbruceidbruce Posts: 6,197
    edited 2020-09-05 17:40
    At this point, all errors have either been fixed or approved. It is now time to install some traces :)

    EDIT: On second thought, I need to do a little more homework and add the necessary vias :( It never ends :(
  • Overlay.zip is now throwing a virus warning. It was fine a couple of days ago.
  • Overlay.zip is now throwing a virus warning. It was fine a couple of days ago.

    I will delete it for now, until I get the a new one made with the new dimensions.
  • Publison

    Perhaps it is a similar issue as I had described in this thread: forums.parallax.com/discussion/172034/problem-with-the-propeller-activity-board-wx-diptrace-files#latest

    Either way, I deleted the zip files which contained programs.
  • PublisonPublison Posts: 12,366
    edited 2020-09-05 22:46
    can you post your Eagle files?
  • I got side tracked by a phone call and now I have someone coming over soon. I will try to get the new templates ready for you as quickly as possible.
  • Publison

    My centers got screwed up somewhere between all the movements. I am trying to get to the bottom of it now.
  • idbruceidbruce Posts: 6,197
    edited 2020-09-06 03:25
    As mentioned in my last post, my centers for the headers got screwed up somewhere along the way. I had to fix alignment in both my current project and in a new set of templates.

    I will be adding the new templates to Post #2 of this thread, in just a minute or two. Here is the corrected board dimensions and center for the headers:

    NEW BOARD DIMENSIONS
    2.2 X 1.65

    NEW HEADER CENTERS
    JP2 - X 0.90 Y 2.05
    JP1 - X 0.165 Y 1.10
    JP3 - X 0.90 Y 0.15

    The funny thing is that I am back to being very close to the dimensions of Parallax's overlay boards :) The length is identical, and I just altered the width by 0.011", and I forget the reason why :)

    Anyhow, after making the necessary corrections, in the following picture, you can see that the layout is starting to look crowded.

    board.jpg
    484 x 557 - 75K
  • A good circuit needs a good schematic. Where is it? Without it you are prone to make mistakes and it only takes one dumb mistake to ruin the pcb, and I've made a ton of dumb mistakes.

    I'm not really liking the use of the overlay module, it is an awkward format with very little usable area and way too many pin headers. Is there a reason you are using it?

  • idbruceidbruce Posts: 6,197
    edited 2020-09-06 04:08
    Peter

    I have attached a picture of my schematic. This schematic was contrived from three different sources 1) DRV8825 datasheet (attached below) 2) The schematic for the DRV8825 Evaluation Board (attached below) 3) The schematic for the DRV8825 Carrier Board from Pololu (attached below)
    I'm not really liking the use of the overlay module, it is an awkward format with very little usable area and way too many pin headers. Is there a reason you are using it?

    Well of course there is a reason :) I have a project in which I need WiFi, BlueTooth, and a stepper driver. The Activity Board WX comes ready with most the stuff I need, except for a WiFi module, a BlueTooth module, and a stepper driver module. The WiFi module that Parallax sells is a perfect fit, it snaps right into the board. The BlueTooth module that Parallax sells, worked well for testing, but it does not fit into my plans, so I must make a BlueTooth module to plug into the Activity Board..... Okay okay let me sum it up.....

    I need everything that I need :) to fit into a small project box, nice and neatly, without designing a very complex board. So my thought is, use the Activity board, with modules that stack on top of each other, to provide everything I need. I may add a RTC... Blah Blah Blah

    schematic.jpg
  • Peter JakackiPeter Jakacki Posts: 10,193
    edited 2020-09-06 04:23
    Here's a suggestion. Skip the pot (coz I hate em) and tie xVREF to an RC from a Prop pin and use duty counter mode to generate a reference voltage between 0 and 3.3V to program the current from software.

    Aren't these modules designed to be soldered together or are you using stacking headers to pull them apart again? It seems to me that if I was not into designing hardware that I would use these modules, especially so for education, but if on the other hand I was designing hardware, then I'd just make what I want how I want.

    When I was using stepper motors I liked using the L6470 for this task as it had a lot more features and was very programmable over SPI bus.

    Now this "Blah Blah Blah", is that the beer talking :)
  • Peter
    Here's a suggestion. Skip the pot and tie xVREF to an RC from a Prop pin and use duty counter mode to generate a reference voltage between 0 and 3.3V to program the current from software.

    I think you may have had a few beers :) Don't forget that you are talking to the electrically illiterate :) That is a nice suggestion for an overall improvement to the driver, however considering that I lack a fair amount of electronics skill, this would be difficult for me to implement. Additionally, at the moment, I have one particular motor that I want to drive, so being able to alter it dynamically is not really something I need, but it is a very good suggestion, and it would be nice.
    Aren't these modules designed to be soldered together or are you using stacking headers to pull them apart again?

    I began this thread with the intention of having stackable modules that could be added or subtracted as needed. There could be many different modules, that could be stacked on top of each other, as I mentioned, BlueTooth, RTC, Compass, I/O expander, stepper driver, the list is endless, until you run out of Propeller pins :) The main problem with stackable modules is the way the Propeller I/O pins attach to the module. If there was some type of routing mechanism, to where you could designate which pins get used by which module, instead of being hardwired, then that would really be something. There is a big difference between the electrical geniuses and the people that just want to play. The modules would be good for the people that just want to play. But in my case, I am doing it to get a product ready for demonstration.
    When I was using stepper motors I liked using the L6470 for this task as it had a lot more features and was very programmable over SPI bus.

    Here we go :) You know I like reading and talking about stepper drivers :) I love stepper motors :) MMMMMMMMM I have so many different stepper driver ICs laying around the house, it is ridiculous :) And I have not done anything with them :( I bought the G251X Gecko drives and it stagnated my desire to toy with the various ICs, because I wanted to build machines. All of my stepper firmware is based upon the 1/10TH microstep of the G251X. I wish the DRV8825 had a 1/10TH microstep :(
    Now this "Blah Blah Blah", is that the beer talking :)

    Do you really have to ask :)
  • It stopped... I wrote an answer to that quote, but it is gone :(
  • Please forgive my whining.... I'll just beer a moment....

    I don't think for a moment that you are in any way electrically illiterate, a bit unsure maybe, but with your skill set you should pick up the basics, if only you could put down that mind-numbing beer :)

    Now when I say an RC circuit, then that's all it is, a series resistor from a Prop I/O connected to a cap to ground. This forms a simple filter that averages the duty cycle of the high-frequency waveform so at a 25% duty cycle produces 1/4 of the 3.3V or 825mV approx. So effectively it's like a DAC. I'd use something like 1K and 0.1uF.

    There are two counters per cog and setting up a counter for duty cycle mode is dead simple and even if you just set it once in software you can do away with a trim pot. But if you can vary the current dynamically you can overdrive the stepper during acceleration and reduce the current when you are stepping slowly etc. This also helps the chip thermally too. Having options, especially when they are so simple never hurts.

Sign In or Register to comment.