This topic contains 8 replies, has 2 voices, and was last updated by Hanno .
January 3, 2012 at 6:13 pm #10667
The schematic for the video frame grabber utilizing the ADC 08100 is different between the Circuit Cellar article & Chapter 7 (CONTROLLING A ROBOT WITH COMPUTER VISION-page 258) of "PROGRAMMING AND CUSTOMIZING THE MULTICORE PROPELLER MICROCONTROLLER:The Official Guide" textbook. Specifically, the magazine article utilizes a 4-bit connection & the textbook utilizes an 8-bit connection.
Obviously, it makes more sense to build the 8-bit version. I wanted to make sure the ViewPort software was configured for the 8-bit schematic wiring. Since the ADC 08100 is SMT…I opted to have Proto Advantage purchase 3 ADC’s directly from Digikey & mount them on TTSOP/DIP adapters. Total cost (from Canada to California) was less than $ 90.00 delivered.
I also noticed that an ADC 08060 (60 MHz. max) is available for about 1/3 cost & appears to be a "drop-in" replacement. I was wondering if you had any experience with this other ADC. The manufacturer’s data sheet also recommends a 10 uF. Tantalum capacitor in parallel with the 0.1 uF. capacitor for noise filtering which I didn’t see on your schematics.
I recently purchased the full-blown version of ViewPort & just received my order from Proto Advantage. I’m very excited about the possibilities of your CV design & ViewPort software. Is there an updated schematic & associated program files available so I can build this correctly?
It’s a lot easier than trying to keep a book open going back & forth. I’ve always highlighted each connection on a schematic as it’s terminated so nothing is skipped and/or miswired. Thanks for your help & continued support…January 11, 2012 at 9:45 pm #11734
While waiting for your response to my previous email…I went ahead & wired the ADC 08100 as follows:
D0-D7 wired to P0-P7 & ADC_Clock_pin wired to P8. From the source code in ViewPort it appears VIDEOSTART connects to D4-D7 for P4-P7 (VIDEOSTART + 3) & ADC_Clock_pin is P8. I did this because the textbook shows D0-D7 going to P16-P23 which are already tied to resistors for TV & VGA connectors in the DEMO boards & Proto boards. P16-P23 are tied to LED’s on the QuickStart boards as well.
P8 is wired to the Microphone on the Demo boards (which I’m not using for this design)…it’s OK on the Proto boards & QuickStart boards. P0-P7 are also wired to pushbutton inputs thru 100K resistors on the QuickStart board, but that’s not a problem either if the pushbuttons aren’t in use. Now…I had an idea for using a "4-bit window" that slides within the 8-bit ADC output.
Since ambient lighting conditions in my application will vary (outside 24/7)…it seemed like the signal level would vary. By testing the VIDEOSTART value…the window could slide left or right to increase or decrease sensitivity. For example, if signal < 5…rotate left to double sensitivity or if signal > 10…rotate right to antennuate signal by half.
I haven’t built the circuit yet to see how it looks in ViewPort, so I don’t know what effects, if any, it would have on the sync signals & their proper detection. Please let me know what you think & how this might be done in modifying the code. Thanks again for all your help…January 16, 2012 at 8:19 am #11733
I look forward to making you successful with video, it’s not easy but very rewarding and fun!
Good job finding someone to do the soldering for you.
Good find on the ADC, no, I have not seen that. Should be fine and allow you to sample at higher frequencies if you need to.
The PropCVCapture is flexible with regards to pins as you’ve found. To stream data to ViewPort only the 4 MSB bits are used to reduce memory usage on the Prop. I’ve found that it’s best to do video step by step- starting with confirming that the ADC is properly reading slowly changing analog values, then measuring the NTSC signal, and then using PropCVCapture to view a video signal.
I published a guide here:
Your schematic and ideas all sound good, keep us posted on your progress/success!
HannoJanuary 16, 2012 at 2:08 pm #11735
Thanks for responding so quickly after your vacation…I really appreciate that! I DL’ed the source code from the Circuit Cellar magazine article last night & DL’ed the PlotSamples.spin code to the Propeller EEPROM & then started ViewPort. I had a proper NTSC signal displaying on the DSO & VIDEO tabs at that time. Since it was midnight here…I got some sleep & I just printed the 6-page documentation for finishing the process.
I’m also awaiting samples from INTERSIL for an ADC with SOIC footprint. I’ve got SOIC/DIP adapter boards & I can hand solder them successfully. Proto Advantage also has "do it yourself" kits with a solder paste mask, adapter PCB & solder paste. I understand many are having success using a household "toaster oven". Once the solder mask allows the solder paste to be placed properly & the IC aligned with tweezers…it’s just a matter of melting the existing solder on the IC pins in the oven causing a reflow & subsequent bonding between IC & adapter board…FYI.
I’m really looking forward to seeing a video image & go from there. By the way, my application involves rotating equipment which has movement in 3-axis so I’m mounting the camera at a transverse top view to capture movement in all three dimensions. Now it’s time to fire up the Expresso Machine & get after it! Thanks again..January 16, 2012 at 9:06 pm #11736
Sounds great! Good luck getting it going.
HannoJanuary 18, 2012 at 7:16 am #11737
Atached is a modified copy of PlotSamples.spin which produces the proper NTSC signal in dso, lsa & video TABS in ViewPort V 4.6.2 I added PropCVCapture but I still don’t have an image on the STREAMING VIDEO window anywhere. I did verify the Propeller bullet selected in the OPENCV window…as noted in the line I added to the spin program. What did I overlook or leave out? Thanks again for your help…January 19, 2012 at 4:45 am #11738
You’re close, but to capture the image you need to run the code in the PropCVCapture object! You’re referencing the object, but not calling any code in it.
You need to allocate some space for the image, ie:
and before you vp.share, you should start the capturing process and register it with the conduit:
See tutorial #11.
HannoJanuary 19, 2012 at 8:27 am #11739
Attached is the code which I modified ver batim. I also included the version of PropCVCapture with the I/O pins configured per my previous schematic. The file does compile but when Dl’ed I get a "Failed to configure on Com10" prompt. I can Dl the previously submitted file & it works & communicates as before. I also noticed in this PropCVCapture a different annotation for registering with ViewPort (vp#HIVIDEO vs. video#HIVIDEO). Trying to use the vp#HIVIDEO format produces errors on compilation.
Another issue I’m having is the ViewPort can’t be reloaded without restarting the computer. It appears a TSR segment is hosing the memory on reload attempts. I’m running Windows XP Home edition-version 2002-service pack 3 for my operating system.
I did find the tutorials under the browse window…I never knew it was there. My version of ViewPort doesn’t produce the welcome screenshot in the manual…displayed on page 14 of the manual. I tried to DL Tutorial # 11 to the propeller & since PropCVCapture was configured for my hardware…I believe it should have worked…but it didn’t either. I need your help getting a downloadable file I can send to the propeller & generate streaming video in ViewPort ASAP…thanks.February 14, 2012 at 7:24 am #11740
I believe you’ve managed to resolve all issues, can you post an update with how you solved it?
You must be logged in to reply to this topic.