Now I ran into funny thing. I switched from 1n4148 diode to red led, and changed the circuit a bit. HSYNC/VSYNC is driven by 1kohm pull-up before the led, and colors are driven after the led with small resistor. The end-result: better picture quality.
I'm personally a bit baffled, ran into this by accident while shuffling parts on breadboard.
Here's schematic:
Anybody has any idea why I get less "ghosting" with this setup?
Now I ran into funny thing. I switched from 1n4148 diode to red led, and changed the circuit a bit. HSYNC/VSYNC is driven by 1kohm pull-up before the led, and colors are driven after the led with small resistor. The end-result: better picture quality.
Anybody has any idea why I get less "ghosting" with this setup?
A RED led is basically a slowish 1.6v or 2v diode (actual drop depends on process used, 1.6v is rarer these days )
That means you have shifted the PIN drive operating point, and the SYNC extraction probably prefers that.
Q: I'd tend to have the same series drive elements in the BLUE channel, so the VGA is more balanced, and anyway, Series LEDs gives a nice 'I'm working' display :)
A RED led is basically a slowish 1.6v or 2v diode (actual drop depends on process used, 1.6v is rarer these days )
That means you have shifted the PIN drive operating point, and the SYNC extraction probably prefers that.
Q: I'd tend to have the same series drive elements in the BLUE channel, so the VGA is more balanced, and anyway, Series LEDs gives a nice 'I'm working' display :)
Thanks for the info, these LEDs are pretty old ones, I'm not 100% sure but I think LTL10223W. LEDs shining on board is nice yes and it works just fine :) BLUE channel is driven from RESET-pin, which is "weak IO" according to datasheet. Together with VGA 75ohm pull-down it seems to not oversaturate, it's pretty much in line with other colors.
Posted by AtomicZombie: Mon. Jan 23, 2017 - 08:50 PM
1
2
3
4
5
Total votes: 0
Great stuff man!
I just received my Tiny5's, and will fire them up next on the free day.
Just a thought on the color / sync sharing...
You may have seen this as well, but it will only work on perhaps 50% of the monitors out there.
The issue is that the pull-up does introduce some voltage back into the analog color, and this is a real issue.
I always test these crazy projects on the 12 monitors I have in my lab, and the sync share only worked on 4 of them.
On the one monitor it did work well on, you could adjust the horizontal position and actually see the light red color being drawn in the blanking area.
Since your monitor does accept this anomaly, go ahead and try this by adjusting horizontal position way out of bounds.
Most monitors have a real issue with color data entering the blanking interval.
This is why I abandoned the shared sync method.
Brad
Jartza wrote:
Hi folks,
Here's my latest, made it a small demo with ATtiny5 (runs with ATtiny4 too, but I couldn't find supplier for those) to carry around in pocket :)
I just received my Tiny5's, and will fire them up next on the free day.
Just a thought on the color / sync sharing...
You may have seen this as well, but it will only work on perhaps 50% of the monitors out there.
The issue is that the pull-up does introduce some voltage back into the analog color, and this is a real issue.
I always test these crazy projects on the 12 monitors I have in my lab, and the sync share only worked on 4 of them.
On the one monitor it did work well on, you could adjust the horizontal position and actually see the light red color being drawn in the blanking area.
Since your monitor does accept this anomaly, go ahead and try this by adjusting horizontal position way out of bounds.
Most monitors have a real issue with color data entering the blanking interval.
This is why I abandoned the shared sync method.
Brad
Hi!
Thanks for kind words. I really hope I would be better in "demo effect coding", but I'm actually not really that much graphics effect oriented, although I would love to make something more visually appealing out of this one day :)
I changed the circuit on purpose to use LEDs and switched the color behind the LED acting as diode because I couldn't get the previous circuit running on all monitors - so far this has worked with every monitor I have access to (7 of them). For the PCB I had to change the color series resistors to 120ohm from 220ohm, as I used different LED than on breadboard (on PCB it is LT Q39G-Q1S2-25-1). Looking at oscilloscope, when color is not active, the voltage pretty much goes to zero, sometimes even a tad on negative side (few tens of milliohms). Maybe you could give my circuit a go with your monitors?
That says the issue is around package. Digikey shows the 8UDFN as active, but the SOT-23-6 as obsolete.
Looks like a move away from gull-wing, to QFN. Could be largest volumes use QFN, and SOT-23-6 is not viable.
Asia suppliers tend to favour Gull Wing, but I've seen a trend for USA vendors to favour QFN.
Time to look around for alternatives ?
Hi,
Ahh, didn't notice it was only single package that gets obsoleted. I think it's pretty understandable. I've been working with 2 customer projects, which were using ATtiny5 (and that was the reason I even had those chips handy, because some left-overs from development boards, in SOT23-6 size) were using UDFN package. I'm pretty sure most products looking for small are REALLY looking for small :)
I changed the circuit on purpose to use LEDs and switched the color behind the LED acting as diode because I couldn't get the previous circuit running on all monitors - so far this has worked with every monitor I have access to (7 of them).
That makes sense, as the LED offset gives more TTL like levels on the sync, so should be more tolerant.
AtomicZombie wrote:
I always test these crazy projects on the 12 monitors I have in my lab, and the sync share only worked on 4 of them.
On the one monitor it did work well on, you could adjust the horizontal position and actually see the light red color being drawn in the blanking area.
Since your monitor does accept this anomaly, go ahead and try this by adjusting horizontal position way out of bounds.
Most monitors have a real issue with color data entering the blanking interval.
Yes, I've found the colour data in Blanking confuses the smart-auto-adjust side of LCD monitors.
If you want perfect centering, it is best to avoid colour in blanking areas, as that is used to decide where to set the margins.
I've also found that black backgrounds (no borders) are not tolerated well, for the same reason, but a 'windows like' background will give better 'auto' operation.
For this type of minimal / marginal work, maybe not-quite-right auto-adjust issues are tolerated ?
If you want perfect centering, it is best to avoid colour in blanking areas, as that is used to decide where to set the margins.
I've also found that black backgrounds (no borders) are not tolerated well, for the same reason, but a 'windows like' background will give better 'auto' operation.
For this type of minimal / marginal work, maybe not-quite-right auto-adjust issues are tolerated ?
Sure, I guess I wouldn't mind much if auto-adjust didn't work perfectly. But I haven't had issues with this tiny PCB and several monitors / beamers I've now tried it with (it's pretty easy to carry this around and test whenever I spot a monitor / beamer with VGA cable, LOL).
Heh, thanks for the suggestions... though this was just a stupid demo that I wanted to make for the smallest MCU I could find - not necessarily physical size, but smallest in resources provided (flash/ram/pins) :)
I now have about 30 attiny5s in my stock, I guess it's more than enough for whatever purposes I want to ever use them :D
Failure was either do to pure IO speed, instruction efficiency, or deterministic cycle counting.
ATiny = 18 MHz cycle exact IO bandwidth
XMega = 34 MHz cycle exact IO bandwidth
Now, this is of course overclocking and acting as nothing more than an address counter for the most part, but that's fast!
AVR also has the ability to fetch bytes from internal SRAM and spit them back out at CLK / 3, which is what I am taking advantage of here.
So far I have found only CPLD / FPGA to be more capable than AVR for making bit banged Video Projects.
PIC32 was the largest disappointment for several reasons that I won't bother to get into in this thread.
There was a uC from Ubicom way back that I enjoyed as well, but it went extinct.
2 new AVRs I would like to see...
XMega1024
XTiny
Brad
No Fate But What We Make!
- Log in or register to post comments
Tophttps://www.iar.com/iar-embedded-workbench/?focus=wbselector#!?device=ATtiny817&architecture=AVR
http://www.mouser.com/search/ProductDetail.aspx?R=0virtualkey0virtualkeyATTINY817XPRO
"Dare to be naïve." - Buckminster Fuller
- Log in or register to post comments
TopIf it's an XMega in a reduced format DIP, then count me in!
Can't find anything real on it yet.
Lately, my main source of enjoyment with electronics has been trying to do way too much with way too little.
Have been overclocking to the edge so often that running at spec feels dirty to me now!
Brad
No Fate But What We Make!
- Log in or register to post comments
TopMade a little progress on hacking the keyboard matrix that will be used for my simple Quark tracker project.
This keyboard had " velocity", which means there were 2 contacts per key. Time between strike was measured.
The quicker the 2 contacts close determined the intensity of the key strike.
In my project, I only want one switch, so I pulled apart the keyboard to look at the matrix.
Ebony and Ivory - torn apart in total disarray.
Under the contact bubble there were 2 contacts and 2 pads. One pad was a little closer than the other.
I intend to jack into the closer contacts, and simply make every key an on or off event.
Dual contacts determine key strike velocity.
I unsoldered the little connectors and ribbon cable to replace it with breadboard friendly wiring.
It took some time to decode the matrix, but I did manage to break it down into single switches and rows (Morpheus would be proud).
My target ATMega164 will require 23 IO to handle the keyboard... no problem, there are 32 IO available.
Hacking into the Matrix... no green digital rain or rotary dial phones required here.
All of my projects begin life on the breadboard, so I plugged the wiring into its own board.
A second board with the ATMega164 will connect beside this one.
Green wires are diode side, red wires are switch side.
I won't know if my matrix decoding is correct until I connect the wires to the AVR and code up a test program.
Sure, using MIDI and C would be easier, but I think it will be more fun to control the bare matrix with assembly!
If time offers, next week I may have enough done to make some music.
The most challenging part will be coming up with a tracker format for my demos.
Demo music will require the following data...
- 3 tracks (one per voice)
- Note on and off duration
- Note Frequency
- Track 1,2,3 Portamento setting
- Track 1 modulation setting
Considering the target is an ATiny, the data is going to have to be very efficient, or possibly compressed.
Fun fun!
Cheers,
Brad
No Fate But What We Make!
- Log in or register to post comments
TopIf I were evil (and present), I would swap two of your green wires without your knowing
If you don't know my whole story, keep your mouth shut.
If you know my whole story, you're an accomplice. Keep your mouth shut.
- Log in or register to post comments
TopJust two?
David
- Log in or register to post comments
TopThat's all it would take.
If you don't know my whole story, keep your mouth shut.
If you know my whole story, you're an accomplice. Keep your mouth shut.
- Log in or register to post comments
TopNo way man... you could rip 'em all out and even resolder a few in the wrong pads.
If it took more than 5 minutes to repair, then I would take up a new hobby!
On my giant 300 IC breadboard, I would get someone to randomly move only a single wire (of thousands), and then see how long it took to debug.
Once you "become one" with a project, there is no glitch that cannot be dialed in on.
I actually have sound coming out of this thing now. It's ugly though, and needs some work!
To simulate the ATiny85 Sound Engine, I dropped the Video Interrupt code into the ATMega164 and tweaked it around.
The goal is the be able to play live and record so that what I hear is exactly what the target is capable of.
Not sure how I will do the "Tracker" part yet though... probably just a 16x2 basic LCD display.
Also putting thought into the music file format. Cannot afford to waste even one bit.
Brad
No Fate But What We Make!
- Log in or register to post comments
TopHi folks!
Just thought to bump again in this old thread.
This is ATtiny5 + 354 bytes of code:
https://www.youtube.com/watch?v=kUYDQ30lnT4
Still a bit of a work-in-progress, but just thought to show it :)
Cheers,
//Jartza
//Jartza
Jari Tulilahti
- Log in or register to post comments
TopHey cool, thanks for sharing!
I didn't even know these little AVRs existed. I am going to order an ATTiny 4 and 5 and see what I can do with it as well.
Getting 8 colors, you must be doing the tri-state trick... using 2 colors to also act as sync pulses.
Drive a diode forward into sync (pulled up via resistor), and let the monitor drive color down with its internal 75 ohm?
Can you program these little 6 pins with the AVRISP-II?
I will order a few and see if my Quark Engine will fit.
Should be able to get an ATTiny 4 or 5 bouncing some checker balls!
It's kind of ironic that you bumped this thread, as I had messed around on this project over the holidays for a bit.
I have the ATTiny-85 now capable of 316x240 VGA with 16 colors. It's crazy what these Tiny chips can do!!
I have been working on a 6502 project called Fusion-6502 lately as my main focus though.
AVR is still by far the best micro-contoller out there.
Not even an ARM can keep up to an overclocked XMega at 64 MHz... and I have tried.
Thanks for the inspiration. great to see others pushing the limits of sanity!
Cheers,
Brad
No Fate But What We Make!
- Log in or register to post comments
TopHi!
Yes, indeed, same trick as yours :) RED/HSYNC, GREEN/VSYNC. Blue comes out from Reset-pin which I disabled, the reset-pin
as output is weaker than other pins, but it can still drive color a-ok. Attiny4 and 5 are basically same chip, except Attiny5 has ADC
which is missing from Attiny4 and I found out that Attiny4 is pretty obsolete / hard to get these days, but Digikey still at least
has Attiny5 in stock.
I use USBAsp myself but I remember AVRISP-II has TPI in it? These tinies use TPI and for to be able to disable Reset, you need high-voltage TPI.
Fortunately TPI and High Voltage TPI has only one difference, which is that during HV programming the RESET is kept at 12V, so I used ST662A
to make a "HV TPI" programmer out of USBAsp - just routed RESET through ST662A and no other changes needed:
Programming control pin = TPI RESET, Flash Memory VPP = Attiny5 RESET.
These tinies are still one of my favorite chips, and your work also has acted as inspiration, even though
I started my own ventures with VGA and stuff parallel to yours :)
Cheers,
//Jartza
//Jartza
Jari Tulilahti
- Log in or register to post comments
TopHi,
Here's the schematic for my attiny5 stuff:
LTC1799 can of course be replaced with standard oscillator can, but I didn't have 12MHz oscillators at shelf...
Cheers,
//Jartza
Jari Tulilahti
- Log in or register to post comments
Tophttp://www.atmel.com/webdoc/avrispmkii/avrispmkii.section.phf_vsd_lc.html (AVRISP MkII, Target Interface, TPI Interface)
"Dare to be naïve." - Buckminster Fuller
- Log in or register to post comments
TopThat thread seems to be about HVSP, not High-Voltage TPI...
Cheers,
//Jartza
Jari Tulilahti
- Log in or register to post comments
TopHi folks,
Now I ran into funny thing. I switched from 1n4148 diode to red led, and changed the circuit a bit. HSYNC/VSYNC is driven by 1kohm pull-up before the led, and colors are driven after the led with small resistor. The end-result: better picture quality.
I'm personally a bit baffled, ran into this by accident while shuffling parts on breadboard.
Here's schematic:
Anybody has any idea why I get less "ghosting" with this setup?
Cheers,
//Jartza
Jari Tulilahti
- Log in or register to post comments
TopComparing the V-I curves of a switching diode and an LED will show the difference in forward voltage drops and LED's curve is a bit less steep.
--
Could try substituting for the 1n4148 a 1A (or greater) silicon diode; same except a rectifier diode (1n4001) for the 1n4148.
A diode's junction capacitance is proportional to its current rating.
Edit : proposed experiment
"Dare to be naïve." - Buckminster Fuller
- Log in or register to post comments
TopA RED led is basically a slowish 1.6v or 2v diode (actual drop depends on process used, 1.6v is rarer these days )
That means you have shifted the PIN drive operating point, and the SYNC extraction probably prefers that.
Q: I'd tend to have the same series drive elements in the BLUE channel, so the VGA is more balanced, and anyway, Series LEDs gives a nice 'I'm working' display :)
- Log in or register to post comments
TopThanks for the info, these LEDs are pretty old ones, I'm not 100% sure but I think LTL10223W. LEDs shining on board is nice yes and it works just fine :) BLUE channel is driven from RESET-pin, which is "weak IO" according to datasheet. Together with VGA 75ohm pull-down it seems to not oversaturate, it's pretty much in line with other colors.
Cheers,
//Jartza
Jari Tulilahti
- Log in or register to post comments
TopHi,
I checked with oscilloscope. Blue signal max is 680mV, red and green are around 730mV max... so pretty close.
Cheers,
//Jartza
Jari Tulilahti
- Log in or register to post comments
TopHi folks,
Here's my latest, made it a small demo with ATtiny5 (runs with ATtiny4 too, but I couldn't find supplier for those) to carry around in pocket :)
https://www.youtube.com/watch?v=WdBNR0JEDcY
Even made a tiny tiny PCB for it:
https://drive.google.com/file/d/0B2dTzW9TMeBxbWlaMVFoQnp4aGc/view
https://drive.google.com/file/d/0B2dTzW9TMeBxM29RUkFkazFKQnM/view
The USB is just for powering, I figured that if there's a VGA capable monitor/beamer around, there's probably also USB ports...
Cheers,
//Jartza
Jari Tulilahti
- Log in or register to post comments
TopYou are completely mad. I bow to your madness.
If you don't know my whole story, keep your mouth shut.
If you know my whole story, you're an accomplice. Keep your mouth shut.
- Log in or register to post comments
TopGreat stuff man!
I just received my Tiny5's, and will fire them up next on the free day.
Just a thought on the color / sync sharing...
You may have seen this as well, but it will only work on perhaps 50% of the monitors out there.
The issue is that the pull-up does introduce some voltage back into the analog color, and this is a real issue.
I always test these crazy projects on the 12 monitors I have in my lab, and the sync share only worked on 4 of them.
On the one monitor it did work well on, you could adjust the horizontal position and actually see the light red color being drawn in the blanking area.
Since your monitor does accept this anomaly, go ahead and try this by adjusting horizontal position way out of bounds.
Most monitors have a real issue with color data entering the blanking interval.
This is why I abandoned the shared sync method.
Brad
No Fate But What We Make!
- Log in or register to post comments
TopHi!
Thanks for kind words. I really hope I would be better in "demo effect coding", but I'm actually not really that much graphics effect oriented, although I would love to make something more visually appealing out of this one day :)
I changed the circuit on purpose to use LEDs and switched the color behind the LED acting as diode because I couldn't get the previous circuit running on all monitors - so far this has worked with every monitor I have access to (7 of them). For the PCB I had to change the color series resistors to 120ohm from 220ohm, as I used different LED than on breadboard (on PCB it is LT Q39G-Q1S2-25-1). Looking at oscilloscope, when color is not active, the voltage pretty much goes to zero, sometimes even a tad on negative side (few tens of milliohms). Maybe you could give my circuit a go with your monitors?
I have even released the source code for the demo: https://bitbucket.org/jartza/pentaveega/src
Cheers,
//Jartza
//Jartza
Jari Tulilahti
- Log in or register to post comments
TopFYI folks:
Attachment(s):
//Jartza
Jari Tulilahti
- Log in or register to post comments
TopBut it has a replacement part, it just looks like microchip put it on an other production line for the house (same size)
- Log in or register to post comments
TopNot according to this link
http://media.digikey.com/pdf/PCN...
That says the issue is around package. Digikey shows the 8UDFN as active, but the SOT-23-6 as obsolete.
Looks like a move away from gull-wing, to QFN. Could be largest volumes use QFN, and SOT-23-6 is not viable.
Asia suppliers tend to favour Gull Wing, but I've seen a trend for USA vendors to favour QFN.
Time to look around for alternatives ?
- Log in or register to post comments
TopHi,
Ahh, didn't notice it was only single package that gets obsoleted. I think it's pretty understandable. I've been working with 2 customer projects, which were using ATtiny5 (and that was the reason I even had those chips handy, because some left-overs from development boards, in SOT23-6 size) were using UDFN package. I'm pretty sure most products looking for small are REALLY looking for small :)
Cheers,
//Jartza
//Jartza
Jari Tulilahti
- Log in or register to post comments
TopThat makes sense, as the LED offset gives more TTL like levels on the sync, so should be more tolerant.
Yes, I've found the colour data in Blanking confuses the smart-auto-adjust side of LCD monitors.
If you want perfect centering, it is best to avoid colour in blanking areas, as that is used to decide where to set the margins.
I've also found that black backgrounds (no borders) are not tolerated well, for the same reason, but a 'windows like' background will give better 'auto' operation.
For this type of minimal / marginal work, maybe not-quite-right auto-adjust issues are tolerated ?
- Log in or register to post comments
TopSure, I guess I wouldn't mind much if auto-adjust didn't work perfectly. But I haven't had issues with this tiny PCB and several monitors / beamers I've now tried it with (it's pretty easy to carry this around and test whenever I spot a monitor / beamer with VGA cable, LOL).
Cheers,
//Jartza
//Jartza
Jari Tulilahti
- Log in or register to post comments
TopMore information via
http://www.atmel.com/about/quality/pcn-eol/pcn-eol-notifications.aspx
enter ATtiny5 for Part Number
results are :
ATTINY5-TS8R, LTB 6/5/2017, replacement is ATTINY5-TSFR
The datasheet states :
The move from Atmel PCN to Microchip PCN completes tomorrow (2017-01-31) :
http://www.atmel.com/Images/SC164003%20-%20Integration%20of%20the%20Atmel%20PCN%20system%20with%20the%20Microchip%20PCN%20system1.pdf
http://www.microchip.com/pcn
http://www.atmel.com/devices/ATTINY5.aspx
"Dare to be naïve." - Buckminster Fuller
- Log in or register to post comments
Toptiny102?
8 leads in SOIC (1.27mm pitch) but that's a larger package than SOT23 (6 leads, 0.95mm pitch)
http://www.atmel.com/devices/ATTINY102.aspx
"Dare to be naïve." - Buckminster Fuller
- Log in or register to post comments
TopHeh, thanks for the suggestions... though this was just a stupid demo that I wanted to make for the smallest MCU I could find - not necessarily physical size, but smallest in resources provided (flash/ram/pins) :)
I now have about 30 attiny5s in my stock, I guess it's more than enough for whatever purposes I want to ever use them :D
Cheers,
//Jartza
//Jartza
Jari Tulilahti
- Log in or register to post comments
TopWhat to do with that spare monitor? ... kiosk ... what's on sale at a small business ....
"Dare to be naïve." - Buckminster Fuller
- Log in or register to post comments
TopThe Microchip PCN :
http://www.microchip.com/mymicrochip/NotificationDetails.aspx?pcn=WE164801
(more changes : less expensive lead finish, tape width reduction)
via
http://www.microchip.com/mymicrochip/Reports.aspx?type=eol
"Dare to be naïve." - Buckminster Fuller
- Log in or register to post comments
TopWhat's the current state-of-the-art for a minimal-hardware character/text mode display from an AVR?
- Log in or register to post comments
TopHi,
How about OctaPentaVeega?
https://github.com/rakettitiede/octapentaveega
Cheers,
//Jartza
Jari Tulilahti
- Log in or register to post comments
TopPages