Senior Member
Here's a photo of the multabela ilda glue board. It can be seen that on bottom is headers for the four belas. Left side is 4 x +-10V analog inputs, 2 x XY analog joystick inputs, and eight button inputs. None of this but one button is needed for playback of shows. The right side is all the trim pots, and the four ilda out ports are across the top.
Senior Member
TL084's aren't becoming obsolete, only the old style packages, slew rates and voltage ratings are.Originally Posted by TheHermit
PCN20230217004.3.pdf
-glue board may equate to prototype-
________________________________
Everything depends on everything else
Senior Member
Sorry for all the confusion caused by using the word "glue". I was thinking of the term "glue logic" which means custom logic circuitry used to interface a number of off-the-shelf integrated circuits. So I was thinking off the shelf ilda projectors, off the shelf bela devices, off ebay analog joysticks, analog signal generators of the users' choice, and a "glue board" to interface it all together. The second part of the joke was that due to my attempt to keep the board field serviceable, it is big enough to trap a rat if it was actually a glue trap.
Roj: Yes, these circuits are summing DC with the amplified or attenuated signal to position it within range, hence all the trim pots.
Senior Member
Does your system have +-15V supply wherever it's needed? CYGN-B page 63 is where I explain my reasoning for the op amp circuits and lasermaster1977 who is leagues above me in understanding this stuff, considers, improves, and approves my approach once I confess that my goal is somewhat different than exactly what I have stated in the notes.
Here attached are the two relevant circuits on the glue board. The PCB version of these hasn't been tested yet.
Senior Member
I call these voltage projection circuits because the goal in my view is taking whatever is given voltage range wise, and getting decent projected range while keeping a safe distance from the desired limits, and doing so in a way that I pretty much understand why it works.
Last edited by Greg; 03-22-2024 at 19:41.
Senior Member
Yes, yes, and no. There is no ILDA anything involved except the physical connectors. I use code the smart people at bela gave me that converts HSV to RGB. The "degrees out of phase" hue value offset for each channel to give pure looking red, yellow, green, and blue when the torque dial is at default was picked through trial and error.
Senior Member
Yeah, no, the belas are 0 to 5V just like what you are using, and only 0 to 3.3 for the beaglebone, else brick. It took some nerve to connect my homemade +-15V circuits to a pile of expensive belas, but everything worked out ok, I guess because I was careful.
Regarding diodes as voltage limiting devices, my opinion may be little more than woo, but I have become skeptical of diodes used for this purpose. I built the RGB RYGB threshold colormod circuit using, I think, a 5.1V Zener as a positive limit and a germanium to limit the negative excursion to -0.3V. The circuit, now relegated to the junk pile, worked but appeared to have a white noise component which I suspect came from the Zener. I don't entirely understand the physics, but I have read that this is an intrinsic characteristic of the device. The effect was kind of interesting when one torqued the beam to a threshold and the lines broke up into rapid sprays of tiny random dots.
Senior Member
"We don't need no stinkin' diodes." to paraphrase a line in Blazing Saddles.
My T4_LS synth board's XY and RGB outputs work just fine for me, as designed. The only way I could produce negative voltage values on the RGB op amp outputs was to incorrectly adjust the trim pot. I'm surprised you haven't sorted that out yet. The RGB offset trim pots must be set to a +6.1875v or thereabouts. What is the method and madness to this?
Assume the trim pot on the non-inverting input is cranked counterclockwise so the wiper and the non-inverting input is at Ground (0v). You should therefore see the following:
Since the +DAC output from the CS42448 is a positive only DC or AC voltage signal with a 3.25v p-p range and is centered on a +2.5v offset, the RGB op amp must have a gain of 1.538 to achieve the desired 5v p-p RGB signal range, but remember it is an inverting gain, so a 15K feedback resistor and a 10K input resistor on the op amp's inverting input (Rf/Rin...15K/10K) gets us a gain of -1.5 resulting in a -4.875v p-p DAC signal swing, centered on a 2.5v * -1.5 gain = -3.75v offset. The most negative the op amp's output will be is half the negative p-p DAC signal plus the negative offset, or -2.4375v plus -3.75v = -6.1875v. To raise the most negative voltage value of -6.1875v to 0v requires adding +6.1875v to the op amp's output through the non-inverting input trim pot. If the trim pot is now rotated clockwise towards the +9v resistor end until the wiper is at 6.1875v then the RGB op amp's output signal swing will be from 0v to +4.875v. AND, the downside to this arrangement is that when the DAC's peak positive output voltage is +4.125, the op amps output voltage will be 0v, and when the DAC's peak least positive output voltage is +.875v the op amps output voltage will be +4.875v.
I still would prefer to have a unity gain inverter on the existing RGB op amp outputs we wind up with positive voltage DAC outputs yielding positive op amp voltage outputs, so digital "on" is analog "on".
Considering that there is no way that the CS42448 +DAC outputs can ever be a negative voltage value, the only possible way for this to happen is in the way the op amp inverter with its voltage offset trimmer is set. And, whether the op amps are powered with + and - 15v or + and - 9v matters not since it is really hard to kill the TL084 resulting in the output to swing to a power rail. What matters is that the op amp supply voltage is sufficient in range to accommodate the signal output voltage ranges needed.
My recollection is that we wound up with the RGB op amp output circuitry like it is because you preferred to reduce the overall parts count and to make the circuit trace layout easier and faster to do.
Prototyping is mostly fun, and how we learn to make things better.
Edit: Since Roj reminded me we had a 2nd inverter of unity gain the offset trim pot on the first inverter requires far less positive voltage on the non-inverting input such that the 1st RGB op amp's output swing is 0 -> -4.875v. This zero to negative output signal range is then inverted by the 2nd inverter to 0 -> +4.875v.
Last edited by lasermaster1977; 03-23-2024 at 18:22.
________________________________
Everything depends on everything else
Senior Member
It was both, but I was exactly quoting Blazing Saddles: https://www.youtube.com/watch?v=Dln7yj8MDWE
It has been so long since I looked at the schematic I forgot. You are right. We did have a 2nd inverter on RGB. (Yea!)
Why not loose the +DAC and use the -DAC output? I have a nagging recollection that if only one output of the two balanced are used then it should be the +DAC out. But...I've slept since then.
I just remember dropping a bunch of resistors in addition to what you mentioned.
________________________________
Everything depends on everything else
Senior Member
Um... For those of us old enough to remember it, that was a paraphrase of a quote from the book and movie "The Treasure of the Sierra Madre". See:
https://en.wikipedia.org/wiki/Stinki...;%20badges.%22
But, this is off topic...
Ron
PS. I'm not technically old enough to "remember" this...
Posting Permissions
Reply With Quote