01:48 <lsneff> Do any of you know of any fpga projects that have managed to control digital micromirror arrays?

01:48 <lsneff> The mirror arrays are pretty reasonably priced, but the driver chip is ridiculously expensive

01:49 <lsneff> The datasheet for the mirror arrays doesn’t really go too far into how to control them

02:18 <tpw_rules> i'm pretty sure they run at a relatively high voltage

02:18 <tpw_rules> the driver chip is as much power supply as it is timing

02:23 <lsneff> Yeah, it looks like the mirrors operate at roughly 30V, but the logic signals are still around 3

02:23 <lsneff> It still seems unreasonable to me to require a driver chip that’s 10 or more times more expensive than the mirror array

02:24 <tpw_rules> what prices are you seeing?

02:27 <tpw_rules> i wonder if you could use old plasma driver chips

02:27 <lsneff> Weird

02:27 <tpw_rules> ?

02:27 <lsneff> Some are like $800 dollars but others are more like $70-100

02:27 <lsneff> And I don’t see a difference between them

02:28 <tpw_rules> this one DLPA200 i found on mouser for $47.

02:30 <lsneff> The micromirror arrays I saw on the ti website seemed to use a different interface than the one on that device

02:31 <lsneff> I guess I didn’t look hard enough

02:33 <lsneff> Wait, does the dlpa200 even control the mirror positions?

02:33 <tpw_rules> it looked to just be a power supply and multiplexer basically

02:37 <lsneff> I see. You still need the DMD controller to send data to the mirror array

02:38 <tpw_rules> well in this situation that would be homebrew fpga logic, no?

02:39 <lsneff> Oh yeah, I’d prefer to do it that way

02:40 <lsneff> There’s just very little documentation on how the talk to the mirrors

02:40 <tpw_rules> i don't imagine there really would be

02:40 <lsneff> And I’m also not sure why the data lines to the mirror array have such a high frequency

02:40 <tpw_rules> i guess my original point is that probably the best power ICs to drive a DMD are made by TI

02:41 <lsneff> I agree

02:41 <lsneff> Thank you for sorting that out

02:42 <tpw_rules> i think also the individual mirrors are some type of pulse density modulated

02:43 <tpw_rules> have you ever driven one of those HUB75 LED panels?

02:44 <tpw_rules> i'm pretty sure it works like that

02:44 <lsneff> I haven’t

02:45 <whitequark> yes, DMDs use temporal dithering to get you color resolution

02:45 <tpw_rules> basically you divide your frame time up into 256 bit times or whatever. then if bit 7 of a pixel is on, you turn on that LED (or DMD mirror) for 128 bit times. then if bit 6 is on, you turn it on for 64 bit times

02:45 <tpw_rules> and so on

02:45 <whitequark> (usually less than 8 bit)

02:46 <lsneff> Ah, interesting

02:47 <lsneff> In this case, I’m driving effectively a static image.

02:48 <alyssa_> lukewarm take, dithering is magic

02:57 <lsneff> I found a controller that’s ~$30. I think I’m happy with that

03:00 <lsneff> It had a clocked rgb pixel input

13:14 <MikeK> I was wondering about a link to get started with Amaranth?

13:14 <MikeK> https://cfu-playground.readthedocs.io/en/latest/crash-course/gateware.html

13:14 <MikeK> I cannot seem to find this webpage, is it still avaiabale?

13:15 <MikeK> I cannot seem to find this web-page, is it still available?

13:15 <MikeK> https://vivonomicon.com/2020/04/14/learning-fpga-design-with-amaranth/

13:15 <MikeK> Thanks!

14:04 <d1b2> <zyp> looks like somebody got too eager with the renaming, this url still works: https://vivonomicon.com/2020/04/14/learning-fpga-design-with-nmigen/

14:06 <whitequark> yeah, a sed accident