Page 1 of 12
help on DC-AC inverter for Public art Project. Thks
Posted: Thu Jul 17, 2008 7:58 am
My name is Federico Muelas, I run a New Media/Public Art studio in New York. Iâ€™m currently working on a Commission for the George Pearl Hall building, the new School of Architecture and Planning at the University of New Mexico.
You can find further info about the project at:
For this project Iâ€™m designing a large format display incorporating 3,700 pixels of 5â€™x5â€™ each made out of NPD-LCD. Iâ€™m using RS422 serial communication to control the units and most likely I will use a MAX335 chip to locally multiplex the signal on the screen.
The problem I have is that NPD-LCD technology requires AC current (although the required voltage is quite low, only 15V ac @ 7.3mA per pixel). So I need to find a reliable and inexpensive way to convert the 15 V DC that Iâ€™m getting out of the MAX335 chips into 15V AC voltage at 60 Hz (or closer) to control each pixel.
Iâ€™m afraid that if I use triacs with optoisolators or Solid State relays to convert the DC current coming out of the MAX335 into AC current I will end having too much electrical noise and ultimately an instable system.
Thank you so so much
Posted: Thu Jul 17, 2008 9:09 am
I'm a little confused as to why you should want to convert DC to AC. As far as I can see, you intend to take the DC output of each MAX335 switch (this chip is in effect just 8 serially controlled on-off switches) and convert to AC. This would mean that you would need a converter for each pixel, all 3700 of them!
Would it not be simpler to feed a 15VAC signal to the output of each MAX335 switch and use the switch itself (this switch will conduct an AC signal) to ether connect or disconnect each pixel element to the AC source. The AC voltage would be derived from a standard iron core transformer driven directly of the mains supply. With 3700 pixels at 7.3mA each, the full on drive current for the complete display would be around 27A or about 400W.
Or have I misunderstood completely?
Posted: Thu Jul 17, 2008 10:02 am
Thanks a lot for your answer. I thought it wasn't possible to feed an AC current through a MAX335 switch. If so that would be the ideal situation of course, because you're right, converting the current individually for each pixel would be crazy.
but looking at the MAX335 diagram I'm confused on what pins you have to use to feed the AC signal without messing with the serial signal coming from the microcontroller (or controlling unit).
Posted: Thu Jul 17, 2008 10:15 am
The switch section of the IC is effectively isolated from the logic side, so the two don't interact. You would connect all the COM pins of all the MAX335 chips to the AC signal, and connect each NO line to individual pixels. I havenâ€™t studied the properties of the display all that well but I would guess that you would connect one terminal to system ground and the other terminal to each of the pixel drives. This would then be very similar to driving a conventional LCD display.
Do you have data on the NPD-LCD display?
Posted: Thu Jul 17, 2008 10:28 am
What Iâ€™m exactly using for the screen are 5 by 5 inches individual squares of NPD-LCD film. This is the film used in privacy windows or Smart Glass , itâ€™s also called 3G Privacy Film and itâ€™s made by Scienstry, Inc.)
Driving Voltage: 50 Â± 5 V_AC, 50 â€“ 60 Hz_
Current: 0.1 ( A/ m2 )
Energy Consumption: 5 ( Watt/m2 )
Response Time: <0> 80,000,000 (Switching Times)
The Generation of film Iâ€™m using can be switched with 15v AC, but there is a newer generation that uses 0.05 ( A/ m2 ) of current and only 10V AC.
Posted: Thu Jul 17, 2008 10:49 am
I went on the web site. Very glossy, zero technical information!
So I'm going to have to guess that it would be driven in the same way as a standard LCD.
The MAX335 can only handle a maximum of +/-15V, so you would probably want to go for the type that operates on 10V AC.
I will need to check that the MAX335 will switch the required current, although I suspect that it should be able to.
Posted: Thu Jul 17, 2008 11:04 am
one question, if I'm using 15v ac on the switches, what DC voltage do I need to put on V+ and V-? 15 and-15?
Posted: Thu Jul 17, 2008 11:35 am
Yes, your supply voltage must always exceed or equal the voltage being switched. So you would need a +15 and -15 supply and +5V for the logic section.
Your AC supply should be arranges so that the common input pin swings both positive and negative with respect to the pixel display common pin (and the GND pin of the MAX335) and must never exceed +/- 15 volts.
Posted: Thu Jul 17, 2008 11:39 am
thanks a lot Rob!,
it seems a pretty nice and easy solution. I'll try and keep the forum posted.
Posted: Fri Jul 18, 2008 3:53 am
I will put together some suggested layout for this, over the coming days, which may give you some ideas.
Posted: Fri Jul 18, 2008 6:27 am
thanks a lot! I really appreciate it
Posted: Fri Jul 18, 2008 7:44 am
federico, what sort of images will you be displaying?
Also, how are you going to mount those pieces of film?
Are they transparent or clear, or can you transition through shades of gray?
Posted: Fri Jul 18, 2008 8:00 am
they will be really simple bitmaps images, you can get different grades of transparency by modifying the voltage, but I don't know if this will eventually damage the film, it is kind of picky about the voltage and current you use.
I'm still working on the mounting system but I'll try a much standardized system as possible.
when current 0ff=opaque
when current on= transparent.
this is the technology they used on the Chanel store in 15 E 57th St if you want to check it out.
Posted: Fri Jul 18, 2008 9:47 am
Using the MAX335, the absolute maximum theoretical time needed to write to every pixel in the display is around 2.5mS. This doesn't take into account the delay for signals to travel the considerable distance of the display (about 25x25feet if it is square). At a guess it would be possible to have a refresh rate of several frames per second in practice.
Question: How are you intending to configure your pixels? 48x77 for example.
Initial thoughts about construction would be to break the display down into manageable, identical units of say 16 pixels, which would be stacked together to make up the full display. Each unit would have it's own power supply and be relatively self-contained, able to be fully tested before installation into the display
Posted: Fri Jul 18, 2008 10:43 am
That's perfect!, I was actually thinking on 4 fr/sec max. to increase the life span of the LCD.
The idea is to divide the screen in 28 panels of 4' x 8' incorporating 200 pixels each (10 x 20). the screen is circular so in total is about 70 x 70 pixels. the panels will be mounted on the wall using a Ventilated curtain wall system
it would be great if I could use one transformer per panel , each pix uses 15v ac @7.3mA (or 117mW) so 1.46Amps per panel if all units on.