Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Photo

DIY Alnitak flat panel

  • Please log in to reply
104 replies to this topic

#51 redtail

redtail

    Lift Off

  • -----
  • Posts: 7
  • Joined: 24 Sep 2016

Posted 02 October 2016 - 07:54 AM

Hi i have it all working now and i am very pleased with your project...thanks.
However sgpro is not switching the light of at end of sequence.
Is there a setting or code change to get it to switch off.
Regards.
Kenny



#52 BGRE

BGRE

    Mercury-Atlas

  • *****
  • Posts: 2,987
  • Joined: 21 Mar 2016
  • Loc: New Zealand

Posted 02 October 2016 - 08:59 PM

I must be missing something because current can't flow backwards in a NPN transistor. The arrow on the schematic symbol shows current flow direction. His first schematic has no power connected
to the panel that I can see. that is the one showing the breadboard. The one with the power transistor shows the collector grounded and current from the panel entering the emmiter.

Not sure what is happening. I'd use the FET and not risk overloading the Arduino port but if it is working it may be OK

The first circuit using the small signal transistor - the leads are (EBC usually on a TO92 case transistor) is correct except that there is no jumper
from the collector to the V+. Later a jumper added but it connects to ground instead of V+.

Then the last diagram reversed the collector and emmiter. So I am not sure how you have it connected. He says a 12 volt center positve power supply connected to the Arduino in a label which is not clear.
erSome fets can sometimus be used with the Drain and source reversed. I wouldn't do it with a power FET as there are protective diodes built into many of them.

Actually you can swap the collector and emitter over and run the transistor in inverse mode.

 The only problems are the reduced current gain and the relatively low emitter base   breakdown voltage which is typically

much lower than the collector base breakdown voltage.



#53 pbunn

pbunn

    Vanguard

  • *****
  • Posts: 2,000
  • Joined: 29 Sep 2010
  • Loc: Spartanburg, SC

Posted 03 October 2016 - 09:06 AM

That is true but is very rarely used.  In a search, I find no real example of its use. Someone stated that some  TTL gate ICs used it , but a former TTL gate designer indicated that he was not aware of it.

 

In this case, A cheap power FET seems the best way .


Edited by pbunn, 03 October 2016 - 09:15 AM.


#54 BGRE

BGRE

    Mercury-Atlas

  • *****
  • Posts: 2,987
  • Joined: 21 Mar 2016
  • Loc: New Zealand

Posted 04 October 2016 - 03:32 AM

It was used in bipolar transistor choppers.
Early TTL gates used the emitter of a multiemitter npn transistor as the input.
With a logic high input the input device is actually biased to operate in inverse mode.
With a logic 0 input the input device is biased to operate in normal mode.
Some later versions of TTL used diode or even pnp input devices
Your TTL designer seems to be unaware of how early TTL gates actually worked.

The real point is that being aware of it is very useful in understanding some circuits where a transistor may have been incorrectly installed with the base and emitter interchanged.

In the sixties connecting specific transistors in reverse and having the the emitter base junction undergo (current limited) avalanche breakdown, leaving the base open allowed the forward biased collector base junction to temperature compensate the emitter base breakdown voltage producing a temperature compensated zener.
It doesnt work with current transistors as the doping levels etc have changed.

#55 redtail

redtail

    Lift Off

  • -----
  • Posts: 7
  • Joined: 24 Sep 2016

Posted 25 October 2016 - 11:50 PM

Sgpro now has an update and it now switches off.



#56 CCDer

CCDer

    Messenger

  • -----
  • Posts: 494
  • Joined: 06 Nov 2015

Posted 06 March 2017 - 12:45 AM

This is really cool! I have a DIY flats box for my C14 that uses one of those PWM knob boxes off ebay but they tend to drift their frequency after its adjusted where you need it. I've already got the Arduino programmed and working with a string of LEDs and I might use it with SGP someday. But for now I'm writing a program to run on a Raspberry PI so it can be wireless and I'll just VNC to the PI to run it. I'll post a link on this thread if anyone is interested. The program will allow X number of cameras for Y number of filters to be configured to default values with labels attached to each setting. Then just a click on the filter label will set the light level. I'm currently running with TheSkyX which has no light box support so this is all I can use for now. But its a whole lot better than having to go from my warm room out to box for each set of flats, not to mention all the times I've had to readjust it and retake flats because of the drift.



#57 CCDer

CCDer

    Messenger

  • -----
  • Posts: 494
  • Joined: 06 Nov 2015

Posted 07 March 2017 - 09:33 PM

Here is a Gambas executable that will control the arduinio for 9 filters x 7 camera/telescope combinations. It creates a default config file on the desktop to save all settings.

 

https://drive.google...iew?usp=sharing

 

I'm using the PI to control my custom light box via a VNC login to the newest Jeremy version of the PI linux. So you'll need a network connection (wireless or not) to the PI. I'd copy this executable to the desktop and just run it from there.



#58 ColoradoBound

ColoradoBound

    Sputnik

  • *****
  • Posts: 43
  • Joined: 15 May 2015
  • Loc: Colorado

Posted 23 May 2017 - 11:17 PM

I just got mine put together and it is working great. I used a MOSFET to keep everything nice and cool. This was about the easiest DIY project I have done. Now if the clouds would just go away I could do more imaging.



#59 pfile

pfile

    Fly Me to the Moon

  • -----
  • Posts: 5,588
  • Joined: 14 Jun 2009

Posted 17 June 2017 - 11:40 PM

has anyone tried this with one of the smaller panels, such as this one?

 

https://www.amazon.c...uct/B01J9B16QM/

 

the smaller ones seem to run on 5V; i guess my question is if you can control the brightness with the input voltage, or if there's some kind of buck/boost converter inside such that PWM of the main voltage would not work to control the brightness.

 

rob



#60 CCDer

CCDer

    Messenger

  • -----
  • Posts: 494
  • Joined: 06 Nov 2015

Posted 18 June 2017 - 10:21 AM

There is no reason you could not find a arduinio compatible device that would work at the lower voltage needed by this panel. That lower voltage arduinio device must have 3.3V switching outputs and you would need to be sure the MOSFET you use would switch at least as low as 3.3 volts. However, to use the more basic arduinio device used in this thread you would need to have a 7-12V power supply and another lower voltage supply for the LED's used in this panel. A "buck" type converter off the 7-12V supply might work to get this lower voltage but if you are "switch" powering a device off of a "switched" powered "buck" supply, you'd want a good sized capacitor on the output of that buck converter to keep it stable.

 

You also may be able to simply add a power resistor in line with the LED's in this panel so you could still use the 7-12 volts to drive them. But that's wasted battery current if your're going portable.

If it were me, I'd try for one 5V power supply to power both arduinio and this panel's LEDs but you may have to experiment with more than one arduinio device to get it working with the software designated in this thread. So you may need to alter the software as well. I've got EE degrees and wrote software all of my career so I see no problems in getting this done but it costs more time and $.

 

The nice thing about staying with the same design in this thread is that it just works.

 

Edit: I just realized I didn't answer your question and went straight to the solution. To answer your quest, no I don't think you will be able to drive that panel directly from its 5V supply input. I'm thinking you'll need to disassemble it to access/isolate the LED's directly and do something like I mentioned above. With that panel's "stepped" brightness control this is a certainty. There probably is a MOSFET with power  inside (off the 5V input supply)  that you could drive directly with a lower voltage arduinio attached to that same power input but you would need to be able to isolate all of it and that could be difficult.


Edited by CCDer, 18 June 2017 - 11:12 AM.


#61 pfile

pfile

    Fly Me to the Moon

  • -----
  • Posts: 5,588
  • Joined: 14 Jun 2009

Posted 18 June 2017 - 09:07 PM

i guess why i'm confused is that the larger, 12V panel also seems to have dimmer buttons, which to me implies that it also has a circuit inside to drive the LEDs with a PWM waveform... so that applying a PWM waveform to the main 12V input would not actually do anything.

 

but - this boils down to reading comprehension - i missed this in Tim's original post:

 

 

To make this work, I first needed to bypass the control circuitry in the flat panel (no pics of this).

 

 

so i guess it should all work out the same with the smaller panel; unfortunately my oscilloscope blew up but otherwise i would be able to measure the output of whatever is driving the LEDs in the smaller panel and emulate that with the arduino + fet. probably time to visit the tektronix refurb store smile.gif

 

rob

 



#62 pfile

pfile

    Fly Me to the Moon

  • -----
  • Posts: 5,588
  • Joined: 14 Jun 2009

Posted 22 June 2017 - 01:29 PM

OK i got it all working - of course the key was to disconnect the LED strips from the built-in controller (duh).

 

i am using the MTP3055VL enhancement mode mosfet to switch a dedicated 5V supply for the panel. i was surprised to measure 1.7V across the drain and source when the PWM was set to what i believe should be 100% duty cycle; i measured 5V at the gate. the MTP3055VL has a protection diode so maybe what i'm seeing is the forward drop across that diode? it would be nice if the panel were a little brighter for narrowband; i had previously tried the panel set to 100% by its internal controller and found the brightness so high that my narrowband flats were sub-1s; with the arduino driving it at 100% duty cycle the flats take 7s, which is actually OK but a little more brightness would be nice.

 

on the other hand, for my LRGB filters i have to have the duty cycle all the way down at around 6% so increasing the voltage might make it impossible to take LRGB flats of reasonably long length.

 

rob



#63 CCDer

CCDer

    Messenger

  • -----
  • Posts: 494
  • Joined: 06 Nov 2015

Posted 22 June 2017 - 05:08 PM

MOSFETs have an Rds value that's the effective drain/source On Resistance. But for your MOSFET that's only about .1 ohms if the Gate is fully driven on. So even at 1amp you'd see only .1 volts dropped D/S.

 

The internal controller for your 5V panel might do a DC/DC conversion to boost the 5V to something higher. If you drive the LED's with the internal controller at 100% (if that can be restored) you may be able to measure the LED voltage with just a voltmeter (if you at least have one of those that's working).  That would verify that 5 volts is the max to directly drive the LEDs. If you can go higher on the LED voltage that gives you another option to tweak the brightness curve.

 

Edit: To clarify, you need to measure what voltage the original circuit puts on the LEDs and compare that to what voltage your circuit puts on the LEDs. Then you can raise that dedicated 5V supply as needed.


Edited by CCDer, 22 June 2017 - 05:43 PM.


#64 pfile

pfile

    Fly Me to the Moon

  • -----
  • Posts: 5,588
  • Joined: 14 Jun 2009

Posted 22 June 2017 - 05:34 PM

yes i agree, this is why i was confused to have measured 1.7V across the drain-source (with my working-as-far-as-i-can-tell fluke meter). since it's an enhancement mode mosfet at least you could explain a high Rds by insufficient gate voltage, but i measured 5V at the gate, which should be enough (though i have not consulted the datasheet). i would be surprised if 5V were not enough to get the Rds down to the milliohm range. this is why i was suspecting the protection diode. i guess this might mean that i have hooked up the FET backwards, but if that were the case then with 5V at the anode of the diode it would just conduct whether or not the gate was energized. i guess i need to read the datasheet.

 

before i took it apart, when the LEDs were still attached to the controller i measured 5V across the output of the controller with the brightness at 100%. so i think the LEDs are probably being run from 5V but yes, without an oscilloscope its hard to be 100% sure of that fact. since the LED power supply is separate from what's powering the arduino i could certainly goose the voltage to get around this problem, but theoretically the fet should not drop so much voltage.

 

rob



#65 CCDer

CCDer

    Messenger

  • -----
  • Posts: 494
  • Joined: 06 Nov 2015

Posted 22 June 2017 - 09:51 PM

Can you see the LED load and verify there are only LED(s) and some limiting resistance? Since you do not know for sure what the original driver circuit was (maybe MOSFET/ maybe not), there may be a reactive element like a capacitor in the load. You'd expect the original driver to use PWM to dim the LEDs but maybe it's something entirely different. Once again, a scope would help solve this.



#66 pfile

pfile

    Fly Me to the Moon

  • -----
  • Posts: 5,588
  • Joined: 14 Jun 2009

Posted 23 June 2017 - 01:21 PM

not sure about any caps; the panel looks to me to be constructed out of two LED bars that appear to have current limiting resistors and LEDs only. the controller is quite small with just one IC if i remember properly; the caps on the PCB look like bypass caps. i have been hesitant to really take the panel completely apart because the wiring is all very fragile.

 

i took a look at some tek factory refurb scopes and predictably they are pretty expensive. i guess i could buy some el-cheapo scope but since it's essentially a one-time purchase i should probably bite the bullet and pay up... or go down to halted and see what they have used.

 

anyway i guess it will work as-is with a minor mystery; my next focus is finding a project box, appropriate connectors, and coming up with some method for strain relief on the panel wires - they are maybe 28 gauge stranded, very thin and weak. at this point the plan looks to be kapton tape :)

 

rob



#67 CCDer

CCDer

    Messenger

  • -----
  • Posts: 494
  • Joined: 06 Nov 2015

Posted 23 June 2017 - 02:51 PM

There's a plethora of scope kits of all types on ebay. If you have a PC, particularly a laptop, I recommend something like this that's not a kit but is still inexpensive:

 

https://www.amazon.c...h/dp/B009H4AYII

 

I bought a 60MHZ DSO2150 for $180 about 10 years ago and it's still working great. This newer model is < $60 but is only 20MHZ but still gets you most of what you need.



#68 macona

macona

    Ranger 4

  • *****
  • Posts: 365
  • Joined: 18 May 2011
  • Loc: Beaverton, OR

Posted 23 June 2017 - 06:02 PM

You can pick up used Tek TDS-340A's for about $100 on ebay. Good scopes. I have used USB scopes and they just kind of suck for anything real time. 



#69 pfile

pfile

    Fly Me to the Moon

  • -----
  • Posts: 5,588
  • Joined: 14 Jun 2009

Posted 23 June 2017 - 11:34 PM

thanks i will check ebay out. i just remembered that i have one of those TI usb scopes but have only used it like once or twice, mainly because i had to set up a PC to use it. i guess i can resurrect that laptop...

 

rob



#70 Henry from NZ

Henry from NZ

    Apollo

  • *****
  • Posts: 1,307
  • Joined: 26 Jan 2014

Posted 23 July 2017 - 02:15 AM

This is a very interesting project. I happen to have all the components except the light panel. Driving and controlling a led panel is a lot easier to implement on an arduino as it does not require working with high voltage inventor required for el panels. However I wonder how even the illumination is going to be with a LED panel compared to an el panel? I have made myself a white LED light box many moons ago and to get even flatness the light source need to be some distance away from the aperture. How does a thin LED panel do in this regards?

#71 CCDer

CCDer

    Messenger

  • -----
  • Posts: 494
  • Joined: 06 Nov 2015

Posted 23 July 2017 - 11:44 AM

This is a very interesting project. I happen to have all the components except the light panel. Driving and controlling a led panel is a lot easier to implement on an arduino as it does not require working with high voltage inventor required for el panels. However I wonder how even the illumination is going to be with a LED panel compared to an el panel? I have made myself a white LED light box many moons ago and to get even flatness the light source need to be some distance away from the aperture. How does a thin LED panel do in this regards?

 

I took the approach of using a circular containment for the LEDs so the light can only bounce around evenly.

 

 
I used a standard strip of white 5050 LEDs off of Ebay. The circular area is constructed of a white melamine back board and a 3.25 inch wide strip of white surfaced roof flashing. The LEDs were simple stuck to this strip but also reinforced them against the flashing wherever the flashing made contact with the box. You can also find the white plastic sheeting on Ebay that's about 1/8" thick that will lay over the circle. In this configuration the light can only even out. I used two sheets. However, the LEDs tend to be on the bright side, even with the PWM set at its smallest setting. If you could find LEDs that are spaced at least twice the distance of the 5050 strips, it would be better. I need to do this or just put electrical tape over every other LED.   
 
Edit: Note that it would also work to use just the right thickness of material over the front but I had a hard time doing this for a 2x2 foot area. Stacking more than a couple of thin sheets causes a bunching up effect over this wide area that produces uneven light. If you could find a uniform piece of white packing material that's the right thickness, that would be great to dim the LEDs.
 
Ok, you got me motivated for an upgrade. I just ordered a set of these:
 
They're 1/2 the density (30 per meter vs 60 per meter) and they're a lower temperature (3000K vs 5000K) so the color balance is better (more red)

Edited by CCDer, 23 July 2017 - 01:03 PM.


#72 AndreaT

AndreaT

    Lift Off

  • -----
  • Posts: 15
  • Joined: 07 Jun 2010
  • Loc: Venice, Italy

Posted 12 August 2017 - 08:46 AM

Hi 

I have all the parts and i am ready to build this but i am confused with the conflicting info in the drawings.

The parts i have are the uno and this MTP3055V N channel 60V 12A 0.15Ohm TO 220 MOSFET's.I also have the light panel.

Is this the right one?

post-242464-0-23618000-1464718412.png.

Is this schematics correct? I`m building an identical device.. but if it`s possible to not burn arduino or panel is better..

 

Most supplyers are closed and need it for next week



#73 CCDer

CCDer

    Messenger

  • -----
  • Posts: 494
  • Joined: 06 Nov 2015

Posted 12 August 2017 - 01:28 PM

I used an LED strip instead of a panel and this schematic,https://learn.adafru...ed-strips/usage serves as a reference for that. It's the first schematic on this page (the second is for a transistor and uses resistors). This schematic shows three MOSFETs for a RGB LED strip but you only need one MOSFET for a White LED strip. You can connect power to the LED strip from board power if that works. But if it takes more current > 1A, you can solder the LED 12V power lead directly to the underside of the Arduinio board at the 12V power plug which is even better. Note that the switched LED Cathode leads are connected to the Drain (PIN 2) of the MOSFETS because the load has resistance and a resistor is usually put at the Drain side because the Source is usually connected to GND (For a N-Channel MOSFET). If the load is connected at the Source side (PIN 3) instead before going to ground, the Vgs needs to be higher in order to switch the MOSFET and this could cause problems. If you've used a prototyping block before, you already know the BLACK wires are all connected together for the R,G,B circuits through the edge bus of the prototyping block so all of the Source (PIN3) MOSFET leads are grounded. Note the pin configurations are the same for the MTP3055 as the IRLB8721 which is what I used.

 

 

So given the above schematic, this is what I would do for a panel. If you have the LED's power leads isolated, say RED for power and  BLACK for ground:

 

Connect PIN3 PWM of the Arduinio to MOSFET Gate (PIN1) (same as your drawing)

Connect the BLACK lead of the LEDs to MOSFET Drain (PIN2) (different from your drawing because the load should be connected to the drain side for best switching performance for a N-Channel MOSFET)

Connect the RED lead of the LEDs to Vin Power of the Arduinio (same as your drawing)

Connect MOSFET Source (PIN3) to Gnd of the Arduinio (different from your drawing because the Source should be connected to ground for best switching performance for a N-Channel MOSFET )

 

Note that if you happen to wire the panel's LED leads backwards then nothing will happen because current can't travel backwards through LEDs (i.e. no damage)

 

So this is the very same wiring that is shown on the Adafruit webpage for the LED strip. The only difference is the BLACK lead of the LEDs is represented as a R,G,B colored wire instead of a BLACK wire.



#74 sloz1664

sloz1664

    Lift Off

  • -----
  • Posts: 8
  • Joined: 10 Jun 2017

Posted 16 February 2018 - 10:16 AM

I have, hopefully, completed the psuedo Altinak light panel as described in this topic. Just one question. When I attempt to adjust the brightness from 0 - 255 in SGPro the lightness level goes from bright to a higher bright. Is this correct? I have added two layers of opaque drafting film, a white sheet and a dark grey sheet onto the panel and I still cannot the light low enough to gather flats for my Atik 383L+, which needs a minimum of 2 secs expoure to overcome the shutter lag.

Any pointers anyone?

 

Steve



#75 sloz1664

sloz1664

    Lift Off

  • -----
  • Posts: 8
  • Joined: 10 Jun 2017

Posted 21 February 2018 - 10:29 AM

I have, hopefully, completed the psuedo Altinak light panel as described in this topic. Just one question. When I attempt to adjust the brightness from 0 - 255 in SGPro the lightness level goes from bright to a higher bright. Is this correct? I have added two layers of opaque drafting film, a white sheet and a dark grey sheet onto the panel and I still cannot the light low enough to gather flats for my Atik 383L+, which needs a minimum of 2 secs expoure to overcome the shutter lag.

Any pointers anyone?

 

Steve

Good news I have found the problem, faulty transistor. Ordered a couple of replacements (just in case). The replacement transistor fitted and the Light Panel works as it should.  




CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.


Recent Topics






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics