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

#26 William Mc

William Mc

    Mercury-Atlas

  • *****
  • Posts: 2,783
  • Joined: 24 Nov 2007
  • Loc: North East Florida

Posted 15 August 2016 - 03:39 PM

Yes that's my goal also.  Automated panel control through the SGP software.

 

BTW, A big thanks to spokeshave for posting this project. It's going to be a real bargain assuming the Arduino doesn't blow again.  :D   The large panel itself was a steal for the price. I have already modded it for full brightness at 12V  ( A fiddly operation itself), and I've have it mounted to the OB wall.   I had actually even mounted the controller onto the pier and thought I had finished just before I blew it up!


Edited by William Mc, 15 August 2016 - 03:44 PM.


#27 pbunn

pbunn

    Vanguard

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

Posted 15 August 2016 - 04:13 PM

If you are not familiar with what an Alnitak flat panel does - it wasn't all that apparent. It looks like a great idea. I understand that the Alnitak turns on and off
and has dimming at 255 (8 bit) levels.

Does the 255 step dimming linearize the very non linear relationship of a typical LED? I would assume that it does? Otherwise the light output would be zero for a large amount of the 0-12volt PWM scale.

Edited by pbunn, 15 August 2016 - 04:14 PM.


#28 spokeshave

spokeshave

    Vanguard

  • *****
  • topic starter
  • Posts: 2,302
  • Joined: 08 Apr 2015

Posted 15 August 2016 - 04:25 PM

pbunn, on 15 Aug 2016 - 5:13 PM, said:

Does the 255 step dimming linearize the very non linear relationship of a typical LED? I would assume that it does? Otherwise the light output would be zero for a large amount of the 0-12volt PWM scale.

The Arduino provides 255 linear levels of output. The output is the full 12V with pulse width modulation so there is no non-linear response by the LEDs. The voltage itself is not changed.The Arduino modulates the output at 490Hz with each 12V pulse width determined by the setting. So the LEDs are either fully on, or fully of 490 times a second with the pulse width determining how long per cycle they are on. At the lowest level the panel is just perceptibly lit if the room is dark. At 255, it is essentially at full brightness.

 

Tim



#29 spokeshave

spokeshave

    Vanguard

  • *****
  • topic starter
  • Posts: 2,302
  • Joined: 08 Apr 2015

Posted 15 August 2016 - 04:28 PM

Yes that's my goal also.  Automated panel control through the SGP software.

 

BTW, A big thanks to spokeshave for posting this project. It's going to be a real bargain assuming the Arduino doesn't blow again.  :D   The large panel itself was a steal for the price. I have already modded it for full brightness at 12V  ( A fiddly operation itself), and I've have it mounted to the OB wall.   I had actually even mounted the controller onto the pier and thought I had finished just before I blew it up!

You might try getting a simple breadboard kit so that you can test the configuration before soldering everything up. That's what I did.

 

Tim



#30 spokeshave

spokeshave

    Vanguard

  • *****
  • topic starter
  • Posts: 2,302
  • Joined: 08 Apr 2015

Posted 16 August 2016 - 07:33 AM

The base of the 3055 looks like a short to the Arduino with no current limiting resistor. Does the Arduino have a current limiting resistor on board?

I've been giving this some thought and doing a bit of research. A number of sources say that a base resistor is not needed since the Arduino is internally limited to 40mA on the PWM pins. However, other sources say that drawing the full 40mA can damage the Arduino, though I have seen no evidence of that on mine. Given the conflicting information, I think it is probably a good idea to add a current-limiting resistor to the base (270 ohm should be fine). If you decide to use a MOSFET, the resistor is not needed.

 

Tim



#31 pbunn

pbunn

    Vanguard

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

Posted 16 August 2016 - 10:47 AM

But a 47 or 100 ohm is a good idea to keep the FET from taking off at an RF parasitic. Especially if the gate lead had long leads (over a couple of inches)

 

I wouldn't want my Arduino port to be going into current limit on every pulse. The base voltage is roughly .7 volts and the Arduino port is probably near 12 volts - so it looks like a near short to the Arduino port.

 

You might measure the Collector to Base voltage with the PWM full on. If you are getting more than .2-.3 volts - the TIP 3055 is not saturating (not fully switching)

 

The current gain on that transistor is specifies as minimum 20 max 70. At higher currents - the gain is lower.

 

.040 x 20  + 0.8 amps - so if the transistor is running at the low end of the current gain spec - 40 Ma can't saturate it   - and a currrent limiting resistor would make it worse.   Not a problem with a FET



#32 William Mc

William Mc

    Mercury-Atlas

  • *****
  • Posts: 2,783
  • Joined: 24 Nov 2007
  • Loc: North East Florida

Posted 20 August 2016 - 12:21 AM

OK tonight I tried again and blew a second arduino.  So I started all over, this time removing every possible ground connection to the metal case.  That seems to have worked, as I haven't yet blown Uno #3.     I'm using the recommended Mosfet, and the Arduino is communicating with SGP. As I move the brightness slider I get flashing tx, rx lights so that all looks great.

 

BUT!  The slider isn't controlling the brightness of the panel?  It appears stuck at full brightness.  If I disconnect the PWM pin connection the panel still stays fully lit?  Is there something I'm missing?  Did the Mofset get fried when I blew the second arduino? Is this the behavior it would produce if blown?

 

Again, I have mostly worked with vacuum tubes, and I'm new to Mosfets / Transistors.  I need some troubleshooting steps. I do have an assortment of test e gear to use.  multimeters / Oscilloscopes etc.


Edited by William Mc, 20 August 2016 - 12:38 AM.


#33 pbunn

pbunn

    Vanguard

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

Posted 20 August 2016 - 09:24 AM

Sounds like the gate circuit is working correctly. You are measuring the pwm signal after is is averaged by the meter. That seems to be working OK and that is the hard part IMO.

You should have 13.8 volts positive tied to the panel V+ and the and the drain pin (also the metal tab) of the mosfet tied to the V- input. The source pin should go to your power supply V-.
It sounds like you may have the tab grounded. If that is the case you would get full brightness.

Think of the FET as a switch - when a pulse is applied to the gate - When the pulse is removed - the switch turns on and shorts the drain to the source - when the pulse is over, the switch turns off. The PWM is a switch turning on and off very fast. It varied thee duty cycle
to change the brightness. The panel is on full each time it gets a pulse - but the time it is on varies to the full time of each cycle.

It seems almost certain that the tab or the drain pin is shorted to V- . If not measure the resistance from drain to source and make sure the fet isn't shorted - not likely.

Also you can test the setup by removing the gate signal. the panel should turn off. Shorting the gate to V+ will turn it on - removing it off. For testing you can use a cheap LED and a 1k resistor in series.
Thantkeeps you from worrying about messing up the panel.


Pat

Edited by pbunn, 20 August 2016 - 09:28 AM.


#34 William Mc

William Mc

    Mercury-Atlas

  • *****
  • Posts: 2,783
  • Joined: 24 Nov 2007
  • Loc: North East Florida

Posted 20 August 2016 - 10:07 AM

Disconnecting the gate signal is not turning off the panel.

 

Tried another mosfet --   The panel still will not turn off with no gate signal?

 

This is baffling!

 

What a minute!  You say the source pin should go to the PS V-?     Mine goes to the v- return from the panel.   I'm following Spokeshaves schematic on page 1.  Is that not correct for a mosfet?

 

Here's my confusion -  You said that the drain should be tied to V-, and also it should not be grounded?  But isn't V- and ground the same thing in this circuit?  I have continuity between the PS V- AND the boards ground.  It's the same connection? How can it connect to V- and not be grounded?

 

 

 

I have gate to PWM signal

 

drain to Ground connection on the UNO board

 

source to V- return from panel

 

PS v+ to panel V+

 

 

 

 

post-242464-0-23618000-1464718412.png


Edited by William Mc, 20 August 2016 - 10:31 AM.


#35 pbunn

pbunn

    Vanguard

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

Posted 20 August 2016 - 12:08 PM

Remove the arduino signal wire and run a jumper from Gate to ground. If the mosfet doesn't cut off it is shorted externally or internally. Remove the mosfet from it's mounting also. Remember the drain is the tab and if it is grounded via the mounting screw, what you are describing would happen,

#36 William Mc

William Mc

    Mercury-Atlas

  • *****
  • Posts: 2,783
  • Joined: 24 Nov 2007
  • Loc: North East Florida

Posted 20 August 2016 - 12:15 PM

I'll try that right now


Edited by William Mc, 21 August 2016 - 06:47 AM.


#37 pbunn

pbunn

    Vanguard

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

Posted 20 August 2016 - 12:19 PM

The drain (and tab) goes to +13.8 volts

The source is grounded and the -13.8 volt supply is grounded

The panel minus lead goes to the drain (it is not grounded.)

The mosfet is acting like a switch in the panel negative power lead.

I think that you have the fet backwards

The signal from the Arduino has the positive side to gate and the negative side to ground - so the negative signal, the source and the 13.8 volt power supply all are hooked together.

the positive arduino signal goes to the gate. the positive power supply lead goes to the drain (or the tab)

The schematic shown the transistor having the emmiter tied to +12 and the current flowing into the emmiter. Not sure what is going on here.

Wire the FET as above and it will work.

Edited by pbunn, 20 August 2016 - 01:00 PM.


#38 William Mc

William Mc

    Mercury-Atlas

  • *****
  • Posts: 2,783
  • Joined: 24 Nov 2007
  • Loc: North East Florida

Posted 20 August 2016 - 12:56 PM

Well there's the issue!   Thanks.   I thought I had asked earlier if I used the same schematic?. I had tried the bipolar one last time and for some reason it now works, so I'm thinking I'll leave it this way for now.  Getting kinda tired of the entire project and I have a nice pile of burnt boards.  Good to know how to wire the Mosfet though in case I need to later, and for others wanting to do this.

BTW, the USB chips on these Uno's seem easy to blow.   For others wanting to do this I suggest buying a large bag of the cheap $2.50 clones.


Edited by William Mc, 21 August 2016 - 06:48 AM.


#39 pbunn

pbunn

    Vanguard

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

Posted 20 August 2016 - 01:09 PM

The earlier schematic is in error. If it were correct - the fet would have worked wired the same. I didn't look at it closely.

The drawing showing the TIP 3055 in detail has the collector connected to ground. That can never work. In an NPN transistor - the collector is ALWAYS positive with respect to the emmiter and current can only flow from collector to emmiter.

There is essentially no difference in the 3055 fet and the tIP 3055 except that the fet has voltage drive.

This motor control is showing an Arduino connected to a motor - but the LED panel would connect the same. The diode is for preventing motor back emf from damaging the transistor and is not needed for a resistive lED type load.

Attached Thumbnails

  • learn_arduino_schematic.jpg

Edited by pbunn, 20 August 2016 - 01:39 PM.


#40 William Mc

William Mc

    Mercury-Atlas

  • *****
  • Posts: 2,783
  • Joined: 24 Nov 2007
  • Loc: North East Florida

Posted 20 August 2016 - 01:55 PM

Thanks,

 

It is curious that my current configuration using the bipolar does seem to be operating with the other schematic?    That being the cause, is there any real benefit to opening the box back up and swithing back to the Mosfet with your schematic?



#41 pbunn

pbunn

    Vanguard

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

Posted 20 August 2016 - 03:25 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.


Some fets can sometimes 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. th

Edited by pbunn, 20 August 2016 - 03:38 PM.


#42 William Mc

William Mc

    Mercury-Atlas

  • *****
  • Posts: 2,783
  • Joined: 24 Nov 2007
  • Loc: North East Florida

Posted 20 August 2016 - 04:18 PM

Couldn't resist.

 

Swapped a Mosfet back in using the new schematic and all is working now!  Whew! All Done.


Edited by William Mc, 20 August 2016 - 11:43 PM.


#43 spokeshave

spokeshave

    Vanguard

  • *****
  • topic starter
  • Posts: 2,302
  • Joined: 08 Apr 2015

Posted 20 August 2016 - 05:21 PM

My apologies. Pbunn is correct - the drawing is in error. The collector and emitter should be swapped. Sorry to have cause you so much trouble.

 

Tim



#44 William Mc

William Mc

    Mercury-Atlas

  • *****
  • Posts: 2,783
  • Joined: 24 Nov 2007
  • Loc: North East Florida

Posted 20 August 2016 - 05:53 PM

My apologies. Pbunn is correct - the drawing is in error. The collector and emitter should be swapped. Sorry to have cause you so much trouble.

 

Tim

 

 

No worries.  I'm grateful you posted this project at just the time I needed it. Strangely the Tipp will work with that reversed wiring. The Fet however will not.

 

The system is currently working fantastic!  I running test sequences in SGP, and the panel is now auto adjusting brightness with each filter change.


Edited by William Mc, 21 August 2016 - 06:50 AM.


#45 William Mc

William Mc

    Mercury-Atlas

  • *****
  • Posts: 2,783
  • Joined: 24 Nov 2007
  • Loc: North East Florida

Posted 01 September 2016 - 09:53 AM

Update:

 

 

 

The panel has been working great.  What a huge money saver this was. A fantastic DIY post!

 

One thing I haven't done is frost the panel as recommended, With it finally working OK, I hate to take the sander to it and risk disaster. I see unevenness in the frosting being a big risk, and it seems it would get smudged easily and be hard to clean.

 

 

How about this Idea. for less than the cost of an orbital sander which I would have to buy, I could order the matte finished, translucent white panel cut to size, and eliminate all risk.  The matte sheet could easily be fastened to the front of the LED panel.

 

Any thoughts on going this route?  Would 92%, or 55% transmission be best?

 

 

 

Here's the sheet I'm talking about.....       https://www.acrylite...73pwordl~p.html



#46 pbunn

pbunn

    Vanguard

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

Posted 01 September 2016 - 11:24 AM

How about a lower cost solution. I plan to place a sheet of drafting mylar over mine. 



#47 William Mc

William Mc

    Mercury-Atlas

  • *****
  • Posts: 2,783
  • Joined: 24 Nov 2007
  • Loc: North East Florida

Posted 01 September 2016 - 11:42 AM

How about a lower cost solution. I plan to place a sheet of drafting mylar over mine. 

 

 

 

 

Where to purchase that?  Have a link?



#48 pbunn

pbunn

    Vanguard

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

Posted 01 September 2016 - 07:31 PM

I have an engineering business and we purchase it by the roll, but I suspect that you can get it online. Also I think reproducible high quality plotter paper would work well.



https://www.amazon.c...t/dp/B000FA7KT4

Amazon has it

Edited by pbunn, 01 September 2016 - 07:34 PM.


#49 redtail

redtail

    Lift Off

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

Posted 29 September 2016 - 12:56 PM

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.


Edited by redtail, 29 September 2016 - 01:31 PM.


#50 redtail

redtail

    Lift Off

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

Posted 01 October 2016 - 02:55 AM

Its ok got it now...brilliant.




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