DIY LED driver for reef lighting

[O2Surplus]

Kevin

This is my first post, but I'm really familiar with the Cat 4101. I think I can help you out. When viewing the cat chips with the pins pointing down and working from left to right- these are the voltages you should see when the driver board is powered up.
Left pin (#1 enable) = 5 volts or a pwm signal
(#2 Vcc ) = 5 volts
center pin (#3 ground) = 0 volts
(#4 Rsesns) = 1.2 volts
Right pin (#5 Led-) = vf of led string

take some measurements and let me know what your missing. We can figure it out from there.

 

[Kevin012]

Kevin

This is my first post, but I'm really familiar with the Cat 4101. I think I can help you out. When viewing the cat chips with the pins pointing down and working from left to right- these are the voltages you should see when the driver board is powered up.
Left pin (#1 enable) = 5 volts or a pwm signal
(#2 Vcc ) = 5 volts
center pin (#3 ground) = 0 volts
(#4 Rsesns) = 1.2 volts
Right pin (#5 Led-) = vf of led string

take some measurements and let me know what your missing. We can figure it out from there.

I have to leave now for 4 houres, but is it not so that from the 24V it is going threw the volt regulator to bring it to 5V, and from there to the CAT chip ?

and becouse of this I should be able to mesure either 24V ore 5V on the voltage regulator ?

at this time, when 24V is on, I cant mesure any voltage on any of the pins of the voltage reg.

----------------------------------------------------------

for the mesurements you mention, I when testing all 5 pins, have to put the black probe on a ground, and with the red probe go over each pin and write down its value ? ore where do I put the probes (I only have basic mesuring skills)

Kevin

 

[O2Surplus]

You've got the right idea. Keep the black probe on a ground point and use the red probe to take measurements. If your not seeing 24 volts at the regulator then that's where we'll start. You can't get 5 volts without the 24 volts being there first.

 

[Gorgok]

Yes, to measure voltage you ground one probe and measure with the other, which one doesn't really matter it just will read negative volts if you have them backwards. Doing that you should get a reading on the input leg of the 5V reg (left leg is input) of 24V. If that is not the case you won't get anything on the right leg or the driver 5V pin.

Either follow the traces (or get us close up clear pictures front and back) to see where the 24V input goes and see if it goes to the regulator, or use a multimeter set to measure resistance and measure with one probe on the input leg of the 5V regulator and the other probe the 24V screw terminals. If you get infinity or really high numbers then its not connected. If no luck, then measure resistance from the input leg of the 5V to the 5V screw terminals to see if those are connected. If those are connected measure from the 5V screw terminals to the 24V screw terminals. Also measure from the 5V screw terminals to the #2 (VCC) pin on the CATs. If still no luck start testing if the 5V regulator is connected to anything. We need to know where it is connected to be able to really get this figured out.

There is a slight chance that if the 5V reg is connected to the 5V terminals, but not to the 24V terminals or any other 24V source and the 5V terminals are not connected straight to the #2 pin of the CATs that you need to give the 5V terminals 24V as well. This is one of those things you have to test and test because just trying this idea out would damage the CATs most likely. The basic idea here is that the board was designed in a way that if you were not using the 5V reg you could give 5V power to the 5V terminals and put a jumper wire in place of the 5V reg to get 5V to the output side. Without the jumper wire the screw terminal would not be connected at all to the CATs.

 

[Kevin012]

Yes, to measure voltage you ground one probe and measure with the other, which one doesn't really matter it just will read negative volts if you have them backwards. Doing that you should get a reading on the input leg of the 5V reg (left leg is input) of 24V. If that is not the case you won't get anything on the right leg or the driver 5V pin.

Either follow the traces (or get us close up clear pictures front and back) to see where the 24V input goes and see if it goes to the regulator, or use a multimeter set to measure resistance and measure with one probe on the input leg of the 5V regulator and the other probe the 24V screw terminals. If you get infinity or really high numbers then its not connected. If no luck, then measure resistance from the input leg of the 5V to the 5V screw terminals to see if those are connected. If those are connected measure from the 5V screw terminals to the 24V screw terminals. Also measure from the 5V screw terminals to the #2 (VCC) pin on the CATs. If still no luck start testing if the 5V regulator is connected to anything. We need to know where it is connected to be able to really get this figured out.

There is a slight chance that if the 5V reg is connected to the 5V terminals, but not to the 24V terminals or any other 24V source and the 5V terminals are not connected straight to the #2 pin of the CATs that you need to give the 5V terminals 24V as well. This is one of those things you have to test and test because just trying this idea out would damage the CATs most likely. The basic idea here is that the board was designed in a way that if you were not using the 5V reg you could give 5V power to the 5V terminals and put a jumper wire in place of the 5V reg to get 5V to the output side. Without the jumper wire the screw terminal would not be connected at all to the CATs.

ok got some measurements:

reading on the input leg of the 5V reg (left leg is input) of 24V. If that is not the case you won't get anything on the right leg or the driver 5V pin. result = nothing



measure resistance and measure with one probe on the input leg of the 5V regulator and the other probe the 24V screw terminals
result:

with the red probe on the 24V+ and the black probe on the input pin of the volt. R, there is nothing to measure

with the red probe on the input pin of the V.R. and the black on the 24V+ it went from - 964 slowly down, and after 30 secounds it was in the +200 and counting

measure resistance from the input leg of the 5V to the 5V screw terminals to see if those are connected

resold: it counted 20,30,40 and then went to 1 .


measure from the 5V screw terminals to the 24V screw terminals

resold: nothing to measure

Also measure from the 5V screw terminals to the #2 (VCC) pin on the CATs

resold: with the black probe on the 5v- terminal and the red on pin 2 of the cat (vin) it gives me 4.8

when having the red probe on the 5v- and the black on pin 2 of the cat, it mesures nothing

hope this info gives you some derection,

i also attached the schematics of the board. I could make pictures of the board, but I think those that know more than I do, can read the diagram better.


overall what is strange is that Rob (thefishman65) build me 3 driverboards, and all 3 have the same proublem, I mailed some other's rob build the drivers for, but everyone that has this board, is aparently on vacation, as I asume it is something structurally)

 

[Gorgok]

Based on the schematics it should work as is, +24V goes to the input pin via traces on the board, the 5V screw terminals are really completely separate and nothing needs a jumper even without the regulator in use.

The only readings that matter with the schematic in hand is the resistance from the voltage regulator to the +24V terminal and the voltage from the vreg to ground. The rest were just ideas that could have been but have no meaning any more. Hmm.

Some more voltage tests you can do though. Test the LED#+ screw terminals to ground and see if each of them get +24V. If the voltage tests read nothing then you probably have a bad connection from the PSU to the board, the screw terminal to the board, or a trace has lifted in the 24V line.

You can also test for resistance again from the input leg of the voltage regulator to the LED#+ terminals (only need to measure one time, which probe goes where doesn't matter). Again these should be really low readings like 0.05Ω to 0.1Ω. If these tests either cannot be measured or have huge numbers then it could be the regulator has a bad connection to the board or a trace has lifted before it joins the 24V line.

 

[Kevin012]

Based on the schematics it should work as is, +24V goes to the input pin via traces on the board, the 5V screw terminals are really completely separate and nothing needs a jumper even without the regulator in use.

The only readings that matter with the schematic in hand is the resistance from the voltage regulator to the +24V terminal and the voltage from the vreg to ground. The rest were just ideas that could have been but have no meaning any more. Hmm.

Some more voltage tests you can do though. Test the LED#+ screw terminals to ground and see if each of them get +24V. If the voltage tests read nothing then you probably have a bad connection from the PSU to the board, the screw terminal to the board, or a trace has lifted in the 24V line.

You can also test for resistance again from the input leg of the voltage regulator to the LED#+ terminals (only need to measure one time, which probe goes where doesn't matter). Again these should be really low readings like 0.05Ω to 0.1Ω. If these tests either cannot be measured or have huge numbers then it could be the regulator has a bad connection to the board or a trace has lifted before it joins the 24V line.

gust to make shore, the test I did above where based on mesuring without powering the unit, so I did not plug anything in !!

and as there is already nothing to be mesured between the volt. reg. and the 24V+ I asume there is my proublem somewhere, not

ohh, and : You can also test for resistance again from the input leg of the voltage regulator to the LED#+ terminals

it went up like 350, 480 870 and at 1400 it went I and nothing anymore

 

[Gorgok]

Yes the resistance tests should be done unpowered. The further testing is just to try and identify where the problem is. If its as simple as the +24V screw terminal or the vreg being loose and needing to be re-soldered then its not a big deal to fix. If its a trace that is broken or lifted or simply not connected then it could be much more serious.

 

[O2Surplus]

Also measure from the 5V screw terminals to the #2 (VCC) pin on the CATs

resold: with the black probe on the 5v- terminal and the red on pin 2 of the cat (vin) it gives me 4.8

when having the red probe on the 5v- and the black on pin 2 of the cat, it mesures nothing

hope this info gives you some derection,

i also attached the schematics of the board. I could make pictures of the board, but I think those that know more than I do, can read the diagram better.


overall what is strange is that Rob (thefishman65) build me 3 driverboards, and all 3 have the same proublem, I mailed some other's rob build the drivers for, but everyone that has this board, is aparently on vacation, as I asume it is something structurally)

Kevin

The 5 volt regulator has to be working since you measured 4.8 volts at the Vcc pin on one of the Cats. According to the schematic - 5volts is fed from the regulator to all the Cat Vcc pins. I think what your missing is a 5 volt signal to the Enable/pwm pins. Connect a jumper wire between the +5 volt connection and all of the PWM input connections. The Leds should now light up. Let us know what happens.

 

[Kevin012]

Kevin

The 5 volt regulator has to be working since you measured 4.8 volts at the Vcc pin on one of the Cats. According to the schematic - 5volts is fed from the regulator to all the Cat Vcc pins. I think what your missing is a 5 volt signal to the Enable/pwm pins. Connect a jumper wire between the +5 volt connection and all of the PWM input connections. The Leds should now light up. Let us know what happens.

The 5 volt regulator has to be working since you measured 4.8 volts

this was on the ohm position, without any power attached to the board !!

Connect a jumper wire between the +5 volt connection and all of the PWM input connections. The Leds should now light up

if you se one of the pictures I placed earlyer, of Rob's driver boards, it shows in deed a jumper, ore something.
http://www.reefcentral.com/forums/showpost.php?p=17221092&postcount=859

but in call it his manual, it never mentiones anything like this to be dun, only he warns: You can leave the 5 volt regulator off and get 5 volts from somewhere else. Or you can use this to power another device. DO NOT connect multiple 5 volt supplies together.

thing gust is that as i'm not to familiar with all this, I dont want to start conecting whires, and than fry the chips, and as Rob is already un rachable for about 2 weeks, I'm glad you guys finaly have some input.

so you think it is safe to jump from the 5v+ conection to the PWM (meaning the PWM gets 2 whires, one isthe jumper, and the other the PWM)

 

[Gorgok]

That jumper is from the +5V to one of the PWM sets, effectively making the LED run full bright. I think its just used for testing.

 

[Kevin012]

That jumper is from the +5V to one of the PWM sets, effectively making the LED run full bright. I think its just used for testing.

in this case it is 100%, but as long as I cant mesure any power on the volt reg (when power is on) than there also cant be any power going past the volt. reg,

I could try it,(to place a jumper) as long as I dont fry any thing, gust plain 24V on the driver board, a jumper on the 5v+ and PWM and a string on one of the string conections.

 

[O2Surplus]

whoops, sorry I didn't realize that you were taking continuity measurements with the board non-powered. I don't normally rely on continuity tests, as the results can be misleading. you need to power up the board and perform voltage tests at the Cats and see what signal you're missing. Like I said earlier- in order for the cats to function you need a 5 volt signal to the #1 & #2 pins, 0volts at #3, 1.2 volts @ #4, and the Vf of your led string at pin #5.

 

[Gorgok]

Nono, don't place jumpers. I know you don't get voltage to the reg, which is the problem. Putting 24V from a jumper to the 5V side will destroy stuff.

The continuity tests are simply an attempt to see why the board doesn't seem to get voltage where it should. It would be easier to test for voltages, but as the first test showed, it never reaches the vreg, so the whole board will not fire.

 

[Kevin012]

Nono, don't place jumpers. I know you don't get voltage to the reg, which is the problem. Putting 24V from a jumper to the 5V side will destroy stuff.

The continuity tests are simply an attempt to see why the board doesn't seem to get voltage where it should. It would be easier to test for voltages, but as the first test showed, it never reaches the vreg, so the whole board will not fire.

kind of what I thought,

strange thing actually is that gust now, when I did the sets again, I realised that the power reg. gets burning hot (and I double checked with a other meter as well, the power supply only gives out 24V at the moment.

as it is 11 now, I print everything we went over, and tomorrow I will re do all things we gust did, to double check things, and with all 3 boards, gust so with a clear mind I walk over the steps again, ans se if I missed something, and to double confirm, my mesurements.

thanks all your help so far (you all)

 

[TheFishMan65]

Gee go on vacation for a week and look what happens. Thanks to all for helping out.

I skimmed thread trying to catch up, but I think the problem is that he is using the wrong 24 volt supply. The board has two 24 volt inputs (designed for one for blue and one for white). Only one supplies the 5 volt regulator. Based on the picture in post 1869 he is using the wrong one.

So with the terminals down supply power to the right side. Scrap the Arduino for now. Connect the PWMs to the 5 volt terminal and LEDs to + an -.

 

[Gorgok]

Ah bah, i knew it had 2 inputs just somehow that never occurred to me that it was the wrong one. It would explain everything though =D.

 

[Kevin012]

Ah bah, i knew it had 2 inputs just somehow that never occurred to me that it was the wrong one. It would explain everything though =D.

hy guys,

so Rob is back, and I will find out how to get it running with him, as he build the drivers. I greatly thank you all so far, if I end up with somemore questions, I now know whome to ask)

ahh, I did found out this morning that hte other 24V conection did work on the regulator, so one of the boards is kind of working (but I have to work with Rob a bit to get it 100%)

2 of the 3 baords have the power reg. turn burning hot, so might be only a broken power reg, but will se later, first try to get one board working, and then go from there.

Kevin

 

[jgwinner]

Heads up to anyone planning to use my driver boards. There is a little problem with the analog dimming part. Apparently the chip supports a third dimming method (RTFDS) and switches to internal power source when there is a resistor (pot) from Iadj to GND (which there is on my board for the voltage divider 5V - 1.24V), so you can never shut off the LEDs (there is about 30mA current when 0 volts are passed). I'm thinking how to take care of that now. Just wanted to let you know.

I guess I can just skip the voltage divider but then there will be "only" about 64 dimming steps from an arduino...

If anyone has ideas I'm all eyes.

Not sure ... sounds like R2 of the board should have a jumper?

64 might be Ok but I think I've heard it's a noticeable 'step'.

I like the idea of an analog and PWM dim; my concept was a plug so I could yank my controller off the light bracket, and have the lights go to a preset brightness while I fiddle with the hardware.

What's RTFDS?

While searching for that, look what I found:

The Effects of Dimming on High-power LEDs

== John ==

 

[terahz]

John, the problem is that we want to be able to do analog dimming from a controller and also be able to turn off the lights by supplying 0 voltage to that pin. As it is, the board will require 2 pins from the controller: one for dimming and another to the EN pin, so when the dim down period is over, you just set the EN pin to LOW and the light is gone. Ideally, however, I'd like this to work with a single pin (only the voltage to the IAdj pin on the IC). That can happen if there is a transistor which turns off the EN connection to VIN if the Iadj voltage is under say 100mV. The only problem is, that I can't find such transistor with cutoff voltage of 100mV that can switch on/of up to 75V

Oh and RTFDS is like RTFM for data sheets