DIY LED driver for reef lighting
- Thread starter der_wille_zur_macht
- Start date
TheFishMan65
New member
I guess I can give it a shot. However since I did not follow the 3 drive board that was done here I will probably need some reviews. I think 3 chips have been discussed, but most decided on the CAT4101 so I will probably only cover that chip. Any specific points that some one needs answered.
mprygh
New member
Summary on just CAT4101 makes sense.
Regarding ON-Semi NCP3065, which is the first part of the thread:
Best is to read the app notes from the on-semi website.
They even have a great spreadsheet there, where you plug in various voltages and
currents you want, and it spits out exact Rsense / caps / inductors to use.
Very handy, and probably workable for others in same frequency range.
Regarding ON-Semi NCP3065, which is the first part of the thread:
Best is to read the app notes from the on-semi website.
They even have a great spreadsheet there, where you plug in various voltages and
currents you want, and it spits out exact Rsense / caps / inductors to use.
Very handy, and probably workable for others in same frequency range.
chrisfowler99
Premium Member
Building my boards now...I have a growing hatred of tiny, surface-mount components.
der_wille_zur_macht
Team RC
Trust me, those are HUGE for SMT. 
I think I wrote some tips earlier in the thread. It really helps to get technique down. First you need a way to hold the parts (fingers don't count). One common method is super fine tweezers. Another method is the "blue tack" method - you use a dental probe or other fine-tipped probe with a tiiiiny little dot of that tacky stuff you use to hang posters.
In either case, once you have a method to hold the part, go around the board and tin one pad for each part.
Then, use your tweezers or pick to hold the part in place. Touch the soldering iron tip to the pad you tinned, and the solder will reflow and hold the part in place. Then, you're good to go - that one pad will hold it down, so you can go solder the other pad "normally."
For the CAT4101 itself, you probably want to do the heatsink tab first. Tin the pad it goes on (on the board). Flip the part over and tin the bottom of the tab. Then hold it in place with your tweezers in one hand, and touch the soldering iron to the exposed edge of the tab with your other hand. Having a high-wattage iron turned to a reasonably high temperature will help to pump in enough heat to make the joint.
I think I wrote some tips earlier in the thread. It really helps to get technique down. First you need a way to hold the parts (fingers don't count). One common method is super fine tweezers. Another method is the "blue tack" method - you use a dental probe or other fine-tipped probe with a tiiiiny little dot of that tacky stuff you use to hang posters.
In either case, once you have a method to hold the part, go around the board and tin one pad for each part.
Then, use your tweezers or pick to hold the part in place. Touch the soldering iron tip to the pad you tinned, and the solder will reflow and hold the part in place. Then, you're good to go - that one pad will hold it down, so you can go solder the other pad "normally."
For the CAT4101 itself, you probably want to do the heatsink tab first. Tin the pad it goes on (on the board). Flip the part over and tin the bottom of the tab. Then hold it in place with your tweezers in one hand, and touch the soldering iron to the exposed edge of the tab with your other hand. Having a high-wattage iron turned to a reasonably high temperature will help to pump in enough heat to make the joint.
chrisfowler99
Premium Member
Ah...tinning the pad. I know I read that somewhere, but forgot about it. Got some fine tweezers I've been using to hold it in place to solder one side, then I do the other once it's held in place.
Just have 22 resistors and 22 CAT4101s to go. :lol:
Just have 22 resistors and 22 CAT4101s to go. :lol:
kcress
New member
Yes, go around and lightly tin the upper right pad - if you're right handed - of every footprint on the board. This includes resistors, caps, whatever. Then as wille suggests use tweezers to place and hold every part as you re-heat the single tinned pad. Use tweezers that could give you a good sharp poke.
Do not pre-tin more than one pin.
Once you have every part pinned down turn the board clockwise 180 degrees and solder all pads facing you along the bottom. Then turn it another 90 degrees. Step and repeat until all are soldered.
When you're done take a magnifying glass - the most powerful one you possess - and examine each and every joint. Some may need a touch-up, some you will have missed completely. Some can look soldered but the solder actually hadn't reached clear up to the pins. (Very annoying to trouble shoot.)
Once it all looks good, if you were smart and used water-cleanable-solder, wash the boards.
To wash them get a tooth brush. Use a pretty stiff one not the really soft ones. Put a squirt of dish soap in a bowl and mix it with about 10 parts HOT water. I use a squirt bottle. Dribble some hot tap water onto the both sides of the board and leave the boards face up. Soaking wet. Wait about 2 minutes. This allows the water to work on the rosin.
Now dip your brush into the soap mix and start scrubbing the boards. Scrub from top to bottom and then turn the board 90 degrees and repeat. Then turn the board 45 degrees and do one last pass. Flip the board and do it again to the back.
Now rinse the board under the tap with water as hot as you can handle.
At this point the savvy Cleaner Person will have a teapot full of realllly painfully warm RO water waiting. Do a generous rinse of both sides of the board.
Otherwise proceed with drying.
Shake off the board then tap it lightly on each edge. You want to jar the water out of all the nooks and crannies.
Once this is done set the boards down on one corner. You want any water that thinks about running down the board to run to a corner - not a side.
That's it. You have to steel yourself to wait for them to dry - 24 hours! The areas under the ICs stay wet a long time as the air doesn't reach those spots very well.
If you are an impatient dork, (like myself), you can speed things up a few ways. You can do the propping up on an oven shelf with the oven set to warm. Less than 140F!!! Or you can set the boards up with a fan blowing diagonally across them. This is sometimes hard because they must stay on their corners. Or, if you have an articulated lamp with an incandescent bulb in it you can arrange it to heat the boards up - all that can easily fit in its cozy glow. Again, not hot, just warm. Any of these speed-up schemes can bring the time down to about 2 hours. If you are really a speed freak you can employ a heat gun very judiciously remembering the air coming out of them is about 900F. That could reduce the time down to about 15 minutes because you'd need to let the board cool 10 minutes after such abuse.
When you're all done you will have nice clean -actually squeaky- clean boards that won't have strange things happening on them, won't capture dust, won't corrode down the road, and look very sharp.
Do not pre-tin more than one pin.
Once you have every part pinned down turn the board clockwise 180 degrees and solder all pads facing you along the bottom. Then turn it another 90 degrees. Step and repeat until all are soldered.
When you're done take a magnifying glass - the most powerful one you possess - and examine each and every joint. Some may need a touch-up, some you will have missed completely. Some can look soldered but the solder actually hadn't reached clear up to the pins. (Very annoying to trouble shoot.)
Once it all looks good, if you were smart and used water-cleanable-solder, wash the boards.
To wash them get a tooth brush. Use a pretty stiff one not the really soft ones. Put a squirt of dish soap in a bowl and mix it with about 10 parts HOT water. I use a squirt bottle. Dribble some hot tap water onto the both sides of the board and leave the boards face up. Soaking wet. Wait about 2 minutes. This allows the water to work on the rosin.
Now dip your brush into the soap mix and start scrubbing the boards. Scrub from top to bottom and then turn the board 90 degrees and repeat. Then turn the board 45 degrees and do one last pass. Flip the board and do it again to the back.
Now rinse the board under the tap with water as hot as you can handle.
At this point the savvy Cleaner Person will have a teapot full of realllly painfully warm RO water waiting. Do a generous rinse of both sides of the board.
Otherwise proceed with drying.
Shake off the board then tap it lightly on each edge. You want to jar the water out of all the nooks and crannies.
Once this is done set the boards down on one corner. You want any water that thinks about running down the board to run to a corner - not a side.
That's it. You have to steel yourself to wait for them to dry - 24 hours! The areas under the ICs stay wet a long time as the air doesn't reach those spots very well.
If you are an impatient dork, (like myself), you can speed things up a few ways. You can do the propping up on an oven shelf with the oven set to warm. Less than 140F!!! Or you can set the boards up with a fan blowing diagonally across them. This is sometimes hard because they must stay on their corners. Or, if you have an articulated lamp with an incandescent bulb in it you can arrange it to heat the boards up - all that can easily fit in its cozy glow. Again, not hot, just warm. Any of these speed-up schemes can bring the time down to about 2 hours. If you are really a speed freak you can employ a heat gun very judiciously remembering the air coming out of them is about 900F. That could reduce the time down to about 15 minutes because you'd need to let the board cool 10 minutes after such abuse.
When you're all done you will have nice clean -actually squeaky- clean boards that won't have strange things happening on them, won't capture dust, won't corrode down the road, and look very sharp.
chrisfowler99
Premium Member
Ok...I've got a weird one.
I've built 12 of the triple driver boards, each with a pair of drivers.
I was checking them over before hooking them up for the first time and 11 of them seem to be ready to test.
The other one, however, has something weird going on that I can't track down. Pin 1 and 3 of both of the CAT4101 chips are ground. On all of the other boards only pin 3 is ground. (actually, even the empty spot has pins one and three as ground, but that's to be expected if the other two are...)
I'm not sure if it's the board or something I did while building it.
Anyone seen this happen?
I've built 12 of the triple driver boards, each with a pair of drivers.
I was checking them over before hooking them up for the first time and 11 of them seem to be ready to test.
The other one, however, has something weird going on that I can't track down. Pin 1 and 3 of both of the CAT4101 chips are ground. On all of the other boards only pin 3 is ground. (actually, even the empty spot has pins one and three as ground, but that's to be expected if the other two are...)
I'm not sure if it's the board or something I did while building it.
Anyone seen this happen?
der_wille_zur_macht
Team RC
Get a magnifying glass and check the whole length of every trace connected to the suspect pin and see if there's a fault in the board. Look over your solder joints and make sure there are no shorts.
And thanks kcress for the soldering notes.
And thanks kcress for the soldering notes.
chrisfowler99
Premium Member
Testing for shorts was how I found this. Wonder if my daughter has a magnifying glass. :lol:
Is it possible (likely?) that the CAT4101 has an internal short that could cause this? Either a bad chip or some damage I caused soldering it?
Is it possible (likely?) that the CAT4101 has an internal short that could cause this? Either a bad chip or some damage I caused soldering it?
kcress
New member
Thanks Willie.
chrisfowler99; Possible? Yes. Likely? Not at all.
You need to get that magnifying glass and look for solder bridging which is most likely.
Also likely but less so is a faulty board fab which happens occasionally. That's where copper wasn't correctly removed during the board's fabrication. There are lots of ways for that to happen.
Also make sure that you didn't cause a solder bridge under or around that chip. That's really common.
You need as much magnification as you can get. A loop is best.
chrisfowler99; Possible? Yes. Likely? Not at all.
You need to get that magnifying glass and look for solder bridging which is most likely.
Also likely but less so is a faulty board fab which happens occasionally. That's where copper wasn't correctly removed during the board's fabrication. There are lots of ways for that to happen.
Also make sure that you didn't cause a solder bridge under or around that chip. That's really common.
You need as much magnification as you can get. A loop is best.
chrisfowler99
Premium Member
I was looking for a solder bridge around either of those chips. I find it somewhat unlikely that I would have bridged the 1st and 3rd pin without including the 2nd.
I'll pick up a good magnifying glass tomorrow and get a better look.
I'll pick up a good magnifying glass tomorrow and get a better look.
Anybody got up to date files for the CAT4101 boards for seeedstudio?
Or if someone is going to order and won't mind adding some I would appreciate it.
Can someone please just list the parts needed for this board if I want to manually dim 2 led strings
independently with a pod?
Or if someone is going to order and won't mind adding some I would appreciate it.
Can someone please just list the parts needed for this board if I want to manually dim 2 led strings
independently with a pod?
Last edited:
der_wille_zur_macht
Team RC
Chris, make sure you inspect the entire length of the trace(s) in question - as kcress suggested the fault may be in the board itself. I've had a bad board here and there from seeed. Or it may be in a solder joint on one of those traces not near the IC pins (i.e. over by the input header).
Gumph, shoot me a PM with your email and I'll send the gerber files to you.
When you say "I want to manually dim 2 LED strings independently with a pod" what exactly do you mean? How many LEDs will be in each string? What's a pod?
Gumph, shoot me a PM with your email and I'll send the gerber files to you.
When you say "I want to manually dim 2 LED strings independently with a pod" what exactly do you mean? How many LEDs will be in each string? What's a pod?
jfarabaugh
In Memoriam
When is someone going to start selling these prebuilt already? hint hint for someone to make a few $
der_wille_zur_macht
Team RC
Personally I have no interest in doing this as a business venture. The exact designs I've released here are released under an open source license that restricts commercial use (creative commons 3.0 BY-NC-SA). The idea is to foster DIY. If someone wants to design and sell their own drivers, they are certainly free to do so, just not exact copies of these designs. 
And, as an aside, if someone were to do that, they'd have to be careful about talking about it openly on the forums to ensure they were following Reef Central's guidelines about commercial posting and sponsorship.
And, as an aside, if someone were to do that, they'd have to be careful about talking about it openly on the forums to ensure they were following Reef Central's guidelines about commercial posting and sponsorship.
Thanks DWZM but I kind of worked out that with shipping and all to me
I am better of bying a driver from sure elecronics on ebay, free shipping.
Only would have to DIY a pwm module. Might pm you for a part list on those as you never
listed the parts on your other thread.
Will move my other Q's to the LED thread as they do not really involve the DIY of a driver.
I am better of bying a driver from sure elecronics on ebay, free shipping.
Only would have to DIY a pwm module. Might pm you for a part list on those as you never
listed the parts on your other thread.
Will move my other Q's to the LED thread as they do not really involve the DIY of a driver.
TheFishMan65
New member
The Summary
The Summary
Well here goes. Sorry actarus I am not going to try and write this in French. Comments (missed questions) and correction welcome. I hope this fairly represent what has been said in this thread.
CAT4101 SUMMARY
I suppose the first place to start is with the DataSheet. The advantage of this part is the few external components needed. They are a resistor and capacitor. The basic layout is on the top of sheet 8 of the datasheet.
BOARD DESIGNS
This thread has multiple drivers in it, but this summary centers on the CAT4101 driver which most seemed to use. I believe there are 2 boards discussed which used the CAT4101. One designed by der_wille_zur_macht which will be referred to as the 3 driver board. This board has probably been used by more people since the boards are easier to find in small quantities. The board I am more familiar with is the one I designed and will be referred to as the 8 driver board. This is the schematic for one driver of the 8 driver board.
I believe the differences are that the 3 drive board does not have a 5 volt regulator and there is no potentiometer for manual dimming.
The three driver version can be found here:
The three driver board does not use a dimming potentiometer for these reasons:
LEDS PER STRING
Because of the voltage limitation most have run the CAT4101 with 6 LEDs in the string. In theory the XP-G might be able to run 7.
PWM DIMMING
Dimming could be done with a 555 timer chip or a microcontroller.
The CAT4101 is limited to 25 volts, so most people have used it with a 24 volt supply. I have noticed a heat problem in some cases and designed my board with 2 inputs for the supply voltage. My plan was to be able to turn them down so that the CAT4101 would not have to waste as much energy. The CAT4101 needs about 1/2 volt over the voltage of the LED string. So if 6 LEDs at 3.5 volts are used the supply can be set to 21.5 volts to produce less heat and let the CAT4101 runs more efficiently.
POWER SUPPLY SIZE
There was some confusion about how large a power supply is needed. Take the number of string times the current * the voltage. This is the amount power (in watts) that are required to drive the circuit. Per kcress's suggestion of 80% safety factor divide this but 0.80 to get the wattage of the power supply needed. This does not mean that the AC circuit will be supplying the same amount of current. Divide by 120 to find out how much current will be drawn for the AC house receptacle.
Example
12 string
0.70 amps
At 24 volts (Ideally you should turn this down.
So the power (watts) is 12 * 0.70 * 24 = 202 watts. So the supply should probably be around 250 watts. This would draw approximately 1.7 amps (probably a little more due to efficiencies) from the AC receptacle.
HOW TO SOLDER
The Summary
Well here goes. Sorry actarus I am not going to try and write this in French. Comments (missed questions) and correction welcome. I hope this fairly represent what has been said in this thread.
CAT4101 SUMMARY
I suppose the first place to start is with the DataSheet. The advantage of this part is the few external components needed. They are a resistor and capacitor. The basic layout is on the top of sheet 8 of the datasheet.
BOARD DESIGNS
This thread has multiple drivers in it, but this summary centers on the CAT4101 driver which most seemed to use. I believe there are 2 boards discussed which used the CAT4101. One designed by der_wille_zur_macht which will be referred to as the 3 driver board. This board has probably been used by more people since the boards are easier to find in small quantities. The board I am more familiar with is the one I designed and will be referred to as the 8 driver board. This is the schematic for one driver of the 8 driver board.
I believe the differences are that the 3 drive board does not have a 5 volt regulator and there is no potentiometer for manual dimming.
The three driver version can be found here:
POTENTIOMETER DIMMING
The three driver board does not use a dimming potentiometer for these reasons:
When I searched the data sheet the only requirement I found was on page 8,"The RSET resistor should have a Kelvin connection to the GND pin of the CAT4101." When I searched for Kelvin Connection was in using a resistor to measure current through voltage. I am not sure this is what they meant (kcress you started this care to chime in?). I can say that I kept my leads short and have not had any problem (but I admit my fixture is not up yet). I also shared the board with some folks and have not heard of any reported problems. So in summary if you use a dimming potentiometer you may want to do a little more research.I think we discussed this back about ten or twenty pages - some time early last spring.
It's not really a big deal, but I'm just nervous about using the feedback pin for external control. I'm projecting a bit but it's just not designed for that. Look in the datasheet at how specific they are about things like trace length and layout for the RSET resistor. Putting a signal that sensitive through a pot (which are generally unstable and inaccurate) makes me nervous. Compare that to the EN/PWM pin, which they expect to be fed from an unknown-quality external signal, and is pretty wide open design-wise.
I'm sure that using RSET for dimming can work just fine, and in an implementation where it's working I wouldn't change anything - but, if designing from scratch, I'd try my best to use EN/PWM instead.
LEDS PER STRING
Because of the voltage limitation most have run the CAT4101 with 6 LEDs in the string. In theory the XP-G might be able to run 7.
PWM DIMMING
Dimming could be done with a 555 timer chip or a microcontroller.
VOLTAGEder_wille_zur_macht said:
The CAT4101 is limited to 25 volts, so most people have used it with a 24 volt supply. I have noticed a heat problem in some cases and designed my board with 2 inputs for the supply voltage. My plan was to be able to turn them down so that the CAT4101 would not have to waste as much energy. The CAT4101 needs about 1/2 volt over the voltage of the LED string. So if 6 LEDs at 3.5 volts are used the supply can be set to 21.5 volts to produce less heat and let the CAT4101 runs more efficiently.
POWER SUPPLY SIZE
There was some confusion about how large a power supply is needed. Take the number of string times the current * the voltage. This is the amount power (in watts) that are required to drive the circuit. Per kcress's suggestion of 80% safety factor divide this but 0.80 to get the wattage of the power supply needed. This does not mean that the AC circuit will be supplying the same amount of current. Divide by 120 to find out how much current will be drawn for the AC house receptacle.
Example
12 string
0.70 amps
At 24 volts (Ideally you should turn this down.
So the power (watts) is 12 * 0.70 * 24 = 202 watts. So the supply should probably be around 250 watts. This would draw approximately 1.7 amps (probably a little more due to efficiencies) from the AC receptacle.
HOW TO SOLDER
