perikaruppan
Member
or i have a 0.25Ohms 5W Resistor, that will give an avg current of 797ma, but this is a higher W resistor. will that help? currently using only a 1W resistor.
DIY LED driver for reef lighting
- Thread starter der_wille_zur_macht
- Start date
perikaruppan
Member
yes it does allow upto 60v but they have an XLS worksheet using which i had planned the specs on it.. this is the link..
http://www.diodes.com/_files/calculators/ZXLD135X_136X_Multi_device_calculator.xls
Used this to change the values to get the specs.
http://www.diodes.com/_files/calculators/ZXLD135X_136X_Multi_device_calculator.xls
Used this to change the values to get the specs.
Last edited:
TheFishMan65
New member
I remember looking at that spreadsheet. I never did figure out how they decided the limit of LEDs to input voltage.
I would leave the resistor where it is for now. If you decide you want more light then change it.
So I would still go with the plan above. Add an LED turn down till the current changes, turn up 1/2 volt and see what the temperature does. If it is low then maybe turn it up a little more and see where it start to get hot. If it is hot at the minimum voltage then we are going to have to come up with something else.
I would leave the resistor where it is for now. If you decide you want more light then change it.
So I would still go with the plan above. Add an LED turn down till the current changes, turn up 1/2 volt and see what the temperature does. If it is low then maybe turn it up a little more and see where it start to get hot. If it is hot at the minimum voltage then we are going to have to come up with something else.
perikaruppan
Member
ok will do that then.. and what do you mean by "If it is hot at the minimum voltage then we are..." What min voltage that you are referring here?
der_wille_zur_macht
Team RC
Rather than specifying a minimum drop across the chip, they limit the maximum duty cycle of the internal switch (i.e. the "on" time) which implies a minimum drop across the chip for a given set of variables. OnSemi and other manufacturers of constant current switching regulators do this too, so it must be commonplace for this type of chip.
SpacedCowboy
New member
I think I mentioned before that I had major problems with the ZXLD chip and heat dissipation, and I don't think I was the only one. Even 3v difference between the input voltage and the voltage that the LEDs wanted to see (ie: number-of-leds x Vf-per-led) was sufficient to cause meltdown fairly quickly, and that's after I had a heatsink attached to the top of the chip...
That's why I switched to the CAT4101. I was actually going to ditch the ZXLD chip for DWZM's design before I found the CAT4101. Both of these other designs are far easier to work with, at least IMHO.
Just my 2p-worth,
Simon.
That's why I switched to the CAT4101. I was actually going to ditch the ZXLD chip for DWZM's design before I found the CAT4101. Both of these other designs are far easier to work with, at least IMHO.
Just my 2p-worth,
Simon.
Skeptic_07
New member
OK, I'm seriously considering selling my meanwells and building my own drivers because I'm having lots of problems getting what I want out of them. Don't get me wrong about meanwells, they are great and make the LED's very bright. Its just that I can't get them to only light the emitters a tiny bit. The low end sensitivity is .. well... non-existent or so it seems.
My objective from the start was to have very long gradual increases and decreases in light. I want to be able to get them to light up minimally (like when you use the diode tester function on a multi-meter) and gradually increase to somewhere like 75% of maximum rated current. Can I accomplish that through this project? Are you guys seeing this kind of low end sensitivity?
sorry if this is in the thread somewhere already, I'm about to go back and re-read it.
My objective from the start was to have very long gradual increases and decreases in light. I want to be able to get them to light up minimally (like when you use the diode tester function on a multi-meter) and gradually increase to somewhere like 75% of maximum rated current. Can I accomplish that through this project? Are you guys seeing this kind of low end sensitivity?
sorry if this is in the thread somewhere already, I'm about to go back and re-read it.
der_wille_zur_macht
Team RC
Depends on which of the three driver circuits you're asking about. The original NCP chip, in the mode I used it (boost) had a very big cutoff, worse than the meanwells. In a buck configuration I'd guess it would perform better, but I never tested it.
The ZXLD - well, the datasheet shows a pretty big cutoff when dimmed via analog signal, and doesn't show anything for PWM so it's anyone's guess.
The CAT4101 - I'm holding out hope that it has a small cutoff at worst when PWM dimmed but I haven't tested it yet. My proto boards are taking forever to get here.
But I think you're on the right track - DIY'ing means you can get what YOU want, not what a manufacturer thinks you want.
TheFishMan65
New member
PWMs should be good for at leat 1% more likely 1 part in 255. So the 4101 should work.
der_wille_zur_macht
Team RC
The problem is that the chip might or might not be able to turn on and off fast enough at such a low duty cycle, which can create weird behavior and a cutoff. I agree that the datasheet makes the CAT4101 look promising in this respect but I've seen weirdness in other drivers so I wouldn't rule it out here.
perikaruppan
Member
and the results...
tested with an additional LED the comparison :
10 LEDS : 3.76 V @ 650ma
11 LEDS : 3.51 V @ 450ma
and @ the 11 LED stage, the heat much much less when compared to that of the 10 LEDs. All this tested over a pretty short period of 60 seconds!!! cuz i didnt want to blow the chip away.
As Fishman suggested, have not tried lowering the Voltage.
Now in which string either with 10 LEDs or 11 LEDs what should i do to get the desired results?
Desired results is driving @ 650+ma with zero heat.
tested with an additional LED the comparison :
10 LEDS : 3.76 V @ 650ma
11 LEDS : 3.51 V @ 450ma
and @ the 11 LED stage, the heat much much less when compared to that of the 10 LEDs. All this tested over a pretty short period of 60 seconds!!! cuz i didnt want to blow the chip away.
As Fishman suggested, have not tried lowering the Voltage.
Now in which string either with 10 LEDs or 11 LEDs what should i do to get the desired results?
Desired results is driving @ 650+ma with zero heat.
der_wille_zur_macht
Team RC
Run it with 11 LEDs, and slowly turn the voltage on your power supply UP until the current in the LED string reaches your desired value. At that very point, you'll be at the optimal point.
TheFishMan65
New member
And then make sure your house doesn't burn down 
. From what Spaced Cowboy said you may still find that it is too hot. But dwzm told you to do what I would tell you do do.
. From what Spaced Cowboy said you may still find that it is too hot. But dwzm told you to do what I would tell you do do.perikaruppan
Member
ok will do that in steps of 1V. and wont that increase the voltage across the LEDS?
Just a question, aren't these constant current drivers, then why does the current vary when the LEDs are increased or decreased? or when the VCC is increased or decreased? should the current be the SAME, is that for what we are using these drivers? CONSTANT CURRENT!
Just a question, aren't these constant current drivers, then why does the current vary when the LEDs are increased or decreased? or when the VCC is increased or decreased? should the current be the SAME, is that for what we are using these drivers? CONSTANT CURRENT!
perikaruppan
Member
and one more question, if it is TOO Hot, will the IC turn OFF or will it BLOW!! i know u said to take care of my roof, but just in case so that i can handle the situation properly at home!!!
der_wille_zur_macht
Team RC
Yes, but you want that at this point.
It is a constant current driver, yes - but it's a BUCK driver. It regulates current by dropping voltage.
To reach a target current, a given LED string needs a certain voltage applied. Your driver starts with an incoming voltage higher than that, and chops it down. So the reason you're seeing a "low" current on the longer LED string right now is because the input voltage is LOWER than what the LED string needs. And since the driver can only reduce voltage (not increase it), it can't solve the problem for you.
Hence, you need to turn the input voltage up until the target current is reached. At that very point, the driver will actually be operating in a constant current mode, and it will be the most efficient point in it's operational range because it'll be the point at which it's dropping the least possible voltage.
der_wille_zur_macht
Team RC
In theory, the IC has a thermal protection circuit that'll shut it off if it gets too hot. That protection has some hysteresis built in, so if you run it in a situation where it's getting too hot, it'll shut itself off, cool down, start back up, heat up, shut itself off, etc. - it'll just cycle on and off as it trips the protection over and over. Clearly, you want to design the system so that this doesn't happen!
perikaruppan
Member
thanks a lot for that detailed explanation.
I think i will revise.. as i have a 48V/7.3A Power supply and was thinking of a 7 Driver System, i.e., 7X12 = 84 LEDs for my tank of 120, i was planning to run it at 48V as per the spec sheet. So might as well start of with 13 LEDs (as in the spec it tells 12LEDS so add 1 more to the string). and i can get another 7 LEDS into the system HOPEFULLY. and reduce the Voltage to get the Current @ 650mA and hopefully LESS heat!!!
I think i will revise.. as i have a 48V/7.3A Power supply and was thinking of a 7 Driver System, i.e., 7X12 = 84 LEDs for my tank of 120, i was planning to run it at 48V as per the spec sheet. So might as well start of with 13 LEDs (as in the spec it tells 12LEDS so add 1 more to the string). and i can get another 7 LEDS into the system HOPEFULLY. and reduce the Voltage to get the Current @ 650mA and hopefully LESS heat!!!
SpacedCowboy
New member
Which is exactly what I saw, running at 0.5A, right up until the magic smoke came out of the input-stage resistor and/or capacitors. It's difficult to tell where the smoke comes from, it appears too fast and the circuit is small.
My experience is that the chip also dumps a whole load of power into the sense resistor when it's voltage isn't well-matched. The resistor was actually hotter than the chip, until it burnt out
Simon.
perikaruppan
Member
since already had it wired for 11 LEDS...
got 650ma At 44.5V but there was heat on the zxld. not burning heat but still heat.. in about a minute of operation!!!.
As cowboy had mentioned i think i need to address the heat problem.
But there is not that much heat @ 500ma!!... but dont want to drive my LEDs at that current..
got 650ma At 44.5V but there was heat on the zxld. not burning heat but still heat.. in about a minute of operation!!!.
As cowboy had mentioned i think i need to address the heat problem.
But there is not that much heat @ 500ma!!... but dont want to drive my LEDs at that current..
