DC-DC converters and pixels.. getting your feet wet?

[AAH]

The reasoning for David replacing the trimpot is that it is an open frame model and they are susceptible to the wiper corroding over time which means that the voltage loses stability and depending on the circuit design the voltage can end up up being anything from zero volts up to the input voltage. Enclosed conductive plastic are my preferred pots.
He doesn't show a replacement pot as he has selected 2 resistors that give the right output voltage for him. Buried under a liberal coating of flux residue you may make out what looks to be 2 0805 sized fixed resistors.

 

[David_AVD]

It's only the first type of converter that I mention replacing the crappy trim pot with the (one) resistor. The multi-turn trim pot on the LM2596 based board is fine.

The LM2596 is an older design, but is most certainly still current. The links to the eBay sellers are the ones I used in the past few weeks.

I have another (higher current) board that I'm yet to test. I'd also like to get one of those sealed units to check the overall quality and efficiency.

 

[davrus]

Sorry David, my mistake - I read your sentence wrong, thinking that you had adjusted the board by replacing the trim pot - but what you actually are saying is that it is better compared to the first board, because it has a better trim pot. (So I did choose a good product )




"The adjustment on this board is via a better quality trim pot. Being multi-turn means it it much easy to adjust and should be more resistant to the set voltage drifting. "

 

[multicast]

As an aside, your stock 300W AC--->DC Switch mode PSU is actually mostly a DC/DC converter..


Yup! The AC is rectified first to a high voltage DC buss, which is then converted to your ( normally ) low voltage outputs..

 

[davrus]

Update on the Solar Fairy Lights.... I cut the wires going to the solar cell, and connected a DC-DC converter to the battery terminals.
The battery was rated at 1.2V, so that is where I started - the lights were not very bright. Increased to 4V and they were very bright, dropped back to 3V and they were bright enough, left it running for a while then pulled the plug and checked to see if anything was hot (not the case). Not even warm. It had been drawing less than an amp.
Considering that in a Light Show, these lights won't be on continuously for long periods, I'm guessing that it will be safe to run them at 3 to 4V

 

[David_AVD]

You could try bypassing all of the original electronics and just feed the LED string with the output of the DC-DC converter.

Start with the converter output set to 2V and wind it up slowly while keeping an eye on the LEDs. Stop when they are as bright as they were on the original solar electronics.

In general the solar sets run the LEDs at a reduced current to get a better run time from each day's charge. This means you can probably run them at a slightly higher brightness than before.

The voltage at which they run correctly will vary depending on the LED colour and whether or not there are resistors built into the LED string wiring.

 

[davrus]

I decided to try your suggestion, and hook the LEDs directly to the DC-DC converter.
First set worked fine. The controller for these had three settings - Off / On / Flashing, and the wires even had marking to indicate which one was positive. They came on at 2V and were bright at about 3V to 4V. (remember, the battery was 1.2V).
Next type of lights, the controller had two switches - On / Off and 1 / 2/ 3 ..... the 1/2/3 switch caused all on / odd on / even on. And it did this with just two wires. I wondered what would happen (was expecting to get no lights) - the wires had no marking, so made a random choice as to which one was +ve. Plugged the power, and the lights came on. Brilliant, I thought. But .... then I realised only half the lights were on. So I switched the wires around, so +ve went to the other wire, and bingo, the lights came on ... but this time it was the other lights !!!
so the puzzle ..... presumably the controller chose which way to send +ve by the 1/2/3 switch. But ..... how does it get all lights on at the same time !!!! Hmmmmm... could it be switching the +ve back and forth across the two wires ? But anyway, my DC-DC converter can't help with that, so I have a choice, half the lights, or use the controller ... or cut all the LEDs and rewire them with all the odd together, and the even on a separate string.


Moral of the story, if you want to do this, only buy the cheapest ones, with a simple on/off switch !! (But the Trade Me / Ebay ads don't tell you that.)


Second moral of the story - why did I even think of going down this road !!! Why didn't I just buy dumb RGB Leds!!!

 

[David_AVD]

The one that can do the odd/even LEDs is swapping the polarity to the LED string. When it's in the "all on" position it's simply doing the swap quicker than you can see so both sets appear to be steady on.

Directly feeding the LED string with a regulator output won't work with those strings as you've found out.

 

[David_AVD]

I've just tested one of the 5V @ 5A potted DC-DC converters and have updated my da-share page.

 

[multicast]

Interesting look at your converters David.. What i've noticed is that they all are pretty much Constant On Time setups with a diode.. The diode is the big power sucking element! 3A @ .4V for example ( assuming they use a Schoty ) its 1.2W..

 

[AussieDoug]

So I have decided to dive into the DC - DC convertors to run my hedge matrix. What I have is a 24v 26amp 600w PSU to 3 x 12v 25amp 300w buck convertors. These will be running a 720 pixels.

 

[AAH]

Just be a bit careful Doug. Assuming the buck converters are 90% efficient you'll only get 540W out of converters assuming the power supply is running at 100% and you have no cable losses. The 90% is possibly being generous but I'm assuming nice efficient ones as they are large. There will be cable losses of course but it will depend on whether the cable run is 1m or 10m obviously.

So I have decided to dive into the DC - DC convertors to run my hedge matrix. What I have is a 24v 26amp 600w PSU to 3 x 12v 25amp 300w buck convertors. These will be running a 720 pixels.

 

[David_AVD]

Another thing to be careful with is that when the input voltage to the DC-DC converter decreases (say due to voltage drop in the cabling), the input current increases.

If the input cabling is too thin, this can cause a snowball effect; more voltage drop -> more current draw -> even more voltage drop - even more current draw, etc.

The end result can be a input cabling overheating, a fuse blowing or the upstream power supply shutting down.

 

[multicast]

Another thing to be careful with is that when the input voltage to the DC-DC converter decreases (say due to voltage drop in the cabling), the input current increases.

If the input cabling is too thin, this can cause a snowball effect; more voltage drop -> more current draw -> even more voltage drop - even more current draw, etc.

The end result can be a input cabling overheating, a fuse blowing or the upstream power supply shutting down.

Any good setup will have a uvlo setup so should the voltage drop to far it will shutdown.

 

[David_AVD]

A lot of the converters we've been talking about here are specified to run with a 10V - 30V input. If you run a nominal 24V input, it's a long way from there down to 10V where it may cut out. Meanwhile, the input current has more than doubled.