# How can I scale amperage?

Discussion in 'General Electronics Discussion' started by kitt001, Jul 29, 2013.

1. ### kitt001

11
0
Jul 29, 2013
Hi All,

Let me preface this with the fact that I am a DIY kind of person, , but no electronics wizard ... I just learn this stuff as I go along; so I apologize if I don't ask this gracefully.

Anyway, I'm working with some LEDs and I have a variable voltage IC that goes from 0-5VDC at roughly 18.5ma max (per channel). I'd like to variably power larger LEDs which run at more than 10X the available amperage. So I need to transfer the voltage and provide a higher amperage to the LEDs without burning up the chips.

I've looked up voltage dividers, voltage pumps and other odds and ends, but since I've never attempted this, I'm having a hard time figuring out if I'm even on the right track ...

Any insight would be greatly appreciated.

2. ### BobK

7,682
1,688
Jan 5, 2010
Hi kitt, and welcome to the forum.

What is a "variable voltage IC?" Part number? Schematic?

What is needed to increase the current is a power supply capable of supplying the current. If the LEDs are to be controlled by an electronic signal, you will need a transistor to switch the higher currents.

Bob

3. ### kitt001

11
0
Jul 29, 2013
After continued searching, I think that so many simaler questions have been asked by people trying to make something out of nothing (more power with some magic circuit), and then abandoned threads because the answer is, 'that's not how it works'; I can't find anything particularly helpful.

So I should clarify that I'm not trying to make power where there wasn't any, I know that I will need to introduce the additional power somehow.

So what I envision is a circuit (or better yet a chip) with two inputs and one output:
in1) variable (0-5v) at 18.5ma constantish
in2) 5v source at perhaps 750ma
ou1) variable (0-5v (mimicing in1)) at 750ma

I have no idea what this would be called, and nothing I've found so far seems to quite fit.

4. ### kitt001

11
0
Jul 29, 2013
Hi Bob .. Thanks for the reply .. I meant to include the chip in the original post because I knew that might be helpful ... oops

The chip is a WS2811 and a datasheet can be cound here http://www.adafruit.com/datasheets/WS2811.pdf

5. ### BobK

7,682
1,688
Jan 5, 2010
It would not be called anything because your specifications do not make sense.

You can specify a voltage or a current but not both for a source. If the voltage is specifed, the current will depend on what load is connected to it. If you specify the current, the voltage must automatically adjust to maintain the current.

So, is the input a voltage from 0 to 5V or is it a current source at 18.5 ma?

Please tell what you are trying to do, without telling us about what kind circuit you think is needed.

Bob

6. ### kitt001

11
0
Jul 29, 2013
I apologize, I thought I was clearer in my original post. ... I'll try again.

I have the ws2811 which has a maximum output per channel of [email protected] (I assume the voltage varies as the outputs are indicated as constant current, but this may be incorrect, perhaps the outputs pulse to control brighness?)

I want to power 5v LEDs that require 750ma ... so somehow I need to get the increased amperage without sending the ws2811 up in smoke.

7. ### BobK

7,682
1,688
Jan 5, 2010
Okay, I have read the datasheet, not easy since the writer apparently does not speak English.

So this is an RBG LED controller that is controlled by a non-standard serial inteface and it runs the LEDs (or LED strings up to 12V) at a constant current when on. It varies the duty cycle to get 256 levels of brightness.

So what are you using to control the chips?

Bob

8. ### BobK

7,682
1,688
Jan 5, 2010
Also, how may of these are you stringing together?

Bob

9. ### kitt001

11
0
Jul 29, 2013
I'm using an Arduino to provide the serial signal to the chips, so my next assumption (please correct me) is that I want to insert something circuit-wise between the channel output (outr/outg/outb) and ground ... so the input side of the chip would be non-impacted.

10. ### kitt001

11
0
Jul 29, 2013
There will be a max of one LED per output channel, but as many as 200 chips (that's yet to be determined)

11. ### kitt001

11
0
Jul 29, 2013
And based on what you've already said, I'm guessing that "varying the duty cycle" is more of a pulsing (for lack of more technical analogy) and not a shift in the voltage? (which invalidates my original assumption)

Sorry, I'm much more of a software guy than I am a hardware guy .. at least at this level of hardware. I'm aware my limitations in this area, so I appreciate the help/education.

12. ### BobK

7,682
1,688
Jan 5, 2010
Great, now we are getting somewhere!

It is an interesting chip for stringing together RGB LEDS, but of course, has the 18.5mA limitation.

To drive higher power LEDs while still using this chip to control the PWM generation, you would just replace the R, G, and B LEDs with resistors and then use the outputs as a signal to control a MOSFET for the high power LEDs. The outputs would go low when the LED is supposed to be on and high when it is off. You would size the resistor to drop 5V at 18.5 mA, which would be 270 Ohms.

But this would waste a lot of power. I am not sure what the chip would do if you used a higher value resistor, say 10K that would drop 5V at only 0.5mA. If it works that way, then fine.

Since the signal is inverted, you would want to use a P channel MOSFET. Connected as follows:

Source to the power supply +
Drain to a current limiting resistor
Other end of resistor to anode of LED
cathode of LED to ground
gate of LED to the OUTR, OUTG, or OUTB with resistor to 5V.

The MOSFET must be a logic level device that will turn on fully at ~4.5V.

Bob

Last edited: Jul 29, 2013
13. ### kitt001

11
0
Jul 29, 2013
Is this what you mean? (part pumbers/resistor values may not be right, still need to research the parts, just making sure I'm on the right track)

#### Attached Files:

• ###### mpu.png
File size:
10.4 KB
Views:
140
Last edited: Jul 29, 2013
14. ### BobK

7,682
1,688
Jan 5, 2010
No, the resistor from ROUT needs to go to V+ so the chip thinks it is driving an LED. The gate can be connected to ROUT directly.

And the IRF9530 is not a logic level MOSFET. It will not turn on enough at 5V to use in this circuit.

Bob

15. ### kitt001

11
0
Jul 29, 2013
Like this?

Also, I knew the 9530 wasn't right, but that's the default type in the circuit lab drawing thing. On that topic though .. I can't seem to find a p-channel logic level mosfet that is not in a surface mount package ... and my soldering skills aren't really on par with that type ... any suggestions on how to find one? ... or is surface mount my only option?

File size:
10.5 KB
Views:
131
16. ### (*steve*)¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥdModerator

25,496
2,837
Jan 21, 2010

1) go here.
2) Select all the "P Channel" options under "FET type"
3) Select all the "Logic Level" options under "FET frature"
4) Select "Chassis mount" and "through hole" under "Mounting type" (you'll have to scroll to the right)
5) Click on Apply Filters

You now have 87 (or thereabouts) to choose from.

You can now select based on features like Id, Vds, etc.

Even a quick eyeball over the first page will indicate some options.

17. ### kitt001

11
0
Jul 29, 2013
Thanks Steve .. that was helpful, but left me with more questions; mosfet's have a lot of parameters to muddy the water.

So I'm pretty sure at this point that the 4.5v "fully on" that Bobk referred to is the Vgs value? is that correct?

If so, that leads me to question 2, as the highest Vgs value I can find is 3.5v, do I need to drop a resister in front of it? and at that point, does it make a difference what the Vgs is, as long as the resistor is sized properly and the rating is 3.5v or under?

Next question is that the Vgs has a power rating choice; 1mA or 250µA ... I'm leaning towards the 250µA, but besides thinking it's more efficiant, I don't really have a good reason. Any chance someone can enlighten me?

Thanks

18. ### BobK

7,682
1,688
Jan 5, 2010
You are completely misunderstanding the Vgs.

The datasheet will have a parameter Vgs threshold. This is the voltage at which the MOSFET just starts to turn on, and is totally useless in picking a device. The current you see is the source to drain current at Vgs threshold. 250uA is not going to power your load.

A logic level MOSFET is one that is fully turned on at about 4V or less. Normal MOSFETS require up to 10V to be turned on.

There is no problem putting a higher voltage than necessary on the gate as long as it does not exceed Vgs max, which is typically 20V.

What you need to do is pick a logic level MOSFET that can handle about 2X the current you need.

The other important parameter is the Rdson, which is the resistance when the MOSFET is fully on. These are inn the milliohm region, the smaller the better. When you look at a datasheet, the Rdson is stated at a particular gate voltage. That is the voltage at which the MOSFET is fully on (or nearly so, if they give more than one)

And finally, there is the gate capacitance or total gate charge. These are indicators of how much current will be needed to charge up the gate when a voltage is applied. The gate draws current only when charging, i.e. it acts like a capacitor. This parameter is really only important if you are switching the MOFSET at a high frequency, which it this application should not be case.

Bob

19. ### BobK

7,682
1,688
Jan 5, 2010
Deleted, had suggested an N channel device.

Bob

Last edited: Jul 31, 2013
20. ### kitt001

11
0
Jul 29, 2013
Thanks for the clarification Bob, I was thinking that Vgs was the 'all the way on' voltage. I did realize that the 250uA is not for the load, I was more thinking that's the power spent to hold the MOSFET open ... sort of like the coil power to a relay. Perhaps that's another incorrect understanding.

As it is ok to apply a higher voltage, is there an advantage to using a MOSFET with a lower Vgs? Just to ensure that it is all the way open and applying as little resistance on the load side as possible? ... or is it a given that logic level mosfets are all fully on 4v?