Pete,kitbots wrote:Thanks for the reply.
The info. for the constant current regulator says it will regulate both current and voltage and I will be supplying it with 5v from a regulated power supply so how could the voltage ever go higher? Is the microprocessor really regulating the current when its powered through the USB or is it simply the regulators in the laptop that is limiting the current? If the latter then surely the constant current constant voltage device would do the same but simply at a higher level. I certainly do not see the brightness change as the amount of text goes up and down when powered through the USB.
To answer your first question... When you are powering from USB, notice that:
- there's no direct power connection between USB and DMD power rails
- the lights go out when too many LEDs are lit at once
Your second answer is:
The laptop regulators will quite happily supply much more current than the USB port will source, which is limited to 0.5Amps. So if any device develops a short circuit, the USB driver chip will sense the overcurrent and rapidly disconnect the USB line from the device, again it's an inbuilt protection method.
Yes, I agree that you don't see a brightness change so let's look at the science by analysing the circuit in depth.
Signal R selects pin 12 of every 595 in the system, which is the stop pin so it has nothing to do with the data control system. It controls the entire board state, either ON or OFF and overrides the nOE pin so it's just used for blanking.
The data could be such that all 512 LEDs appear to be on at once, but is this really so?
Signals A and B select one of the four quadrants to be addressed, a maximum of 128 LEDS. The two lines are decoded by the 138, which is set up as a 2 - 4 line decoder. These outputs each feed two 4952 row driver chips, which further restrict the output to two rows so that only 32 LEDS can be addressed at any one time.
Signal nOE is also connected to every 595 on pin 13. It is an 8 bit wide tristate latch and current sink, which latches the data on the 595 and sinks the "LED Data Bus" for want of a better description.
This "LED data bus" is a matrix of 32 LEDS powered by the 4952 row driver chips and because they are latches the outputs are ON or OFF until instructed otherwise. Each of the 595s 8 outputs drives 4 LEDS but only one will light because each of the 4 LEDs is powered by a different row. Moreover, only one of the two possible rows are selected which means that it is possible to address a single LED, which we all know is true.
Does this mean that only one LED per board is switched on at any one time? No, if it were so then the USB port could deliver enough current to power 25 LEDs at full brightness, provided that no other USB device was connected.
It is latching that's important, not addressing. Each latch can sink up to 8 pixels concurrently and since there are 4 latches per quadrant then 32 pixels per quadrant are possible and there being 4 quadrants there are a maximum of 128 pixels that can be lit at any one time, per board.
If each LED is at full brightness and the max current that a single LED can handle comfortably is 20ma then the max current setting will be 128 pixels x 20ma = 2.56 Amps per board.
In summary, there are several important points to take away.
- That it requires a serious amount of energy to drive large arrays
- That each of the LEDs has a constant current sink and thats why you don't see a brightness change.
- That the existing current sink is not variable and therefore not useful as a control unless we could somehow control the length of time that the 595s were latched in the ON state
Preregulating a current source defeats the purpose and it has never worked, in my experience. The idea is to allow the voltage to float so that a low current drain on the supply lets the voltage rise such that at virtually 0 Amps the voltage reaches peak and it sags as current rises.
You can, however, post regulate the voltage by using avalanche techniques so that the voltage that the device sees is never greater than its capacity to handle it. As you can imagine, making such a current source, while possible is not an easy task and one that few would tackle, especially when there are easier methods of getting the job done.
This might be an answer to a maiden's prayer for many people...
I was having a discussion with Stryker last night about controlling brightness on the nOE port. It appears that there may be a positive outcome from his work, within the next few days.
See this thread viewtopic.php?f=26&t=3435