PDA

View Full Version : Aussie 595



Rocket
01-22-2012, 07:32 PM
This is my adaptation of the Olsen 595 that I use for direct control of 24V strings of LEDs.

I am sure it has been mentioned before that in Australia most if not all currently available lights are powered from a low voltage adapter, most of which are 24 volt.

I supply the 595 board from a single 24 volt D.C. supply that has sufficient capacity for all the strings attached. Using a single D.C. supply eliminates the need for having multiple power adapters plugged in. Because the whole board uses low voltage D.C. it is possible to control each string of LEDs with a FET driven directly from the output of the 595 ICs which eliminates the need & expense of SSRs.

I have made this version of the 595 board with 56 N-Channel FETs that switch low when on and 8 P-Channel FETs that switch high when on. My idea is to use the P-Channel FETs to enable banks of LEDs that are connected to the N-Channel FETs. For example I could have 8 arches connected in parallel that each consist of 56 channels and have the common anode from each arch connected to one of the P-Channels. It would then be possible to enable any or all of the arches to display the pattern being output on the 56 N-Channels. Effectively this will give me 448 channels (8 x 56).

LightUp
01-22-2012, 09:03 PM
Will your scheme also do dimming, or strictly ON/OFF as the original Olsen 595?
What will control the 8 P-channel FETs?

Rocket
01-22-2012, 10:55 PM
It is ON/OFF the same as the original Olsen 595. All I have done is replace the SSRs with FETs and mounted them on the same board as the 595s. A Ren-C should work fine for dimming. I am currently in the process of trying that now.

My thought was to have the P-channel FETs assigned as a channel in Vixen, when that channel is on it will enable the strings attached to it. Im not sure how this would work in Vixen 3 but I dont think it will be a problem.

djulien
01-23-2012, 10:36 PM
The bank-switching has interesting possibilities for other configurations also. Do you have a circuit diagram and part#s for the FETs?

don

David_AVD
01-24-2012, 05:42 AM
Have you tested the bank switching (P-Channel) outputs yet?

Rocket
01-24-2012, 07:28 AM
I dont have a circuit diagram in a suitable format for posting. Its the standard 595 configuration with the N-channel FETs driven directly from the output of the 595 and and the P-channel FETs driven via a ULN2803 to invert the gate input (they need a low signal on the gate to turn on). Its not clearly visable in the picture but the P-channel FETs have a pull up resistor attached to their gate to turn them off.
I dont think the choice of FET is critical as long as its ratings are sufficient for the load you wish to connect to it. I chose (mainly based on the price) BS170 N-channel FETs that are rated 500mA and FQP17P10 P-channel FETs that are rated at 16A.

I have tested the operation of all channels on the bench with a few short strings of LEDs and so far it works fine. I am yet to connect a Ren-C to it to test for dimming, the switching speed of the FETs may then need to be considered.

P. Short
01-24-2012, 01:20 PM
I think that you need to look at more than just the current ratings for the transistor. You need to look at the curves for the on-region characteristics and for the on-resistance variation to make sure that the part will actually pass the desired current as a function of the VGS voltage as well as making sure that the on-resistance isn't too high (as a function of the VGS). You should also look at the curves to see how these vary with temperature, and bear in mind that the curves are only 'typical' characteristics.

Bottom line, I'm not sure how well those BS170's would survive a 0.5A load with a VGS of 4.0 or 4.5V.

David_AVD
01-24-2012, 05:31 PM
Phil is right, the BS170 FET is a bad choice. From memory (I looked it up last time someone suggested it), the Ron spec is woeful.

alijutt8
04-02-2012, 09:36 AM
sounds good to me will come back soon and leave further comments