PDA

View Full Version : cheap transistors to switch dc leds



smartalec
02-20-2010, 11:35 PM
hi all
i was reading my circuit design from 10yrs ago, an this is how i was switching my leds an incan lights on a dc source.

i think i copyed the circuit correctly.
http://doityourselfchristmas.com/forums/picture.php?albumid=138&pictureid=700

will this work, im guessing it would.. i just cant remember how much current i was drawing from it..
i know i did'nt have heatsinks an some started to fry with the heat.

the main reason i was thinking transistors is its cheaper than the dcssr's an can handle more current than the ULN series of IC's.

could this be used to control them icle lights people are talking about?
thanks alec.
http://doityourselfchristmas.com/forums/picture.php?albumid=138&pictureid=700

P. Short
02-21-2010, 12:09 AM
Did you forget to include an attachment or link?

smartalec
02-21-2010, 12:32 AM
Did you forget to include an attachment or link?

sorry i thought the pic displayed in my message.
here's the link then
http://doityourselfchristmas.com/forums/picture.php?albumid=138&pictureid=700

if i remember correctly the BD139 can handle up to 80v at 1.5amp, if im correct.

smartalec
02-21-2010, 03:12 AM
here's the pcb view of the board
http://doityourselfchristmas.com/forums/picture.php?albumid=138&pictureid=701

Matt_Edwards
02-21-2010, 05:16 AM
In fact the SOA is a BD139 would limit the DC current to about 150mA @ 40v
using PWM will allow higher currents.

I think you are missing base Resistors for Q1 -Q4.

How about the Darlington TIP120 that is capable of 1Amp @ 40v? only 69 cents each too @futurlec.com.au

smartalec
02-21-2010, 07:39 AM
In fact the SOA is a BD139 would limit the DC current to about 150mA @ 40v
using PWM will allow higher currents.

I think you are missing base Resistors for Q1 -Q4.

How about the Darlington TIP120 that is capable of 1Amp @ 40v? only 69 cents each too @futurlec.com.au

i did orginaly use TIP41C but went for the bd139 cause it was common

an i looked over my circuit diagram an it looks like i ran the bc639 straight off the 4017 chip output.

so would such a design work for the renard64?, since the ULN's seem like they could be too low power.

P. Short
02-21-2010, 12:44 PM
The schematic looks a little bit funny, it might be worth taking a closer look at it. For example, the collectors of the bd139 transistors are connected to ground. And I'm more used to seeing the load connected to the collector of a transistor in this sort of circuit.

Just by 2 cents...

Entropy
02-21-2010, 01:01 PM
Lack of base resistors is likely to cause your first resistors in the chain to fry.

Might want to consider wiring the two transistors per channel as a Darlington pair - then you won't need the +5V and would just need base resistors for the first resistors in the chain.

P. Short
02-21-2010, 04:52 PM
Wiring the transistors up as a darlington pair will more or less double the saturation voltage, therefore double the power dissipation (loss). Since +5V is available on the normal SSR input connectors, that shouldn't be a problem.

smartalec
02-21-2010, 05:31 PM
i remember the main reason we used the bc639 before the bd139, is cause when the outputs started flashing fast, the output voltage would drop by heaps.
please remember the bd139 is a subsutute for the TIP41c, which ended up hard to source.
the reason for switching the GND is apparently its easer than switching the positive, so i was told.
the pic is my orginal chaser, where i got the transistor idea from.

n1ist
02-21-2010, 08:02 PM
The question isn't about switching the low side, it is about the transistors being upside down...
Redrawing the schematic in the more usual orientation may make the issue clearer.

smartalec
02-21-2010, 08:29 PM
The question isn't about switching the low side, it is about the transistors being upside down...
Redrawing the schematic in the more usual orientation may make the issue clearer.

did i stuff up an read the pin outs for the transistors wrongly..
bugger, i'll have to check it out.

does you circuit work, or was that the way mine is wired..

thanks for pointing that out.
alec.

n1ist
02-21-2010, 09:04 PM
Mine was just yours redrawn in a more traditional layout. This is what I usually use for a light driver, though I am more likely to use an n-fet due to the lower drop.
/mike

smartalec
02-21-2010, 09:27 PM
Arr thanks for that, the reason 2 transistors were used, is cause when i was chasing the lights fast, there would be dimming on all.
so now this circuit should be correct then?

edit = added board


Mine was just yours redrawn in a more traditional layout. This is what I usually use for a light driver, though I am more likely to use an n-fet due to the lower drop.
/mike

dlovely
02-21-2010, 09:49 PM
Arr thanks for that, the reason 2 transistors were used, is cause when i was chasing the lights fast, there would be dimming on all.
so now this circuit should be correct then?

Diagram says +12v - +40v, will this handle 5v also for the Super Strips? Also what can this handle per channel?

Seems this is designed for the same results I was looking for. No need for many designs of the same thing so long it can handle 5v @ 450ma / channel.

smartalec
02-22-2010, 01:52 AM
it should also work at 5v, an its around 1amp@30v output, so maybe 1.5amp at 5v, have'nt tryed it on 5v only 12 as a min.

Entropy
02-24-2010, 09:14 PM
Arr thanks for that, the reason 2 transistors were used, is cause when i was chasing the lights fast, there would be dimming on all.
so now this circuit should be correct then?

edit = added board

That sounds more like a power supply droop or unexpected interaction between channels (drawing too much current from the controller for instance).

The problems I see with the current design:

The is no base resistor for the first transistor per channel in your circuit. The base resistor for the second one will limit total current through this transistor, but the current distribution (how much comes from +5v and how much from the controller) will be semi-unpredictable.

Also, the emitter and collector of your power transistor are reversed. This will sort of work, but not optimally. Typically bipolar transistors use asymmetric doping and layer thicknesses so that they won't work as designed if you swap emitter and collector.

You are correct in that it is usually easier to sink current with transistors than to source it - especially if you want to control a high voltage with a lower voltage. You are wiring your load and power transistors as a current sink - it's just the power transistors are reversed.

smartalec
02-25-2010, 04:21 AM
what value resister would you reckon i would need for the first transistor, when its going to be connected to the renard 64's.

power supply not handling the load, (yer theats possable) it fried a 100v cap i was trying to use an a filter on the DC..

this year i have made an brought a rock solid 24v 20amp suppy, a computer PSU for the +5v, an a scrappy 6.3v tranformer to do the ZC, all housed in a wooden box with AMP an Volt meters, so i know whats going on.
the way i thought the circuit worked was i have a transistor switching a higher capacity transistor, then switch's the Gnd supply of the load.

i had the circuit designed for me to run off 4017 counter ic's, to switch like 1.5a an 2amp loads, lost a couple of power transistors cause of no heatsink with a 50c roomtemp as well to cook them nicely..
i ran DC to all my AC lights, no probs.


That sounds more like a power supply droop or unexpected interaction between channels (drawing too much current from the controller for instance).

The problems I see with the current design:

The is no base resistor for the first transistor per channel in your circuit. The base resistor for the second one will limit total current through this transistor, but the current distribution (how much comes from +5v and how much from the controller) will be semi-unpredictable.

Also, the emitter and collector of your power transistor are reversed. This will sort of work, but not optimally. Typically bipolar transistors use asymmetric doping and layer thicknesses so that they won't work as designed if you swap emitter and collector.

You are correct in that it is usually easier to sink current with transistors than to source it - especially if you want to control a high voltage with a lower voltage. You are wiring your load and power transistors as a current sink - it's just the power transistors are reversed.

Entropy
02-25-2010, 12:29 PM
What's the ZC for???

smartalec
02-25-2010, 04:38 PM
What's the ZC for???
zero crossing, so it knows when to start dimming the load,

budude
02-25-2010, 05:10 PM
zero crossing, so it knows when to start dimming the load,

This is a DC control device correct? If so, you just need to configure the controller for PWM - no ZC is required.

Entropy
02-25-2010, 05:22 PM
zero crossing, so it knows when to start dimming the load,

I thought you were dimming DC???

smartalec
02-26-2010, 05:51 AM
im am dimming dc.
just got the renards working off the computer power supply an a little 6.3v AC tranformer to do the ZC zero crossing, would'nt work with out it.
i got the pic's pre-programmed, no pic programmer as of yet..

so within the next couple of weeks, i'll be all sorts of dc switching/dimming for leds

Entropy
02-26-2010, 10:20 PM
I'm not an expert on Renards, but you should never need a ZC signal when dimming DC. (I think for Renards this means using slightly different firmware when dimming DC?)

The emitter of your power transistors should go to ground, and the collector to your load. You have it the other way.

As to resistors for the other transistors........

It may be best to eliminate the resistor between the emitter of the first transistor and the base of the power transistor. Instead, have:
One resistor to set base current to the first transistor
One resistor from +5v to the collector of the first transistor to set main current through that transistor (which will then consist of the majority of the base current through the power transistor)

I don't have datasheets for the transistors with me so I can't quite calculate the exact resistor values.

Even though a Darlington pair configuration might be *slightly* more power efficient, for any reasonable load, the simplicity of a Darlington pair will offset the power dissipation gains...

P. Short
02-26-2010, 10:30 PM
You can set a compile-time option in the Renard firmware so that you can omit the ZC signal (set the so-called DC_build option to 1). You must provide a ZC signal if you do not set that option to '1', and the pwm signal will by synchronized to the ZC signal. In this latter case you can control either a DC or AC load, as determined by the SSR.

P. Short
02-26-2010, 11:34 PM
It's a bit late in the process, but the STN1360 transistors look interesting. They appear to have a high gain and low saturation voltage to the point where a single transistor stage would be feasible up to perhaps 1/2A output current.

smartalec
02-27-2010, 04:57 AM
You can set a compile-time option in the Renard firmware so that you can omit the ZC signal (set the so-called DC_build option to 1). You must provide a ZC signal if you do not set that option to '1', and the pwm signal will by synchronized to the ZC signal. In this latter case you can control either a DC or AC load, as determined by the SSR.

i went the easy way an got the pic's pre programmed from the supplyer of the renard64 boards.
easer than buying something an learning another thing.
plus he was aussie, so i like buying in australia

smartalec
02-27-2010, 05:54 AM
It's a bit late in the process, but the STN1360 transistors look interesting. They appear to have a high gain and low saturation voltage to the point where a single transistor stage would be feasible up to perhaps 1/2A output current.

any chance you could scribble out a diagram that would work?:D