You'd scribble it on the side here, but there's not enough room in the margins :^)
Say, can you give us a hint? I bet 'we' can figure it out.
George H. (who will have to go look at the I-V curve of the lnd150 again.)
I'll draw it up and post it, but the thing I described devolves to a negative resistor.
-- John Larkin Highland Technology, Inc Science teaches us to doubt. Claude Bernard
DF file of the schematic.
Use a small series sense resistor between the LED and the FET source. Two resistors across the LED provide a tap for the gate to provide an appropria te bias to essentially reduce the voltage required on the sense resistor.
A transistor and three resistors compared to two identical transistors and two resistors. I prefer the minimum BOM approach with three line items and it seems to work better and is very insensitive to which transistors you u se. Why specify an additional part on the BOM when it will be used so seld om? Exactly the sort of boutique part I try to avoid unless it buys someth ing useful in the design.
-- Rick C. -+- Get 1,000 miles of free Supercharging -+- Tesla referral code - https://ts.la/richard11209
Huh.. OK.. I've got a vague memory of a capacitor feedback circuit that looked like a negative resistance. Does this gizmo oscillate? (It's like 'What's my line', Soupy Sales has the next question. :^)
George H.
PDF file of the schematic.
po resistors across the LED provide a tap for the gate to provide an appropr iate bias to essentially reduce the voltage required on the sense resistor.
d two resistors. I prefer the minimum BOM approach with three line items a nd it seems to work better and is very insensitive to which transistors you use. Why specify an additional part on the BOM when it will be used so se ldom? Exactly the sort of boutique part I try to avoid unless it buys some thing useful in the design.
Interesting. I ran this simulation and the LND150 seems to require a fair amount of head room on Vds. Rather than working through feedback like the PNP design I used, this design is really just biasing the gate to a level s o the characteristics of the FET control the current. Once Vds reaches abo ut 1V the current stabilizes. Below that level the change in voltage acros s the current sense resistor can't ever turn on the FET fully regardless of Vds. So the circuit really doesn't even need the sense resistor unless th e FETs vary enough that a fixed bias to the gate won't be accurate enough.
Trying it with 1 ohm for the sense resistor works pretty much the same. Eq ual value resistors dividing the LED voltage give 1 mA of current. So one transistor and two resistors of the same value.
LED_drive.asy
Version 4 SHEET 1 880 680 WIRE 320 16 208 16 WIRE 496 16 320 16 WIRE 208 48 208 16 WIRE 496 112 496 16 WIRE 160 128 96 128 WIRE 208 176 208 144 WIRE 208 208 208 176 WIRE 496 224 496 192 WIRE 208 320 208 288 WIRE 288 320 208 320 WIRE 336 320 288 320 WIRE 208 352 208 320 WIRE 336 432 336 320 WIRE 96 464 96 128 WIRE 144 464 96 464 WIRE 208 464 208 432 WIRE 208 464 144 464 WIRE 208 496 208 464 WIRE 208 592 208 576 WIRE 336 592 336 480 WIRE 336 592 208 592 WIRE 208 608 208 592 FLAG 208 608 0 FLAG 496 224 0 FLAG 288 320 LED FLAG 144 464 Vg FLAG 320 16 Vd FLAG 208 176 Vs SYMBOL res 192 192 R0 SYMATTR InstName R1 SYMATTR Value 300 SYMBOL res 192 336 R0 SYMATTR InstName R2 SYMATTR Value 70K SYMBOL res 192 480 R0 SYMATTR InstName R3 SYMATTR Value 150K SYMBOL AutoGenerated\\TLDR5800 336 480 R90 WINDOW 0 2 -36 VTop 2 WINDOW 3 -42 -60 VBottom 2 SYMATTR InstName D1 SYMBOL voltage 496 96 R0 WINDOW 123 0 0 Left 0 WINDOW 39 0 0 Left 0 SYMATTR InstName V1 SYMATTR Value PULSE(0 5 500ms 1 1 500ms 3) SYMBOL LND150 192 96 R0 SYMATTR InstName U1 TEXT 544 312 Left 2 !.tran 4
LND150.asy
Version 4 SymbolType BLOCK LINE Normal 16 0 16 48 LINE Normal 8 -4 -16 0 LINE Normal 8 4 8 -4 LINE Normal -16 0 8 4 LINE Normal 16 0 8 0 LINE Normal -16 32 16 32 LINE Normal -16 40 -16 -40 LINE Normal 16 -32 16 -48 LINE Normal -16 -32 16 -32 LINE Normal -24 32 -32 32 LINE Normal -24 -32 -24 32 WINDOW 0 25 -16 Left 2 WINDOW 3 25 16 Left 2 SYMATTR Value LND150 SYMATTR Prefix X SYMATTR ModelFile LND150_SPICE.lib PIN 16 -48 NONE 8 PINATTR PinName D PINATTR SpiceOrder 1 PIN -32 32 NONE 8 PINATTR PinName G PINATTR SpiceOrder 2 PIN 16 48 NONE 8 PINATTR PinName S PINATTR SpiceOrder 3
LND150_SPICE.lib
.SUBCKT LND150 1 2 3
LED_drive.plt
[Transient Analysis] { Npanes: 1 { traces: 9 {524290,0,"V(vg)"} {524291,0,"V(vs)"} {524292,0,"V(led)"} { 524294,0,"V(vd)"} {589836,0,"V(Vs,LED)"} {589833,0,"V(Vs,Vg)"} {524295,0,"V (Vd,Vs)"} {524296,0,"V(Vd,LED)"} {68157445,1,"Ix(D1:1)"} X: (' ',1,0,0.4,3.5) Y[0]: (' ',1,-0.5,0.5,5) Y[1]: ('m',1,-0.0001,0.0001,0.001) Volts: (' ',0,0,1,-0.5,0.5,5) Amps: ('m',0,0,1,-0.0001,0.0001,0.001) Log: 0 0 0 } }-- Rick C. -++ Get 1,000 miles of free Supercharging -++ Tesla referral code - https://ts.la/richard11209
This is it:
It only needs three parts, but negative resistors are perverse, so it might not start up. So it needs R2A or R2B or something to kick it off.
It only loses millivolts.
-- John Larkin Highland Technology, Inc Science teaches us to doubt. Claude Bernard
e:
a PDF file of the schematic.
rop
s,
Yes, it is very low drop out and the current will be fairly constant. But it misses on one important mark. It doesn't set the current very well. Th e current is
Id = hfe(Q2) * (Vd - Vbe(Q1)) / R1
While Vd and Vbe will be fairly predictable, hfe(Q2) has a range of nearly
2:1. Design the circuit for 1 mA through the LED and it can vary almost fr om 0.75 mA to 1.5 mA without regard to any other conditions.The issue is not so much that the current sense resistor needs to be out of the diode current path, but that it needs to be compared to a lower voltag e than Vbe. A Schottky diode can be added between the LED cathode and the emitter of Q1. A small current sense resistor can be added between the cat hode and the base. This will offset the Vbe voltage so the sense resistor will only see the difference ~0.25V. Then not only will the current not v ary with the applied voltage, it can be set by the value of the current sen se resistor. That would be an improvement over the circuit I'm currently u sing only requiring a diode be added. Still, I don't think it is worth it. The two transistor + 2 resistor circuit I'm using only starts to lower th e current when the applied voltage gets down to 2.2 volts.
-- Rick C. +-- Get 1,000 miles of free Supercharging +-- Tesla referral code - https://ts.la/richard11209
Damned ingenious Sir! Violates classical teaching about relying on beta variation but modern more tightly specified devices mean you could get away with it sometimes.
piglet
Yes, breaking old-fart holy rules is worth doing for its own sake, if just to hear them squeal. Nobody will ship it back for a refund if an LED is a little brighter or dimmer than they expected.
There are a few more precise opamp versions of the self-reference LED thing, but they are boring.
Beta limiting can be good, like for instance when you have a 3.3 volt supply and want to stuff some current into a 3 volt bandgap.
I do often put an LED in series with a depletion fet running at Idss. It stays on at constant brightness until the supply caps are almost totally discharged, then winks out.
-- John Larkin Highland Technology, Inc Science teaches us to doubt. Claude Bernard
If no one will return the product for the LED varying in brightness, the ultimate circuit is a resistor. One part, not so critical specs, easy to buy. Win-win-win.
-- Rick C. +-+ Get 1,000 miles of free Supercharging +-+ Tesla referral code - https://ts.la/richard11209
Thanks! George H.
I still have heaps of germanium transistors. I'm looking for uses for them, and this might be one of them. For once the nice leakage current can be put to good use, such that R2A can be missed. Moreover the saturation voltage is even less.
Groetjes Albert
-- This is the first day of the end of your life. It may not kill you, but it does make your weaker. If you can't beat them, too bad. albert@spe&ar&c.xs4all.nl &=n http://home.hccnet.nl/a.w.m.van.der.horst
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.