1. Home
  2. Electronics Design
  3. LED driver
  4. Page 12

LED driver

Piglets run for cover when Titans clash.

piglet

Reply to
piglet
Loading thread data ...

As designed, C1 charges up to 27 volts before the entire mess turns on. It would work with a higher voltage LED.

One interesting static state is:

5 ua at (I)

Q1-Q2 latched, about 0.6 volts across the pair

100 millivolts dropped across R3

About 0.6 Vbe on Q4, microvolts lost in R4

Q4 on, but no collector current because...

Not enough voltage available to make significant current in D1

Not enough drop in R7 to turn on Q6

So it's stable in that state, PNPN ON but the LED is off.

It gets more interesting when (I) is ramped up from there. I think the PNPN will not unlatch.

--

John Larkin                  Highland Technology Inc 
www.highlandtechnology.com   jlarkin at highlandtechnology dot com    

Precision electronic instrumentation
Reply to
John Larkin

Why should you run for cover?... you got it right! Be proud that you didn't let BS overcome your judgment. ...Jim Thompson

-- | James E.Thompson | mens | | Analog Innovations | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | San Tan Valley, AZ 85142 Skype: skypeanalog | | | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | | E-mail Icon at

formatting link
| 1962 | I love to cook with wine. Sometimes I even put it in the food.

Reply to
Jim Thompson

Fantasy is not engineering. My updated post yesterday to my website did a sweep from 5uA. Did you bother to read it? I doubt it. You're frozen in your opinion. That's commonly defined as poor engineering.

Subthread terminated. ...Jim Thompson

--
| James E.Thompson                                 |    mens     | 
| Analog Innovations                               |     et      | 
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    | 
| San Tan Valley, AZ 85142     Skype: skypeanalog  |             | 
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  | 
| E-mail Icon at http://www.analog-innovations.com |    1962     | 
              
I love to cook with wine.     Sometimes I even put it in the food.
Reply to
Jim Thompson

Your sweep avoided entering the hangstate. That doesn't mean the state doesn't exist, or that the state couldn't be entered in real life. I don't want a safety-related circuit that has a failure latchup state.

PUT it into that state, and then slowly ramp up the current. Do you deny that the described state exists?

Post an LT Spice net list, so other people can play with it. All you post is PDFs that show the cases you want to show.

(How's that killfile project coming along?)

--

John Larkin                  Highland Technology Inc 
www.highlandtechnology.com   jlarkin at highlandtechnology dot com    

Precision electronic instrumentation
Reply to
John Larkin

Q3-4 can't latch, because of R4. R4 means Q3 needs .6Vbe/330R = 1.8mA to turn on. Since the supply is limited to 1 to avoid latching.

There may be a Vcc that meets that condition.

There are three interacting loops. Complicated.

Cheers, James

Reply to
dagmargoodboat

There are three distinct latch zones:

Low uA, Q1-Q2 on, neither Q4 nor Q6 has collector current

Mid current, D1 now conducting a little, Q6 still off

More current, Q6 now limiting the voltage across C1.

Something like that.

Yup. Too complicared.

--

John Larkin                  Highland Technology Inc 
www.highlandtechnology.com   jlarkin at highlandtechnology dot com    

Precision electronic instrumentation
Reply to
John Larkin

Jim,

Suppose Q1-2 is latched. What prevents it supplying enough current through R3->Q4>->Q6 to gobble all the available icc (thus charging current), leaving the blinker stuck?

Cheers, James Arthur

Reply to
dagmargoodboat

I can force that condition to exist, but haven't been able to cause it to happen unless manually forced to that state (switch Open on t=1, for example). So I still say it's hysteria. ...Jim Thompson

--
| James E.Thompson                                 |    mens     | 
| Analog Innovations                               |     et      | 
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    | 
| San Tan Valley, AZ 85142     Skype: skypeanalog  |             | 
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  | 
| E-mail Icon at http://www.analog-innovations.com |    1962     | 
              
I love to cook with wine.     Sometimes I even put it in the food.
Reply to
Jim Thompson

You can't guarantee that the circuit will never get into that state, but it only needs to do it once. Some interesting power supply woopie-do, or an EMI spike, could do it. Yeah, an EMI spike (I'm expecting kilovolts per ns not far away) could retrigger Q1-Q2 at just the right time. Once hung up, R7-Q6 makes a negative feedback loop that limits the max possible LED current to about 20 uA.

I don't like circuits that have hang states. It's safer to design them out, rather than assume they can't possibly be entered.

My mosfet schmitt is simple, has no hangs, and all the input current flows through the LED, with no shunt failure paths. It's the only circuit I've seen so far that looks bug-free.

--

John Larkin                  Highland Technology Inc 
www.highlandtechnology.com   jlarkin at highlandtechnology dot com    

Precision electronic instrumentation
Reply to
John Larkin

I don't think so, with the resistors across each Vbe.

You're so wonderful >:-}

But, IIRC, it doesn't meet your power budget?? ...Jim Thompson

--
| James E.Thompson                                 |    mens     | 
| Analog Innovations                               |     et      | 
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    | 
| San Tan Valley, AZ 85142     Skype: skypeanalog  |             | 
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  | 
| E-mail Icon at http://www.analog-innovations.com |    1962     | 
              
I love to cook with wine.     Sometimes I even put it in the food.
Reply to
Jim Thompson

~210us pulsewidth... is that visible? I guess so, the rate is basically continuously on. ...Jim Thompson

--
| James E.Thompson                                 |    mens     | 
| Analog Innovations                               |     et      | 
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    | 
| San Tan Valley, AZ 85142     Skype: skypeanalog  |             | 
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  | 
| E-mail Icon at http://www.analog-innovations.com |    1962     | 
              
I love to cook with wine.     Sometimes I even put it in the food.
Reply to
Jim Thompson

Your circuit has negative d.c. feedback via the MOSFET, and positive d.c. feedback via R6. If the MOSFET negative overwhelms the fairly highly-attenuated R6 feedback, the gadget hangs unless the phase delays save it (which they might).

A speed-up cap across R6 would avoid that.

My two-transistor circuit was a d.c. amplifier that's off below ~50V. Above 50Vcc, it self-biases linear, with positive a.c. feedback.

I don't see any hangs there if the a.c. feedback is over unity.

5Kv --- .---------------. | | R3 | .---.i1 | 2.2k | | |200uA | | | V | | | '---' | LED V ~~> | | --- +---------' | | +------------+-----. | | | | | R1 R2 |/ Q1 | | +-----47K---+---3.3k----| R6 | --- C3 | | |>. 2M --- 470nF --- C1 | | | | --- 2uF --- C2 '>| R5 | | | --- 470nF |----3.3k--+-----' | | /|Q2 | === | | | '--------------+ .---' Z1 | ^ 48v .-. | R4 | | | 220 '-' === | === 5kV Blinker Blink rate = 2.5Hz iLED(avg) = 120uA

Still needs the magic current source.

I've got a few analyses of ye olde traditional multivibrators. They're ideally two cross-coupled amps, d.c.-biased linear, with positive a.c. feedback. Some of the old bias schemes can saturate one transistor, its gain falls below unity, and the oscillator hangs. (Your preferred collector-to-base d.c. bias is more forgiving.)

It's amazing how complex two transistor circuits can be.

Cheers, James

Reply to
dagmargoodboat

And, with the initial post's requirement of ~1W, that is, 22Meg, it doesn't begin working until ~415V... thus the 3Meg, "I can do anything I want" bit >:-} ...Jim Thompson

--
| James E.Thompson                                 |    mens     | 
| Analog Innovations                               |     et      | 
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    | 
| San Tan Valley, AZ 85142     Skype: skypeanalog  |             | 
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  | 
| E-mail Icon at http://www.analog-innovations.com |    1962     | 
              
I love to cook with wine.     Sometimes I even put it in the food.
Reply to
Jim Thompson

All the LED and overhead current flows through R7. As the voltage drop across R7 hits 0.65 or so, Q6 turns on and diverts the current away from the LED. It's a stable negative feedback loop that limits the LED current.

No, I'm careful.

If I go with a 2KV supply, which I probably can, 3M is OK for the resistor. For

5KV, 25M resistor roughly, the mosfet circuit is easily scaled; just gotta buy more exotic high-meg resistors. Total available energy would be low, so I could use a 6V or 12V multichip LED to get a better flash. 
--

John Larkin                  Highland Technology Inc 
www.highlandtechnology.com   jlarkin at highlandtechnology dot com    

Precision electronic instrumentation
Reply to
John Larkin

A hard dump of 1 uF at 10 volts into a white Cree make a nice fat flash.

I've ordered some high voltage white LEDs to play with, too.

--

John Larkin                  Highland Technology Inc 
www.highlandtechnology.com   jlarkin at highlandtechnology dot com    

Precision electronic instrumentation
Reply to
John Larkin

Easily fixed.

--

John Larkin                  Highland Technology Inc 
www.highlandtechnology.com   jlarkin at highlandtechnology dot com    

Precision electronic instrumentation
Reply to
John Larkin

I think it's OK. The negative feedback discharges C4, which it's supposed to do. But when VK hits about 4 volts, M3 snaps off. Actually, delays keep it discharging down to near zero, but that's all good.

If you force-sweep the voltage at VK, and look at the current, it's very nice.

Yikes, that one is hard to think about.

Great, that means I don't have to think about it!

Yeah, that one has uglier waveforms but always oscillates. The "classic" mv can hang with both transistors saturated.

Exactly. Three is even worse.

--

John Larkin                  Highland Technology Inc 
www.highlandtechnology.com   jlarkin at highlandtechnology dot com    

Precision electronic instrumentation
Reply to
John Larkin

I was watching it in PSpice, "marching waveforms" in LTspice lingo... it does "hang" at one particular voltage where the current balance.

...Jim Thompson

--
| James E.Thompson                                 |    mens     | 
| Analog Innovations                               |     et      | 
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    | 
| San Tan Valley, AZ 85142     Skype: skypeanalog  |             | 
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  | 
| E-mail Icon at http://www.analog-innovations.com |    1962     | 
              
I love to cook with wine.     Sometimes I even put it in the food.
Reply to
Jim Thompson

Actually, it limits the LED current to zero.

--

John Larkin                  Highland Technology Inc 
www.highlandtechnology.com   jlarkin at highlandtechnology dot com    

Precision electronic instrumentation
Reply to
John Larkin

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.