Ackerman Agreement Current Limiter

There seems to be some interest in the LED current limiter I needed for a r ecent design. There is another current limiter to charge the supercap that powers the LED and sounder. I thought I'd paste it here for those who mig ht be interested. It was a post from piglet that gave me the idea. I'm su re it's an obvious circuit to many here, but I'm impressed with how well it works. As the supercap charges from zero it holds the current rock steady until the limit of drop out voltage is reached at less than 0.6 V. It's a hard circuit to beat for the simplicity. It's more sensitive to the incom ing power supply rail, but that's ok in this design. Again, I'm just shoot ing for something better than a resistor.

I used a FET with a model from the vendor, so the symbol is autogenerated a nd placed in an autogenerated directory. You may have to tweak the symbol to locate the model file since it isn't on your hard drive... er, I mean my hard drive.

Comments welcome.

formatting link

If you don't want to futz with the LTspice simulation there is a PDF file o f the schematic.

formatting link

Bonus points if you can tell me what the Ackerman agreement is from.

--
  Rick C. 

  - Get 1,000 miles of free Supercharging 
 Click to see the full signature
Reply to
Ricketty C
Loading thread data ...

BTW, I discovered a setting in the LTspice control panel for the line width used in the drawings and also the waveform display. The default 1 setting can be virtually impossible to see in a PDF. Even on the display a 2 is a significant improvement. I'm currently using 3.

--
  Rick C. 

  + Get 1,000 miles of free Supercharging 
 Click to see the full signature
Reply to
Ricketty C

I think my circuit was not so much a current regulator as a switch that permitted unlimited load current after the initial capacitor charge-up surge. But glad my contribution helped you.

Art of Electronics has no reference to Akerman circuits? :)

piglet

Reply to
piglet

a recent design. There is another current limiter to charge the supercap that powers the LED and sounder. I thought I'd paste it here for those who might be interested. It was a post from piglet that gave me the idea. I' m sure it's an obvious circuit to many here, but I'm impressed with how wel l it works. As the supercap charges from zero it holds the current rock st eady until the limit of drop out voltage is reached at less than 0.6 V. It 's a hard circuit to beat for the simplicity. It's more sensitive to the i ncoming power supply rail, but that's ok in this design. Again, I'm just s hooting for something better than a resistor.

ed and placed in an autogenerated directory. You may have to tweak the sym bol to locate the model file since it isn't on your hard drive... er, I mea n my hard drive.

le of the schematic.

Yes, they were different circuits with different purposes, but in trying to understand your circuit the idea that it could be modified to produce a cu rrent limited occurred to me.

There is no Akerman circuit. This was something in an (old) movie and at s ome point I wrote it into the settings for printing PDF files. So it shows up as the embedded title.

It's funny that the "Akerman agreement" was a big deal in the movie and men tioned more than once, but Google searches don't reveal its nature. Maybe I'm not spelling it correctly.

I've been simulating alternative components in LTspice and found that the t ool has a large number of issues. The most recent issue was trying to swap models using a copy of a previous component symbol. I changed every acces sible mention of the previous device and got an error saying it can't find "the old component name". I went through every step and could not find the error. Everything pointed to the new model file... everything I could se e.

The pointer to the model file is in the symbol. I edited the symbol by cli cking through to it from the schematic where it is used. That updates the file, but not the schematic. I even created a small test design, but copie d the part from the larger schematic, so the fault remained! In the end th e helpful guys from the LTspice group told me to check the netlist. They c ertainly don't make it easy to reuse symbols.

LTspice is not a tool for the faint of heart.

--
  Rick C. 

  -- Get 1,000 miles of free Supercharging 
 Click to see the full signature
Reply to
Ricketty C

"My God, it's full of parts!"

Why not a depletion-mode MOSFET with a resistor?

Best regards, Piotr

Reply to
Piotr Wyderski

I don't think he can afford the ~1 V drop of the d-fet. (I hear you responding that the above thing has a Vbe drop.. which I guess he can afford.) I'm not sure about the Schottky diode. Is that to reduce temperature effects?

George H.

Reply to
George Herold

le of the schematic.

What's your circuit and what are the specs? In particular, what is the dro p out voltage? That might be a viable solution, but I like to keep my BOM minimal and every one of these parts are used elsewhere in the design. Rip ping out the circuit in its entirety won't reduce the BOM at all.

I can't even find a P channel depletion mode MOSFET and an N channel part r equires a series resistor between the part and ground making the supercap i solated. Having separate grounds for the power source and the alarm circui t is not advisable since the alarm has a signal passing to the MCU for the charge state of the supercap.

I just don't think a depletion mode device is practical in this circuit.

--
  Rick C. 

  -+ Get 1,000 miles of free Supercharging 
 Click to see the full signature
Reply to
Ricketty C

file of the schematic.

The diode is to prevent back current. By putting in the Vsense loop its vo ltage drop is not added to the Vbe which is less than 0.6 volts. At that p oint the cap is pretty well charged and the current can taper off without p roblem. I don't know that 1 volt would be enough for the depletion mode de vice, but I see no reason to add another part to the BOM just to save using one transistor and a two resistors. Every part in this circuit would eith er remain or still be used elsewhere in the circuit. The real killer is th e lack of p-channel depletion mode devices.

--
  Rick C. 

  +- Get 1,000 miles of free Supercharging 
 Click to see the full signature
Reply to
Ricketty C

formatting link

See "Constant current diode"

At low voltage, probably worse than that of your circuit. At higher it would be just right to maintain your current. What are *your* specs?

This is one constraint.

Why would you need a P-channel DMOS?

You asked for comments and here is one. It is practical, yet may not be useful.

Best regards, Piotr

Reply to
Piotr Wyderski

Vbe + the Schottky and the sim say that Vin=5.25V, which gives 2.75V of headroom. This is a lot.

Best regards, Piotr

Reply to
Piotr Wyderski

You didn't specify that constraint in your initial message, but said "It's a hard circuit to beat for the simplicity."

It is actually not hard at all, "my" circuit is 3x simpler. I am not counting the diode, as you may want it to prevent the current flowing backwards.

Why would you need one? This constant current diode is a floating

2-terminal device. Feel free to use it on either rail. It would work equally well with a P nad N DMOS, as well as with a P and N JFET, which are easily available.

Best regards, Piotr

Reply to
Piotr Wyderski

every one of these parts are used elsewhere in the design. Ripping out the circuit in its entirety won't reduce the BOM at all.

rt requires a series resistor between the part and ground making the superc ap isolated.

.

I think you can see from the circuit that the intent is to charge the super cap as fast as possible without an initial excessive current. Maybe that i s not clear. I'm used to seeing this in context.

Yes, I asked for comments and I did consider depletion mode devices and I d on't see where it is better in any useful way. The parts I found have the several problems I outlined. Just saying, "Why not a depletion-mode MOSFET with a resistor?" doesn't tell me what you had in mind. I'm picturing a c ircuit that performs poorly compared to this simple two transistor circuit. There is also a soft start feature which I expect to be very useful.

Is there something about the depletion mode device that is better? Am I mis sing something about it?

BTW, telling me to read may be a "comment", but not an overly useful commen t.

I had the same problem only worse trying to drive the LED. That circuit se ems to be pretty optimal too, using two PNP transistors and three resistors . I suppose a depletion mode FET and a resistor might be fewer parts, but it would not control the current as well over the voltage range required.

That's not criticism, that's just feedback.

--
  Rick C. 

  ++ Get 1,000 miles of free Supercharging 
 Click to see the full signature
Reply to
Ricketty C

This is a supercap charger, not a NIST reference 1A standard. And you are wrong, the CSD based on a D-MOS is very stable and predictable. Actually, I am using it as an active bleeder for a 900V capacitor.

You have asked for simplicity and here it is. 2 parts, or 3 if you want to keep the diode. You have 6, not counting the diode and the output capacitor.

You have asked for a diagram and there is the required diagram. What is missing to make you consider this picture useful?

formatting link

This is not true, it works well at least to kV range. With cascoding you can go as high as you wish. At low voltage it is equivalent to two resistors connected in series, the DMOS is fully open there.

No problem, but here is mine.

Best regards, Piotr

Reply to
Piotr Wyderski

Ricketty C wrote:

I don't have a good model for a low-resistance DMOS, so used a standard LTSpice JFET instead and scaled the components to match your time constants. It works excellent, given its simplicity.

If your charging current requirements match I_DS_SAT of some cheap D-MOSFET, which wouldn't be particularly hard given the application, you can even further relax the regulation quality and remove the resistor.

I am using DN2530N8-G instead of PolyFuses, as it does not wear off.

Best regards, Piotr

Version 4 SHEET 1 1320 680 WIRE 144 -400 112 -400 WIRE 224 -400 208 -400 WIRE 336 -400 320 -400 WIRE 432 -400 416 -400 WIRE 448 -400 432 -400 WIRE 448 -352 448 -400 WIRE 288 -336 288 -352 WIRE 432 -336 432 -400 WIRE 432 -336 288 -336 WIRE -48 112 -80 112 WIRE 112 112 112 -400 WIRE 112 112 32 112 WIRE 144 112 112 112 WIRE 176 112 144 112 WIRE 192 112 176 112 WIRE 288 112 256 112 WIRE 320 112 288 112 WIRE 448 112 400 112 WIRE 496 112 448 112 WIRE 576 112 496 112 WIRE 704 112 672 112 WIRE 736 112 704 112 WIRE 768 112 736 112 WIRE 736 128 736 112 WIRE 448 144 448 112 WIRE -80 160 -80 112 WIRE 144 208 144 112 WIRE 320 208 144 208 WIRE 400 208 320 208 WIRE 544 208 496 208 WIRE 592 208 592 160 WIRE 592 208 544 208 WIRE 736 208 736 192 WIRE -80 256 -80 240 WIRE 320 272 320 208 WIRE 416 272 320 272 WIRE 592 272 592 208 WIRE 592 272 496 272 WIRE 320 336 320 272 WIRE 416 336 320 336 WIRE 592 336 592 272 WIRE 592 336 480 336 WIRE 592 368 592 336 WIRE 592 464 592 448 FLAG 704 112 Vbb FLAG 736 208 0 FLAG 176 112 Vin FLAG -80 256 0 FLAG 544 208 M1g FLAG 496 112 M1s FLAG 592 464 0 FLAG 288 112 Vs FLAG 448 -288 0 SYMBOL voltage -80 144 R0 WINDOW 3 -66 230 Left 2 WINDOW 123 0 0 Left 0 WINDOW 39 37 74 Left 2 SYMATTR Value PULSE(0 5 10ms 20ms 50ms 62s 100s) SYMATTR SpiceLine Rser=0 SYMATTR InstName V1 SYMBOL res -64 128 R270 WINDOW 0 32 56 VTop 2 WINDOW 3 0 56 VBottom 2 SYMATTR InstName R99 SYMATTR Value 0.01 SYMBOL polcap 720 128 R0 WINDOW 3 24 56 Left 2 SYMATTR Value 2.5 SYMATTR InstName C1 SYMATTR SpiceLine V=10 Irms=255m Rser=0.36 Lser=0 mfg="Nichicon" pn="UPR1A331MPH" type="Al electrolytic" SYMATTR Description Capacitor SYMATTR Type cap SYMBOL res 304 128 R270 WINDOW 0 32 56 VTop 2 WINDOW 3 0 56 VBottom 2 SYMATTR InstName R1 SYMATTR Value 1.2 SYMBOL res 576 352 R0 WINDOW 0 35 49 Left 2 WINDOW 3 33 79 Left 2 SYMATTR InstName R3 SYMATTR Value 100K SYMBOL pnp 496 144 R90 WINDOW 0 35 62 VLeft 2 WINDOW 3 67 9 VRight 2 SYMATTR InstName Q1 SYMATTR Value 2N2907 SYMBOL cap 480 320 R90 WINDOW 0 0 64 VBottom 2 WINDOW 3 0 -3 VBottom 2 SYMATTR InstName C2

SYMBOL res 512 256 R90 WINDOW 0 2 106 VBottom 2 WINDOW 3 2 44 VBottom 2 SYMATTR InstName R2 SYMATTR Value 1.6Meg SYMBOL schottky 192 128 R270 WINDOW 0 33 35 VTop 2 WINDOW 3 0 42 VBottom 2 SYMATTR InstName D1 SYMATTR Value PMEG4010BEA SYMATTR Description Diode SYMATTR Type diode SYMBOL PMPB13UP 592 144 M270 WINDOW 0 45 -24 VLeft 2 WINDOW 3 69 23 VLeft 2 SYMATTR InstName M1 SYMBOL schottky 144 -384 R270 WINDOW 0 33 35 VTop 2 WINDOW 3 89 37 VBottom 2 SYMATTR InstName D2 SYMATTR Value PMEG4010BEA SYMATTR Description Diode SYMATTR Type diode SYMBOL polcap 432 -352 R0 WINDOW 3 24 56 Left 2 SYMATTR Value 0.25 SYMATTR InstName C3 SYMATTR SpiceLine V=10 Irms=255m Rser=0.36 Lser=0 SYMATTR Description Capacitor SYMATTR Type cap SYMBOL res 432 -416 R90 WINDOW 0 0 56 VBottom 2 WINDOW 3 32 56 VTop 2 SYMATTR InstName R4 SYMATTR Value 15 SYMBOL njf 224 -352 R270 SYMATTR InstName J1 SYMATTR Value J310 TEXT 88 320 Left 3 !.tran 150s TEXT 368 368 Left 2 ;Soft Start Cap TEXT -88 296 Left 2 ;Simulaton\nOnly TEXT 32 56 Right 2 ;Power Source TEXT -88 -104 Left 2 ;Vbe - Vf is threshold for current limit across R1\nAt 205 mA, \nVbe = 0.57V, \nVf = 0.324V, \nV(R1) = 0.245V TEXT 296 -24 Left 2 ;M1 Ppeak = 940 mW,\nM1 - 2 Wmax on 1 sq inch (6 mm^2) \ncopper and tinned

LINE Normal 64 112 80 96 LINE Normal 80 128 64 112 LINE Normal 80 112 96 96 LINE Normal 96 128 80 112 LINE Normal 576 96 512 32 LINE Normal 80 16 192 80 RECTANGLE Normal 48 352 -144 32 1 RECTANGLE Normal 864 512 -240 -176

Reply to
Piotr Wyderski

The little LND150 Supertex parts get ohmic, around 1K, at low voltages. The LED will get a bit dimmer at low supply voltage. That's not lethal.

I recently invented a cute LED current limiter, where the voltage drop across the LED is the voltage reference for its own current. 4 parts,

8 cents, nearly zero headroom.
--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  
 Click to see the full signature
Reply to
John Larkin

The charging is to get as much voltage on the supercap as possible. 2 volt s is the minimum level the rest of the circuit will operate when power is l ost and the supercap discharges.

So the headroom goes down to zero as the supercap charges. The issue is at what point does the charging drop off and turn into a trickle? As I've sa id, the present design charges at full rate up to within 0.57 volts of the input power. That's pretty good without adding opamps and such. Also, the 1 volt drop of the depletion mode FET doesn't include the diode. Can that be inserted in the Vsense leg with the resistor like I am doing? I suppos e so.

Where do you get the 1 volt number? Are there parts that will work at lowe r voltages at 200 mA current? I'm sure there is a relation between current and voltage. A part that passes 10 mA at 1 volt would pass more current a t a lower voltage, right?

I've dug around and have not found a part that would work. Thanks for the input.

--
  Rick C. 

  --- Get 1,000 miles of free Supercharging 
 Click to see the full signature
Reply to
Ricketty C

I don't see where it is better in any useful way. The parts I found have the several problems I outlined. Just saying, "Why not a depletion-mode MO SFET with a resistor?" doesn't tell me what you had in mind. I'm picturing a circuit that performs poorly compared to this simple two transistor circ uit. There is also a soft start feature which I expect to be very useful.

I don't understand your comments. When I say it works poorly, I mean it ha s too much drop out voltage. By "too much" I mean more than the circuit I currently have. I don't know for sure how low I can make the drop out volt age with a depletion mode FET, I haven't found one that looks like it would work as well as what I have.

missing something about it?

Ok, that's great.

mment.

Yes, I understand how it works although I was turned around for a bit think ing it had to be ground referenced. I meant something realistic in having a part number perhaps. Everything I've found would not be suitable and the re aren't tons of choices.

it would not control the current as well over the voltage range required.

You are looking at the wrong end. I'm talking about when the cap charges u p and the drop out voltage limits the current. The larger drop out means t he current will taper off earlier.

Ok, thanks for the info.

--
  Rick C. 

  --- Get 1,000 miles of free Supercharging 
 Click to see the full signature
Reply to
Ricketty C

With a depletion mode device there is no such point, as there is none with a regular resistor. You will be gradually reaching your R_DS_ON and the device will become more and more ohmic.

This is the typical level of V_GS, i.e. drop on the series resistor where you get very decent regulation. If you want to rely solely on the saturation current of the device, which I very often do, you don't need the resistor.

Sure, but you may not like the price. However, if you want to go to the extreme just to see what's possible, check the UJ3N065025K3S: 25mOhms.

At lower voltages it is just equivalent to a resistor.

How about DN2450N8-G? 10Ohm max. Or IXTY1R6N50D2, 2.3Ohm max.

Best regards, Piotr

Reply to
Piotr Wyderski

OK, 25mV per amp.

Best regards, Piotr

Reply to
Piotr Wyderski

rickle?

passes 10 mA at 1 volt would pass more current at a lower voltage, right?

the input.

The IXTY1R6N50D2 might do the job. The IXTT16N10D2 would be better, but I am going to need a power switch to control a 12 volt line to a motor at 5.5 amps and I can use the PMPB13UP for that.. well, the PMPB20XPE actually wh ich has replaced the PMPB13UP in the last few days just because of this req uirement. I can't see having two, so similar parts on the BOM.

One of the requirements is for the components to be "readily available" any where. This is vague and some aspects of the design (not this board) are d eviating from that pretty oddly. They will need a plastic membrane to cove r the front panel, display, etc. That is going to be custom. Can that be made everywhere? I don't know.

Thanks for the insights into depletion mode devices. I've learned a few th ings.

--
  Rick C. 

  --+ Get 1,000 miles of free Supercharging 
 Click to see the full signature
Reply to
Ricketty C

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.