Yeah, but every time I looked at a digital solution I found that things fall apart when you need cycle-to-cycle performance at much more than
100kHz. Same for the loops, many of mine have to operate at 50kHz BW or higher and there's often two or three loops operating simultaneously. Even a serious uC runs out of coal real fast then.
Many do these days. Same with OVP schemes that occasionally backfire. Sometimes I goose them in a way that they think everything is in the green. For example, by putting the FB node on an artificial pedestal. Then it never goes into foldback. If you don't have a precision reference available for that you can concoct a simple circuit that sources current into the FB pin when the FB voltage wants to drop towards the threshold where foldback would kick in. If an opamp from a quad-pack happens to be vacant that makes things easy. But even a stiff voltage source and a diode is sometimes enough to fool the chip into permanent non-foldback. One the FB voltage is above a limit where the diode quits conducting things look like normal.
Sure, but it seems silly to add an opamp active diode to fool a chip that's already too complex.
TI has a DSP-core digital power controller with picosecond PWM resolution. I'll ask Rob what the part number is.
You could also give a cheap uP a little help around the edges, like a gate driver that shuts off when peak current is sensed, something like that. One advantage of designing everything yourself is that you can understand it; lots of complex analog chips are mysterious about what they actually do. Like an opamp on the block diagram with three inputs... what's that supposed to do?
--
John Larkin Highland Technology Inc
www.highlandtechnology.com jlarkin at highlandtechnology dot com
Precision electronic instrumentation
Picosecond-resolution Digital Delay and Pulse generators
Custom timing and laser controllers
Photonics and fiberoptic TTL data links
VME analog, thermocouple, LVDT, synchro, tachometer
Multichannel arbitrary waveform generators
Nah, foldback makes a lot of sense in many situations. For example, I just designed a unit that also has to supply a couple amps to miscellaneous external 12V gear. Small displays and such, and if the regulator were capable of dumping 2A into a shorted out electrolytic, changes are there would be a serious *KAPHOOF* situation. If I wanted no foldback for some reason all I need to do that is place three parts. Two resistors and one diode. About 50 millibucks worth.
I'll
That is nice. As long at it has the horsewpower for a reasonably fast loop (as in tens of kHz) and won't require the San Onofre power plant to be fired up again in order to operate.
See? There the digital situation already requires extra hardware, including extra ICs. There is no way to properly drive a big fat switcher from a 5V logic device of any kind. Because you need +10V (or better 12-15V) to turn the power FETs on and -5V or so to turn the off with gusto. Else the efficiency will suffer. All this logic-level driving is no good when efficiency is paramount.
Pretty soon you'll find out that leading-edge blanking just isn't in the cards with a uC, its internal comparator has the acceleration of an Isetta, some high priority interrupt keeps trampling on the loop performance, and it can't prevent a loud bang when the regulator goes into CCM and the inductor core saturates within a microsecond or so.
That almost drove people nuts when we had transistors on the schematic of an IC design that had numerous collectors.
That's how I designed the boost converter on my first mass-produced design in the early 90's, less the transformer (used only an inductor). But I used a BJT instead of the 7002 because that saved another penny. No kidding, and we also had the inductor made in Asia because then it cost less than a catalog part from a distributor.
I like your use of a B source, to plot efficiency. Must remember that one.
I rehashed it a bit, to include "proper" 74HC14s (actually a 74HC .lib file that purports to come from LT, but I can't remember where I got it from), and a real-life Zetex 39V zener.
Little if any difference.
--
"For a successful technology, reality must take precedence
over public relations, for nature cannot be fooled."
(Richard Feynman)
You could also get a cheap inverting gate driver with Schmitt input, runs that as oscillator. If it is one without UVLO you could feed it directly off of your 9V input.
We stock LM5112, a pretty cool part. That would work. It might turn out that there is some fet with nice specs but prefers more gate drive. Might need another resistor to futz the duty cycle.
But it costs over a dollar! An AC14 is 10 cents!
--
John Larkin Highland Technology Inc
www.highlandtechnology.com jlarkin at highlandtechnology dot com
Precision electronic instrumentation
Picosecond-resolution Digital Delay and Pulse generators
Custom timing and laser controllers
Photonics and fiberoptic TTL data links
VME analog, thermocouple, LVDT, synchro, tachometer
Multichannel arbitrary waveform generators
It should cost less than 50c. Where do you buy them? Other chips:
formatting link
And, if you calculate the energy savings over the next 200 years you'll have it all back. When it has to be ral chep I use a CD40106 and a PNP/NPN follower. Some seciton of the 40106 can be rented out.
Simulation slows to a crawl as the converter starts up. Not so much a slow, as a delayed, start. Surge current is substantially the same, just delayed.
--
"For a successful technology, reality must take precedence
over public relations, for nature cannot be fooled."
(Richard Feynman)
This is what I'm seeing. Without the soft-start, the input current peaks at about 1.6 amps.
formatting link
This run took a couple of minutes on my 5-year old HP. The soft-start doesn't affect the sim speed much; I get around 13 PPM of real time during chargeup.
--
John Larkin Highland Technology Inc
www.highlandtechnology.com jlarkin at highlandtechnology dot com
Precision electronic instrumentation
Picosecond-resolution Digital Delay and Pulse generators
Custom timing and laser controllers
Photonics and fiberoptic TTL data links
VME analog, thermocouple, LVDT, synchro, tachometer
Multichannel arbitrary waveform generators
Didn't slow down much here. But yesterday the fan speed cranked itself to the hilt. I took the PC apart and found an incredible amount of dog hair and dust in the cooling fins for the processor. One of our Labradors has a habit of curling himself up in the small space under my PC workplace.
Here is a version with a better soft-start:
Version 4 SHEET 1 1920 1076 WIRE 208 48 96 48 WIRE 96 144 96 48 WIRE 240 144 96 144 WIRE 320 144 240 144 WIRE 464 144 400 144 WIRE 528 144 464 144 WIRE 656 144 608 144 WIRE 816 144 768 144 WIRE 960 144 896 144 WIRE 1040 144 960 144 WIRE 1152 144 1104 144 WIRE 1200 144 1152 144 WIRE 1328 144 1280 144 WIRE 1344 144 1328 144 WIRE 1360 144 1344 144 WIRE 96 176 96 144 WIRE 768 208 768 144 WIRE 240 240 240 144 WIRE 320 240 240 240 WIRE 464 240 464 144 WIRE 464 240 384 240 WIRE 1152 240 1152 144 WIRE 1200 240 1152 240 WIRE 1328 240 1328 144 WIRE 1328 240 1264 240 WIRE 1328 288 1328 240 WIRE 656 304 656 144 WIRE 720 304 656 304 WIRE 768 304 720 304 WIRE 96 320 96 256 WIRE 656 336 656 304 WIRE 768 336 768 304 WIRE 960 336 960 144 WIRE 304 384 272 384 WIRE 400 384 368 384 WIRE 1328 400 1328 352 WIRE 656 432 656 416 WIRE 144 448 96 448 WIRE 272 448 272 384 WIRE 272 448 224 448 WIRE 304 448 272 448 WIRE 400 448 400 384 WIRE 400 448 368 448 WIRE 768 480 768 416 WIRE 960 480 960 416 WIRE 272 512 272 448 WIRE 304 512 272 512 WIRE 400 512 400 448 WIRE 400 512 368 512 WIRE 432 512 400 512 WIRE 560 512 512 512 WIRE 608 512 560 512 WIRE 96 576 96 448 WIRE 160 576 96 576 WIRE 272 576 272 512 WIRE 272 576 224 576 WIRE 304 576 272 576 WIRE 400 576 400 512 WIRE 400 576 368 576 WIRE 656 592 656 528 WIRE 768 592 768 544 WIRE 960 592 960 544 WIRE 1216 624 1152 624 WIRE 1344 624 1296 624 WIRE 1360 624 1344 624 WIRE 96 672 96 576 WIRE 160 672 96 672 WIRE 272 672 160 672 WIRE 464 672 400 672 WIRE 560 672 464 672 WIRE 688 672 624 672 WIRE 720 672 688 672 WIRE 1152 672 1152 624 WIRE 1360 688 1360 624 WIRE 272 704 272 672 WIRE 464 704 464 672 WIRE 96 720 96 672 WIRE 400 784 400 672 WIRE 400 784 320 784 WIRE 96 832 96 784 WIRE 272 832 272 800 WIRE 1152 832 1152 752 WIRE 1360 832 1360 752 WIRE 240 912 144 912 WIRE 368 912 320 912 WIRE 464 912 464 784 WIRE 464 912 432 912 WIRE 464 928 464 912 WIRE 464 1056 464 1008 FLAG 1344 144 VP FLAG 656 592 0 FLAG 96 320 0 FLAG 720 304 DRAIN FLAG 560 512 GATE FLAG 768 208 0 FLAG 960 592 0 FLAG 1152 832 0 FLAG 1344 624 EFF FLAG 768 592 0 FLAG 1360 832 0 FLAG 96 832 0 FLAG 160 672 rc FLAG 272 832 0 FLAG 464 1056 0 FLAG 688 672 VP FLAG 1328 400 0 FLAG 208 48 9V FLAG 144 912 9V SYMBOL ind2 512 160 R270 WINDOW 0 -41 56 VTop 2 WINDOW 3 -53 53 VBottom 2 SYMATTR InstName L1 SYMATTR Value 3µ SYMATTR Type ind SYMATTR SpiceLine Rser=0.06 SYMBOL schottky 1040 160 R270 WINDOW 0 -41 30 VTop 2 WINDOW 3 -49 27 VBottom 2 SYMATTR InstName D1 SYMATTR Value 10MQ060N SYMATTR Description Diode SYMATTR Type diode SYMBOL voltage 96 160 R0 WINDOW 0 48 55 Left 2 WINDOW 3 54 83 Left 2 WINDOW 123 0 0 Left 2 WINDOW 39 0 0 Left 2 SYMATTR InstName V2 SYMATTR Value 9 SYMBOL nmos 608 432 R0 WINDOW 0 -55 -55 Left 2 WINDOW 3 -101 -22 Left 2 SYMATTR InstName M1 SYMATTR Value Si7218DN SYMBOL ind2 800 160 R270 WINDOW 0 -32 56 VTop 2 WINDOW 3 -41 57 VBottom 2 SYMATTR InstName L2 SYMATTR Value 27µ SYMATTR Type ind SYMATTR SpiceLine Rser=0.06 SYMBOL res 528 496 R90 WINDOW 0 77 56 VBottom 2 WINDOW 3 84 57 VTop 2 SYMATTR InstName R3 SYMATTR Value 10 SYMBOL res 1184 160 R270 WINDOW 0 -38 31 VTop 2 WINDOW 3 -11 86 VBottom 2 SYMATTR InstName R6 SYMATTR Value 1m SYMBOL cap 1264 224 R90 WINDOW 0 68 59 VBottom 2 WINDOW 3 42 -8 VTop 2 SYMATTR InstName C3 SYMATTR Value 100m SYMBOL res 304 160 R270 WINDOW 0 -37 24 VTop 2 WINDOW 3 -11 88 VBottom 2 SYMATTR InstName R8 SYMATTR Value 1m SYMBOL cap 384 224 R90 WINDOW 0 73 60 VBottom 2 WINDOW 3 47 -14 VTop 2 SYMATTR InstName C6 SYMATTR Value 100m SYMBOL res 944 320 R0 WINDOW 0 -54 67 Left 2 WINDOW 3 -58 98 Left 2 SYMATTR InstName R9 SYMATTR Value 500 SYMBOL cap 944 480 R0 WINDOW 0 -46 35 Left 2 WINDOW 3 -50 68 Left 2 SYMATTR InstName C7 SYMATTR Value 30p SYMBOL bv 1152 656 R0 WINDOW 3 -557 141 Left 2 WINDOW 0 -83 97 Left 2 SYMATTR Value V= LIMIT (0, 100 * I(R6) * V(VP) / ( I(R8) * 9 ), 100 ) SYMATTR InstName B1 SYMBOL res 1312 608 R90 WINDOW 0 70 60 VBottom 2 WINDOW 3 78 61 VTop 2 SYMATTR InstName R1 SYMATTR Value 1K SYMBOL res 752 320 R0 WINDOW 0 53 70 Left 2 WINDOW 3 51 100 Left 2 SYMATTR InstName R2 SYMATTR Value 200 SYMBOL cap 752 480 R0 WINDOW 0 60 34 Left 2 WINDOW 3 56 72 Left 2 SYMATTR InstName C1 SYMATTR Value 200p SYMBOL cap 1344 688 R0 WINDOW 0 -37 63 Left 2 WINDOW 3 -43 94 Left 2 SYMATTR InstName C2 SYMATTR Value 50n SYMBOL Digital\\schmtinv 160 512 R0 SYMATTR InstName A1 SYMATTR Value2 Vhigh=5 Vt=2.5 Vh=0.5 SYMATTR SpiceLine Td=5n SYMBOL cap 80 720 R0 WINDOW 0 44 60 Left 2 WINDOW 3 37 92 Left 2 SYMATTR InstName C4 SYMATTR Value 500p SYMBOL res 128 464 R270 WINDOW 0 88 58 VTop 2 WINDOW 3 74 57 VBottom 2 SYMATTR InstName R4 SYMATTR Value 2K SYMBOL Digital\\schmtinv 304 448 R0 SYMATTR InstName A2 SYMATTR Value2 Vhigh=5 Vt=2.5 Vh=1 SYMATTR SpiceLine Td=5n SYMBOL Digital\\schmtinv 304 384 R0 SYMATTR InstName A3 SYMATTR Value2 Vhigh=5 Vt=2.5 Vh=1 SYMATTR SpiceLine Td=5n SYMBOL Digital\\schmtinv 304 512 R0 SYMATTR InstName A4 SYMATTR Value2 Vhigh=5 Vt=2.5 Vh=1 SYMATTR SpiceLine Td=5n SYMBOL Digital\\schmtinv 304 320 R0 SYMATTR InstName A5 SYMATTR Value2 Vhigh=5 Vt=2.5 Vh=1 SYMATTR SpiceLine Td=5n SYMBOL nmos 320 704 M0 WINDOW 0 77 0 Left 2 WINDOW 3 63 35 Left 2 SYMATTR InstName M2 SYMATTR Value 2N7002 SYMBOL res 448 688 R0 WINDOW 0 47 64 Left 2 WINDOW 3 40 94 Left 2 SYMATTR InstName R5 SYMATTR Value 10K SYMBOL zener 560 688 R270 WINDOW 0 50 33 VTop 2 WINDOW 3 -13 35 VBottom 2 SYMATTR InstName D2 SYMATTR Value 1N5369B SYMATTR Description Diode SYMATTR Type diode SYMBOL cap 1312 288 R0 WINDOW 0 -61 48 Left 2 WINDOW 3 -59 81 Left 2 SYMATTR InstName C5 SYMATTR Value 5µ SYMBOL res 640 320 R0 WINDOW 0 47 70 Left 2 WINDOW 3 41 102 Left 2 SYMATTR InstName R7 SYMATTR Value 1m SYMBOL res 448 912 R0 WINDOW 0 47 64 Left 2 WINDOW 3 40 94 Left 2 SYMATTR InstName R11 SYMATTR Value 10K SYMBOL cap 368 928 R270 WINDOW 0 32 32 VTop 2 WINDOW 3 0 32 VBottom 2 SYMATTR InstName C9 SYMATTR Value 0.33µ SYMBOL res 224 928 R270 WINDOW 0 32 56 VTop 2 WINDOW 3 0 56 VBottom 2 SYMATTR InstName R10 SYMATTR Value 10k TEXT 1064 432 Left 2 !.tran 0 15m 0 10n uic TEXT 1040 488 Left 2 ;PRETTY DUMB BOOST TEXT 1072 544 Left 2 ;JL June 29 2013 TEXT 648 104 Left 2 !K L1 L2 0.95 TEXT 192 624 Left 2 ;74HC14 TEXT 296 1016 Left 2 ;SOFT TEXT 296 1040 Left 2 ;START TEXT 160 504 Left 2 ;OSC TEXT 696 168 Left 2 ;1:3
Cute, but the oscillator runs a bit at powerup. The zener forward conducts and clamps the 2N7002 gate voltage to VP+0.6. Not lethal, since VP jumps up and fixes things pretty quickly. With my 4000 uF load, it might be a problem.
With 4000 uF, my soft-start would have to be pretty slow. Maybe we'll do it digitally after all.
I'm in a situation where the customer is in a frenzy to get this thing done. We quoted 5 weeks ARO for a first article, so they waited a few weeks and then cut a PO stating 4 weeks. Two weeks later, we have no spec hard enough to design around, and no answers to basic questions, like how does this thing mount? How big can it be? What's the orientation of the connector? What's the required output, into what load?
So I play with the power inverter now and then, and rest up for the panic.
--
John Larkin Highland Technology Inc
www.highlandtechnology.com jlarkin at highlandtechnology dot com
Precision electronic instrumentation
Picosecond-resolution Digital Delay and Pulse generators
Custom timing and laser controllers
Photonics and fiberoptic TTL data links
VME analog, thermocouple, LVDT, synchro, tachometer
Multichannel arbitrary waveform generators
That's better. Initial current draw at switch start down to ~1.2 amps.
Runs faster, too, similar to the original circuit.
I was playing with something similar, hold the switch off by pushing an artificial feedback voltage, then let it ramp down to meet the real feedback coming up. Feedforward, overtaken by feedback.
--
"For a successful technology, reality must take precedence
over public relations, for nature cannot be fooled."
(Richard Feynman)
Bingo! It was the Skyworks schottky model that was causing the problem. It's a 2V mixer diode (BV=3), not suitable. What model were you using? I was using the "official" model per Skyworks datasheet.
Using a Diodes inc. 1N5711, I now get what you got.
Your SMS 7621 model must differ from mine.
Must admit I prefer Joerg's soft start, though;-)
--
"For a successful technology, reality must take precedence
over public relations, for nature cannot be fooled."
(Richard Feynman)
I just pulled it out of the LT Spice list. I'm not sure where it came from. The devices aren't sorted alphabetically, so I tend to pick the first thing that might work.
The SMS7621 is actually good to maybe 6 volts in real life. 0.22 pF. We use a lot of them.
It does have the startup burst bug.
--
John Larkin Highland Technology Inc
www.highlandtechnology.com jlarkin at highlandtechnology dot com
Precision electronic instrumentation
Picosecond-resolution Digital Delay and Pulse generators
Custom timing and laser controllers
Photonics and fiberoptic TTL data links
VME analog, thermocouple, LVDT, synchro, tachometer
Multichannel arbitrary waveform generators
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.