Good Cheese - Student Edition

...

Nice plots

What was the ESR of the two output caps?

I expect that a few 10's of mOhms of resistance in series with the output caps would improve the stability.

kevin

Better late than never, I suppose.

That's a lot of work to do for no pay.

Here's the sim I used. The wide pulse is a step of Vref, and the little one inside is a 100 mA load step. Nothing rings much, and the load step bumps are just about 5 millivolts. It's not as good at light loads, but I don't have light loads.

If I were selling LDOs as general-purpose products, I'd work on it more. But all this has to do is power three chips (FPGA and two Pericom PCIe equalizers) and it works fine for that.

As I've mentioned before, putting a resistor in series with R3 damps it nicely and increases the HF gain around the loop. 2K is pretty good. I didn't do that because I didn't need to, and I didn't have time to make a project out of this... there are lots of other things going on in this box. We try to keep our BOM down, both number of parts and number of different parts. If I ever do a rev B board, maybe I'll add some 2K resistors; they are line-items on the BOM already.

But show us your feed-forward circuit.

--

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

Naaaah! It was educational. I have always had the benefit of lossy caps (on-chip) to produce the zero. But some recent chips are such "honkers" that external caps are needed (and customers in my world resist external components whenever possible). You've provoked me into thinking my way thru this problem. I'll probably be able to thank you later for pushing me toward yet another patent :-)

Huh? I see the pre-load now is 2 Ohms... 600mA!!!, R4 adds a zero... pretty much standard for compensating LDO's... add dissipation... AND you changed all the parts...

Where are the originally specified MC33072 and ZXMN6A25G? I have the models, why don't you?

And you now have a transient load that's only 16% of steady-state, and it looks like you tweaked the TR/TF, PW and PER for "prettiest picture".

I think you're lying. Makes one wonder how many other things you've lied about... I suspect MOST of your claims.

I didn't lie about killfiling you. The filter works perfectly. However I do admit to peeking periodically to see what QUACK claims you had been making :-)

Then I decided to rejoin the fray and systematically kick your ass ever time you go QUACK, QUACK!

Will do. If I crack this problem, I'll announce it to the world... after I patent it ;-) ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| 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.

Of course I changed parts in the sim, lots of times. One has to explore the load range and margins and stuff. The 2 ohms isn't the pre-load, it's the load, or rather a possible load.

I suppose you have to have sims that are exactly what will go into your chips. I don't have to do that. I can play with a simulation enough to be confident that my board will work, but I don't have to match them exactly. And if that regulator did have stability problems, which it didn't, I could change parts on the board, which you can't.

This board has three fiber-input timestampers with 62 ps LSB; about 50 digital delay generators, 16 of which are 1 ns resolution with fiber laser outputs. It interfaces to a PLC, a wafer scanner, and 16 energy digitizers. It has an ARM cpu and a big FPGA and PCI Express and RS232. There's a high-power arbitrary waveform generator. There are currently about six guys writing VHDL and C and PowerBasic code for the application, which may continue approximately forever. All these things are stuff that you don't do. The LDO was well under 0.1% of the design effort and it's good enough.

I don't have those models, so I just picked similar parts from the LT Spice library. It works.

No, I hacked some reasonable values and didn't change them. I did tweak the times of the load steps for pretty graphing, but all the transients are settled so that doesn't matter.

Lying about what? The PDF schematic is a sheet from the actual product, and the pictures of the PCB are real, and it works. We've shipped about 8 so far, and the customer is happy. This is one box of six that we furnish for this project. They all worked with rev A PCBs.

What fraction of your chip designs work first pass?

No self control. You even cheat on yourself.

[snip]

I know you are lying... but that's SOP for you. The drawing you posted may well be what you postulated... but it's been tweaked on the board... as drawn, it sings like a sonnuva-bitch... just like you ;-)

I'll post my simulation later. Why did you need to add a zero in your compensation? Liar! Liar! Pants on fire :-)

100% "work", I never had any DOA except for process house bungles, which they make good on. But I'm proud of my 99.75% record of meeting specification, first pass.

What fraction of your PCB's don't need a tweak?

Always some smug ass remark. But I never lie... it's much easier to keep track of truth than obfuscation.

You ought to be ashamed of yourself. You are indeed NOLA trailer trash. ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| 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.

I just noticed that "quack-quack slippery bastard" slipped us some old drawings and simulations... he didn't attempt to simulate what he claimed worked. Wonder if "quack-quack slippery bastard" would allow some independent party, say Spehro, trace the board "quack-quack slippery bastard" claims works and verify (and post) what _exactly_ is on the PCB?

I doubt it will ever happen. ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| 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.

It does not. I scoped it, under load, from powerup, through FPGA config, all the way through PCI Express activity. It's fine. I changed no values from the original BOM. I did of course sim it with a range of parts values and loads, to get a feel for how it behaved.

It doesn't oscillate, and you are wrong. It's a little bit underdamped, but not enough to matter, not enough to affect the transient load bump.

Do you mean the capacitive ESR? Because caps have ESR. The caps in the sim are a pretty good approximation of what's on the board, including bypasses on the power pours near the loads, the 3 uF thing. You can't neglect them just because they aren't on the "power supply" page of the schematic.

You once said "I don't cook." Then when you later posted about cooking, and I asked, you said "I lied." I thought that was very strange at the time. What engineer doesn't cook?

What's on the PCB is what's on the PDF schematic that I posted. I didn't change the BOM from the original layout, at least not this part.

Sure, buy Sphero a plane ticket, and put him up at W or the Four Seasons, and let him check the boards. Include dinner at Boulevard, recently rated the best restaurant in the USA.

Change C3 to 5n and place a 5k in series with C3

Sim that, I think you'll like the improvement.

I maybe late on the scene, you may have already discovered this.

But this will play with the loop gain..

I get no more than 5 mv ripple with the current source on the load and no ripple with input transient.

This is based from the print you showed. Of course, simulations lie like hell, so what ever. :)

Jamie

Pomposity at its finest :-)

Would you agree to Spehro stopping by the next time he's in the bay area?

And that Spehro can verify all components _are_as_you_posted_originally? ... not with the added zero and load change as you posted today?

...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| 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.

Jim Thompson wrote, in response to other junk,

"... and news brought to you here on the sub-etha wave band, broadcasting around the galaxy around the clock," squawked a voice, "and we'll be saying a big hello to all intelligent life forms everywhere ... and to everyone else out there, the secret is to bang the rocks together, guys. And of course, the big news story tonight is the sensational theft of the new Improbability Drive prototype ship by none other than Galactic President Zaphod Beeblebrox. And the question everyone's asking is ... has the big Z finally flipped? Beeblebrox, the man who invented the Pan Galactic Gargle Blaster, ex-confidence trickster, once described by Eccentrica Gallumbits as the Best Bang since the Big One, and recently voted the Wort Dressed Sentinent Being in the Known Universe for the seventh time ... has he got an answer this time? We asked his private brain care specialist Gag Halfrunt ..." The music swirled and dived for a moment. Another voice broke in, presumably Halfrunt. He said: "Vell, Zaphod's jist zis guy you know?" but got no further because an electric pencil flew across the cabin and through the radio's on/off sensitive airspace. Zaphod turned and glared at Trillian

- she had thrown the pencil.

"Hey," he said, what do you do that for?"

Trillian was tapping her fingers on a screenful of figures.

"I've just thought of something," she said.

"Yeah? Worth interrupting a news bulletin about me for?" "You hear enough about yourself as it is."

"I'm very insecure. We know that."

"Can we drop your ego for a moment? This is important."

"If there's anything more important than my ego around, I want it caught and shot now."

-- The Hitchhiker's Guide to the Galaxy, by Douglas Adams: Excerpt from Chapter 12

Not that better LDOs aren't good things to post about--they're better than a lot of the other stuff round here--but the Godzilla vs Rodan bit gets a bit old, you know?

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC
Optics, Electro-optics, Photonics, Analog Electronics

160 North State Road #203
Briarcliff Manor NY 10510
845-480-2058

hobbs at electrooptical dot net
http://electrooptical.net

[snip]

I grill steak, boil eggs, and, with my wife's new gadgetry, I can even poach them. I can also make myself a package of Ramen soup. Why does that cause you such heartburn? Or is your "concern" all just diversionary subterfuge, as I suspect? ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| 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.

Mind your own business, and I read the book eons ago :-) ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| 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.

Well, we've eaten there, and it's pretty good. My attorney occupies the penthouse above the restaurant, and shares the space with Willy Brown. Or, actually, lets Willy have an office for occasional use. The view from the conference room is impressive.

Sure. He's stayed with us before, and he's always welcome. But you are being silly.

The caps on the board have ESR, and I guessed at their value when I did the sim. Of course I didn't show the ESR on the PCB schematic. And the chips *are* loads. Why would I build a semi-monstrous LDO if it didn't have loads?

If I change the ESR in the sim to 1u ohm, or to 0.1 ohms, the transient response doesn't change enough to notice. I did guard-band that in simulation, to make sure it didn't matter much. It's slightly underdamped at zero ESR. It is critically damped at just a tad over

0.1 ohms.

You seem confused by the fact that the PCB schematic is not identical to the simulation schematic. In your business, I suppose it has to be.

--

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

On Sun, 20 May 2012 21:25:37 -0400, Jamie wrote:

I tried 2K in series with 1n, and it damped it nicely.

5n + 5K rings badly in my simulation.

Here's the version with the 2K tweak:

Version 4 SHEET 1 1256 680 WIRE 144 -176 32 -176 WIRE 608 -176 416 -176 WIRE 144 -144 144 -176 WIRE 608 -112 608 -176 WIRE 416 -48 416 -176 WIRE 144 -32 144 -64 WIRE 32 0 32 -176 WIRE 608 0 608 -32 WIRE 0 16 -32 16 WIRE 144 32 64 32 WIRE 192 32 144 32 WIRE 224 32 192 32 WIRE 368 32 304 32 WIRE -144 48 -208 48 WIRE 0 48 -144 48 WIRE 32 80 32 64 WIRE -208 96 -208 48 WIRE 144 128 144 96 WIRE 288 144 240 144 WIRE 416 144 416 48 WIRE 416 144 368 144 WIRE 512 144 416 144 WIRE 608 144 512 144 WIRE 768 144 608 144 WIRE 416 176 416 144 WIRE 512 176 512 144 WIRE 608 176 608 144 WIRE 768 192 768 144 WIRE -208 208 -208 176 WIRE -32 240 -32 16 WIRE 48 240 -32 240 WIRE 144 240 144 208 WIRE 144 240 48 240 WIRE 240 240 240 144 WIRE 240 240 144 240 WIRE 512 272 512 240 WIRE 416 288 416 240 WIRE 608 288 608 256 WIRE 768 320 768 272 WIRE 512 384 512 352 FLAG 144 -32 0 FLAG 32 80 0 FLAG -208 208 0 FLAG 416 288 0 FLAG 608 0 0 FLAG 512 384 0 FLAG 608 288 0 FLAG 608 144 OUT FLAG -144 48 IN FLAG 192 32 AMP FLAG 48 240 FB FLAG 768 320 0 SYMBOL cap 400 176 R0 SYMATTR InstName C1 SYMATTR Value 3µ SYMBOL cap 496 176 R0 SYMATTR InstName C2 SYMATTR Value 20µ SYMBOL cap 128 32 R0 WINDOW 0 49 42 Left 2 WINDOW 3 50 73 Left 2 SYMATTR InstName C3 SYMATTR Value 1n SYMBOL Opamps\\LT1492 32 -32 R0 WINDOW 0 -76 -33 Left 2 WINDOW 3 -104 3 Left 2 SYMATTR InstName U1 SYMBOL res 320 16 R90 WINDOW 0 -45 59 VBottom 2 WINDOW 3 -36 60 VTop 2 SYMATTR InstName R1 SYMATTR Value 100 SYMBOL res 384 128 R90 WINDOW 0 72 52 VBottom 2 WINDOW 3 78 52 VTop 2 SYMATTR InstName R2 SYMATTR Value 2K SYMBOL nmos 368 -48 R0 SYMATTR InstName M1 SYMATTR Value FDC637AN SYMBOL voltage 608 -128 R0 WINDOW 0 55 39 Left 2 WINDOW 3 53 82 Left 2 WINDOW 123 0 0 Left 2 WINDOW 39 0 0 Left 2 SYMATTR InstName V1 SYMATTR Value 1.5 SYMBOL voltage -208 80 R0 WINDOW 3 -161 182 Left 2 WINDOW 123 0 0 Left 2 WINDOW 39 0 0 Left 2 SYMATTR Value PULSE(1.1 1.15 100u 1u 1u 200u) SYMATTR InstName V2 SYMBOL voltage 144 -160 R0 WINDOW 0 61 23 Left 2 WINDOW 3 60 60 Left 2 WINDOW 123 0 0 Left 2 WINDOW 39 0 0 Left 2 SYMATTR InstName V3 SYMATTR Value 12 SYMBOL res 592 160 R0 WINDOW 0 58 43 Left 2 WINDOW 3 64 76 Left 2 SYMATTR InstName R3 SYMATTR Value 2 SYMBOL res 496 256 R0 WINDOW 0 47 62 Left 2 WINDOW 3 41 98 Left 2 SYMATTR InstName R4 SYMATTR Value 0.025 SYMBOL current 768 192 R0 WINDOW 123 0 0 Left 2 WINDOW 39 0 0 Left 2 SYMATTR InstName I1 SYMATTR Value PULSE(0 0.1 175u 200n 100n 50u) SYMBOL res 128 112 R0 SYMATTR InstName R5 SYMATTR Value 2K TEXT 88 288 Left 2 !.tran 0.001 TEXT -456 -136 Left 2 ;TEM2 LDO REGULATORS TEXT -416 -88 Left 2 ;JL Sep 12 2011 TEXT -392 -48 Left 2 ;+ May 20, 2012

--

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

You snipped the claim that you never lie.

--

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

[snip]

You're on the right track, Jamie. Just play around with the values. (It helps to just Bode the NMOS follower by itself to get a clue about what poles it adds :-)

I can get > 57° Phase margin from 100mA to 1Amp that way.

Still doesn't solve all the issues. But, in my IC world, I think I can build a "tracking compensator" for the NMOS pass device (since it's on my chip) and eliminate the effects of the moving output pole. ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| 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.

The LDO dilemma is that, for merchant ICs, the load capacitance and ESR can be all over the place. In my case, I know pretty much what the load will be, so I can compensate for that.

But the LM8261 drives any c-load. The idea is simple: only have one pole.

--

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

The idea is even simpler... use bipolar, as in your example. But that doesn't scale well into the ampere range (at least efficiently) when you need an RRIO amplifier or an LDO.

We have a problematic world that wants to make everything (CHEAP) in CMOS. It's been quite a few years now since any of my customers would opt for BiCMOS, which is the best of both worlds. ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| 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.