LTspice, a great program, but that UI!

I like it because it doesn't need a power supply. Must be solar powered or something.

What the difference between e and e2? Just polarity? The HELP doesn't say.

--

John Larkin         Highland Technology, Inc 

lunatic fringe electronics

There's more to it than simply finding the cheapest alternative. A big part of electronic design is minimizing risk, and to do that I often use LTspice to simulate parts of the circuit. Sure I look at the alternatives before choosing a part, but being able to or not able to simulate a circuit is a big factor. Another factor that sometimes matters a lot is time.

Once LTspice and another free LT tool called FilterCad was very helpful for our company - we make modules for ANR headsets, everything was going great until suddenly we started getting a lot of returns - bad speakers that would pass factory tests but rattled when our customer assembled them. They were not amused. Needed to make a speaker tester and fast! To test a speaker for distortion I needed to output test tones at various frequencies then filter the output from a sense mic with a brick-wall high-pass filter set to twice the test frequency. With FilterCad I typed in my requirements and it spit out a complicated circuit using a LTC1068. No way I could have derived all those values on my own with math in the allotted time let alone test it - I needed a working solution NOW. LTspice also had this part, so simulated it and found that I needed to actively drive the bias pin or it leaked lows, but otherwise the circuit did what I needed to do, just needed to supply it with a clock that was 100x more than the pass frequency.. used an ancient 74HC4059 programmable divider and a FF to turn my fixed clock into a variable clock.

Unfortunately for LT I just needed one chip and a spare for the factory tester, but it saved us $thousands and possibly our jobs. If I ever need something like that again there it is. For the most part I use LTspice to simulate generic stuff that doesn't need LT parts (unless I want to, I use bunches of LTC1050 low-offset opamps), but if faced with something intense like that filter I'll take a design with a LT part that I can simulate over something untested any day, even if it costs a bit more. I'm sure LT sells lots of parts through their free tools - sure it's advertising but it's also damn useful.

Terry

Am 13.03.2017 um 18:00 schrieb Jim Thompson:

I'm afraid I won't be in that area in the near future. I don't know MicroSim. But I know some other more or less expensive professional software... This video about OrCad for example

is surely not optimized for speed but it shows major lacks in OrCad. You must draw every wire one by one, exactly hit the pins and sometimes OrCad ends the wire in free air (2:32). That's what I call BS. In LTSpice I would just place the components and draw a closed wire loop through all of them and they are connected. Definitely faster. And inserting the Switch (3:03), he has to delete the wire, place the switch and draw another wire. I would just drop the switch over the wire and it is connected.

not sure about that :-). Can you explain how MicroSim makes it faster?

Robert

P.S. just for fun, I tried an astable multivibrator. Can you beat 1:32?

Version 4 SHEET 1 880 680 WIRE -112 -96 -272 -96 WIRE 32 -96 -112 -96 WIRE 208 -96 32 -96 WIRE 320 -96 208 -96 WIRE -112 -48 -112 -96 WIRE 32 -48 32 -96 WIRE 208 -48 208 -96 WIRE 320 -48 320 -96 WIRE -272 48 -272 -96 WIRE -112 80 -112 32 WIRE -64 80 -112 80 WIRE 32 80 32 32 WIRE 32 80 0 80 WIRE 80 80 32 80 WIRE 208 80 208 32 WIRE 208 80 160 80 WIRE 224 80 208 80 WIRE 320 80 320 32 WIRE 320 80 288 80 WIRE -112 112 -112 80 WIRE 320 112 320 80 WIRE 80 160 160 80 WIRE 80 160 -48 160 WIRE 160 160 80 80 WIRE 256 160 160 160 WIRE -272 272 -272 128 WIRE -112 272 -112 208 WIRE -112 272 -272 272 WIRE 320 272 320 208 WIRE 320 272 -112 272 WIRE -112 288 -112 272 FLAG -112 288 0 SYMBOL npn 256 112 R0 SYMATTR InstName Q1 SYMATTR Value BC847A SYMBOL npn -48 112 M0 SYMATTR InstName Q2 SYMATTR Value BC847A SYMBOL res -128 -64 R0 SYMATTR InstName R1 SYMATTR Value 1k SYMBOL res 16 -64 R0 SYMATTR InstName R2 SYMATTR Value 47k SYMBOL res 304 -64 R0 SYMATTR InstName R3 SYMATTR Value 1k SYMBOL res 192 -64 R0 SYMATTR InstName R4 SYMATTR Value 47k SYMBOL cap 0 64 R90 WINDOW 0 0 32 VBottom 2 WINDOW 3 32 32 VTop 2 SYMATTR InstName C1 SYMATTR Value 100n SYMBOL cap 288 64 R90 WINDOW 0 0 32 VBottom 2 WINDOW 3 32 32 VTop 2 SYMATTR InstName C2 SYMATTR Value 100n SYMBOL voltage -272 32 R0 SYMATTR InstName V1 SYMATTR Value 5 TEXT -306 312 Left 2 !.tran 100m startup uic TEXT 432 -48 Left 2 ;1:32 draw the schematic (including values) TEXT 432 -16 Left 2 ;1:47 Simulation on screen

Am 13.03.2017 um 13:22 schrieb Kevin Aylward:

How did you do it? I never had problems here. Works like a charm.

:-)

Sometimes, when I have to insert something, I simply drag the block a bit away and throw the components in.

When I hack something into LTSpice, I normally first draw the schematic and then set the values for the components. When I need an OpAmp, i hit the component button, type op07 and I have it. Don't care from which library it comes. Some people may prefer it different.

cool.

"r" is already resistor :-)

I suppose the original program did not come from the Windows world. It is like if you use some 'original Linux' programs that have been ported to Windows. Some things drive you crazy when you are not used to it.

Robert

...

Yes, somewhere in the control panel (the hammer symbol).

here it works. Maybe it has been fixed.

I remember I have changed some (at least the dark blue, which is hard to see on black) and they're still there.

I fully agree but I have got used to the ones I need and live with it. Especially with text fields, every program behaves different. If you want a new line, in some you press return, others want ctrl or alt return :-(

Robert

Strawman. It's *never* a 30 minute task to evaluate alternative parts.

Well, it might be for passives and simple discretes. But for a switcher or something, I generally get an eval board and do some testing... cap load stability, noise, chip temperature rise, all that. Might take a day.

I usually test opamps, linear regs, things like that.

--

John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com

MicroSim created the original PSpice. You're looking at what OrCAD could do to f..k it up.

When OrCAD came along and bought out MicroSim I refused to play along and switch to OrCAD Crapture.

Fortunately, so did many others. So even up into many years after the Cadence acquisition, MicroSim PSpice Schematics was "supported".

But "support" was a joke. Cadence was systematically removing features. Fortunately PSpice has always been .INI file based. So an engineer at ON Semi and I colluded and dug back thru our archives and restored the original .INI

Support was a joke, nothing was being improved at each maintenance update, so I quit paying maintenance at v15.7.0.p001

Fortunately I had, long ago, been provided a blanket license, since I had been one of the main antagonists over the years, finding all the bugs and quirks, and causing features to be added that suited my IC design needs.

After an initial purchase of $8,000 in 1987 and maintenance over the years, I've got about $40K invested.

It's been a good investment... I've designed at least 60 chips with it, everyone of which came straight out working to spec.

(Though I've got to admit I probably have around 80 chip designs with almost as good a success rate that were done before CAD ;-)

I won't bother duplicating your example... I want a live competitor... I love to watch "mouthers" sweat >:-} ...Jim Thompson

--
| James E.Thompson                                 |    mens     | 
| Analog Innovations                               |     et      | 
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    | 
| STV, Queen Creek, AZ 85142    Skype: skypeanalog |             | 
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  | 
| E-mail Icon at http://www.analog-innovations.com |    1962     | 

     Thinking outside the box... producing elegant solutions.

Nonsense. It is hardly any more work for a competent engineer to go and get a model from a competitor, and *also* simulate it LTSpice.

Sure, it has *some* value, to include LT models directly, but not a lot, imo

To wit, LTSpice does not guarantee that it will be uses just for LT parts. I wajor, that of the 3,000,00 downloads, only a low % actually use it to buy LT parts, and buy other vendors parts instead. Most of those, probably decide to go into banking once they finish their B.S. E.E. anyway. That's the bit many seem to be missing here.

So, ones needs to know, how many lazy, incompetent, drunk... engineers there are that won't do their job and use LTSpice to check out for an optimum part, and not be bribed by freebees.

In this world of 7 billion people, Mike is a big a nobody as anyone of us.

Interestingly I just tried in Google:

"spice" "kevin aylward" 4,180 hits

"spice" "mike engelhart" 1080 hits

:-)

And no, the only rock star in this NG, is the one in the purple shirt playing the blue guitar with cream fingerboard...

-- Kevin Aylward

- SuperSpice

But could you really sleep at night, knowing that you didn't even *try* to find another part? What if the engineer in the next cubicle decided to do so, and then showed to your boss that there was a better and cheaper part?

A key point is that LTSpice runs *any* PSpice/Spice3 model from any vendor, so its a pop to the competitions website to download their model and run it in LTSpice

The point being, LTSpice does not force, or ensure people use it to buy

*only* LT parts. Its just a bit more, in your face advertising, and that his limited value to people like engineers, who almost invariable evaluate based on objective values.

My estimate is that LTSpice is used to buy more of other vendors products, then their own. LT parts are expensive.

-- Kevin Aylward

- SuperSpice

Probably, but company bosses are not necessarily known to be good the people that actually make them their money.

"Talbert handling the fabrication and Widlar handling the design, they ruled the world and led the world in linear integrated circuits for a couple of decades."[

"Widlar and Talbert realized that the founders of Fairchild did not intend to share their windfall profits with the designers."

-- Kevin Aylward

- SuperSpice

Not at all

That is not the point. What part of "check out" and "exists" did you miss?

The only reason any competent engineer, that is designing for major production runs, is not going to the most basic of "does anther part even exist", is if he has no idea that an alternative might exist. i.e. he was a clueless engineer. Dah..Gee,. looks like I have to go with this..dah...

What I can say, is that if I were hiring engineers, hiring an engineer that is bribed by a freebe bit of kit from a vendor, such that he wont even attempt to find a more optimum part, would not be at the top of my resume list.

-- Kevin Aylward

- SuperSpice

That's why I wrote "better option"

Sure, no one has denied that, but are you, personably, so incompetent that you are not able to load in the competitions models form their websites in reasonable time?

I doubt it.

So, sure, its useful to have the model includes, but its is trivial detail, only of value to those engineers that probably would be better employed basket weaving.

I am not talking about the minor, one off, shit, lets just see if this works bit. I am referring to standard, professional long term product design, where one *IS* subjected to a whole host of constraints like, second sourcing, availability, cost, performance etc...

I just don't accept you achieve that by simply using parts, bribed to you, by their inclusion in a freebee spice program.

And illustrates part of the argument I am making.

-- Kevin Aylward

- SuperSpice

I don't drink much. A small beer, or maybe a rum+coke, two or three times a week.

Super Spice 53,000,000 hits, but not about circuit simulators.

LT Spice 747,000 hits, but on topic.

--

John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com

[snip] [snip]

Therein lies the rub... every frickin' manufacturer is heading toward encrypted models that only run on their own version of simulator, or, like Microchip, have PhD monkeys rolling out models using tables or IF statements that hardly ever converge. ...Jim Thompson

--
| James E.Thompson                                 |    mens     | 
| Analog Innovations                               |     et      | 
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    | 
| STV, Queen Creek, AZ 85142    Skype: skypeanalog |             | 
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  | 
| E-mail Icon at http://www.analog-innovations.com |    1962     | 

     Thinking outside the box... producing elegant solutions.

They probably have the notion that they'll lock you into using only their parts.

For me, it's more that they drive me away from using their parts, but hey, I'm a known weirdo.

--

Tim Wescott 
Wescott Design Services 
http://www.wescottdesign.com 

I'm looking for work -- see my website!

uhh... what other switched capacitor active filter chip even has a spice model? LTspice is also a behavioral simulator (why it's so fast for switchers) and LT took the time to write behavior simulations for almost all their analog chips. Other companies, except for standard parts like opamps, not so much. Not that I need specific models for simple stuff like that.

It is not trivial when under the gun. And I suck at making baskets.

My case was a one-off, true, but it was an extremely important one-off and the solution ended up working so well that if I did need something like that again I would not hesitate to reuse it.

Don't have to accept that because I don't, 99% of the stuff I use LTspice for does not involve using LT parts. I used that as an example of how it can matter - just because it was a one-off for me *that time* doesn't mean a similar problem is always a one-off, could have just as easily been for mass production.

The claim was that LT sells lots of chips by giving away free tools, you refuted that claim, I told a story about how once LTspice saved my bacon. I needed a filter with 0db at center more than 60db cut at half that frequency, LT gave me the tools to do that with very little effort (more time to design the rest of the stuff and write the code) and the end result worked perfectly. I did look at alternatives, LT's solution was the best by far. Sure it's not a cheap chip, but for what it did, the price wasn't bad.. ~$12 single for a precision

70db/octave filter with practically no ripple. How many I needed personally is not relevant, DK stocks thousands of them so someone is using the part. I have no doubt that LT sells oodles of parts because the tools they give away make the process of using the parts easy and relatively risk-free.

That said, I still wish it had a better GUI :-) When I get a chance I'm evaluating SS, it does seem to have better measurement options.. it's not either/or...

Terry

There was nothing wrong with the OrCAD DOS products - it's hard to think how the SDT and PCB products could be improved upon. When they moved to Windows - different story.

Kevin, you might be surprised at the number of analogue engineers who design by stitching together application notes. I would say it's the great majority. I never really appreciated the fact until I went freelance and was exposed to dozens of design teams, but the reality is that only a low percentage of the engineers doing board and system level design are competent in their craft.

What I usually see is that if an engineer manages to cobble together something in LtSpice (usually from the supplied examples) that meets the technical specs of whatever it is they've been tasked to do, then the design stage is over.