I used something called "Electronic work bench" for a while. (not good at all) and then some spice thing that was part of a pcb layout package... Maybe Boardmaker?? Anyway for me LTspice was a big step up. Perhaps the previous were the equivalent of a golf cart. Which makes a Citroen look speedy.
Thing is, many people did not live a sheltered life yet they like LTSpice.
Then they encounter what just happened to a neighbor when he wanted to drive to work. The engine sounded like an M18 Hellcat tank and a red light came on. Not the foggiest what the code could mean. The shop revealed that it was a really bad one. Bad for the bank account. He had to let a nice truck go.
formatting link
When I had my first "real" car after the 2CV rusted out from underneath me I was taken aback.
Why can't any of the "modern" cars achieve the 50mpg my 2CV could on regular unleaded? Well, now some can but it took them more than another half century to get there.
Why can't the seats be taken out in seconds at the beach or to play a card game on the plugged up autobahn like we used to do?
Why can't I transport a large upright fridge with the new car?
Why can't I lift and lock the suspension at full upper peg with a piece of wood to get across a muddy field?
Why can't I take off a front fender for service where that only took the stock crank and 10 seconds on the Citroen?
And so on. Andre Citroen and his engineers were among the smartest that ever lived. Quality, however, well, let's not go there ...
When I get more current out of the output of an opamp than current into the power pin, I get pretty leery of the whole thing. Or transistor models working "normally" at absurd extremes. I've seen too many of these sorts of problems to trust any models past pretty trivial sims. I never had any such problems with chip-level sims. If the models weren't correct, no chip would work.
This is different. Say you want to drive an actuator with a switcher chip and the drive is highly nonlinear but must be zippy. If you can't have access to the innards of the loop sections in the chip you can't do it. LTC gives you that inside in a way that it can be simulated fast. Some (few) others give you a complete model but then every simulation run takes hours or days. Most manufacturers give you ... nothing, and then it can't be done with sufficient confidence. Just because it works on the bench does not mean you have an ECO-releasable design. This is where LTSpice comes in a lot for me and afterwards LTC enjoys extra sales.
Companies do not hire consultants to do jobs that can be done by more or less copying an app note solution. Most hire them when they run out of ideas on how to proceed. I make my living that way.
This is where their excellent support comes in. When doing something unorthodox I ask them. Sometimes they then have to defer to their IC designers but I get my answer. The engineer who supported me over 20 years ago is still there. Try that with another company.
LTC products are too expenisive for most mass production projects but for tough mission-critical stuff they are a good source.
It does when the simulator don't have scripting :-(
Cadence LG is quite user friendly, when it works. Place the voltage source, and press the button. However, it has a basic fault, which I am assured their team is now fixing.
The dumb $hits had this:
"unable to calculate loop phase margin and gain margin because the system is unstable". Dah...
As in , when wanting to know how much extra gain one has in an *oscillator*, despite the LG function working, the software says, I am just not going to tell you what the result is.
They had no idea that that knowing the gain for an unstable system is of fundamental importance in designing oscillators. This is 2017. It is simply staggering how they can miss something as trivial as this.
I had to write a script to calculate the data from the LG data that was actually available.
The also said that were were going to implement another feature request of mine.
In designing 3rd overtone oscillators, one needs to know what the gain is large enough, say +6dB at the "C-Mode" desired frequency, and that the gain at the "B-Mode" (10% close) is below 0dB, say -6dB.
So, one needs to pop out the loop gain at all 0 deg, 360 deg, 720 degs etc. Sometimes there are multiple 0 deg phase points, only one wanted.
I have manually wrote scripts that do this, but its a pain.
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.