LTSpice Updates One File at a Time!

I never let LTspice do an update from the app itself after it stopped working several years ago. I'm sure it works again but not great as you are finding ?

I download the .exe install file from analog.com and let it overwrite (recommended method) and then I always have that particular installation exe file. This makes is easy to go backwards to an earlier version if necessary.

boB

I'm not worried about the .exe file being updated. If there's something in my circuit that depends on a particular version of LTspice, I'd rather know about it and figure out where the problem is, than to stick with an old version, possibly covering bugs in my design.

The problem I had was with updating the model files. I expect this was mostly a matter of new models being added. ADI has a lot of catch up to do in creating models for all their parts. It's great they are doing this and not ditching LTspice. I guess, for once, a company buying out another company sees the value in the things that were done there.

Anyone know if ADI is ditching much LT product? I know many here have talked about how LT was happy to keep old parts alive, even with very low volumes.

Good question.

I try not to use their parts, althought they are good, because they come with a high price, usually.

boB

Yeah, I remember a conversation here some years ago, where a few people who loved their parts didn't mind paying more, because the parts were so good and they didn't get obsoleted very often.

I was working on an open source, volunteer project a while back and a guy designed a power supply board. He obviously never saw an LT part he didn't like. I think the cost was two or three times what it could have been built for because of all the LT on board. They had a basic, buck switcher for $5. He justified it by claiming "they work". I guess that's ok if you have no price constraints. I did design in a National buck switcher once. It didn't work, I couldn't find anything in the data sheet that would explain it. National support got back to me and explained it was not designed for the voltage I was using it at (don't recall if it was the input or output voltage). I asked where that's specified in the data sheet and he said, "It's documented on the web page". Really? Data sheets aren't where you put such specifications???

The only other part I tried to use that was a total screw up was a CP Clare telephony CODEC with high voltage isolation using caps and a high frequency switching rate. They also had a companion chip to provide digital outputs to manage the different circuit details for different jurisdictions. The digital control chip was essentially an opto isolated shift register, but with no load signal. As you loaded the configuration, the bits shifted across the outputs and played all hell with the signalling, including the off hook control! The CODEC mostly worked ok, but had 0 dB PSRR. Yes, ZERO dB. I didn't realize that until the customer sent a unit back with a noise issue. The real problem was they were powering it through a lap set up with high impedance wiring, but every mV on the power rail (from the DSP looping) was conducted into the audio line. It took a couple hundred uF to deal with that. I've never seen an audio part with ZERO dB of power supply rejection. I have to spit when I say the name CP Clare.

It updates fine for me, but it thinks it owns the screen and you can't get other stuff done. I start updates and go away for an hour.

Mike says that LT Spice has 1000x the users of any paid-for circuit simulator. I believe that. It probably increased the sales price of LTC to ADI by a billion or two.

Some of their parts are ballpark competitive, RS485 transceivers for example. I was surprised.

This isn't the whole story. If you are designing for high volume production, price matters a lot.

If you are designing for small volume production, the cost of getting the design developed and working reliably can be higher than the cost of parts over the years that that design gets built and shipped. Conservative design with parts that are unlikely to give you trouble during development can save you more money by shortening the development cycle than you can ever save on the cost of the parts you end up shipping.

If the cost of development is your price constraint, it's the right choice to make. Volunteer projects do have to minimise the time spent on design, and they have no idea how many examples will end up getting built.