|
LTspice 24.1 is beta is available now:
NOTE: As requested by some in this group and elsewhere, the installer installs two executables: LTspice 24.1 Beta and LTspice 24.0.12. There will be icons for both on your desktop. Both LTspice 24.0.12 and 24.1 can be used on the same machine, but don't use both on the same schematics at the same time.
What’s New in LTspice 24.1 beta
- Improved performance and convergence
- Add .savestate and .loadstate simulator directives to save and restore the complete transient simulation condition
- Add .option debugtran to add report in log file to help identify convergence challenges
- Enhanced netlist syntax checking
- Add support for directory hierarchy for symbol search paths and schematic directory
- Add support for environment variables in Settings > Search Paths
- Add "All" symbol directories view in Place Component dialog
- Changed default integration method to Trapezoidal
- Remember waveform compression settings (under Settings > Compression) between LTspice invocations
- Always accept “3k4” notation as 3.4k (no longer optional)
- Allow device flags to be set to zero for ease of parameterization
- Updated FRA example fra_eg1
- Updated Help documentation
- Bug fixes
--
Michael Stokowski
LTspice Team
Analog Devices Inc.
|
On Wed, Nov 6, 2024 at 06:36 AM, mstokowski wrote: * *Add support for directory hierarchy for symbol search paths and schematic directory*
* *Add support for environment variables in Settings > Search Paths* Does this allow to get rid of the forced OneDrive rooted top-directory? -marcel
|
> Does this allow to get rid of the forced OneDrive rooted top-directory?
Nothing is forced to be kept on OneDrive. The top directory for the ADI symbols/models has been in %LOCALAPPDATA%\LTspice since the introduction of version 17.1.
Previously, in version XVII, the ADI symbol/model directory was in Documents\LTspiceXVII, and thus OneDrive. So if you still have that LTspiceXVII directory, it is left over from XVII.
LTspice 24.* look for some user files in Documents\LTspice by default, but that path can be changed in Settings > Search Paths.
-Ben
|
On Wed, Nov 6, 2024 at 10:36 AM, mstokowski wrote:
What’s New in LTspice 24.1 beta
- Improved performance and convergence
I noticed a regression while running the example circuit for the LT4295.
This example runs to completion in LTspice 24.0.12 with default settings, though it is slow (about 65 us/s on my machine) after the switcher starts switching.
This same example immediately goes to Def Con 2 at 489.119 ms and makes no further progress as far as I can see when using the LTspice 24.1 beta.
|
On Wed, Nov 6, 2024 at 10:36 AM, mstokowski wrote:
- Updated Help documentation
There is a bad link in the new Help for LTspice 24.1 beta.
On the page at file:///C:/Users/<user name>/AppData/Local/Programs/ADI/LTspice/LTspiceHelp_beta/settings.htm
the link labeled SPICE points to file:///C:/Users/<user name>/AppData/Local/Programs/ADI/LTspice/LTspiceHelp_beta/GeneralPurposeSchematicDrivenSPI.htm
rather than the page at file:///C:/Users/<user name>/AppData/Local/Programs/ADI/LTspice/LTspiceHelp_beta/spice.htm
|
On Wed, Nov 6, 2024 at 10:36 AM, mstokowski wrote:
- Updated Help documentation
The updated Help for LTspice 24.1 at file:///C:/Users/<user name>/AppData/Local/Programs/ADI/LTspice/LTspiceHelp_beta/spice.htm still shows and describes the option for "Accept 3K4 as 3.4K"
|
On Wed, Nov 6, 2024 at 10:36 AM, mstokowski wrote:
- Updated Help documentation
The new Help file at file:///C:/Users/<user name>/AppData/Local/Programs/ADI/LTspice/LTspiceHelp_beta/symbolandlibrarysearchpaths.htm has a typo at:
Symbol Search Paths
Additional search paths you would like to LTspice to use
"to LTspice to" should be "LTspice to"
|
On Wed, Nov 6, 2024 at 12:36 PM, mstokowski wrote:
- Changed default integration method to Trapezoidal
That is a noteworthy change, and somewhat puzzling. It means abandoning the Default "Modified Trap" method that was one of the hallmarks of LTspice for many years, one of the things that set it apart from every other SPICE program. And yet it could occasionally be more a hindrance than an improvement.
Me, I loved Modified Trap, and yet I switched my LTspice from Modified Trap to ordinary Trap a few years ago, but I keep a watchful eye out for signs of Trap Oscillation or Trap Ringing which is what Modified Trap eliminates obscures. I chose Trap instead of Modified Trap so that I can see every artifact that comes along.
Personally, I think I would have preferred having an .OPTION to select between Trap and Modified Trap. There is none, so the only way to change it was by going into Settings / Control Panel.
Another noteworthy change:
- Remember waveform compression settings (under Settings > Compression) between LTspice invocations
I find that somewhat troublesome, because simulations won't run the same every time you fire it up. I like it when you know exactly what you're getting. While it is true that waveform compression can be more trouble than it's worth, this makes controlling it somewhat random.
Regards,
Andy
|
On Thu, Nov 7, 2024 at 02:32 PM, Andy I wrote: I find that somewhat troublesome, because simulations won't run the same every
time you fire it up. I like it when you know exactly what you're getting.
While it is true that waveform compression can be more trouble than it's
worth, this makes controlling it somewhat random. Sometimes, waveform compression distorts in a way that is not immediately expected (a small straight piece in a nicely curved trace). It is a useful option when file size is really hurting, but nowadays, and in the future, who cares? -marcel PS: IMHO, compression is a less of a problem than modified trap gone. The standard TRAP may not look nice, but it shows the error is oscillating (so when you see it, it is certainly bad), while modified trap is purely cosmetic and suggests that the waveform is accurate everywhere and that there are no discontinuities.
|
On Wed, Nov 6, 2024 at 11:18 AM, Voegeli, Benjamin wrote: LTspice 24.* look for some user files in Documents\LTspice by default, but
that path can be changed in Settings > Search Paths. My problem was/is that the default could/can not be overruled, and so I constantly have to copy changed symbols and subs/libs. The change removes a tiresome problem when having more than one workplace. -marcel
|
On Thu, Nov 7, 2024 at 08:32 PM, Andy I wrote:
would have preferred having an .OPTION to select between Trap and Modified Trap.
Andy - Good point, there should be a .option to select modified trap. We will add support for ".options method modtrap" in the next revision.
The anti-trap ringing has caused a lot of confusion, as it is simply a post-processing that hides it from the user. For one, this can cause errors in circuits without trap ringing (see "Integration Methods" in the help if you're curious) and it can hide problems from the user, for example if there is trap ringing on the gate of a MOSFET causing it to switch on and off -- the user can't see the oscillation on the gate but they can see the troublesome downstream effects. Hence, modified trap remains available but is no longer the default.
-Ben
|
On Thu, Nov 7, 2024 at 08:32 PM, Andy I wrote:
- Remember waveform compression settings (under Settings > Compression) between LTspice invocations
I find that somewhat troublesome, because simulations won't run the same every time you fire it up. I like it when you know exactly what you're getting. While it is true that waveform compression can be more trouble than it's worth, this makes controlling it somewhat random.
Note that waveform compression settings are now remembered in version 24.1, for the exact reason of making it consistent every time LTspice opens. In previous versions, users would sometimes disable waveform compression, only to be surprised to find that it reset to enabled the next time they opened LTspice. This came up often. I'll also note that waveform compression can be configured with .options if you need to configure schematics a certain way. Nowadays, available storage space is sufficient for most simulations, so the motivation for compression is reduced, and problems due to its lossy nature are fairly common. Better to utilize a temporary directory for raw files, and limit saved signals and/or turn on compression when necessary.
-Ben
|
On Fri, Nov 8, 2024 at 01:41 AM, <mhx@...> wrote:
Sometimes, waveform compression distorts in a way that is not immediately
expected (a small straight piece in a nicely curved trace). It is a useful option
when file size is really hurting, but nowadays, and in the future, who cares?
Indeed, We frequently needed to remind LTspice users to turn off waveform compression, especially when they run those simulations where it makes a big difference.
I'll also note that Mike Engelhardt abandoned waveform compression in his new QSPICE program.
What bothers me (about Beta version 24.1) is that someone might enable or disable it and not be aware of it, and then that setting "sticks" until they change it back again. That strikes me as undesirable. (But you can't make everyone happy.)
PS: IMHO, compression is a less of a problem than modified trap gone. ...
Interesting, and not what I see. I see many more inquiries about compression artifacts, than those related to Modified Trap. And the ones about waveform compression tend to look more serious.
... The
standard TRAP may not look nice, but it shows the error is oscillating (so when
you see it, it is certainly bad), while modified trap is purely cosmetic and
suggests that the waveform is accurate everywhere and that there are no
discontinuities.
The other thing Modified Trap does, is to suggest that sharp waveform edges have intermediate values. For a long time, I (and others) thought that LTspice made sharp edges have two intermediate points, at about 25% and 75% of the new value, making them appear slightly rounded (S-shaped) instead of sudden or abrupt. It turns out that was entirely due to the Modified Trap post-processing.
Andy
|
On Fri, Nov 8, 2024 at 12:01 PM, Voegeli, Benjamin wrote:
On Thu, Nov 7, 2024 at 08:32 PM, Andy I wrote:
would have preferred having an .OPTION to select between Trap and Modified Trap.
Andy - Good point, there should be a .option to select modified trap. We will add support for ".options method modtrap" in the next revision.
If that is done, it needs the complementary command as well -- something to force non-modified Trapezoidal. Currently, ".options method=trap" gives you EITHER Trapezoidal or Modified Trap, whichever one had been selected already. But it provides no way to change from Modified Trap to regular Trap. You would need yet one more .OPTIONS choice.
The anti-trap ringing has caused a lot of confusion, as it is simply a post-processing that hides it from the user. ...
It can.
On a personal note, the Help page that describes it (added after Mike Engelhardt moved on) was extremely helpful. I am very grateful to whomever added that page.
I also never understood the change (circa 2019) to omit CERTAIN nodes from that post-processing when Modified Trap is selected. IMO, if Modified Trap is used, it ought to be uniformly applied to ALL nodes, not just 95% of them. (Perhaps it is worth revisiting that change?)
Andy
|
On Fri, Nov 8, 2024 at 12:17 PM, Voegeli, Benjamin wrote:
On Thu, Nov 7, 2024 at 08:32 PM, Andy I wrote:
- Remember waveform compression settings (under Settings > Compression) between LTspice invocations
I find that somewhat troublesome, because simulations won't run the same every time you fire it up. I like it when you know exactly what you're getting. While it is true that waveform compression can be more trouble than it's worth, this makes controlling it somewhat random.
Note that waveform compression settings are now remembered in version 24.1, for the exact reason of making it consistent every time LTspice opens. ...
But that might not happen!
Let's say I want to always not use waveform compression. So I set it up that way. Then, one day, I run a simulation that re-enables it. The next day, I resume my other simulations, and a week or a month or year later I discover - to my surprise - that LTspice has been using waveform compression ever since that one simulation weeks earlier.
That is not desirable behavior, in my opinion.
It could also go the other way, where I want to use waveform compression, but later I run a single simulation that disables it, and now I'm stuck with that until I notice, and change it back.
... Nowadays, available storage space is sufficient for most simulations, so the motivation for compression is reduced, ...
Agreed.
I suppose it's possible that enabling waveform compression MIGHT let some simulations run faster, by not having to save as much data to the drive, especially if it is a slower HD.
... and problems due to its lossy nature are fairly common. Better to utilize a temporary directory for raw files, and limit saved signals and/or turn on compression when necessary.
Definitely.
Andy
|
Add a flag box on the schematic page to show
the setting? Could work for other 'saved' settings, too.
On 2024-11-08 18:05, Andy I wrote:
On Fri, Nov 8, 2024 at 12:17 PM, Voegeli, Benjamin wrote:
On Thu, Nov 7, 2024 at 08:32 PM, Andy I wrote:
- Remember waveform compression settings (under
Settings > Compression) between LTspice
invocations
I find that somewhat troublesome, because simulations
won't run the same every time you fire it up. I like it
when you know exactly what you're getting. While it is
true that waveform compression can be more trouble than
it's worth, this makes controlling it somewhat random.
Note that waveform compression settings are now remembered
in version 24.1, for the exact reason of making it
consistent every time LTspice opens. ...
But that might not happen!
Let's say I want to always not use waveform compression. So
I set it up that way. Then, one day, I run a simulation that
re-enables it. The next day, I resume my other simulations, and
a week or a month or year later I discover - to my surprise -
that LTspice has been using waveform compression ever since that
one simulation weeks earlier.
That is not desirable behavior, in my opinion.
It could also go the other way, where I want to use waveform
compression, but later I run a single simulation that disables
it, and now I'm stuck with that until I notice, and change it
back.
... Nowadays, available storage space is sufficient for
most simulations, so the motivation for compression is
reduced, ...
Agreed.
I suppose it's possible that enabling waveform compression
MIGHT let some simulations run faster, by not having to save as
much data to the drive, especially if it is a slower HD.
... and problems due to its lossy nature are fairly
common. Better to utilize a temporary directory for raw
files, and limit saved signals and/or turn on compression
when necessary.
Definitely.
Andy
--
OOO - Own Opinions Only
Best Wishes
John Woodgate
Keep trying
|
Perhaps better added to the notification bar
at the foot of the screen.
On 2024-11-08 18:28, John Woodgate
wrote:
Add a flag box on the schematic page to show
the setting? Could work for other 'saved' settings, too.
On 2024-11-08 18:05, Andy I wrote:
On Fri, Nov 8, 2024 at 12:17 PM, Voegeli, Benjamin wrote:
On Thu, Nov 7, 2024 at 08:32 PM, Andy I wrote:
- Remember waveform compression settings
(under Settings > Compression) between
LTspice invocations
I find that somewhat troublesome, because
simulations won't run the same every time you fire it
up. I like it when you know exactly what you're
getting. While it is true that waveform compression
can be more trouble than it's worth, this makes
controlling it somewhat random.
Note that waveform compression settings are now remembered
in version 24.1, for the exact reason of making it
consistent every time LTspice opens. ...
But that might not happen!
Let's say I want to always not use waveform compression.
So I set it up that way. Then, one day, I run a simulation
that re-enables it. The next day, I resume my other
simulations, and a week or a month or year later I discover -
to my surprise - that LTspice has been using waveform
compression ever since that one simulation weeks earlier.
That is not desirable behavior, in my opinion.
It could also go the other way, where I want to use
waveform compression, but later I run a single simulation that
disables it, and now I'm stuck with that until I notice, and
change it back.
... Nowadays, available storage space is sufficient
for most simulations, so the motivation for compression is
reduced, ...
Agreed.
I suppose it's possible that enabling waveform compression
MIGHT let some simulations run faster, by not having to save
as much data to the drive, especially if it is a slower HD.
... and problems due to its lossy nature are fairly
common. Better to utilize a temporary directory for raw
files, and limit saved signals and/or turn on compression
when necessary.
Definitely.
Andy
--
OOO - Own Opinions Only
Best Wishes
John Woodgate
Keep trying
--
OOO - Own Opinions Only
Best Wishes
John Woodgate
Keep trying
|
John Woodgate wrote:
Perhaps better added to the notification bar at the foot of the screen.
Things are too easily missed when they are in the notification line at the foot of the screen. For years, LTspice announced the "Alternate" server at the bottom, yet I almost never saw it.
It would be better if it was flashing red or orange.
|
I agree about the use of colour, but I am not
so keen on continuous flashing. Perhaps flash for some seconds,
then not.
On 2024-11-08 18:57, Andy I via
groups.io wrote:
John Woodgate wrote:
Perhaps better added to
the notification bar at the foot of the screen.
Things are too easily missed when they are in the
notification line at the foot of the screen. For years, LTspice
announced the "Alternate" server at the bottom, yet I almost
never saw it.
It would be better if it was flashing red or orange.
--
OOO - Own Opinions Only
Best Wishes
John Woodgate
Keep trying
|
Perhaps it's not normal, but I prefer the old rule. I would rather
add directives on a schematic that requires it. I don't think I've
ever changed any of the compression settings from the Control Panel
from default since curiosity made me have a look just to see what
they did. Control Panel settings tend to be forgotten. I would like
every setting to be changeable by a schematic directive, but not all
are.
--
Regards,
Tony
On 08/11/2024 19:28, John Woodgate
wrote:
Add a
flag box on the schematic page to show the setting? Could work for
other 'saved' settings, too.
|
John Woodgate
Tony Casey