On Sat, 04 Mar 2017 09:23:27 +0000, Kevin Aylward wrote:
> As far as development in XSpice, if anyone wants my MS VC++ full code of
> my XSpice version for free, just ask. I consider my value added is my
> GUI.
That's my main bitch about LTSpice: the GUI. The circuits look butt-ugly
IMO. I did test a lot of other flavours of Spice out back in the day and
they nearly ALL had much nicer user interfaces than LT; some really
pretty. Still, you can't complain too much about something that works
pretty damn well and costs nothing!
> "Gerhard Hoffmann" wrote in message
> news:ehukgkFsp9lU1@mid.individual.net...
>
> Am 03.03.2017 um 21:28 schrieb Kevin Aylward:
>> "Jim Thompson" wrote in message
>
>>
>> Legal disclaimer... personal opinion follows...
>>
>>>> LT management must have their heads up their ass... more and more of
>>>> their models will run ONLY on LTspice, some because of encryption,
>>>> some because of proprietary functions not found in any other Spice
>>>> variant.
>>
>>> Agreed.
>
>>As much as I like LTspice, and I use it a lot, methinks in the long run
>>we do ourselves a disservice with it.
>
>>Everybody attaches a small rucksack to the original Spice, renames it
>>somewhat and then sleeps on it like the dragon on his pile of gold.
>>And no one but LT makes any impact in user land, no one gets rich, and
>>no one can collect the good ideas and can concentrate them in one open
>>version so the next bright guy can build upon that.
>
>>LTspice leeches the blood out of any other development.
>
> I agree. There is next to no incentive for any company to develop a
> spice for non ic design applications.
>
>>Maybe I should move on to NGspice or Xspice or QUCS, as an investment
>>into the future.
>
> NGSpice is a very good choice. It includes XSpice anyway. Its got
> quite a lot of good new stuff. There are even smutterings of adding
> PSS.
I believe KiCad now integrates NGSpice.
> I wrote SS to give me personally, in part, affordable ic design
> features. Worst Case corner support on its own, for me, is
> justification to never use LTSpice. Nothing I do has any meaning
> without ensuring it will work over all process corners.
>
> As far as development in XSpice, if anyone wants my MS VC++ full code
> of my XSpice version for free, just ask. I consider my value added is
> my GUI.
>
> -- Kevin Aylward
> http://www.anasoft.co.uk - SuperSpice
> http://www.kevinaylward.co.uk/ee/index.html
>
--
John Devereux
Reply by Kevin Aylward●March 4, 20172017-03-04
"Gerhard Hoffmann" wrote in message
news:ehukgkFsp9lU1@mid.individual.net...
Am 03.03.2017 um 21:28 schrieb Kevin Aylward:
> "Jim Thompson" wrote in message
>
> Legal disclaimer... personal opinion follows...
>
>>> LT management must have their heads up their ass... more and more of
>>> their models will run ONLY on LTspice, some because of encryption,
>>> some because of proprietary functions not found in any other Spice
>>> variant.
>
>> Agreed.
>As much as I like LTspice, and I use it a lot, methinks in the long run
>we do ourselves a disservice with it.
>Everybody attaches a small rucksack to the original Spice, renames it
>somewhat and then sleeps on it like the dragon on his pile of gold.
>And no one but LT makes any impact in user land, no one gets rich, and
>no one can collect the good ideas and can concentrate them in one open
>version so the next bright guy can build upon that.
>LTspice leeches the blood out of any other development.
I agree. There is next to no incentive for any company to develop a spice
for non ic design applications.
>Maybe I should move on to NGspice or Xspice or QUCS, as an investment
>into the future.
NGSpice is a very good choice. It includes XSpice anyway. Its got quite a
lot of good new stuff. There are even smutterings of adding PSS.
I wrote SS to give me personally, in part, affordable ic design features.
Worst Case corner support on its own, for me, is justification to never use
LTSpice. Nothing I do has any meaning without ensuring it will work over all
process corners.
As far as development in XSpice, if anyone wants my MS VC++ full code of my
XSpice version for free, just ask. I consider my value added is my GUI.
-- Kevin Aylward
http://www.anasoft.co.uk - SuperSpice
http://www.kevinaylward.co.uk/ee/index.html
Reply by Gerhard Hoffmann●March 3, 20172017-03-03
Am 03.03.2017 um 21:28 schrieb Kevin Aylward:
> "Jim Thompson" wrote in message
>
> Legal disclaimer... personal opinion follows...
>
>> LT management must have their heads up their ass... more and more of
>> their models will run ONLY on LTspice, some because of encryption,
>> some because of proprietary functions not found in any other Spice
>> variant.
>
> Agreed.
As much as I like LTspice, and I use it a lot, methinks in the long run
we do ourselves a disservice with it.
Everybody attaches a small rucksack to the original Spice, renames it
somewhat and then sleeps on it like the dragon on his pile of gold.
And no one but LT makes any impact in user land, no one gets rich, and
no one can collect the good ideas and can concentrate them in one open
version so the next bright guy can build upon that.
LTspice leeches the blood out of any other development.
Maybe I should move on to NGspice or Xspice or QUCS, as an investment
into the future.
There is enough to do: s-parameters, large signal noise analysis,
contemporary transistor models, electromagnetics, strip lines,
harmonic balance, functional blocks, using CUDA etc. There is so
much that AWR, Genesys, ADS & friends can do and that Spice cannot.
have a good night,
Gerhard
Reply by Kevin Aylward●March 3, 20172017-03-03
"Jim Thompson" wrote in message
news:9c3jbcpcn0mrtlumcl82akd1l24bn4huts@4ax.com...
On Fri, 3 Mar 2017 14:51:37 -0000, "Kevin Aylward"
<kevinRemovAT@kevinaylward.co.uk> wrote:
[snip]
>
>I did notice something a bit strange with the LT user licence. It only
>prohibits use by "...semiconductor manufactures..." .
>
>>This seems a bit of a cockup on their part. Surely one would expect that
>>they would prohibit fabless semiconductor companies or individuals using
>>it
>>to design competing chips, much as what you are now at liberty to do :-)
>
>
Legal disclaimer... personal opinion follows...
>LT management must have their heads up their ass... more and more of
>their models will run ONLY on LTspice, some because of encryption,
>some because of proprietary functions not found in any other Spice
>variant.
On Fri, 3 Mar 2017 14:51:37 -0000, "Kevin Aylward"
<kevinRemovAT@kevinaylward.co.uk> wrote:
[snip]
>
>I did notice something a bit strange with the LT user licence. It only
>prohibits use by "...semiconductor manufactures..." .
>
>This seems a bit of a cockup on their part. Surely one would expect that
>they would prohibit fabless semiconductor companies or individuals using it
>to design competing chips, much as what you are now at liberty to do :-)
>
>
>-- Kevin Aylward
>http://www.anasoft.co.uk - SuperSpice
>http://www.kevinaylward.co.uk/ee/index.html
Legal disclaimer... personal opinion follows...
LT management must have their heads up their ass... more and more of
their models will run ONLY on LTspice, some because of encryption,
some because of proprietary functions not found in any other Spice
variant.
I guess they figure to sell only to amateurs ?>:-}
...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.
Reply by Kevin Aylward●March 3, 20172017-03-03
"Jim Thompson" wrote in message
news:3eihbclmf89nltv6b4hajo19ot4jdb30nt@4ax.com...
>
>>
>>>Spice in general isn't that wonderful if you don't have foundry models.
>>>It's good enough to be useful, at least for discrete designs and LTC
>>>switchers, and it produces pictures for customers to >show to their
>>>bosses
>>>and for posting on SED. Fortunately most of the stuff I need to simulate
>>>is discrete.
>
>>And I have foundry models... this latest job is UMC, 150nm ;-)
>
>Yeah, it is a real problem without good foundry models.
>
>I have collected quite a few sets now for all the processes I have used in
>my day job. Not something I can pass out with SS though.
>Same here. I have literally libraries for at least 60 process
>variations.
I did notice something a bit strange with the LT user licence. It only
prohibits use by "...semiconductor manufactures..." .
This seems a bit of a cockup on their part. Surely one would expect that
they would prohibit fabless semiconductor companies or individuals using it
to design competing chips, much as what you are now at liberty to do :-)
-- Kevin Aylward
http://www.anasoft.co.uk - SuperSpice
http://www.kevinaylward.co.uk/ee/index.html
Reply by Kevin Aylward●March 3, 20172017-03-03
"Jim Thompson" wrote in message
news:3aihbcllbvsfkm1odhoh8q8r4b8rs8chpj@4ax.com...
On Thu, 2 Mar 2017 16:26:52 -0500, Phil Hobbs
<pcdhSpamMeSenseless@electrooptical.net> wrote:
>On 03/02/2017 09:54 AM, Jim Thompson wrote:
>> On Thu, 2 Mar 2017 05:10:59 -0800 (PST), pcdhobbs@gmail.com wrote:
>>
>>>> And I have foundry models... this latest job is UMC, 150nm ;-)
>>>
>>> You can do 150 nm litho with a crayon.
>>>
>>> Cheers
>>>
>>> Phil Hobbs
>>
>> 150 NANO-meters ??
>>
>> ...Jim Thompson
>>
>Last process I worked on was 22nm, and that makes me a grandpa. ;)
>
>My nanoantennas were made on a 30-nm Leica e-beam writer, and that was
>getting on for a decade ago now.
>
>Cheers
>
>Phil Hobbs
>Small feature size is useless for precision Analog... VOS =
>12/sqrt(W*L) (for my current process), so it goes down when area goes
>up.
Exactly. Also, 1/f noise does the same.
Currently using 0u18 process and I rarely use less than about 1u gate length
for the analog bits. Typically, I use 0.18u for things like the output
device of an LDO, where it don't matter, and you want it to be small as
possible.
-- Kevin Aylward
http://www.anasoft.co.uk - SuperSpice
http://www.kevinaylward.co.uk/ee/index.html
Reply by ●March 2, 20172017-03-02
>>> 150 NANO-meters ??
>>>
>>>
>>Last process I worked on was 22nm, and that makes me a grandpa. ;)
>
>>My nanoantennas were made on a 30-nm Leica ebeam writer, and that was
>>getting on for a decade ago now.
>
>Small feature size is useless for precision Analog... VOS =
>12/sqrt(W*L) (for my current process), so it goes down when area goes
>up.
Transistors stopped getting better at about the 65 nm node. Now the game is to use more transistors to counteract the dopant atom statistics (which is probably what makes Vos hard to control). At 7nm, even logic levels aren't that easy to control.
;)
Cheers
Phil Hobbs
Reply by Jim Thompson●March 2, 20172017-03-02
On Thu, 02 Mar 2017 18:41:03 -0700, Jim Thompson
<To-Email-Use-The-Envelope-Icon@On-My-Web-Site.com> wrote:
>On Thu, 2 Mar 2017 16:26:52 -0500, Phil Hobbs
><pcdhSpamMeSenseless@electrooptical.net> wrote:
>
>>On 03/02/2017 09:54 AM, Jim Thompson wrote:
>>> On Thu, 2 Mar 2017 05:10:59 -0800 (PST), pcdhobbs@gmail.com wrote:
>>>
>>>>> And I have foundry models... this latest job is UMC, 150nm ;-)
>>>>
>>>> You can do 150 nm litho with a crayon.
>>>>
>>>> Cheers
>>>>
>>>> Phil Hobbs
>>>
>>> 150 NANO-meters ??
>>>
>>> ...Jim Thompson
>>>
>>Last process I worked on was 22nm, and that makes me a grandpa. ;)
>>
>>My nanoantennas were made on a 30-nm Leica e-beam writer, and that was
>>getting on for a decade ago now.
>>
>>Cheers
>>
>>Phil Hobbs
>
>Small feature size is useless for precision Analog... VOS =
>12/sqrt(W*L) (for my current process), so it goes down when area goes
>up.
>
> ...Jim Thompson
Dimensions are mv/um ;-)
...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.