Anybody get spice to install on Linux?
in [Linux Help]
|
From: Ohmster on 23 Nov 2007 17:55 root <NoEMail(a)home.org> wrote in news:2Fy1j.5$iX1.2(a)newsfe02.lga: > Ohmster <root(a)dev.nul.invalid> wrote: >> I used to have a program called "Electronic Workshop" and it was >> great. You could put transistors, resistors, and all kinds of >> electronic parts on a board and then "run" the circuit to see how it >> would behave. It seems that some other fellow was so impressed with >> Electronic Workshop that he started a thing called "spice" which is >> the same thing only free for Linux. I downloaded the tarball, >> extracted it, did the ./configure, the make, moved all docs to >> "doc-pak" and then ran checkinstall "checkinstall �R make install". >> The thing failed miserably, were at make(5) by the time the thing >> gave up. [..] > > I see that you might have a working solution from the > other responses to your post. Your original problem sounds > like a problem with checkinstall. If the original make > worked then you should be able to install via make install. > Checkinstall had/has problems with some versions. I think > the problem is with recent versions of tar. No, I had a lot of errors during the make. They tend to number as the errors accumulate, I was up to like (5) I think. It was like the "make" did not make at all, at least I did not see anything to indicate that anything was made. No new binaries created or anything like that. I was not expecting checkinstall to work after that but gave it a shot anyway, no dice. Most of the errors were like "Entering so so directory, so and so not found, nothing to do, exiting" and that sort of thing. From my experience, the dreaded tarballs are alright for fairly simple programs like iftop, but when you start making larger programs that require tons of specific libraries, they often fail. Either I don't have the library on my system or the application was written only for a specific version of the lib. Also the file system layout maybe differrent. Fedora really does not use the /opt directory but many of these programs do and they expect to find certain things there. I really like the rpm files with Yuma because Yuma will find the package, install it, find out of you have any conflicting libs or programs that have to go or be updated, find what dependencies you need and include them in the install process, and all you have to do is agree, yes or no. These tarballs give no such help, they just fail with cryptic messages that could take weeks to either fix or find out that you have to give up. I hate that and tend to shy away from tarballs for that reason. I would have liked to try spice too but the nice fellow, John Hasler, gave me a whole list of similar stuff to try and I first tried gnucap at the top of the list but found it was a CLI program and I had not idea of how to work it, then found there is a gui front end for it called gspiceui but could not find an rpm to install it with and gave up. I then found qucs on the web, that is on John's list, and it works pretty good. I tried the win32 binaries first on my XP machine (Has a bigger monitor and faster processor) and it works quite well. What I liked about Electronic Workshop is that you could build a circuit or load a sample, then "run" it and put a scope wherever you wanted and see the output, put a meter where you want and read the voltage and current, and it was all really neat so you could print a parts list and actually build the thing you dreamed up. qucs don't really seem to have bench tools that you can connect to the circuit to take readings and I miss that a lot. For example, you build a one transistor amplifier, common emitter, then inject a 20mV p-p sinewave, 1Khz signal into the base and see the scope displaying a 2 Volt p-p 1Khz signwave on the collector, resulting in a gain of 100 . Very useful. Then you could change the frequency of the input to 100Hz or 20Kz and see what the gain is and where it falls off so that you could ad, remove, or adjust the circuit to give you the spectrum you needed and could then build the thing and it would work. qucs will give you a simulate but then offer up a graph and a chart, showing how it did. Not bad I guess, I could use that, but it would take some getting used to. I cannot (so far as I an tell) put a voltmeter on a component and watch it, just have to specify the generator parameters and then go over the output report. Qucs is free on win32 as well so I will play with it here for a while and see if it meets my needs. Or else, off to delete qucs and try something else. John was very good with supplying me such a vast list to try and so I will be busy with it for a while. Cheers, -- ~Ohmster | ohmster /a/t/ ohmster dot com Put "messageforohmster" in message body (That is Message Body, not Subject!) to pass my spam filter. |