Spice simulation

[klewis]

Originally Posted by akis

You drive around with your foot on the brake pedal until the brakes feel like mush and there is smoke and a very bad fishy smell coming out of the wheel arches :-)

I think they smell more like burnt hair than fish.

[akis]

Ahhh I see. I do hope to be able to find some time to do some serious testing this weekend on a variety of trnasistors and then I will report here. But my bench supply only goes up to 2 A :-)

[Terry Demol]

Originally Posted by akis

Ahhh I see. I do hope to be able to find some time to do some serious testing this weekend on a variety of trnasistors and then I will report here. But my bench supply only goes up to 2 A :-)

akis,

I checked a bunch of my bjt models for ft and Hfe vs Vce and Ic during the
ft discussion. Most of the ft's were pretty close but overall the hfe versus
vce is pretty poorly represented. This was most prominent with high volatge
devices, at much higher vce's, say >100V, hfe was much higher than I've
ever measured.

At the end of the day, I think it pays to probe through your simmed circuit
and get a feel for how sensitive the results, for example distortion, are to
hfe variation of various devices.

Still I agree with you, it would be nice to have more accurate models.

I'd be interested in your results and thanks for the effort.

T

[PB2]

Originally Posted by syn08

There is none that I am aware off. The closest that I have used is Keithley's LabTracer software and Series 2600, Series 2400, and Model 6430 SourceMeters. Expensive stuff...

Wondering, syn08, if you have any experience with the Agilent 4156C; any cautions, issues, etc. with the unit?
Agilent | 4156C Precision Semiconductor Parameter Analyzer

[laid]

Hi all!

After smashing my head against the wall for quite some time, I have been finally able to get the Altium Designer working correctly with Koren tube models.
As this has been quit a hazzle, I put together a how-to and included some actual simulation files as well.
Please feel free to send me comments and corrections, so I can acommodate all the wisdom to the document!

The article itself can be found at

Modeling tube circuits with Altium Designer.pdf

Regards,

Andrus.

[fotios]

In all Spice models of MJE182 that i have download from On Semi site (PSpice, Saber, Spice2, Spice3) its parameter "Saturation Current" is given as "IS = 1e-17". When i put this value in the corresponding box of parameter table of the EDSpice Simulator that i use, is translated automatically (and correctly) in IS = 0.01 femto Amperes. This value in my eyes (i am not so expertized like you in Spice Simulation) seems to be very small for a transistor with Ic=3A. For example, similar NPN transistors like BD137 with Ic=1.5A has IS = 48.15 femto Amperes. And the complementary PNP MJE172 has IS = 53.76 femto Amperes.
Can you help me please?
Thanks

Fotios

[sawreyrw]

Fotios,

As a rule of thumb Is=2E-15*Ic_max. So 1E-17 is way too low. It's too bad, but you can find a lot of SPICE transistor models that aren't very good.

Rick

[keantoken]

For those who can figure out how to use it, SpiceMod can be downloaded here:

Steve Ekblad's Free Audio Software and On-Line Enclosure Design

Found it by accident on Google.

Many models available from the popular vendor sites sport a SpiceMod header, but I'm guessing the models are all made in batches and so an incorrectly used or configured SpiceMod spits out almost entirely wrong models. Most models in my experience don't give the device any justice and, stranger, I've seen the exact same models used for parts that are not alike at all.

- keantoken

[keantoken]

On the conversation about transistor testing:

With some work, I'm sure we could come up with a circuit designed to keep a transistor at constant dissipation/temperature while ramping up whichever parameters we're testing.

So say we want to extract an Hfe vs. Ic curve for a given device. We want to keep the device at a constant temperature so we decide to keep the dissipation at a steady 10W. Vce for the test will be 10V.

Average Ic for this would be 1A. So if we send the test current in pulses, with a duty cycle varying so as to keep an average of 1A, we will keep an average dissipation of 10W. The higher the frequency the better.

Duty cycle=Iavg/Itest

Are there any outstanding issues with this approach? The device should probably be left for a warm up period so temperature can stabilize.

- keantoken

[unclejed613]

tnx for the link to audio software. i'm running linux, so i'll report back and let you know how the programs run under dosbox or wine.