Installing and using LTspice IV (now including LTXVII). From beginner to advanced.

It was pointed out to me that my DC voltage source was upside down. Kind of like, "Sir, your fly is down." I found the problem. I had put the numdgt=12 and plotwinsize=0 in the transient command which Spice didn't like. When I removed it the graph came back. I've attached two more JPEG's. "FFT With" is with a specified time range.
 

Attachments

  • FFT without.jpg
    FFT without.jpg
    163.2 KB · Views: 171
  • FFT with.jpg
    FFT with.jpg
    163.1 KB · Views: 166
Hey Guys, on the question of a constant current source. I've read that a CCS has a high rate of ripple rejection and therefore has a high impedance to an AC signal. In Spice does the current object have the same effect? The reason I ask is that in the preamp when I change the value of the plate resistor for the second stage the gain changes. But if the CCS were actually a very high AC impedance it shouldn't should it? Does Spice see the CCS as high impedance to AC?
 
AX tech editor
Joined 2002
Paid Member
Thanks guys, your advice helped a lot. Again. I'm posting an FFT of the preamp design. While my question isn't specifically a Spice question it arises from Spice and should be easy to answer. The run is set at .5 volt input at 1k with a 30 volt output swing. My question is this, there is a large bump in distortion at 140k. Is that distortion parasitics? And is there way to eliminate it? Or does it even need to be eliminated?

You missed this?

http://www.diyaudio.com/forums/soft...ice-iv-beginner-advanced-106.html#post5353280

Jan
 
I try to model a unipolair class d amp, but it does not what I want, strange output signals for example and uge currents. looks like fases are wrong, see picture 1 and 2 for comparison. such a simple topologie and not see what is wrong, I shame myself.

maybe someone sees the problem and give hint, building schematic as with the paper, did not work, something needs to me done..

Maybe when I get tip, I see it, can be a slight small error, who most of the time is not seen. but I think it is signal input fases is wrong, putting the sinusoidal input not 180 Degree but in fase into the two comparators I get the signal you can see on pic 1 who is the right one.

i did see also some pieces where two in fase input sinusoidals but with one plus and one minus offset and placed inside a triangle is also used. mucho possibillities, but can it be simmed.

thanks.
 

Attachments

  • ScreenHunter_882 Mar. 01 20.21.jpg
    ScreenHunter_882 Mar. 01 20.21.jpg
    161.9 KB · Views: 258
  • ScreenHunter_883 Mar. 01 20.50.jpg
    ScreenHunter_883 Mar. 01 20.50.jpg
    387.1 KB · Views: 253
  • ScreenHunter_885 Mar. 01 20.55.jpg
    ScreenHunter_885 Mar. 01 20.55.jpg
    118.5 KB · Views: 263
A transient plot is showing the instantaneous value of the waveform at a point in time. The RMS (root mean square) value represents the effective value of the waveform over the course of one complete cycle. A sine wave which is symmetrical above and below the X-axis will have an average value of zero, but such a sine wave voltage will still dissipate power in a resistor that is placed across it, just as though there was a DC voltage across that resistor. The RMS value of the AC voltage can be thought of as the equivalent DC value for the waveform. For a sine wave, the RMS value is the peak value divided by the square root of two (or the peak value multiplied by 0.707). You can determine the peak value of the waveform from the graph.

I should add that if there is a DC component to the sine wave (i.e., if it is not centered about the X-axis), that DC component will also add to the dissipation of a resistor placed across that voltage source.
 
Last edited:
P/N FET models Rd and Rs

I hope someone else has noticed that the Rd and Rs model parameters seem to have been swapped. The following PJF and NJF models give exactly the same but complementary characteristics. This is LTSpice from a few gens back maybe it has been fixed. The errors are large for high Idss FET's.

.model NCH NJF(Beta=50.0m Vto=-.50 Lambda=37.300E-3 Rs=7.5000 Rd=1.0)

.model PCH PJF(Beta=50.0m Vto=-.50 Lambda=37.300E-3 Rd=7.5000 Rs=1.0)
 
LTspice XVII disabling waveform "drag" with Ctrl depressed?

I went ahead and installed the latest LTspice ("XVII") last night. For the most part, it looks and feels pretty much as the IV version.

One very irritating "feature" i've stumbled across, though, is that whenever the waveform plot window is highlighted, my mouse cursor is within its confines and i press down Ctrl (in preparation to click on the name of a waveform to have it measured), it drags the plots along :headbash:

Then after i notice that, i have to let go of Ctrl and wait for the waveforms to get re-drawn, then re-zoom, then place the mouse cursor on the waveform name before pressing Ctrl then clicking on it. :redhot: :no:

The older version doesn't behave like that. Anyone have any idea if and how this can be disabled? :grumpy:
 
Administrator
Joined 2007
Paid Member
I don't know of a way to disable the plot window from being dragged around but in reality its just 'different' rather than a problem imo. You soon get used to not pressing Ctrl until the cursor is clear of the viewing area.

Scott... I'm only running LTXVII these days but enlighten me :) Where are the default (as in no type number) models located ? I can't seem to find them in LTXVII.