In Linux my calculator allows me to enter ** instead of ^. I think this is great feature and I hope some pedant never removes it as it makes using the calculator so much faster. I can do exponentiation all from the numpad. If I want a square root I can just do
10**0.5 = sqrt(10)
If I want to convert peak to RMS I just do
10V*2**-0.5 = 10V * 1/sqrt(2)
If I want to do the reciprocal, instead of 1/10 I do
10**-1 = 1/10
I no longer have to use so many bloody parenthesis or use the keyboard to do a square root.
Oh and also, to convert db to integer I do
1.122**-120db = 20 * log(-120db) (to the 2nd decimal place)
Since LTspice also uses **, I can move directly from the calculator to LTspice.
10**0.5 = sqrt(10)
If I want to convert peak to RMS I just do
10V*2**-0.5 = 10V * 1/sqrt(2)
If I want to do the reciprocal, instead of 1/10 I do
10**-1 = 1/10
I no longer have to use so many bloody parenthesis or use the keyboard to do a square root.
Oh and also, to convert db to integer I do
1.122**-120db = 20 * log(-120db) (to the 2nd decimal place)
Since LTspice also uses **, I can move directly from the calculator to LTspice.
Jan and I were approaching calculations from different angles. Here's a modification of Jan's to go with post 13 I think
SINE(0 0.1 {f} 0 0 0 {totalcyc})
.param dlycyc=920 numcyc=80 FFT=262144
.param f=4000
.param totalcyc=dlycyc+numcyc
.param simtime=numcyc/f+dlytime
.param dlytime=dlycyc/f
.param timestep=(simtime-dlytime)/FFT
.tran 0 {simtime} {dlytime} {timestep}
where if not obvious
totalcyc = cycle total
numcyc = cycles to sample
dlycyc = cycles before analysis
simtime = total sim time
timestep = time step
dlytime = time before analysis
Lokks about right?
O, is it impossible to delay .ac analysis sweeps? I'm having trouble with sims of a known working tube preamp sim
SINE(0 0.1 {f} 0 0 0 {totalcyc})
.param dlycyc=920 numcyc=80 FFT=262144
.param f=4000
.param totalcyc=dlycyc+numcyc
.param simtime=numcyc/f+dlytime
.param dlytime=dlycyc/f
.param timestep=(simtime-dlytime)/FFT
.tran 0 {simtime} {dlytime} {timestep}
where if not obvious
totalcyc = cycle total
numcyc = cycles to sample
dlycyc = cycles before analysis
simtime = total sim time
timestep = time step
dlytime = time before analysis
Lokks about right?
O, is it impossible to delay .ac analysis sweeps? I'm having trouble with sims of a known working tube preamp sim
There was some talk about LM317 regulators (in an other thread) that inspired me to make the attached simulation, if interested have a look 
See: https://www.diyaudio.com/forums/the...wtorch-preamplifier-iii-1754.html#post5769744
See: https://www.diyaudio.com/forums/the...wtorch-preamplifier-iii-1754.html#post5769744
Last edited:
What i was troubleshooting was just newbie stuff, and the classic EE issue of when one small thing is off everything is. So I isolated my ef86 ckt, made sure small signal inout was 1, put the ef86 duncan model in the asc file and bingo. finally some love. the step cmd is darn tooting' cool!
funny that the .include didn't work
funny that the .include didn't work
Attachments
Keep in mind Osvaldo, you may have to change all the "^" to "**" for LTSPICE. it took me a week of head banging to finally ask about that.
I really want to thank Molly and KEANTOKEN for giving a solid guide to do some fairly advanced things with just a few tutorials. I’m working on a power supply with the suggestions here and that mosfet with squarewave and it seems pretty unforgiving. I think if I had the circuit in the real world I would probably “get away” with designs that failed that test. I used a simple sine wave to speed up texts for a bit and some passed then failed with square wave. I learned the hard way that that I need to switch back and forth from AC to trans analysis. I got so tunnel vision with AC tests with impedance results that would never be stable over time.
Is there any way to test impedance in .trans mode?
Thanks again guys. You rock!
I really want to thank Molly and KEANTOKEN for giving a solid guide to do some fairly advanced things with just a few tutorials. I’m working on a power supply with the suggestions here and that mosfet with squarewave and it seems pretty unforgiving. I think if I had the circuit in the real world I would probably “get away” with designs that failed that test. I used a simple sine wave to speed up texts for a bit and some passed then failed with square wave. I learned the hard way that that I need to switch back and forth from AC to trans analysis. I got so tunnel vision with AC tests with impedance results that would never be stable over time.
Is there any way to test impedance in .trans mode?
Thanks again guys. You rock!
Exactly as in a real lab. (Assuming you mean output impedance of power supply).
Insert an ac current into the output, measure the ac voltage that results on the output. Then plot Vout/Iinsert, that's your Zout.
Jan