ostripper said:
Hi OS,
The issue of spectrum analysis in general using the FFT is a complex one, and much of that subject is out of my area. In addition, the LTspice FFT capability is not user-friendly. But if your interest is just to simulate distortion, you might have a look at this series of posts.
In short, the easiest way to do this is the following:
Put the directive
.options plotwinsize 0
on the schematic to disable data compression, which messes up the distortion figures. Then put a directive something like this on the schematic:
.four 20kHz 19 v(out)
This assumes the output signal is on a node named "out", your input sine wave signal is 20 kHz, and you wish to include 19 harmonics in the calculation (overkill). ".four" is for "Fourier". To see the results, do a "view, SPICE error log" after the transient simulation and you'll see them in text form. It's kind of a kluge, but it's the easiest way.
Also, you'll need to mess around with the maximum time step allowed in the transient sim. To small and the sim runs extremely slow. Too large will cause the simulated distortion to be large. For further info, see the ".four" documentation under "dot commands" in the LTspice docs.
A little more "pain for the brain" Thanx..Had a little problem with frugal 2 when overdriven, but solved
it. "Sticking" and asymmetrical clipping were quite evident
with the 2sa1381/3503 unless current was dropped to 2ma..
but those good old mje340/50,s saved the day...
An externally hosted image should be here but it was not working when we last tested it.
---20k w/3v input note note symmetrical clip and no "sticking"
on bottom of wave----
The frugal 2 had such high OLG anyway (80DB) the "OLD"
MJE devices had no problem with it. 1 added resistor (will
update) and the widlar buffer had to go (saturated the 2sc3503
hard) being replaced by a standard CM.
The 2 VAS loading resistors (47k) help as well to "suck" the charge out of the VAS CM tranny after clipping. Just had to address this
issue as no doubt somebody (especially) me will abuse
this amp.
OS
You might want to have a look at how Bob Cordell solved this problem in his amp, whose front end has a similar topology to FA. That article can be found here. See figure 8 for details.
The 1381/3503 models I've seen take into account "quasi-saturation" (soft saturation instead of the "hard knee" in the output curves), so these will likely be more accurate to real-world behavior than models that do not. The 340/350 data sheets I've seen leave almost all parameters needed to get SPICE models unspecified - so that's something to consider.
The 1381/3503 models I've seen take into account "quasi-saturation" (soft saturation instead of the "hard knee" in the output curves), so these will likely be more accurate to real-world behavior than models that do not. The 340/350 data sheets I've seen leave almost all parameters needed to get SPICE models unspecified - so that's something to consider.
Seconded - all the MJE340/350 models i have seen never simulate this sticking behaviour, and I'm more inclined to belive it is simply poor modelling than an indicator of what the devices themselves would do. If you are using the KSA1381/KSC3503 models from Fairchild, they are very accurate.
To be honest, I wouldn't worry about clipping unless it sticks hard and latches up, or inverts, or something very nasty. If your amp is clipping - turn the volume down!!
Alternatively you could put a detection circuit in for this, and if the amp keeps clipping hard, treat it as a fault and shut off the output relay.
To be honest, I wouldn't worry about clipping unless it sticks hard and latches up, or inverts, or something very nasty. If your amp is clipping - turn the volume down!!
Alternatively you could put a detection circuit in for this, and if the amp keeps clipping hard, treat it as a fault and shut off the output relay.
I do not use MJE340/350.
BD139/140 is an improvement from those.
If I need more voltage/power than BD139/140 can take
next step up is MJE15030/31.
Nothing Exotic or Fancy
I know .. but this works!
I have seen MJE340/350 used as cascoding transistor, at times.
Is this a good use of MJE340 and MJE350?
BD139/140 is an improvement from those.
If I need more voltage/power than BD139/140 can take
next step up is MJE15030/31.
Nothing Exotic or Fancy
I know .. but this works!I have seen MJE340/350 used as cascoding transistor, at times.
Is this a good use of MJE340 and MJE350?
This is quite evident, as I was "fixing" the vas, I could notice
varying degrees of saturation in both the OP waveform
and in the current pulses within the VAS itself.
Twice as much data in the .model file.. "soft knee"
I know mje340/50 models and datasheet suck..
They are good for current sources and Vbias use..
I did get a better Sim (amp) by all this fussin'. Here is what I did..
An externally hosted image should be here but it was not working when we last tested it.
The main changes are D8,c17/c18. D8 I found in a 90's "HH"
amp schema.. it did wonders both for pre- overdrive performance
(2v input) all the way to "gross" (2.4v+) hard clipping.
C17/18 actually "softens" the distortion when overdriven (rounded
corners) One can see this in the current pulses
within the VAS.
and by adjusting r18balance can be achieved,
c12 actually lets you tweak the phase margin
..Simpler but seems to work better, and has absolutely no
effect on OLG/UG/phase M..
I do not worry..I just fix it..
This amp will end up as a "party amp" ,I sure dont want any tweeters to burn 
on it'saccount...
BTW , Andy .. thanks for that R. Cordell paper.. good stuff.
OS
Yikes.
I'd suggest going back to the original circuit and trying something along the lines of Bob Cordell's clamp circuit - at least for the lower half of the VAS.
If the collector of the lower VAS transistor is clamped to a voltage that's above its base (even a few tenths of a Volt will do), saturation will not occur in that device. Then you won't get the big inrush of base current, so you will have attacked the problem at the source, rather than trying to minimize the effect after it occurs. Then the recovery from clipping is determined by the speed of a very fast switching diode.
I'd suggest going back to the original circuit and trying something along the lines of Bob Cordell's clamp circuit - at least for the lower half of the VAS.
If the collector of the lower VAS transistor is clamped to a voltage that's above its base (even a few tenths of a Volt will do), saturation will not occur in that device. Then you won't get the big inrush of base current, so you will have attacked the problem at the source, rather than trying to minimize the effect after it occurs. Then the recovery from clipping is determined by the speed of a very fast switching diode.
FA3 is here..!!
Thanks for ideas and info (B cordell).. Andy. Went to work
and solved problem totally..(almost).
You must be... FA1 is basically Doug self's "blameless" or the"Quasi'
amp ...proven "beat to death" design. FA2 is just "symasym"
on steriods, same basic topology... what spice do you use..?? (I spiced the original's to compare to the
FA1-2 just as a baseline to fall back on...)
Close, Pots of coffee ,Core 2 quad with 3 spice softWarez
and a pile O' parts (wish I had JD or herb)
New one is nice (FA3) , many more trannies
(7 VAS/ 2 CCS/4 CM-LTP), but cheaper ones, (.12c)
fairchild sells good clones of mpsa XX's with a good model.
A good model has saturation data/ even thermal in some cases.
(unlike 340/350 data).
BTW , "crash the sim" means to much data (complex oscillations,
other sim variables), bad data (from
screwups), non working sim, or stupid, incomplete data..etc.
(from lazy companies)
Feel free to comment on "3"
OS
An externally hosted image should be here but it was not working when we last tested it.
Thanks for ideas and info (B cordell).. Andy. Went to work
and solved problem totally..(almost).
You must be... FA1 is basically Doug self's "blameless" or the"Quasi'
amp ...proven "beat to death" design. FA2 is just "symasym"
on steriods, same basic topology... what spice do you use..??
(I spiced the original's to compare to theFA1-2 just as a baseline to fall back on...)
Close, Pots of coffee ,Core 2 quad with 3 spice softWarez
and a pile O' parts (wish I had JD or herb)
New one is nice (FA3) , many more trannies
(7 VAS/ 2 CCS/4 CM-LTP), but cheaper ones, (.12c)
fairchild sells good clones of mpsa XX's with a good model.
A good model has saturation data/ even thermal in some cases.
(unlike 340/350 data).
BTW , "crash the sim" means to much data (complex oscillations,
other sim variables), bad data (from
screwups), non working sim, or stupid, incomplete data..etc.
(from lazy companies)
Feel free to comment on "3"
OS
I've been using Fairchild's models for 2SA1381/2SC3503, MJL15032/33 from OnSemi, have tried both MJL3281A/1302A from OnSemi and FJL4215/4315 from Fairchild (they're 2SA1943/2SC5200 equivalents). BD139 model from Fairchild.
I've a 2SC1845 model, not sure where from, but have also tried 2SC1815 (from Fairchild) and 2SC2240. No major differences - I still get bad THD and prominent 2nd and 3rd harmonics.
What's throwing me, is that I've had good results from all of these models previously.
I've a 2SC1845 model, not sure where from, but have also tried 2SC1815 (from Fairchild) and 2SC2240. No major differences - I still get bad THD and prominent 2nd and 3rd harmonics.
What's throwing me, is that I've had good results from all of these models previously.
I can't imagine what is wrong you should try FA3 a lot of the
saturation problems are gone with it.. I will make a board for it
soon.. it sims SO much better than 2.. FA1 is just the blameless/
Quasi ,if it doesn't sim right , I don't know what to tell you..
OS
BTW try adjusting the Vbias (just like on the real thing..Xover distortion..)
.MODEL 2SA1381 pnp
+IS=5.5544e-14 BF=148 NF=1 VAF=580
+IKF=0.2163 ISE=2.0546e-15 NE=1.5 BR=1.892
+NR=1 VAR=100 IKR=0.187544 ISC=2.04807e-09
+NC=1.5 RB=10.18 IRB=2.0e-6 RBM=0.02
+RE=0.62 RC=3.572 XTB=0.907 XTI=3
+EG=1.206 CJE=9.572e-12 VJE=0.748 MJE=0.371
+TF=8.0312e-10 XTF=1 VTF=10 ITF=0.01
+CJC=1.147e-11 VJC=0.541 MJC=0.329 XCJC=0.9
+FC=0.5 CJS=0 VJS=0.75 MJS=0.5
+TR=1e-07 PTF=0 KF=0 AF=1
*$
.MODEL 2SC3503 NPN
+ IS =2.0893E-14 BF =101.5 NF =1.0 BR =7.655
+ NR =1.007 IBC =2.0893E-14 IBC =2.0893E-14
+ ISE =4.3652E-14 NE =1.5 ISC =1.2598E-9
+ NC =2.0 VAF =717.25 VAR =13.16 IKF =0.2512
+ IKR =0.0832 RB =2.98 RBM =0.001 IRB =0.001
+ RE =0.5305 RC =0.9 QCO =0.05 RCO =50.1187
+ VO =2.476 GAMMA =1.8231E-7 CJE =6.6039E-11
+ VJE =0.7017 MJE =0.3253 FC =0.5
+ CJC =6.6072E-12 VJC =0.5 MJC =0.2439
+ XCJC =0.6488 XTB =1.4089 EG =1.2129
+ XTI =3.0
*$
.MODEL Qmje15033 pnp
+IS=7.51228e-10 BF=134.35 NF=1.25737 VAF=12.5778
+IKF=1.88497 ISE=7.74267e-12 NE=3.34528 BR=5.14173
+NR=1.47488 VAR=1.4505 IKR=7.47186 ISC=3.25e-13
+NC=4 RB=4.37743 IRB=0.1 RBM=0.1
+RE=0.000332989 RC=0.381218 XTB=0.223027 XTI=1
+EG=1.05 CJE=3.06005e-09 VJE=0.64838 MJE=0.352991
+TF=4.78203e-09 XTF=1.50001 VTF=1.00006 ITF=0.999988
+CJC=3.00101e-10 VJC=0.600019 MJC=0.409916 XCJC=0.8
+FC=0.534975 CJS=0 VJS=0.75 MJS=0.5
+TR=1e-07 PTF=0 KF=0 AF=1
* Model generated on Feb 1, 2004
* Model format: PSpice
.MODEL Qmje15032 npn
+IS=3.7344e-10 BF=86.8313 NF=1.23974 VAF=31.5491
+IKF=9.1678 ISE=9.2499e-12 NE=3.28127 BR=5.59346
+NR=1.33161 VAR=2.1791 IKR=5.15023 ISC=4e-13
+NC=4 RB=9.54492 IRB=0.1 RBM=0.1
+RE=0.000568481 RC=0.0931741 XTB=0.737036 XTI=1.04983
+EG=1.206 CJE=3.05969e-09 VJE=0.648491 MJE=0.352663
+TF=4.94819e-09 XTF=1.50001 VTF=1.0001 ITF=0.999982
+CJC=3.00108e-10 VJC=0.600021 MJC=0.40991 XCJC=0.8
+FC=0.534651 CJS=0 VJS=0.75 MJS=0.5
+TR=1e-07 PTF=0 KF=0 AF=1
* Model generated on Feb 1, 2004
* Model format: PSpice
.MODEL Qnjw0302g pnp
+IS=5.16751e-16 BF=114.657 NF=0.895716 VAF=50.2189
+IKF=6.409 ISE=3.9641e-15 NE=4 BR=1.47167
+NR=0.923324 VAR=255.567 IKR=6.34299 ISC=3.96408e-15
+NC=2.82194 RB=2.66347 IRB=0.1 RBM=2.0828
+RE=0.0001 RC=0.0652395 XTB=1.45322 XTI=1.08126
+EG=1.05 CJE=2.14504e-09 VJE=0.4 MJE=0.376227
+TF=2.16864e-09 XTF=1000 VTF=843.737 ITF=501.348
+CJC=5e-10 VJC=0.95 MJC=0.251547 XCJC=1
+FC=0.8 CJS=0 VJS=0.75 MJS=0.5
+TR=1e-07 PTF=0 KF=0 AF=1
.MODEL Qnjw0281g npn
+IS=4.36849e-12 BF=98.1488 NF=1.01332 VAF=37.9046
+IKF=9.71849 ISE=1e-16 NE=1.8326 BR=0.79921
+NR=1.09994 VAR=339.743 IKR=5.77305 ISC=1e-16
+NC=2.71592 RB=2.74892 IRB=0.33289 RBM=2.74892
+RE=0.000344671 RC=0.03203 XTB=1.7742 XTI=1.12262
+EG=1.20598 CJE=3.66793e-09 VJE=0.74806 MJE=0.85
+TF=2.27115e-09 XTF=1000 VTF=912.955 ITF=296.602
+CJC=5e-10 VJC=0.95 MJC=0.270858 XCJC=0.98254
+FC=0.8 CJS=0 VJS=0.75 MJS=0.5
+TR=1e-07 PTF=0 KF=0 AF=1
saturation problems are gone with it.. I will make a board for it
soon.. it sims SO much better than 2.. FA1 is just the blameless/
Quasi ,if it doesn't sim right , I don't know what to tell you..
OS
BTW try adjusting the Vbias (just like on the real thing..Xover distortion..)
.MODEL 2SA1381 pnp
+IS=5.5544e-14 BF=148 NF=1 VAF=580
+IKF=0.2163 ISE=2.0546e-15 NE=1.5 BR=1.892
+NR=1 VAR=100 IKR=0.187544 ISC=2.04807e-09
+NC=1.5 RB=10.18 IRB=2.0e-6 RBM=0.02
+RE=0.62 RC=3.572 XTB=0.907 XTI=3
+EG=1.206 CJE=9.572e-12 VJE=0.748 MJE=0.371
+TF=8.0312e-10 XTF=1 VTF=10 ITF=0.01
+CJC=1.147e-11 VJC=0.541 MJC=0.329 XCJC=0.9
+FC=0.5 CJS=0 VJS=0.75 MJS=0.5
+TR=1e-07 PTF=0 KF=0 AF=1
*$
.MODEL 2SC3503 NPN
+ IS =2.0893E-14 BF =101.5 NF =1.0 BR =7.655
+ NR =1.007 IBC =2.0893E-14 IBC =2.0893E-14
+ ISE =4.3652E-14 NE =1.5 ISC =1.2598E-9
+ NC =2.0 VAF =717.25 VAR =13.16 IKF =0.2512
+ IKR =0.0832 RB =2.98 RBM =0.001 IRB =0.001
+ RE =0.5305 RC =0.9 QCO =0.05 RCO =50.1187
+ VO =2.476 GAMMA =1.8231E-7 CJE =6.6039E-11
+ VJE =0.7017 MJE =0.3253 FC =0.5
+ CJC =6.6072E-12 VJC =0.5 MJC =0.2439
+ XCJC =0.6488 XTB =1.4089 EG =1.2129
+ XTI =3.0
*$
.MODEL Qmje15033 pnp
+IS=7.51228e-10 BF=134.35 NF=1.25737 VAF=12.5778
+IKF=1.88497 ISE=7.74267e-12 NE=3.34528 BR=5.14173
+NR=1.47488 VAR=1.4505 IKR=7.47186 ISC=3.25e-13
+NC=4 RB=4.37743 IRB=0.1 RBM=0.1
+RE=0.000332989 RC=0.381218 XTB=0.223027 XTI=1
+EG=1.05 CJE=3.06005e-09 VJE=0.64838 MJE=0.352991
+TF=4.78203e-09 XTF=1.50001 VTF=1.00006 ITF=0.999988
+CJC=3.00101e-10 VJC=0.600019 MJC=0.409916 XCJC=0.8
+FC=0.534975 CJS=0 VJS=0.75 MJS=0.5
+TR=1e-07 PTF=0 KF=0 AF=1
* Model generated on Feb 1, 2004
* Model format: PSpice
.MODEL Qmje15032 npn
+IS=3.7344e-10 BF=86.8313 NF=1.23974 VAF=31.5491
+IKF=9.1678 ISE=9.2499e-12 NE=3.28127 BR=5.59346
+NR=1.33161 VAR=2.1791 IKR=5.15023 ISC=4e-13
+NC=4 RB=9.54492 IRB=0.1 RBM=0.1
+RE=0.000568481 RC=0.0931741 XTB=0.737036 XTI=1.04983
+EG=1.206 CJE=3.05969e-09 VJE=0.648491 MJE=0.352663
+TF=4.94819e-09 XTF=1.50001 VTF=1.0001 ITF=0.999982
+CJC=3.00108e-10 VJC=0.600021 MJC=0.40991 XCJC=0.8
+FC=0.534651 CJS=0 VJS=0.75 MJS=0.5
+TR=1e-07 PTF=0 KF=0 AF=1
* Model generated on Feb 1, 2004
* Model format: PSpice
.MODEL Qnjw0302g pnp
+IS=5.16751e-16 BF=114.657 NF=0.895716 VAF=50.2189
+IKF=6.409 ISE=3.9641e-15 NE=4 BR=1.47167
+NR=0.923324 VAR=255.567 IKR=6.34299 ISC=3.96408e-15
+NC=2.82194 RB=2.66347 IRB=0.1 RBM=2.0828
+RE=0.0001 RC=0.0652395 XTB=1.45322 XTI=1.08126
+EG=1.05 CJE=2.14504e-09 VJE=0.4 MJE=0.376227
+TF=2.16864e-09 XTF=1000 VTF=843.737 ITF=501.348
+CJC=5e-10 VJC=0.95 MJC=0.251547 XCJC=1
+FC=0.8 CJS=0 VJS=0.75 MJS=0.5
+TR=1e-07 PTF=0 KF=0 AF=1
.MODEL Qnjw0281g npn
+IS=4.36849e-12 BF=98.1488 NF=1.01332 VAF=37.9046
+IKF=9.71849 ISE=1e-16 NE=1.8326 BR=0.79921
+NR=1.09994 VAR=339.743 IKR=5.77305 ISC=1e-16
+NC=2.71592 RB=2.74892 IRB=0.33289 RBM=2.74892
+RE=0.000344671 RC=0.03203 XTB=1.7742 XTI=1.12262
+EG=1.20598 CJE=3.66793e-09 VJE=0.74806 MJE=0.85
+TF=2.27115e-09 XTF=1000 VTF=912.955 ITF=296.602
+CJC=5e-10 VJC=0.95 MJC=0.270858 XCJC=0.98254
+FC=0.8 CJS=0 VJS=0.75 MJS=0.5
+TR=1e-07 PTF=0 KF=0 AF=1
Yes , MJL simmed FA2 and got nearly immeasurable H3/5,
built it before me..But he uses multisim...
Down load FA3 ..it really "whups" FA2
it's in the same folder....(killer PSRR) and you can "slam" it with
4V inputs ,clips absolutely sym/clean (cascoded)..
I had "glitches" with ksa1381 model,,too..
OS
built it before me..But he uses multisim...
Down load FA3 ..it really "whups" FA2
it's in the same folder....(killer PSRR) and you can "slam" it with
4V inputs ,clips absolutely sym/clean (cascoded)..
I had "glitches" with ksa1381 model,,too..
OS
Ahh... PSRR, there is a thread on it now I think.
good to test for since we do not always make monoblocks.
{Each rail with 2 sines out of sync 120-500hz + 3khz SW for nice
garbage supply sim) Fa2 did fairly good with this, but the "new one" has 90DB + PSRR. This looks just like what I see on a scope,
exactly.. except for occasional transients..
OS
good to test for since we do not always make monoblocks.
An externally hosted image should be here but it was not working when we last tested it.
{Each rail with 2 sines out of sync 120-500hz + 3khz SW for nice
garbage supply sim) Fa2 did fairly good with this, but the "new one" has 90DB + PSRR. This looks just like what I see on a scope,
exactly.. except for occasional transients..
OS
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.