That did it, PSRR of -82 on ltp and -88 at OP..
Measured it with 2V/120 ripple at rails..
OS
An externally hosted image should be here but it was not working when we last tested it.
Measured it with 2V/120 ripple at rails..
OS
Hello to all here. I have been following this thread from the beginning. I haven't posted since I naively offered my help to ostripper back toward the beginning IF he was close to me here outside Nashville. Well, he isn't real close and judging from what I have seen since, he doesn't need my help. He seems to be doing very well on his own with the input he has received here.
So why am I posting again instead of just reading after all this time? Well, I have been following the simulations with the different programs. I use simulation as a tool also, but not any of the ones mentioned so far. I have noticed that most people here use the same few programs which leads me to my thread.
http://www.diyaudio.com/forums/showthread.php?s=&threadid=134619
I had great difficulty simulating this circuit. The first 3 programs I used simply would not show an output. I went with the 4th program (ICAPS) because it gave me results that matched what I was measuring in a working unit. This was about 8 years ago and I know the software has gotten better, but a number of people responding to my design have claimed that it simply will not simulate. You guys seem to have figured out most of the snags. If you are interested in taking a shot at this simulation, and yet another amp design, I would appreciate your input - good or bad. I don't want to thread jack, but this discussion has turned heavily to simulation, so I hope this is an appropriate place and time to ask for input. I will leave it up to ostripper to say if the follow up on this is posted here, on my thread or a new thread about simulation of amp circuits.
Steve
So why am I posting again instead of just reading after all this time? Well, I have been following the simulations with the different programs. I use simulation as a tool also, but not any of the ones mentioned so far. I have noticed that most people here use the same few programs which leads me to my thread.
http://www.diyaudio.com/forums/showthread.php?s=&threadid=134619
I had great difficulty simulating this circuit. The first 3 programs I used simply would not show an output. I went with the 4th program (ICAPS) because it gave me results that matched what I was measuring in a working unit. This was about 8 years ago and I know the software has gotten better, but a number of people responding to my design have claimed that it simply will not simulate. You guys seem to have figured out most of the snags. If you are interested in taking a shot at this simulation, and yet another amp design, I would appreciate your input - good or bad. I don't want to thread jack, but this discussion has turned heavily to simulation, so I hope this is an appropriate place and time to ask for input. I will leave it up to ostripper to say if the follow up on this is posted here, on my thread or a new thread about simulation of amp circuits.
Steve
Go to my WWW,.. all LT .asc files from Frugal 1/2/3 are there,
[electronics/projects/audio amp/frugalamp].
also in my [electronics/software] latest LT4 and all of
the tranny models are in zip form...
You are just over the cumberlands (I-40 is out my door), steve..!!
BTW, I will sim yours and post here all designs are game to me...
OS
BTW, I will sim yours and post here all designs are game to me...
OS
That was what I was hoping for. I HAVE been able to sim this circuit but many, not all, have not. I'm looking for reasons their sims may have failed.
I'm on the far side of Nashville from you. Still a pretty good drive unless you have a good reason. I-40 is also near me. If you find 840 on the map (a new map) I am 2 miles from the south end - at least where it ends until all the law suites have been settled.
I will be looking forward to your results.
Steve
OS
That was what I was hoping for. I HAVE been able to sim this circuit but many, not all, have not. I'm looking for reasons their sims may have failed.
I'm on the far side of Nashville from you. Still a pretty good drive unless you have a good reason. I-40 is also near me. If you find 840 on the map (a new map) I am 2 miles from the south end - at least where it ends until all the law suites have been settled.
I will be looking forward to your results.
Steve
Here is the "low down" on transistor models. Using old models,
(pre 2002) there is little info on saturation,etc. I have noticed
a real improvement (realism..to the working amp),
in simulating with current devices/models
(2005+) So, I will model your amp with both the original
2SC/A japanese devices and modern KSA/MJE (current-mouser)
to confirm my hypothesis.
Half done..tonight Krill will be dissected ..
OS
(pre 2002) there is little info on saturation,etc. I have noticed
a real improvement (realism..to the working amp),
in simulating with current devices/models
(2005+) So, I will model your amp with both the original
2SC/A japanese devices and modern KSA/MJE (current-mouser)
to confirm my hypothesis.
Half done..tonight Krill will be dissected
..OS
Os I see you back to blameless, dont worry this topology can be made to sound as good as most others. It seems to me you were first running a race for THD figures later clipping behaviour and now PSRR, with this topology its easy getting good results.
What a run around, but interesting indeed, something spectacular is bound to come out of this once youve combined the three problem matters.
What a run around
, but interesting indeed, something spectacular is bound to come out of this once youve combined the three problem matters.
All the tests.. all better than expected..did it right the first time.
-100db at unitygain with NO lead comp.
Crazy low distortion..H3 seems to be cancelled out..
Very adequate PSRR.
20khz is good as well.. it even does 40+
For the hell of it, I did 22pf Cdom to see WTF..and..
RINGING...
Oh well, it's not a 30 transistor amp.. so I am pleased.
OS
An externally hosted image should be here but it was not working when we last tested it.
-100db at unitygain with NO lead comp.
An externally hosted image should be here but it was not working when we last tested it.
Crazy low distortion..H3 seems to be cancelled out..
An externally hosted image should be here but it was not working when we last tested it.
Very adequate PSRR.
An externally hosted image should be here but it was not working when we last tested it.
20khz is good as well.. it even does 40+
For the hell of it, I did 22pf Cdom to see WTF..and..
An externally hosted image should be here but it was not working when we last tested it.
RINGING...
Oh well, it's not a 30 transistor amp.. so I am pleased.
OS
Front end is like this Apt power amp by Tomlinson Holman
http://www.diyaudio.com/forums/showthread.php?s=&threadid=70261&perpage=25&highlight=&pagenumber=1
Schematic is in this PDF (2.9 MB).
http://www.diyaudio.com/forums/showthread.php?s=&threadid=70261&perpage=25&highlight=&pagenumber=1
Schematic is in this PDF (2.9 MB).
ostripper said:
OS,
Nice work.
This latest version looks very good. I like the VAS push pull drive.
I have seen this arrangement done on a commercial amp but
I can't remember which one and it is certainly not common.
WRT the simulated THD, I always find 20kHz into 4 ohms works
best for seeing the higher harmonics - these are the ones that
are the real challenge.
Try 20kHz, 30V or 40V pk, 4ohms.
cheers
Terry
You are a true "buff" Andy.. right on the money.. I have his papers and all the APT amp schemas and just had to play..
The VAS is cool in this topology.. by tweaking the bias
and the nested feedback many desirable traits can be obtained.
(soft clipping, even more stability. ability to manipulate the phase
margin)
Very good offset right off the bat (<1mv) extremly fast
and stable... I like it..
I (ahem) have uCAP 9 now, am simming on it but scared to post
(ahem) results..thanks for link..
OS
The VAS is cool in this topology.. by tweaking the bias
and the nested feedback many desirable traits can be obtained.
(soft clipping, even more stability. ability to manipulate the phase
margin)
Very good offset right off the bat (<1mv) extremly fast
and stable... I like it..
I (ahem) have uCAP 9 now, am simming on it but scared to post
(ahem) results..thanks for link..
OS
OK , I slammed the sim with 10k/4R 110v p-p (full just before
clipping) I now see a lot more harmonics, but considering
that I am pushing it .. not bad (better than what I saw on the APT thread ) new transistor technology is cool/
for the test I timestepped 131k/ms (loaded down my quad core fully..100%) used the Andy C. sine generator, and added
some 120hz ripple to the rails as well
.An externally hosted image should be here but it was not working when we last tested it.
I want to go further and add a cascode above the LPT
to try out a fet dual in smd form (already simulated this)
and build it by the first week of 2009 .
OS
Hi os, Ive also used holman topology, it works very well. Theres a modification that can be made within that topology that will make THD go even further down with no effect on the spectrum, now it would be interesting if you discovered it, its a high end amp I sell for car audio. Whats the reason for you going into fets for input, what works better ??. I use jfet cfps, better performance and sound is good.
All done.. late night tweaks and CFP input.. Ready
to build...can't wait!!
New input stage brought unity gain point out to 1.17Mhz
but phase margin is same at 101 dg. . Most notable change
is square wave response 50k looks as good as 20k did
before as the amp is "faster" now.
And finally I got rid of that sticking and clipping problem that has
made for bad dreams.. C7/9 seem to totally
eliminate this errata only on this topology..
I tried it on FA1/2 , no beans....
[ 3v input +-55v symmetric clipping.. all freq. ]
Very happy with this one..
OS
input.. Readyto build...can't wait!!
An externally hosted image should be here but it was not working when we last tested it.
New input stage brought unity gain point out to 1.17Mhz
but phase margin is same at 101 dg. . Most notable change
is square wave response 50k looks as good as 20k did
before as the amp is "faster" now.
And finally I got rid of that sticking and clipping problem that has
made for bad dreams..
C7/9 seem to totallyeliminate this errata only on this topology..
I tried it on FA1/2 , no beans....
An externally hosted image should be here but it was not working when we last tested it.
[ 3v input +-55v symmetric clipping.. all freq. ]
Very happy with this one..
OS
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.