Thank you.C2C.. As I am leaving the ability to thermally couple
the 6 to-126's "open" , I'll try it both ways. 🙂
What I decided is to make the krill OPS "modular" on 1 -3" X 6"
board , and the voltage stage separate from it. For picky
people (even me , sometimes) one could run from 2 totally
separate PS's (antek http://www.antekinc.com/AN-3245.pdf )
for the voltage stages and a monster for the current amps
http://www.antekinc.com/AN-6440.pdf .
Extra pads on the OPS board will allow for led/or diodes w/cap
(better CCS) and the "baby" board will have the flatpack
diodes + bias trim. "room to play" on every square inch.. 😀
OS
anatech said:
Well the painters cleared out today so I hope to get back to building this weekend. Depending on what SWMBO comes up with next.
🙄
traderbam said:Hope this helps.
Sorry, but I can't really describe circuits accurately without using some maths.
I admire your energy on this. A coupe of points. It is hard to imagine the same bias trim would work in getting the bias transistors both saturated "just right" and at the same time bias the output at 25mA. The currents you show are actually the inputs to a darlington with no resistor across the output bases, so at a beta of 100 they are only 2.5uA. In one or the other direction that is the only change you get. That is in one direction the most you can do is turn off the 2.5uA.
The saturation behavior you use in your analysis is basic and I would expect any simulator to replicate it. I revived andy_c's output stage and saw nothing unusual, one bias transistor stays hard saturated over the entire range of trim, when the output stage is biased at 20W or so it is very nice and linear.
Thanks. A labour of love.scott wurcer said:
The LTSpice model are pretty poor at this. The thing you might want to measure is the betas of the the two bias transistors, rather than their Vces. Do this at low f. In real transistors I would expect the Vces to be in the same ballpark; it depends just how complementary they are.
Brian
PS: I found a couple of transistors that have a more convincing turn-on characteristic. I did a curve plot using LTSpice. Still not brilliant, but much better then MJE340/350.
For these transistors in andyC's model I set Rbias to 5500 for approximate THD null:
*ZETEX ZTX458 Spice Model v2.0 Last Revised 14/01/2005
*
.MODEL ZTX458 NPN IS=5.32E-14 NF=0.992 BF=210 IKF=0.5 VAF=1050
+XTB=1.4 ISE=2.1E-14 NE=1.385 NR=1.05 BR=16.3 IKR=1.8 VAR=99
+ISC=6.42E-12 NC=1.25 RB=0.5 RE=0.224 RC=0.134 CJC=8.5E-12
+MJC=0.3966 VJC=0.4332 CJE=122E-12 TF=1.66E-9 TR=16E-6
+QUASIMOD=1 RCO=70 GAMMA=1.6E-7 VO=10
*$
*ZETEX ZTX558 Spice Model v2.0 Last Revised 14/01/2005
*
.MODEL ZTX558 PNP IS=7.8443E-14 NF=0.99774 BF=200 IKF=0.5 VAF=349
+ISE=3.35E-14 NE=1.689 NR=0.99784 BR=3.4 IKR=0.15 VAR=82 ISC=9.42E-12
+NC=1.05 RB=0.133 RE=0.5725 RC=0.748 CJC=17.6E-12 MJC=0.5932
+VJC=0.9135 CJE=110E-12 TF=1.13E-9 TR=160E-6 XTB=1.4 QUASIMOD=1
+RCO=65 GAMMA=1.1e-5 VO=10
*$
traderbam said:
I'll try that this weekend, we will converge (so to speak) on some kind of answer.
Hi Scott,
That was pretty funny! Even I got it.
Hi Joshua,
But, where is it for sure Thomas?
-Chris
🙂
That was pretty funny! Even I got it.
Hi Joshua,
That's what I suspected if t was located across the series feedback element.
But, where is it for sure Thomas?
-Chris
anatech said:Hi Scott,
🙂
That was pretty funny! Even I got it.
-Chris
It was Fr. Crowley my Jesuit English teacher junior year of high school who spoke in puns, I never lost the inclination.
From post 1115:
from Steve Dunlap's post #203:
and from post #242 when he was questioned on the purpose:
ostripper said:
The thermal considerations are "stumping me"
on one hand I see carlo's board with no heatsinking , but
my fingers tell me
😕
I'm not sure what VAS I will use on the "full up" Krill.
from Steve Dunlap's post #203:
and from post #242 when he was questioned on the purpose:
Hi Scott,
I suffered greatly in high school. Since then, everyone who knows me suffers a like fate. 😉
-Chris
I suffered greatly in high school. Since then, everyone who knows me suffers a like fate. 😉
-Chris
traderbam said:
That is why I wrote (in a following post):
.....(assuming, for the sake of the argument, the bias generator performs as advertised without the capacitor).
If I can't convince anyone who believes that this circuit is non-switching of the most obvious and easilly pointed-out boo-boo then there is little point in going any further.
Some people here have even proclaimed that the bias voltage is not dynamic (fixed) while continuing to insist that this OPS is non-switching.
= appropriate smilie.
I have not emailed or communicated with Steve, but he said several times that he wanted to give this design to the DIY community. He also said that he didn't expect to have any boards but found a number of them that if you email him he might sell to you. Given this, it seems quite ok to design ur own board of the Krill and make it available.
It seems likely that if Steve didn't want this he would have made that clear at the start, and/or said so when the subject has been brought up.
Glen, you have a high level of expertise, it would be beneficial to most readers here if you would illuminate your concerns regarding the cap and the issue of switching.
_-_-bear
It seems likely that if Steve didn't want this he would have made that clear at the start, and/or said so when the subject has been brought up.
Glen, you have a high level of expertise, it would be beneficial to most readers here if you would illuminate your concerns regarding the cap and the issue of switching.
_-_-bear
But Glen others are trying to understand/explain the workings of the circuit with an open mind - you just keep referring to the switching aspect. You may be correct!
Can you do what Bear asked & give a full explanation of how you see the circuit working given your level of expertise - this would be of benefit to all. And then maybe how the switching issue is handled?
Can you do what Bear asked & give a full explanation of how you see the circuit working given your level of expertise - this would be of benefit to all. And then maybe how the switching issue is handled?
anatech said:Hi Scott,
🙂
That was pretty funny! Even I got it.
Hi Joshua,
That's what I suspected if t was located across the series feedback element.
But, where is it for sure Thomas?
-Chris
Chris Re C2 - Sorry for the delay - I was putting a desk together for the spousal unit. Great looking beast on the outside - nice carved scroll work - heavy etc. On the inside it's a cheap POS held together by cheap low density particle board. The legs are secured to the base by what appears to be some #8 wood screws that couldn't hold up 10 pounds - 5 kilo - if their life depended on it. Looks like I'll be doing some bracing work with some decent SOLID wood and metal brackets tomorrow.
I regards to C2's location in the circuit - I was able to find my "box of stuff" amongst the others (Raiders of the Lost Ark warehouse scene) and dug out the multimeter. I'll do a ring out tomorrow - it's been a long day and I'm bushed. Time to relax a bit.
Hey - I just realized that we are talking about C2 - and that I'm often referred to as C2C. Hmmmm - something a little strange about looking around for part of me with a multimeter.
Attachments
Hi,
if working as intended, this circuit could at most eliminate switching distortion but not crossover distortion. Switching distortion can be easily and reliably prevented by class AB biasing without the strongly nonlinear transfer characteristic of this arrangement.
http://www.diyaudio.com/forums/showthread.php?s=&threadid=139796
if working as intended, this circuit could at most eliminate switching distortion but not crossover distortion. Switching distortion can be easily and reliably prevented by class AB biasing without the strongly nonlinear transfer characteristic of this arrangement.
http://www.diyaudio.com/forums/showthread.php?s=&threadid=139796
So Lumba, you are saying that you know how the circuit works (please tell) & that the same effect can be achieved by increasing the bias, so this is a complete waste of time?
Why didn't you show us on page one and save the 47 pages of postings?
Why didn't you show us on page one and save the 47 pages of postings?
John,
Criticisms on strictly engineering grounds are not too relevant until the fat lady sings. This amplifier, for all its elegance, may also sound wonderful, in which case further explanation is welcome, or it may sound no better than any other well engineered AB, in which case it is interesting, but not a watershed in design.
Until more people hear it, and can post their views on the sound quality, judgements are ill advised. I recall similar comments about the Dartzeel, which was drubbed on this forum, but which Stereophile raved about. Who knows? No one here evidently bought one... the question remains unanswered.
As a general rule, engineering criticism without listening tests is not a valid means of assessment for market acceptance. Such judgements might be valid with respect to component choices, reliability and build quality, however. The biggest factors in market acceptance are marketing, price, and sound quality.
I might add that many amps, including my own, have been roundly condemned here, but many hundreds of AKSAs have been sold, and reviews are nearly all positive; this is the acid test, I believe, not the engineering merit, which interestingly is often subjective itself. But heck, I was wrong once before, in 1959....
Cheers,
Hugh
Criticisms on strictly engineering grounds are not too relevant until the fat lady sings. This amplifier, for all its elegance, may also sound wonderful, in which case further explanation is welcome, or it may sound no better than any other well engineered AB, in which case it is interesting, but not a watershed in design.
Until more people hear it, and can post their views on the sound quality, judgements are ill advised. I recall similar comments about the Dartzeel, which was drubbed on this forum, but which Stereophile raved about. Who knows? No one here evidently bought one... the question remains unanswered.
As a general rule, engineering criticism without listening tests is not a valid means of assessment for market acceptance. Such judgements might be valid with respect to component choices, reliability and build quality, however. The biggest factors in market acceptance are marketing, price, and sound quality.
I might add that many amps, including my own, have been roundly condemned here, but many hundreds of AKSAs have been sold, and reviews are nearly all positive; this is the acid test, I believe, not the engineering merit, which interestingly is often subjective itself. But heck, I was wrong once before, in 1959....
Cheers,
Hugh
Lumba Ogir said:
What you wrote there refers to the basic 3 classes, however, I'm not sure to what degree it's applicable to Steve's Krill design, which looks unique. I still don't understand fully how this circuit works and I'd appreciate explanation.
- Status
- Not open for further replies.