I’ll make some further measurements, perhaps those suggested by Syn08. They might be the simplest and quickest for me to implement. I won’t get to these today as I leave tomorrow for a short driving vacation. I also want be sure that I have optimized the output stage the best I can. It makes no sense to do make these measurements without doing so.
Hugh Dean had an excellent post earlier in this thread (#1349). He suggested that some of the issues that have come up would be resolved once 10 working models are built. Out of 1500+ posts in 4 months, we have two working prototypes from which data have been posted. I have one, jmmartins has another. I realize that other prototypes have been built. I doubt there are many building either the complete Krill amp or just the output stage who really care if the output switches or if the distortion is less than 0.005%. One is building to evaluate the Krill 50 for a home theater. Another just simply loves to build and listen to new amps. Another sees its application with a tube voltage gain stage. I’m in this small crowd too. I’m building it to better understand how it works – as others have - and to evaluate it as a potential output stage for a late 1970’s era amplifier that I was fond of and would like to resurrect for my own enjoyment.
I’ll do a few more measurements with hope that this thread can move on. Maybe they will help resolve this, maybe not. They are a distraction to me because I am close to having a 150+ W Krill amp to listen to. Distortion measurements will be made and posted once I finish the updates on my Cordell analyzer, but that’s a ways off yet. Please recognize that I have no particular interest in whether or not the output switches. (added in edit) I will post results only if I can reproduce them and defend them. I won't post results that are inconclusive or irreproducible and hope that people accept that. (end edit)
I suspect that favorable listening remarks will do more to inspire people to build this amp or output stage than non-switching or sub-0.01% THD considerations. And, going back to Hugh’s post, more working models benefit all of us. For me, building is only a fun path to the end of enjoying music. The output stage is inexpensive: less than $20 for the parts for one channel at 50W other than power supply and heatsink. It is not difficult to build nor is it not as component-critical as has been suggested. Steve has used 2 different sets of complementary transistors for the TO-220s in the output that I know of. I have used 3 different pairs and even mixed transistor types without problems. The Krill output works with a voltage amplifier ranging from Steve Dunlap’s design, through an op amp (ostripper), a Borbely-designed front-end to a vacuum tube (the only variation among these 4 I have not personally built).
I hope more people build the Krill amp or output and have some fun with it. And fun could be the act of building, using it for listening to music or home theater, or just to document experimentally and discuss how it works - or doesn't. Each of these activities can lead to important contributions here.
Hugh Dean had an excellent post earlier in this thread (#1349). He suggested that some of the issues that have come up would be resolved once 10 working models are built. Out of 1500+ posts in 4 months, we have two working prototypes from which data have been posted. I have one, jmmartins has another. I realize that other prototypes have been built. I doubt there are many building either the complete Krill amp or just the output stage who really care if the output switches or if the distortion is less than 0.005%. One is building to evaluate the Krill 50 for a home theater. Another just simply loves to build and listen to new amps. Another sees its application with a tube voltage gain stage. I’m in this small crowd too. I’m building it to better understand how it works – as others have - and to evaluate it as a potential output stage for a late 1970’s era amplifier that I was fond of and would like to resurrect for my own enjoyment.
I’ll do a few more measurements with hope that this thread can move on. Maybe they will help resolve this, maybe not. They are a distraction to me because I am close to having a 150+ W Krill amp to listen to. Distortion measurements will be made and posted once I finish the updates on my Cordell analyzer, but that’s a ways off yet. Please recognize that I have no particular interest in whether or not the output switches. (added in edit) I will post results only if I can reproduce them and defend them. I won't post results that are inconclusive or irreproducible and hope that people accept that. (end edit)
I suspect that favorable listening remarks will do more to inspire people to build this amp or output stage than non-switching or sub-0.01% THD considerations. And, going back to Hugh’s post, more working models benefit all of us. For me, building is only a fun path to the end of enjoying music. The output stage is inexpensive: less than $20 for the parts for one channel at 50W other than power supply and heatsink. It is not difficult to build nor is it not as component-critical as has been suggested. Steve has used 2 different sets of complementary transistors for the TO-220s in the output that I know of. I have used 3 different pairs and even mixed transistor types without problems. The Krill output works with a voltage amplifier ranging from Steve Dunlap’s design, through an op amp (ostripper), a Borbely-designed front-end to a vacuum tube (the only variation among these 4 I have not personally built).
I hope more people build the Krill amp or output and have some fun with it. And fun could be the act of building, using it for listening to music or home theater, or just to document experimentally and discuss how it works - or doesn't. Each of these activities can lead to important contributions here.
Having built a working prototype krill (50watt, Homebrew pcb) and liked the result, I'm currently halfway through building a stereo version. PSU and chassis done, PCB's finalised but still to be etched. about a week should see it working, though given commitments it may extend a bit.. Now pass that soldering iron 
Measurements will not be reliable - a so-so sig gen and an old Hameg analog scope.
Measurements will not be reliable - a so-so sig gen and an old Hameg analog scope.
PH104 said:
Absolutely agreed with almost everything you said. I am sure that almost every builder is looking at Krill from a subjective perspective and there's nothing wrong with that, as much as I don't have a problem if somebody finds Bybee devices improving the sound.
But wrapping everyhting in a shell of pseudo-science, shoveling the concept down everybody's throat, and selling it as the best thing since sliced bread is a completely different story and boils down to anything else but subjective audio. Likewise, rejecting engineering principles is almost ok, mystifying them without even the benefit of ignorance is again a different story.
Steve Dunlap said:
Ok.
Krill performs as good (or, subjectively, as bad) as a well engineered blameless. Objective performance wise, there's nothing that makes the Krill special. Subjectively, I'll let other decide, to me there's nothing special as well.
Don't worry for JC, he seems to enjoy the kicks. As many other around, he seem to enjoy self victimizing
Let's see what we're saying (on topic) about the measurement of switching vs non-switching?
Syn08 suggests putting a resistor in series with the collector and I think looking at the current through it...?
I'm unclear why not just look through the emitter resistor?
We can't use just one channel of a standard scope because the "other" wire going to it is at ground, so a two channel scope can be used as Syn08 noted to get around this problem...
But if it were non switching the outputs then the bias current would never be driven down to nil... isn't this what one wants to see?
I'm not sure but I think Scott is suggesting putting DC on the input (not a lot of it) and running the voltage up and down through zero to see if the outputs of the opposite polarity to the DC (assuming input to the non-inverting input of the amp here) are not cut off... a steady state test...
Of course I may be all upside down on this...
_-_-bear
Syn08 suggests putting a resistor in series with the collector and I think looking at the current through it...?
I'm unclear why not just look through the emitter resistor?
We can't use just one channel of a standard scope because the "other" wire going to it is at ground, so a two channel scope can be used as Syn08 noted to get around this problem...
But if it were non switching the outputs then the bias current would never be driven down to nil... isn't this what one wants to see?
I'm not sure but I think Scott is suggesting putting DC on the input (not a lot of it) and running the voltage up and down through zero to see if the outputs of the opposite polarity to the DC (assuming input to the non-inverting input of the amp here) are not cut off... a steady state test...
Of course I may be all upside down on this...
_-_-bear
syn08 said:
Why are spending so much time and energy on something that is so mundane in your eyes?
This is all in your head - can you show where this heinous act was perpetrated?syn08 said:
Why don't you go waste your time on the Blowtorch thread as you were advised and let this mundane design die a death - your contribution here is tiresome.
syn08 said:
I tried this. I have a .82 ohm resistor between output collector and the positive rail. Done on my Krill facsimile.
This look about right? top trace is (inverted) channel 2 at the positive rail, bottom trace is channel 1 at the collector and middle trace is both added.
Attachments
All there traces have to be directly coupled and the offsets spot on (zero)
It would be much easier to just use one channel by floating the CRO ground on the amplifiers positive rail. You can do this by by either floating the CRO mains earth (preferably with an isolation transformer) or floating the amplifiers earth/ground along with the signal generator. What are you using for a signal generator? If it is a battery powered unit the latter would be the easiest test method.
It would be much easier to just use one channel by floating the CRO ground on the amplifiers positive rail. You can do this by by either floating the CRO mains earth (preferably with an isolation transformer) or floating the amplifiers earth/ground along with the signal generator. What are you using for a signal generator? If it is a battery powered unit the latter would be the easiest test method.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.