Krill - The Next Generation

[Steve Dunlap]

I am not using the latest version, but here is the web site. I can not justify the cost of upgrading.

http://www.intusoft.com/

[andy_c]

Thanks Steve.

Some of these simulators are really expensive. I'm glad there's free SPICE simulation software available. Looks like Intusoft have a free demo version available that handles up to 20 components.

[scott wurcer]

Originally Posted by Steve Dunlap

Doesn't even a "real" class A amp do that also?

Hi Steve,

I hope you are feeling well. I'm taking a back seat on this stuff when, I guess, people are invested in a different way of looking at things. Anyway, here is a picture trying one more time to show what I am talking about. This is the simple voltage transfer funtion error of a simple CEF (Vin - Vout) for two bias levels and their derivatives. The distortion is related to the derivative once this output stage is put into a conventional op-amp configuration. The lower plot on the right is biased near the "Bernie Oliver" point where the translinear Vin vs Vout function becomes linear. In the first case there is considerable distortion without any switching necessary. Class "A" just keeps flattening this transfer funtion till the distortion matters less and less.

Where I come from crossover distortion has nothing to do with switching off of one side or the other of an output stage.

[Steve Dunlap]

Originally Posted by andy_c

Thanks Steve.

Some of these simulators are really expensive. I'm glad there's free SPICE simulation software available. Looks like Intusoft have a free demo version available that handles up to 20 components.

Back when I acquired mine, none of the free or cheap ones would model my circuit correctly. A couple simply showed no output signal at all. Fortunately for me, the company I was working for had three different SPICE programs I could choose from. I liked the Intusoft the best because at the time it had more features and a library of over 30K symbols. The head of engineering ask 3 of us working for him to take (legal and licensed) copies of all our software home so we could work from home at night and on weekends. They let me keep the software when I was laid off because they hired me back as a consultant.

[Steve Dunlap]

Where I come from crossover distortion has nothing to do with switching off of one side or the other of an output stage.

Hi Scott,

I wasn't aware that that I had said anything to contradict that. I was a little more excitable a few months back and I could have said things without thinking them through. The doctors have cut my Prednisone back by 50mg a day so I am somewhat more civel now. I am sorry if I was harsh with you and some others here. Roid rage is real.

[Bigun]

Hi Steve, I'm afraid I don't understand some of the posts above and haven't had luck simulating the Krill output myself - I use LTspice. My skills are improving slowly.

Anyhow, I was going to ask a different question, i.e. not one about whether devices are on or off, but, assuming that Krill works as you think, what next ?

Your amp has many people saying how nice it sounds. Is that the end ? Do you see ways in which this topology / biassing can be extended, perhaps to run at lower current through the driver stages, or built with non-BJTs ? simplifications for lower powers etc.

Just being curious...

[andy_c]

Originally Posted by Bigun

Hi Steve, I'm afraid I don't understand some of the posts above and haven't had luck simulating the Krill output myself - I use LTspice. My skills are improving slowly.

I've posted the Krill LTspice simulation files in this post earlier in the thread. Ostripper contributed a lot to that (actually most of it), and I forgot to thank him earlier. Thanks OS! Just unzip it to a directory and run it. All you need to do is probe on the nodes you're interested in. It's ready to run and completely self-contained.

Steve hasn't posted his simulation files yet though.

[ostripper]

Thanks OS!

your welcome... but I must really thank you ,Andy... for all the preliminary LT "coaching". Not only can I make any DIY amp work with any device (and sound good) , but I am making $$ using LT to dissect commercial amps.
BTW - Steve .. cool website .. I like the 200 watter !!

OS

[Steve Dunlap]

[QUOTE]Steve hasn't posted his simulation files yet though.[/QUOTE

I took a day off. I zipped the file since .cir isn't one of the approved file types.

[Steve Dunlap]

Originally Posted by Bigun

Hi Steve, I'm afraid I don't understand some of the posts above and haven't had luck simulating the Krill output myself - I use LTspice. My skills are improving slowly.

Anyhow, I was going to ask a different question, i.e. not one about whether devices are on or off, but, assuming that Krill works as you think, what next ?

Your amp has many people saying how nice it sounds. Is that the end ? Do you see ways in which this topology / biassing can be extended, perhaps to run at lower current through the driver stages, or built with non-BJTs ? simplifications for lower powers etc.

Just being curious...

A lot of people have had trouble simulating the Krill, including me at first. Many of the earlier spice programs gave no results at all, god or bad.

Whether the Krill works as I think it does or not, it does work. I went in several directions with this circuit including high powered pro amps. I made several improvements to the schematics that improved performance also. My original goal was to build a very affordable and pleasant sounding amp. I also made high end amps because the demand was there. I had planned to show how this evolved over time, but it seems it became mired in controversy so I gave up. If the circuit works as I think, non BJTs would not bee a good choice due to their very low gate currents.

The circuit can be run at lower currents, but will switch before it reaches rated output. It still sounds good in my opinion. As for simpler, it would be difficult to make it any simpler than what I presented first. That schematic is on my web site.