C11 bootstrap could be better driven, either by VOut or Q2E.
That does assume you are driving Q2B through Q12EC like an
Allison or Aleph. Are you trying to load Q1C with R23 1K, and
drive Q2B through the electrolytic path in parallel with Allison?
Is this part of a plan I don't yet grasp, or a simple miswire?
That does assume you are driving Q2B through Q12EC like an
Allison or Aleph. Are you trying to load Q1C with R23 1K, and
drive Q2B through the electrolytic path in parallel with Allison?
Is this part of a plan I don't yet grasp, or a simple miswire?
Last edited:
Hi Ken. I didn't notice that schematic blip. It IS intentional, but I haven't wired it up that way in real life. When wired this way, the current through R21 acts as push-pull and helps Q12 drive Q2, or at least acts as a better CCS. Wired with Q2E as reference, it does the opposite.
I've mentioned in other places how bootstraps can be improved upon, yet usually aren't.
- keantoken
I've mentioned in other places how bootstraps can be improved upon, yet usually aren't.
- keantoken
I've done some more rigorous listening tests, and again have not been disappointed.
This amp seems to have a hypnotizing affect! it puts you to sleep and doesn't try to wake you up. It seems more emotionally involving, and natural. My brain seems to trigger somewhere and ride the wave. I find myself adjusting the headphones to get closer to the sound, whereas listening directly from the soundcard I'm always worrying about it being too loud. It seems to sing... I don't think I'm biased towards my own creation, since (though I was afraid to say it) I thought it sounded pretty bad at times (usually very harsh). Remember I was listening in mono, and my reference was in stereo. I leveled the playing field by getting a mono-stereo VST plugin. I thought at first the soundcard had improved resolution, but I think it was the harshness that gave me that impression. With the amp I hear the same things, though they don't jump out unnaturally.
The PCB continues to give me hell.
- keantoken
This amp seems to have a hypnotizing affect! it puts you to sleep and doesn't try to wake you up. It seems more emotionally involving, and natural. My brain seems to trigger somewhere and ride the wave. I find myself adjusting the headphones to get closer to the sound, whereas listening directly from the soundcard I'm always worrying about it being too loud. It seems to sing... I don't think I'm biased towards my own creation, since (though I was afraid to say it) I thought it sounded pretty bad at times (usually very harsh). Remember I was listening in mono, and my reference was in stereo. I leveled the playing field by getting a mono-stereo VST plugin. I thought at first the soundcard had improved resolution, but I think it was the harshness that gave me that impression. With the amp I hear the same things, though they don't jump out unnaturally.
The PCB continues to give me hell.
- keantoken
kt, can you please post the schematic of the amp that you actually built?
BTW, a couple of days ago I almost started a layout for your amp then got side tracked again with more experiments with my lna. But you keep making me curious with this amp, so I might just start the pcb layout soon.
BTW, a couple of days ago I almost started a layout for your amp then got side tracked again with more experiments with my lna. But you keep making me curious with this amp, so I might just start the pcb layout soon.
Gah, not again. R12 and R10 should be 100 ohms. 25 ohms gives funny shaped overdrive characteristics.
There are a few margin-type things I have excluded in order to get the prototype ready fast. For instance, Q12 may need stability compensation although I haven't had problems. If you're making a board I would like to discuss the few extra things that should go on it "just to be sure".
- keantoken
There are a few margin-type things I have excluded in order to get the prototype ready fast. For instance, Q12 may need stability compensation although I haven't had problems. If you're making a board I would like to discuss the few extra things that should go on it "just to be sure".
- keantoken
Sure, we can take it offline if you want. I don't have any 2sc4370s, but I got some bd139. I'm ikoflexer at gmail.com
Should Q4 be thermally bonded to Q10? I'm not sure why Q4 is oriented
so as to be a diode curved load for Q8. Unless its some sort of current
mirror thingie together with VBE Q10... But wouldn't a CCS type loading
have made more open loop gain?
so as to be a diode curved load for Q8. Unless its some sort of current
mirror thingie together with VBE Q10... But wouldn't a CCS type loading
have made more open loop gain?
Hello keantoken
Here is some links of high resolution audiophile music files to download, most of them are free.
Bye
Gaetan
High resolution audio downloads, audiophile audio downloads, hd tracks
Ryan Adams Live at Das Haus on 2006-10-17 : Free Download & Streaming : Internet Archive
Index of /samplerdownload
Brandenburg Concertos - ?eský rozhlas D-Dur
Mobile DJ Downloads - Complete DJ Library of Digital Music Downloads - 100% Legal
Index of /music
Chandos Records Classical Music CDs and MP3 Downloads OnLine
http://www.bowers-wilkins.com/display.aspx?infid=777
http://www.audioworld.com/cgi-bin/dir/jump.cgi?ID=2511
Design w Sound Registration Form
High Resolution Music DOWNLOAD services
Downloads on the Up and Up ? High Quality Music Downloads | Home Entertainment
Last edited:
Q4 might be considered the third buffer to the VAS Q1. So it's like a double buffered VAS. Just think of Q4 as a complimentary EF buffer... This was originally added because I wanted to experiment with MOSFETs in the VAS, but that turned out to be a bad plan... Haha lol.
I didn't plan to have thermal bonding, and temperature change here will have a very negligible affect compared to the tempco induced by Q9.
Q4 gives Q10 a very low source impedance, which boosts usable bandwidth. Or so the theory goes. At the least it was comforting to know that bandwidth would probably be the least of my problems. Bandwidth is important if you want open-loop plots like the ones I posted.
Hey Gaetan! Thank you for the links.
- keantoken
I didn't plan to have thermal bonding, and temperature change here will have a very negligible affect compared to the tempco induced by Q9.
Q4 gives Q10 a very low source impedance, which boosts usable bandwidth. Or so the theory goes. At the least it was comforting to know that bandwidth would probably be the least of my problems. Bandwidth is important if you want open-loop plots like the ones I posted.
Hey Gaetan! Thank you for the links.
- keantoken
Hello all.
I found a 68p cap stuck in my foot this morning (my last one!) so I decided to upgrade C6. So far it's stable into 10nF.
Here is the schematic of the prototype sitting on my bench.
C5 and R14 look like a Zobel, but their actual purpose is stability compensation.
- keantoken
I found a 68p cap stuck in my foot this morning (my last one!) so I decided to upgrade C6. So far it's stable into 10nF.
Here is the schematic of the prototype sitting on my bench.
C5 and R14 look like a Zobel, but their actual purpose is stability compensation.
- keantoken
Attachments
Last edited:
Something's oscillating but it's NOT my amp!
Look at my soundcard output!
20mV DC offset, and oscillating! NOW I know where all the ultrasonic trash is coming from!
I'll try getting a shorter cable from radioshack, not one of those spiral ones I have now, it must be a perfect resonator.
- keantoken
Look at my soundcard output!
20mV DC offset, and oscillating! NOW I know where all the ultrasonic trash is coming from!
I'll try getting a shorter cable from radioshack, not one of those spiral ones I have now, it must be a perfect resonator.
- keantoken
Attachments
Last edited:
Well, the new cable didn't help any. This will certainly foil any measurements I intend to make.
- keantoken
- keantoken
Last edited:
Hello
Did you try with another source to see if it give a clean sine out ?
Or it's your sound-card who do that problem.
You could do a simple sine-square wave oscillator to do tests.
Bye
Gaetan
The amp gives a clean sine well into the hundreds of KHz, with my signal generator. My signal generator is not suitable for audio however.
I think, if I build two of these high-Z Allison buffers, I'll be able to measure differential voltage directly, and thus take visual distortion measurements and directly verify operation.
- keantoken
I think, if I build two of these high-Z Allison buffers, I'll be able to measure differential voltage directly, and thus take visual distortion measurements and directly verify operation.
- keantoken
Attachments
Actually, here is the actual circuit I might try to build. That is 268megaohms at 10KHz!!! (this is with the 1meg trimmer disconnected) Of course this value is optimistic because of the current sources, but it gives one an idea of what is possible.
So it should be an excellent electrometer. 😉
The big boost comes from the effective cascode of the input transistors, which removes early voltage and nullifies collector capacitance. This is better performance than I would hope to achieve with a Jfet buffer. Not sure how noise pans out.
- keantoken
So it should be an excellent electrometer. 😉
The big boost comes from the effective cascode of the input transistors, which removes early voltage and nullifies collector capacitance. This is better performance than I would hope to achieve with a Jfet buffer. Not sure how noise pans out.
- keantoken
Attachments
With 100k CCS impedance, input impedance is still 2megaohms. But it is possible to get way higher than that with just a 2-transistor CCS. I have two options: options 1 has high impedance. Option 2 has lower impedance but less imprint at HF.
My main concern is stability. It is hard to get the circuit stable without limiting HF. I think I have a solution but the bench will be the arbiter...
- keantoken
My main concern is stability. It is hard to get the circuit stable without limiting HF. I think I have a solution but the bench will be the arbiter...
- keantoken
- Status
- Not open for further replies.