John Curl's Blowtorch preamplifier part II - Page 331

Joshua_G · 4th January 2010, 09:52 PM

John,
Few posts back I asked you about servos. Your response would be highly appreciated.

john curl · 10:12 PM

Well, I will summarize about servos, rather go into extended length with them. Two examples have been shown already, and they, or variations of them are what I use almost exclusively. You can also create a servo that controls offset by using the current passing through its V+ and V- power terminals, but I won't discuss it here.
The most serious problem with servos seems to come from the effective bandwidth of the servo, usually being too high, so that the servo sometimes dances along with the music, OR it tries to make symmetric a non-symmetric low frequency waveform as far as its short term DC offset is concerned. Therefore, my advice is to keep your servos very sluggish, even to the point where first turn on takes an observable time to settle.
The next strong suggestion is to not tightly couple the IC op amp very tight to the input of the main amp or preamp. This will mean a build out resistor of some magnitude from the output of the servo to the main amp to make an attenuator. Make the servo work for fixing the offset, for example 3V offset on the servo amp, might be 30mV at the input of the main amp. This way the servo IC, which always has some defects, is buffered from the main amp stage. This also says that even if you could easily servo out an almost completely unmatched assemblage of input parts, if you connected the servo directly to the input, you would do better with a limited effective servo range, and better matched parts at the input, since then you can have more buffering between the servo, which should only be used to maintain a 0 DC output, and nothing else.

Joshua_G · 4th January 2010, 10:52 PM

Thank you, John.
The question I asked was:
If I remember correctly, the servos designs by you which were copied, or emulated here, do 2 different things. 1 is to compare the outputs DC levels to the ground level and keep the outputs DC at ground level. The second is to compare end equalize the DC level between the 2 outputs.

Now, once both outputs DC level is kept at the ground DC level, why is it necessary to compare and equalize the DC level between the 2 outputs?

john curl · 4th January 2010, 11:02 PM

The servos are kind of stupid. One can match the outputs to each other. The other can find ground. You have to work with them both to see what happens.

bear · 4th January 2010, 11:42 PM

Quote:

If you guys truly wanted to test my opinion then you would give me a system to beat. I can't beat a system at the fraction of the cost if I don't have something to work from.

I can't resist this challenge.

Let's start with one single component, before we get to a system approach (which is far more complex). Ok? Can you do that?

Ummm... let's think really hard about it and come up with some single component to "beat".

Ah ha! By jove, I've got a capital idea!

How about a preamp??

Yeah, a preamp. Now which one?? Ummm... ok, how about a BLOWTORCH?

Beat a BLOWTORCH, Key.

Your turn. Show us what you've got.

_-_-bear

bear · 4th January 2010, 11:47 PM

Quote:

Originally Posted by john curl

The servos are kind of stupid. One can match the outputs to each other. The other can find ground. You have to work with them both to see what happens.

So would that mean that if one was not at absoloute ground then it would lift the other off absolute ground... perhaps pull it down - who is the master and who is the slave? ...or are you suggesting that ground is not where it seems to be WRT an output stage - some sort of Vdrop issue??

Joshua_G · 4th January 2010, 11:54 PM

Quote:

Originally Posted by john curl

The servos are kind of stupid. One can match the outputs to each other. The other can find ground. You have to work with them both to see what happens.

I don't get. Once one servo keeps both outputs DC at ground level, why should we bother compare them?

pinkmouse · 5th January 2010, 12:08 AM

Bear, leave it alone.

Quote:

Originally Posted by bear

...Your turn. Show us what you've got...

Terry Demol · 5th January 2010, 12:15 AM

Quote:

Originally Posted by PMA

Wouldn't it be nice to bring anything constructive here??

My suggestion - let's discuss idle current of the input JFEts.

Zero tempco point or as close to as possible.

Zero tempco Vgs = Vp+0.63V where Vp is the "pinch off" voltage (g-s)

For the Toshiba high gm types it is very close to Idss.

SY · 5th January 2010, 12:31 AM

Quote:

Originally Posted by Joshua_G

I don't get. Once one servo keeps both outputs DC at ground level, why should we bother compare them?

Half the potential offset. Of course, if the opamps are decent (and they don't have to be fast, just low offset), that's probably not important.

4th January 2010, 10:10 PM	#3302
john curl diyAudio Member Join Date: Jul 2003 Location: berkeley ca	Well, I will summarize about servos, rather go into extended length with them. Two examples have been shown already, and they, or variations of them are what I use almost exclusively. You can also create a servo that controls offset by using the current passing through its V+ and V- power terminals, but I won't discuss it here. The most serious problem with servos seems to come from the effective bandwidth of the servo, usually being too high, so that the servo sometimes dances along with the music, OR it tries to make symmetric a non-symmetric low frequency waveform as far as its short term DC offset is concerned. Therefore, my advice is to keep your servos very sluggish, even to the point where first turn on takes an observable time to settle. The next strong suggestion is to not tightly couple the IC op amp very tight to the input of the main amp or preamp. This will mean a build out resistor of some magnitude from the output of the servo to the main amp to make an attenuator. Make the servo work for fixing the offset, for example 3V offset on the servo amp, might be 30mV at the input of the main amp. This way the servo IC, which always has some defects, is buffered from the main amp stage. This also says that even if you could easily servo out an almost completely unmatched assemblage of input parts, if you connected the servo directly to the input, you would do better with a limited effective servo range, and better matched parts at the input, since then you can have more buffering between the servo, which should only be used to maintain a 0 DC output, and nothing else. Last edited by john curl; 4th January 2010 at 10:12 PM.

Thread Tools	Search this Thread
Show Printable Version Email this Page	Search this Thread: Advanced Search

New To Site?	Need Help?
Register to Participate Search Privacy Statement Contact Us	Frequently Asked Questions Did you forget your password? Mark Forums Read


	Please consider donating to help us continue to serve you. Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving

4th January 2010, 09:52 PM	#3301
Joshua_G diyAudio Member Join Date: Nov 2008 Location: Israel	John, Few posts back I asked you about servos. Your response would be highly appreciated.

4th January 2010, 10:52 PM	#3303
Joshua_G diyAudio Member Join Date: Nov 2008 Location: Israel	Thank you, John. The question I asked was: If I remember correctly, the servos designs by you which were copied, or emulated here, do 2 different things. 1 is to compare the outputs DC levels to the ground level and keep the outputs DC at ground level. The second is to compare end equalize the DC level between the 2 outputs. Now, once both outputs DC level is kept at the ground DC level, why is it necessary to compare and equalize the DC level between the 2 outputs?

4th January 2010, 11:02 PM	#3304
john curl diyAudio Member Join Date: Jul 2003 Location: berkeley ca	The servos are kind of stupid. One can match the outputs to each other. The other can find ground. You have to work with them both to see what happens.

Currently Active Users Viewing This Thread: 3 (2 members and 1 guests)
john curl, riccoryder