John Curl's Blowtorch preamplifier part II - Page 73 - diyAudio
Go Back   Home > Forums > Member Areas > The Lounge

The Lounge A place to talk about almost anything but politics and religion.

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
Reply
 
Thread Tools Search this Thread
Old 11th August 2009, 07:52 AM   #721
Bonsai is offline Bonsai  Taiwan
diyAudio Member
 
Bonsai's Avatar
 
Join Date: Jul 2003
Location: Taipei, Taiwan
Anatek,

ok - point taken. Its hard to evaluate your own creation objectively.



On a separate note, I was trawling around on the web and found some parts that might be of interest wrt this subject.

They are SMD, but, if you solder carefully its quite easy to use SMD components. I am going to try to order some and try them out

PMP5201G Dual PNP Transistor in SOT353 - Hfe match to within 2% and Vbe match to within 2mV (NXP)

NPN counterpart is PMP4201G with same matching (NXP)

Of course, for a fully fully symmetrical input stage, NPN and PNP must match, but it should be easier if half the job is already done.

I also noticed they have matched NPN and PNP mirror pairs.

Other suppliers to look at are Rohm and probably On.
__________________
bonsai
Amplifier Design and Construction for MUSIC! http://hifisonix.com/
  Reply With Quote
Old 11th August 2009, 09:21 AM   #722
AndrewT is offline AndrewT  Scotland
diyAudio Member
 
Join Date: Jul 2004
Location: Scottish Borders
Quote:
Originally posted by anatech
Hi Andrew,.............................
Hi Anatech,
I see, as I knew all along, that we are in substantial agreement.

I use pair matching nowadays, REF + DUT, thermally coupled together.
That's the best way "I" can control temperature.

I also use CCS in the common tail.
If I have 7mA through the CCS and the two matched collector resistors drop the same voltage then I know that the transistors, both DUT & REF, are passing 3.5mA. That allows me to log the results for all my DUTs.
But usually the collector currents are not the same. So I batch them into similar groups.
Swap out the ref to a new ref in the group and retest. this finds the matched Vbe sets.
Now, this seems to be the point you have missed, I try to find matched Vbe that are close for hFE. In this respect we do the same. But I do it in the other order.
Why?
Because I found almost no correlation by selecting hFE first.

Yes, I add in a pair of resistors (an hFE jig) to measure base currents when doing the hFE comparisons, but I must still do this with REF & DUT + CCS otherwise temperature changes dominate the source of measurement error.


Finally, many are saying here that Vbe is almost the same when they check their DUTs.
How do they know Tj and Vce and Ic and Ib and Pq are being held constant and can they be sure that each Vbe that they measure has all these parameters held at the same values.

Take a pair of hFE matched DUTs.
Connect the bases and emitters and check how close the collector currents track each other when one KNOWs that Vbe is identical. I was surprised the first time I did this. That same day I was following precisely the jig that Anatech recommended and then deliberately shorted out the base resistors. All the matching went AWOL. One or two mVbe difference makes a big difference to the collector current. Even tenths of a mVbe make an easily measurable difference.
__________________
regards Andrew T.
  Reply With Quote
Old 11th August 2009, 09:39 AM   #723
forr is offline forr  France
diyAudio Member
 
forr's Avatar
 
Join Date: Dec 2004
Location: Next door
About distorsion generated by transistors mismatch in a differential input pair, it was dealt by Eric F. Taylor in Wireless World, August and Spetember 1977. Available through email.

In the same article, Taylor proposed a new feedback circuit to avoid common mode distorsion, the input signal being introduced in series with the NFB network, the non-inverting input of the amplier being grounded. The scheme has been recently re-discovered for a mike preamp in the JAES.
  Reply With Quote
Old 11th August 2009, 11:36 AM   #724
diyAudio Member
 
jan.didden's Avatar
 
Join Date: May 2002
Location: Great City of Turnhout, Belgium
Blog Entries: 7
Quote:
Originally posted by forr
About distorsion generated by transistors mismatch in a differential input pair, it was dealt by Eric F. Taylor in Wireless World, August and Spetember 1977. Available through email.

In the same article, Taylor proposed a new feedback circuit to avoid common mode distorsion, the input signal being introduced in series with the NFB network, the non-inverting input of the amplier being grounded. The scheme has been recently re-discovered for a mike preamp in the JAES.

Forr,

I believe this scheme is a textbook scheme for an inverting opamp and isn't anything special or inventive.

jd
__________________
If you don't change your beliefs, your life will be like this forever. Is that good news? - W. S. Maugham
Check out Linear Audio!
  Reply With Quote
Old 11th August 2009, 04:33 PM   #725
diyAudio Member
 
Joshua_G's Avatar
 
Join Date: Nov 2008
Location: Small village, Israel
Quote:
Originally posted by anatech

I have mentioned this before, and I have posted the circuit a couple times as well.
-Chris

Thanks.
Can you please re-post the circuit?
  Reply With Quote
Old 11th August 2009, 06:36 PM   #726
diyAudio Member
 
john curl's Avatar
 
Join Date: Jul 2003
Location: berkeley ca
I wish to say that I have been at a loss of words to comment on this approach to device matching. For engineers, I would recommend re-reading R.G. Meyers' book 'Analog Integrated Circuits' pp.227-263, or a similar text normally taught in upper division engineering. With all due respect to egalitarianism, that is why engineers go to engineering school in the first place, to come to understand the tradeoffs in matching of active devices and its side effects. The very idea of NOT matching input devices and expecting good results with quad input circuits is almost absurd, but it has been done in practice with the early Ampzilla amps, for example. When I made my first comp diff input stage in 1968, more than 40 years ago, I used matched duals and beta matched the pairs. I was lucky that they were available to me, but why bother with such an elaborate input stage, WITHOUT MATCHING?
Serious manufacturers, like Audible Illusions, went to even greater lengths to get virtually perfect matched pairs. I know, because I had to find the matches, about 20 years ago.
Today, with gain fixed at 28.2 by THX, direct coupling is almost impossible without extreme matching, beyond what is discussed here, so far, and servos become both useful and cost saving in both commercial and amateur efforts.
Now please understand: I am NOT saying that you have to have an engineering degree to properly deal with these challenges. I don't have an engineering degree. You ONLY have to work at an ENGINEERING LEVEL to be most effective. That means, at least, reading up and learning the engineering level approaches to device matching.
  Reply With Quote
Old 11th August 2009, 06:47 PM   #727
diyAudio Member
 
jan.didden's Avatar
 
Join Date: May 2002
Location: Great City of Turnhout, Belgium
Blog Entries: 7
Hi John,

I remember an article in Audio Amateur, many years ago, from Walt Jung reviewing that Ampzilla and yes, finding that the unbalance in the input stage caused excess distortion.
There were some harsh exchanges with Jim Bongiorno who brushed it off, then Walt retorting that he didn't like to be brushed off for correcting a design oversight...

I was aghast; at that time I still thought all audio designers were a friendly jolly good bunch..

Sorry for OT, but I just remembered it.

jd
__________________
If you don't change your beliefs, your life will be like this forever. Is that good news? - W. S. Maugham
Check out Linear Audio!
  Reply With Quote
Old 11th August 2009, 07:26 PM   #728
anatech is offline anatech  Canada
diyAudio Moderator
 
anatech's Avatar
 
Join Date: Jun 2004
Location: Georgetown, On
Hi bonsai,
Yes, the SM parts are where we will find the next generation of "cool parts". I've been working with SM circuitry for a while now, and I'm pretty comfortable designing that way. I still haven't learned the schematic capture and layout stuff yet. There are many reasons for that and I may need some help there. I used to simply sit down and draw out the schematic, then generate three or four board layouts as I optimized things. The old stuff used to work the first attempt mostly.

I already have some SM matched and complimentary transistor packages. Fairchild is another source I think. It's the next big thing that by now has almost passed us by, so unless you want to be "that guy" who can't deal with PC board and solid state equipment, it's time to dive in and run with it. I think heat and higher voltages are the things to watch out for. The "high" voltages in SM land are not very high from our perspective, adjust!

Hi Andrew,
Yes, I agree. We are mostly saying the same thing. However, there are a couple points I would like to comment on ...
Quote:
Now, this seems to be the point you have missed, I try to find matched Vbe that are close for hFE.
No, I didn't miss that. hFE is the more variable parameter, so I match for that. vBE doesn't matter that much, and I'll explain why ...
Quote:
That same day I was following precisely the jig that Anatech recommended and then deliberately shorted out the base resistors. All the matching went AWOL. One or two mVbe difference makes a big difference to the collector current. Even tenths of a mVbe make an easily measurable difference.
That is completely expected Andrew. You just placed the transistors in a situation they do not see in real life. When you design any audio stage, there are generally some highish resistances in series with each base. Without this resistance, the vBE understandably becomes far more important. That is exactly the reason I designed the matching jig the way I did. It was to put the parts under test into a situation that mimicked the circuit they would be put into.

Andrew, show me any of your designs that work well with no series resistance in the base circuits. If you did create something like that, the maximum input voltage might well be in the tens of mV, rather than half a volt or so. The input resistance serves to convert the signal voltage into a signal current.

-Chris
__________________
"Just because you can, doesn't mean you should" my Wife
  Reply With Quote
Old 11th August 2009, 07:54 PM   #729
anatech is offline anatech  Canada
diyAudio Moderator
 
anatech's Avatar
 
Join Date: Jun 2004
Location: Georgetown, On
Hi Forr,
Quote:
About distorsion generated by transistors mismatch in a differential input pair, it was dealt by Eric F. Taylor in Wireless World, August and Spetember 1977. Available through email.
I'm not familiar with that. Could you give me some more information on how I can read Mr. Taylor's article?

After reading many, many posts, and some email exchanges, I can see that I should not have dismissed "Wireless World" and "Audio Express". At the time, I gave up on magazines after "Electronics World" died before my subscription was over, and "Radio Electronics" ran to being a computer technology publication. I gave up on magazines after that.

Hi Joshua,
Quote:
Can you please re-post the circuit?
Sure, this is the first part, attached.

Hi John,
Quote:
I wish to say that I have been at a loss of words to comment on this approach to device matching. For engineers, I would recommend re-reading R.G. Meyers' book 'Analog Integrated Circuits' pp.227-263, or a similar text normally taught in upper division engineering.
Good lord sir, I do appreciate the references, but I'd love to see you try to explain your viewpoint as well. References are normally cited to back up a statement. Where's your statement?
Quote:
With all due respect to egalitarianism, that is why engineers go to engineering school in the first place, to come to understand the tradeoffs in matching of active devices and its side effects.
I will disagree with you here. I went to Ryerson and some friends of mine went to Waterloo (I was accepted there as well). What they learned was far more abstract. We learned the same information, but we were forced to apply it as well. I don't like the math behind things because I found that my friends who didn't understand the concepts hid behind the equations and forced any arguments to be fought on those grounds. Like a debating trick. From where I sit, translating concepts to the world of math is bloody dangerous if you do not completely understand the subject. If the subject is not well understood, improper weight is given to some variables and others are missed completely. I learn the math I need simply because I use it, the stuff I don't use often, I look up. Much safer that way. You can not baffle anyone if you talk using engineering terms and concepts. Nothing to hide behind.

BTW, I'm not suggesting you are hiding behind anything John. However, I would like to keep the discussion real and understandable for everyone.

Quote:
The very idea of NOT matching input devices and expecting good results with quad input circuits is almost absurd, but it has been done in practice with the early Ampzilla amps, for example. When I made my first comp diff input stage in 1968, more than 40 years ago, I used matched duals and beta matched the pairs. I was lucky that they were available to me, but why bother with such an elaborate input stage, WITHOUT MATCHING?
That's is what I have said. The concept is stone simple. Given that the concept is so basic and undeniable, why do you think many manufacturers actually release products that do not measure up? This is one of the points I was attempting to bring to your attention, nicely though. This does not take an engineer to understand, it only requires some basic common sense, right?

Quote:
Today, with gain fixed at 28.2 by THX, direct coupling is almost impossible without extreme matching, beyond what is discussed here, so far, and servos become both useful and cost saving in both commercial and amateur efforts.
I don't really agree with that statement. True, you have to work to a stipulated gain, but there are many compound diff. pairs used that provide the required gain and match well. Of course, we could simply use an op amp for the front end as well. What I do agree with is the concept that it is a cost issue. People prepared to DIY (rather than call them amateurs) with the skills will create things that perform better. I think that the term "amateur" is used to denote someone of lesser skill, and that is clearly wrong if you define it that way. For me, the term "amateur" means that the person does not do this for a living or regular paycheck. The term is sullied, so I would refrain from using it.

Quote:
You ONLY have to work at an ENGINEERING LEVEL to be most effective. That means, at least, reading up and learning the engineering level approaches to device matching.
I completely agree 100%. Learn the parts before learning how to use them. Now, it's time to transfer the knowledge from those who have a good understanding to others who would like a better understanding. By this, I am not saying I am correct. That's why I'm questioning my beliefs, to correct errors in what I think I know.

-Chris
Attached Images
File Type: jpg n matcher (medium).jpg (49.5 KB, 711 views)
__________________
"Just because you can, doesn't mean you should" my Wife
  Reply With Quote
Old 11th August 2009, 07:55 PM   #730
anatech is offline anatech  Canada
diyAudio Moderator
 
anatech's Avatar
 
Join Date: Jun 2004
Location: Georgetown, On
Hi Joshua,
The P channel or PNP version ...
Attached Images
File Type: jpg p matcher (medium).jpg (46.6 KB, 658 views)
__________________
"Just because you can, doesn't mean you should" my Wife
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off



New To Site? Need Help?

All times are GMT. The time now is 04:27 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2