Transistor gain matching - diyAudio
Go Back   Home > Forums > Amplifiers > Solid State
Home Forums Rules Articles diyAudio Store Gallery Wiki Blogs Register Donations FAQ Calendar Search Today's Posts Mark Forums Read

Solid State Talk all about solid state amplification.

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 17th May 2016, 12:37 AM   #1
Dave L is offline Dave L  United States
diyAudio Member
 
Join Date: Dec 2015
Location: Central coast of California
Default Transistor gain matching

Hello all,
I am in the process of (slowly) restoring a Pioneer SX-1250 receiver. My understanding is that two of the transistors- part number 512-KSA992FBU, NPN 120V 0.5A 120 MHz need to be gain matched. I only purchased enough for the restoration and know that I may need to buy many more, but my questions have to do with the gain matching process. I bought a cheap (ETEKCITY) DVM with hFE.
So, my questions are:
1. Is a cheap DVM acceptable for this testing or should I build a circuit for the testing that may be more accurate?
2.What should the range be for matching with these transistors as a percentage? I intend to keep the receiver and have 56 year old ears.
3. Where is the hFE range listed on the data sheets for transistors?
This is my first restoration and a new hobby for me. I have a much better DVM but without hFE capacity. I hope to have this as a long term hobby.
I appreciate your patience with this newby's question.
Dave
  Reply With Quote
Old 17th May 2016, 01:43 AM   #2
anatech is offline anatech  Canada
diyAudio Moderator
 
anatech's Avatar
 
Join Date: Jun 2004
Location: Georgetown, On
Hi Dave,
Your ears are younger than mine young feller! And I can hear just fine and enjoy equipment running at it's best. So age doesn't really have a lot to do with anything.

You need to match transistors in diff pairs as closely as you can. Your Hfe test in the meter, if used with care, only provides a pre-selection for the real matching process. The match is critical to distortion levels. The bad news is that beta is very temperature sensitive. The good news is that this characteristic forces you to do things properly.

So how to make the temperature term drop out of the list of variables? It's done the same way that amplifier designers use the diff pair. Create a long tailed pair using a current source. Stick your pre-selected transistors into sockets while they are touching and shield them from ambient air flow. I find s foam cover works well, then use a small box to block air currents. You can run the pair at the target tail current, or even differing currents. Higher currents will settle before lower currents will. So a 6 mA tail will settle faster than a 1 mA tail current. It'll be more stable too.

What to measure? Well, again you're in luck. You measure the difference between the two collectors - a null measurement. AS long as your cheap meter indicates zero with a short across it's terminals, you're good to go. Record the transistor matches and their differences. No matter how you do this, some time is required to allow thermal equilibrium between the transistors to be achieved.

Install your matched pairs similarly to how you test them. I use some thermal compound between the parts, then heat shrink the two together. Works like a charm.

So what is the benefit? More stable DC offset, and lower DC offset. A perfect match will generate the amount of DC offset that is designed into the amplifier. So that's good, but not good enough for the amount of labour involved. The other huge plus is the lowest distortion that circuit is capable of (as long as the outputs are also matched). The diff pair is where the output is subtracted from the input to generate an error signal. So a perfect match will send a more perfect error signal down the line which translates into low distortion. If they are not matched, the subtraction is flawed and you do not receive the distortion performance you paid for in the first place.

The more transistors you buy to test, the greater your percentage yield will be. So it pays to buy many. Often the same transistors can be used as substitutes for many others, so your personal hoard of matched parts will be used. Try to select low noise, high gain examples. Some at 50 or 60 volts and some at 120 volts NPN and PNP will cover your bases pretty well. I've done this for years, and it really works well.

-Chris
__________________
"Just because you can, doesn't mean you should" my Wife
  Reply With Quote
Old 17th May 2016, 02:01 AM   #3
diyAudio Member
 
Join Date: May 2003
Location: quebec
You may read Leach documentation ...

The Leach Amp - Part 2

you will find usefull information ... if you have enough time read the other pages linked to this one ..
  Reply With Quote
Old 17th May 2016, 02:20 AM   #4
anatech is offline anatech  Canada
diyAudio Moderator
 
anatech's Avatar
 
Join Date: Jun 2004
Location: Georgetown, On
You're missing the one problem that prevents all the other common attempts at getting good matches.

Temperature variation!

If you can't control this, you can't match a darned thing. Sorry, but this is true.

Experiment: Stick a transistor in your favorite beta measuring device. Let it settle and note the reading. Now blow on it - what happens? Try pinching it between your (normally warm) fingers. What happens?

So you're going to get good numbers trying to match parts this way? I really don't think so.

-Chris
__________________
"Just because you can, doesn't mean you should" my Wife
  Reply With Quote
Old 17th May 2016, 03:51 AM   #5
T117 is offline T117  Russian Federation
diyAudio Member
 
T117's Avatar
 
Join Date: Apr 2016
There is such a circuit for selection of transistors for h21e and check UceR.
It is clear from all electrical circuits.
Attached Images
File Type: jpg attachment.jpg (120.6 KB, 322 views)
File Type: gif check UceR.GIF (13.7 KB, 317 views)
File Type: gif check UceR.GIF (143.6 KB, 311 views)
File Type: gif selection of transistors for h21e.GIF (13.8 KB, 308 views)

Last edited by T117; 17th May 2016 at 04:01 AM.
  Reply With Quote
Old 17th May 2016, 04:23 AM   #6
anatech is offline anatech  Canada
diyAudio Moderator
 
anatech's Avatar
 
Join Date: Jun 2004
Location: Georgetown, On
Hi T117,
These instruments still do not regulate or balance the temperature of the unit under test. The actual goal is to match, or balance, a pair of transistors. When the semiconductor manufacturer generates curves for transistors at Tc=25C, they control the case temperature to 25C. When they say ambient temperature is 25C, they allow the parts to settle into the controlled temperature, then run the tests. Since this is unreasonable for the average person, I designed a jig that balances two transistors as a pair causing the temperature term to be common to both and therefore drop out.

If you can't either control the junction temperature, or cause it to become unimportant, you can not match transistors. I'm very sorry, but this is the truth.

To prove my point, serious matching of transistors generally requires matching on the wafer and co-mounting the die in a single package. Or, as in the case of the LM394 (I thinkthe # is right), the two die are probed in the wafer and are installed as one piece into a package. The LM394 takes this further and is actually made up from many small transistors in parallel and mounted in a single unit onto the package. If you can't do that, then the best you can do is balance the two devices against each other, then tie their cases together to test them for a match. Once matched, you fasten the two parts together, then make the ambient the same for them (heat shrink tubing). You could add a shroud or cap as long as they were not running too warm.

Why match the temperature? Because semiconductors are extremely sensitive to temperatures. So if you can't control the temperature, or balance two devices in the same thermal environment, you can not match them effectively.

Once the circuit and PCB are finalized, this device will be available in the DIYAudio store. I feel it will really help our members to be able to obtain extremely tight matching. I routinely match transistors so close that I have to also match the degeneration resistors to prevent throwing the match out. They can be extremely tight. Measured and listening performance is always improved by doing this if the amplifier has a good design.

-Chris
__________________
"Just because you can, doesn't mean you should" my Wife
  Reply With Quote
Old 17th May 2016, 04:42 AM   #7
T117 is offline T117  Russian Federation
diyAudio Member
 
T117's Avatar
 
Join Date: Apr 2016
Quote:
Originally Posted by anatech View Post
Hi T117,
These instruments still do not regulate or balance the temperature of the unit under test. The actual goal is to match, or balance, a pair of transistors. When the semiconductor manufacturer generates curves for transistors at Tc=25C, they control the case temperature to 25C. When they say ambient temperature is 25C, they allow the parts to settle into the controlled temperature, then run the tests. Since this is unreasonable for the average person, I designed a jig that balances two transistors as a pair causing the temperature term to be common to both and therefore drop out.

If you can't either control the junction temperature, or cause it to become unimportant, you can not match transistors. I'm very sorry, but this is the truth.

To prove my point, serious matching of transistors generally requires matching on the wafer and co-mounting the die in a single package. Or, as in the case of the LM394 (I thinkthe # is right), the two die are probed in the wafer and are installed as one piece into a package. The LM394 takes this further and is actually made up from many small transistors in parallel and mounted in a single unit onto the package. If you can't do that, then the best you can do is balance the two devices against each other, then tie their cases together to test them for a match. Once matched, you fasten the two parts together, then make the ambient the same for them (heat shrink tubing). You could add a shroud or cap as long as they were not running too warm.

Why match the temperature? Because semiconductors are extremely sensitive to temperatures. So if you can't control the temperature, or balance two devices in the same thermal environment, you can not match them effectively.

Once the circuit and PCB are finalized, this device will be available in the DIYAudio store. I feel it will really help our members to be able to obtain extremely tight matching. I routinely match transistors so close that I have to also match the degeneration resistors to prevent throwing the match out. They can be extremely tight. Measured and listening performance is always improved by doing this if the amplifier has a good design.

-Chris
You're absolutely right. But. Selection for һ21е reduces INPUT OFFSET CURRENT, and we are interested in, in addition, the INPUT OFFSET VOLTAGE. To reduce this last to be measured and Ube, which often has a variation.


In fact, a high degree of matching of the transistors in the amplifiers is not as important. The servo helps to eliminate the effects of unbalance in the input stage.Remain to know what happens with distortion. About this I have no data.
  Reply With Quote
Old 17th May 2016, 04:52 AM   #8
anatech is offline anatech  Canada
diyAudio Moderator
 
anatech's Avatar
 
Join Date: Jun 2004
Location: Georgetown, On
Hi T117,
It's not the DC offset we are worried about. It is how effective the diff pair is in comparing the input with the signal output to cancel distortion. It doesn't take much unbalance to affect how the amplifier sounds (for some), and I can certainly measure the difference. Look at the residuals from a good THD meter with a spectrum analyzer and you will see. That's assuming the amplifier is of good enough quality so that it matters.

This is a very real effect. There are app notes and books that support this if experimentation isn't enough. Even a thought experiment should show this to be true. The diff pair is crucial to how the amplifier operates.

-Chris
__________________
"Just because you can, doesn't mean you should" my Wife
  Reply With Quote
Old 18th May 2016, 04:47 AM   #9
Dave L is offline Dave L  United States
diyAudio Member
 
Join Date: Dec 2015
Location: Central coast of California
Thank you all for taking the time to help with my transistor matching problem. I will take temperature into consideration and build the recommended circuit for testing rather than using the cheap DVM I bought.

Dave
  Reply With Quote
Old 18th May 2016, 11:57 AM   #10
T117 is offline T117  Russian Federation
diyAudio Member
 
T117's Avatar
 
Join Date: Apr 2016
Quote:
Originally Posted by anatech View Post
Hi T117,
It's not the DC offset we are worried about. It is how effective the diff pair is in comparing the input with the signal output to cancel distortion. It doesn't take much unbalance to affect how the amplifier sounds (for some), and I can certainly measure the difference. Look at the residuals from a good THD meter with a spectrum analyzer and you will see. That's assuming the amplifier is of good enough quality so that it matters.

This is a very real effect. There are app notes and books that support this if experimentation isn't enough. Even a thought experiment should show this to be true. The diff pair is crucial to how the amplifier operates.

-Chris
Hi, Chris.
Without the use of special monolithic pairs from the problems you described, to get rid of difficult. The THD of the cascade depends on the scope changes the current of the transistor. Therefore, a more easy way - by increasing the gain voltage of the other cascades to force the transistors of the input differential stage to work with the least possible change in the current. I do, based on this understanding, the results are still good.
  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


Similar Threads
Thread Thread Starter Forum Replies Last Post
Transistor matching JaredC79 Construction Tips 1 6th March 2016 09:25 PM
Transistor matching diy didi Solid State 2 19th January 2012 03:01 PM
Transistor matching graham-r Parts 1 6th May 2009 08:21 PM
transistor matching? millwood Solid State 8 20th December 2003 05:41 PM
transistor matching weinstro Parts 28 6th July 2003 03:52 PM


New To Site? Need Help?

All times are GMT. The time now is 03:44 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2016 DragonByte Technologies Ltd.
Copyright 1999-2016 diyAudio

Content Relevant URLs by vBSEO 3.3.2
Wiki