[sdiy] Re: Transistor Matching

[Tim Ressel]

Yo,

I dug up some 0.1% resistors. After adjusting the
circuit to make up for the different values of the new
resistors I got the offset down to about 12mV. Again,
the method of determining the offset is to reverse the
trannies under test and take the difference.

--tr

--- Tim Ressel <madhun2001 at yahoo.com> wrote:
> Hi Grant,
> 
> Thanks for the tip. After looking at the datasheet,
> I'm wondering how good this part would be. The top
> of
> datasheet claims 0.5% accuracy, but later it clams
> 1%
> precision and 1% matching. It is the matching number
> we want. I believe that one can do better with
> discrete. With 0.1% resistors I can get 0.2%
> matching.
> And if I has a really good ohmmeter to further match
> the resistors, I could get much better numbers.
> 
> By the way, do you know how close transistors need
> to
> be? What is a good matching number to shoot for?
> 
> --tr
> 
> --- Grant Richter <grichter at asapnet.net> wrote:
> > There is a really great chip from Burr-Brown (now
> > TI) called the REF200. It
> > is two floating 100 uA precision constant current
> > sources and a current
> > mirror in one package:
> > 
> >
>
http://focus.ti.com/docs/prod/folders/print/ref200.html
> > 
> > While not cheap, this is a WAY good chip for
> > building a transistor matching
> > rig. The Moog schematic is just a constant current
> > source/sink (for PNP/NPN)
> > and you put a volt meter across Vbe. You can match
> > transistors at any
> > current. Some people use 1 mA, I've had good luck
> > with 100 uA which is a
> > decade below max current into a CA3080 pin 5.
> > 
> > If you put the 100 uA current source inside a full
> > wave rectifier, you have
> > a bi-directional current source (see the REF200
> > datasheet). You can take a
> > transistor socket with a BDCS and a DPDT switch
> and
> > match PNP or NPN in the
> > same socket with the same current source.
> > 
> > I'll try to find time to draw a schematic. It's
> just
> > the same as the Moog
> > circuit only using the REF200 instead of the
> op-amp
> > controlled transistor.
> > 
> > Get a large sheet of conductive foam, stick paper
> > labels over it with mv
> > marked from 560 to 640 (depending on room
> > temperature). Use plastic tweezers
> > to handle the transistors and sort 100 each PNP
> and
> > NPN into the correct
> > bins at one sitting. Then just grab two from the
> > same bin and they are
> > matched.
> > 
> > The REF200 is just great to have around. For
> > instance, a REF200, 555 and a
> > cap makes a nice fixed frequency linear ramp
> > oscillator (for a dither
> > oscillator).
> > 
> > 
> > > From: Tim Ressel <madhun2001 at yahoo.com>
> > > Date: Mon, 9 Dec 2002 17:02:54 -0800 (PST)
> > > To: synth-diy at dropmix.xs4all.nl
> > > Subject: [sdiy] Transistor Matching (was moog
> high
> > pass flter)
> > > 
> > > Yo,
> > > 
> > > Question: How close does the trannies have to be
> > > matched for say a ladder filter? is 6mV close
> > enough?
> > > 
> > > By the by, I breadboarded, for lack of a better
> > name,
> > > a differential transistor matcher. I used a
> CA3096
> > to
> > > make both an NPN and PNP matcher. The idea is
> > this:
> > > you can run two transistors at the same time,
> and
> > tie
> > > them together to keep them at the same
> > temperature.
> > > Also the voltmeter now reads the difference
> > directly.
> > > Seems to work pretty well.
> > > 
> > > --tr
> > > 
> > > 
> > >
> __________________________________________________
> > > Do you Yahoo!?
> > > Yahoo! Mail Plus - Powerful. Affordable. Sign up
> > now.
> > > http://mailplus.yahoo.com
> > > 
> > 
> 
> 
> __________________________________________________
> Do you Yahoo!?
> Yahoo! Mail Plus - Powerful. Affordable. Sign up
> now.
> http://mailplus.yahoo.com


__________________________________________________
Do you Yahoo!?
Yahoo! Mail Plus - Powerful. Affordable. Sign up now.
http://mailplus.yahoo.com