Matching Transistors with DCA55/DCA75 and others

[ungifted]

I'm waiting for DCA75 http://www.peakelec.co.uk/acatalog/dca75-dca-pro.html (DCA55 is already purchased) and thinking about their abilities for matching transistors for various purposes.
Well, obviously DCA55 is good for measuring germanium transistors (Hfe and leakage) for guitar pedals.  It's ok.
What shall I expect from DCA75 with ability of curve tracing and measuring jfets etc? I've read through some topics here about matching/selecting transistors, can anybody share his experience of matching/selecting transistors with DCA55/75 (or Chinese transistor testers) for DOA, selecting compementary pairs, or perhaps possibility of matching pairs to substitute MAT02 and so on?
DCA55 can measure Hfe, Vbe, leakage (for germanium). DCA75 can measure jfets and have curve-tracing function. Does it really so important under what conditions we make measurements and where it's better to use curve tracing?

 

[ungifted]

Ok, lets narrow the question:
Matched transistors in DOA: very often we can see in the parts list something like - BC550C - 2pcs (matched) - does it mean two transistors matched by Hfe and Vbe? The sаme story in matching pairs for a differential input stage (good CMRR achivement).
Do we simply measure Hfe and Vbe and select pairs? If we use multimeter or DCA55 it's usually 2.5mA (2-3V) when measuring Hfe, and about 5v when measuring Vbe. DCA75 allows more accurate measurements with higher voltage but do we really need it?

When i read about measuring germanium transistors for FuzzFace clones we can find opinion that only measurements under 9v is the accurate one (RG Keen and Small Bear schemas).
So what is your experience - do we need matching and if we make matching, what instruments we can use?
Many thanks to everyone!

 

[SirNickity]

I would love to see some discussion on this, too.  The learning curve gets pretty steep once we start talking about component variations, particularly non-linear variations.

 

[Monte McGuire]

The best 'matching' jig is to put the device into essentially the same circuit as what you'll use it in, and then see how it performs. If that's difficult, then you can sometimes use a tester but center your focus on (for example) base currents and collector currents that circle the operating point your circuit will operate them within.

I have not extensively played with the DCA unit that I have, but when you plug it into a computer via USB, you can choose some drive and output characteristics to get a set of output curves that could allow one to select devices close to how a specific circuit will operate the devices. So, in that case, you don't necessarily have to build a surrogate circuit to match the devices, but one could still get very appropriate measurements of a pile of devices, and probably allow one to select some good pairs or quads or whatever is needed.

There will always be limitations on discrete devices, such as poor thermal tracking etc. that will put discretes at a disadvantage, but still, within that limitation, it should be possible to measure and select devices that are much closer than usual to each other and much closer to "ideal" for that circuit.

 

[pucho812]

I got a DCA75 for work to match transistors. It was a faster way to do things  as previously we used this analog sencor transistor checker. The DCA75 has worked quite well for us.    I have had great results matching fets that  have the same IDSS current.  So far it has been working great and I have yet to hear complaints.

 

[Jarno]

I have both the DCA55 and 75, I bought the 75 because of the improved accuracy, but upon reading the manual, there is a catch.
Although Vbe is indicated quite accurately, in the manual a larger measurement inaccuracy is mentioned. So, it is not quite as accurate as the indicated value on the display suggest.
Vbe is indicated in millivolts (or actually 0.001V) but the total accuracy (for Vbe) is +/- 0.006V, a tolerance range of a whopping 12mV. Ideally, you'd like transistors to be matched better than that.

Still, it is an awesome tool, just be aware of it's limitations.

 

[Whoops]

Normally people match just by the hfe value, but that happens because most DIY'ers dont have equipment to do curve tracing.

If you want to Match Tubes and transistors more accurately curve tracing is needed, and you have the ability to do so now.

I have an DCA75 unit, I love it

 

[Jarno]

I am not sure curve tracing the parts will help, since there still quite a significant measurement error, regardless if you read a value on the display, or curvetracing a part.
Like I said, use it, but beware of the limitations (as goes for each and every measurement tool).

 

[diydidi]

the Atlas meters are great for matching small signal transistors. I match for Vbe and Hfe regardless if they will be used as LTP or mirrors.
For output stage matching of BJT's and Mosfets, I use a home made jig to match at higher currents.  Match Vbe to 1% and gain within 10%.
Outputs are usually matched in the region of the amps idle current if they are to be parallel connected with ballast resistors.

 

[benb]

Ok, lets narrow the question:
Matched transistors in DOA: very often we can see in the parts list something like - BC550C - 2pcs (matched) - does it mean two transistors matched by Hfe and Vbe?

Which one, Hfe or Vbe? at what Vce and at what Ic (or Ib)?

The Moog transistor matcher (google it, there are several schematics and they appear to jibe) uses 10V Vcb (thus about 10.6V Vce) and 100uA Ie (which is within 1 percent of Ic at >100 Hfe) and measures the resultant Vbe.

I would guess there's a correlation between  Hfe and Vbe, depending on which of the other parameter(s) are controlled, but I'd have to try and see.

 

[dustbro]

Hey guys,
I just picked up a DCA75 for matching FETs.  Seems like a really cool device.... but I have no experience with it. Can someone post a suggested way to use the software for curve tracing and organizing of FETs?
thanks!!