P channel fet matcher

Help Support GroupDIY:

Spencerleehorton

Well-known member
Joined
May 12, 2012
Messages
3,261
Location
Felixstowe, Suffolk, UK
Hi All,

Trying to match a few p channel fets and wanted to make a simple circuit to match them.

Has anyone got a link please to a good circuit, I seem to only find n channel fet matchers?
 

ruffrecords

Well-known member
Joined
Nov 10, 2006
Messages
13,518
Location
Norfolk - UK
I am no semiconductor expert but, in the best Dr. Who tradition, if you take an N-channel circuit and reverse the polarity of the energy source. it should work for P-channel. :cool:

Cheers

Ian
 

abbey road d enfer

Well-known member
Moderator
Joined
Jan 22, 2008
Messages
12,622
Location
Marcelland
Hi All,

Trying to match a few p channel fets and wanted to make a simple circuit to match them.

Has anyone got a link please to a good circuit, I seem to only find n channel fet matchers?
If your concern is matching Vgs and Gm, you can use such a thing:
I use it everyday for appairing transistors and capacitors. Very good $25 investment.
Of course it depends what degree of matching you need.
My view on the subject is that a good design should not need superlative matching; it should be tolerant or provide means of adjusting to tolerances.
 

JohnRoberts

Well-known member
Staff member
Moderator
Joined
Nov 30, 2006
Messages
20,484
Location
Hickory, MS
The last time I matched JFETs was for a phono preamp kit back in 1980(?). Despite being used in the Cohen (cough) topology that should hold the device current relatively stable I measured a distortion difference from using unmatched JFETs with different Vgs pinch off voltages. A bit of a surprise to me.

What I did (not saying its the only or best way) was to try to mimic the actual circuit the devices would be used in for measurement accuracy. For my case a phono preamp front end (basically common source but with a couple k resistor in series with the source to ground. I purchased something like 1k of JFETs and then tested them with gates grounded (0v), 2k resistor in series with source to ground, and drain to appropriate PS rail. I was testing Nch so for Pch test with drain to opposite rail. I basically let them all stabilize at room temperature and wrote down the Vgs on the paper envelopes I put them in. When it came time to sell the kits I sorted through all the envelopes for pairs with the closest Vgs voltages and stapled those bags together in pairs. The preamp used 2 JFETs per side.

I still have some of these JFETs in my back lab, but that's life in the kit business. :unsure: Not very remarkable JFETs but OK for a circa 1980 phono preamp (4.5nV/rt HZ IIRC).

JR

PS; I used a zero insertion force socket so I wouldn't heat the JFETs with my hand's heat while measuring.
 

Rob Flinn

Well-known member
Joined
Jun 3, 2004
Messages
4,501
Location
Sussex, UK
Question based on that there's no such thing as a stupid question. I use a peak dca transistor tester. It measures a bunch of parameters: Vgs (off), Vgs(on), gfs, Idss, Rds. Which of these parameters should I be matching for using fets in a compressor ?
 

JohnRoberts

Well-known member
Staff member
Moderator
Joined
Nov 30, 2006
Messages
20,484
Location
Hickory, MS
It kind of depends on the circuit design, but I will ASSume gate pinch-off voltage (Vgs off?, not sure what Vgs on is). For left and right channel tracking similar Vgs characteristics could be useful. Some designs are looking for a ballpark Vgs range.

Some of those metrics don't seem that germane to compressor design.
-Rds (on) is the on resistance when turned hard on. Useful perhaps for mutes.
-Idss is current with 0V Vgs. High Idss generally correlates with lower noise.
-Gfs (i had to look that one up) I didn't find a clear definition but it seems related to forward transconductance.

JR
 

ruffrecords

Well-known member
Joined
Nov 10, 2006
Messages
13,518
Location
Norfolk - UK
Question based on that there's no such thing as a stupid question. I use a peak dca transistor tester. It measures a bunch of parameters: Vgs (off), Vgs(on), gfs, Idss, Rds. Which of these parameters should I be matching for using fets in a compressor ?
Do you have the regular model or the Pro?

Cheers

ian
 

Rob Flinn

Well-known member
Joined
Jun 3, 2004
Messages
4,501
Location
Sussex, UK
Do you have the regular model or the Pro?

Cheers

ian
I have the pro model. I can get it to curve trace if I hook it up to my laptop, but the software doesn't seem to have an obvious way to make comparisons. I emailed PEAK about it but they didn't respond.
 

ruffrecords

Well-known member
Joined
Nov 10, 2006
Messages
13,518
Location
Norfolk - UK
I have the pro model. I can get it to curve trace if I hook it up to my laptop, but the software doesn't seem to have an obvious way to make comparisons. I emailed PEAK about it but they didn't respond.
The two key parameters for FET matching are IDSS and the pinch off voltage both of which the pro model will measure so you should be good to go. I have been toying with buying one myself for some time. Do you think it is worth the money?

Cheers

Ian
 

Rob Flinn

Well-known member
Joined
Jun 3, 2004
Messages
4,501
Location
Sussex, UK
The two key parameters for FET matching are IDSS and the pinch off voltage both of which the pro model will measure so you should be good to go. I have been toying with buying one myself for some time. Do you think it is worth the money?

Cheers

Ian
It's quick & easy to use. In the main I just use it for testing transistors when repairing stuff for people. It does this well, but has a whole load more to it that generally I don't use, like the curve tracing for example. I guess there are probably cheaper boxes that would do what I mainly use it for. I wish they would make the software so that you could curve trace a bunch of devices & it would log the parameters maybe in spreadsheet format to make matching easier.
 

Spencerleehorton

Well-known member
Joined
May 12, 2012
Messages
3,261
Location
Felixstowe, Suffolk, UK
Well using a simple fet tester circuit I seem to have 4 matches quads from 21 fets, 2n3820.
When I say matched I mean with 10% of each other.
Just needed to reverse polarity for p channel as Ian said.
2 resistors 100R and 1meg and a 47uf cap with 3 way switch.
 

ruffrecords

Well-known member
Joined
Nov 10, 2006
Messages
13,518
Location
Norfolk - UK
Well using a simple fet tester circuit I seem to have 4 matches quads from 21 fets, 2n3820.
When I say matched I mean with 10% of each other.
Just needed to reverse polarity for p channel as Ian said.
2 resistors 100R and 1meg and a 47uf cap with 3 way switch.
Don't forget if you have an electrolytic cap in there you will need to reverse its polarity also.

Cheers

IAn
 

Latest posts

Top