Dumb question about "matched transistors"

[Locrian99]

It’s always the shipping…

 

[BryGuy]

Just found this video helpful

Haven’t seen one of Paul’s videos in a while. He seems to have stopped posting them for whatever reason. Really good stuff though.

 

[andare]

Haven’t seen one of Paul’s videos in a while. He seems to have stopped posting them for whatever reason. Really good stuff though.

Yes, so much so that that's Eric, the other owner of the company who's been making videos on the channel instead of Paul. Something about Paul needing to take care of other parts of the business and Eric being more knowledgeable on technical matters.
This channel was my gateway drug into pedal making. I hope Paul resurfaces soon.

 

[MichaelW]

I'm using a DCA75 with SMD adapter.

I tested a few MMBFJ201s similar to how you did it, then tested them again with the SMD adapter and got completely different measurements.

Ok, so I have another dumb question.

I soldered 8 of the 10 MMBF2n5457's I had tested earlier with my "pencil eraser" jig (same batch as the ones I sent you) on adapter boards and measured them. Not getting anything close to -1.4v. They were all in the positive numbers....around .4xx to .8xx range.

I'm a bit confused about this JFET matcher board.

This is what the board looks like. Note that it only has 2 unmarked "DMM Test Points".



Here's the schematic...



Again, it only says DMM1 and DMM2.

I've been assuming that DMM 1 is where I should have the COM lead of my meter and DMM2 is where I am measuring the voltage with the red lead of my DMM. Obviously, flipping them gives me negative readings.

Now I'm confused, do I have this backwards?

 

[fig]

I think you need 2 DMMs No, that’s another method, sorry.

I think you my have the leads reversed yes. You are reading the voltage/current from the opamp output that is trying to look like a 10kΩ .

 

[MichaelW]

Ok, I think I may be doing this backwards. Looking again at RG Keen's article and also looking at tagboard effects layouts for a Vero based matcher, I think COM or negative lead of my DMM should be at the DMM2 testpoint of the schematic above. Ugh....stretching my brain here about stuff I don't understand. Can someone confirm? TIA!

 

[Deperduci]

I see where you might be getting confused, the pin 3 of IC ALSO being listed as positive.. other than the +/_ are the numbers the same when you test either way? If you're getting the same readings numerically but only pos/neg are swapped.. I wouldn't worry to much, the last time I built a pge board I clipped my negative to the "hot" pin of the power jack and the numbers still matched regardless of the +/- inidcator, but that might be my odd ball luck

 

[MichaelW]

Yah that’s what’s confusing me. But I think I got it now after looking at some of the other tag board layouts.

 

[MichaelW]

Ok just for sh*ts and giggles, I built it out today. Tried the MMBF5457's, didn't work, very light phasing.
Then I tried 4 random MMBFJ201's, slightly more phasing (I think) but still doesn't work.
Sigh, guess I'll have to wait for the matched set. Oh well......
(This was with a 200k trimmer, I couldn't find a 250k.)

 

[MichaelW]

ok, more dumbassery abounds.....I forgot that the 2n5952 pinout was reverse from the 5457's and J201's.
I turned the 201's around and it works now. Actually pretty decent phasing. (Randomly chosen MMBFJ201's)
I'll try the MMBF5457's again oriented correctly later today and compare.

 

[RobertJay]

Hey all, is this still a place to chat about this? I too am in the same boat. I have about 25 UNMATCHED 5952's, 12 of which went into a BYOC Lil phaserpcb to crappy results. I want to test them to match them and just don't know if I HAVE TO build the tester, us the KEEN tester I made for Germanium tests, or if I can simply use mt TC-1 component tester. When I do use the mini tester, I get 2 readings listed as what looks like @Vg. One reading =202mV and another that says 421mV. Actually here's what the screen reads for a test...
Idss=2.5mA
Id=0.29mA
@Vg=201mV
Id=0.00mA
@Vg=421mV
Are any of these values what I need to match my quad sets? Why do I have a tester and NOT know what it's telling me? I'll look these readings up in case this thread is dead. Thank!

 

[giovanni]

You need to use the RG Keen method. Chuck posted a really good explanation: https://forum.pedalpcb.com/posts/178703/

 

[DeadAirMD]

Man oh man do I love a slow phaser into an overdrive

 

[darwin999]

Just updating this for people that arrive at this old post through search.

Obviously, one popular use of matched 2N5952's is for a phaser. I stumbled upon the following useful information in Aion FX's build docs for their Apex (MXR Phase 45) and Zenith (MXR Phase 90) phasers, so I thought I'd share some selected excerpts here.

• Most TO-92 through-hole JFETs have been discontinued, and of the few that are still active, none are very well-suited to the Phase 45 / Phase 90 circuits. There are still plenty of SMD types available, though. The Toshiba 2SK208-GR and Fairchild/ONSemi MMBF5485 (2N5485) are great choices, though matching can be difficult due to the SMD's small size.
• For the MXR Phase 90 phaser, the JFETs in the four phase stages must be matched for the phasing effect to occur. The JFETs are used as variable resistors controlled by the gate voltage. When used as soft switches, the gate voltage changes make the source-drain channel resistance switch between very low ("on"), in the hundreds of ohms, and very high ("off"), in the megaohms. However, for soft switches, the exact “on” and “off” resistance is not important. But for the case of phasers, we need a set of JFETs that have nearly the same channel resistance when the same gate voltage is applied. This way, all four of the phase-shift stages will respond the same way to the LFO voltage.
• This means that for phasers, the JFETs must be measured and sorted. Historically, most manufacturers graded JFETs according to their VGS(off) parameter, the voltage at which the channel resistance is highest, also called pinch-off voltage. However, some luminaries in the DIY scene suggest that it’s better to match phaser JFETs to their VGS(10k) parameter, the voltage at which the channel resistance is 10k - which is within the normal operating range of a phaser. This leads to better matching, since the correlation between VGS(off) and VGS(10k) isn’t always exact.
• The most straightforward way of measuring JFETs is to use an electronic component tester such as the Peak Atlas DCA75. This will give you all of the standard datasheet parameters of a particular JFET, including VGS(off). These testers won’t give you the VGS(10k) value, but it’s very simple to breadboard a circuit that can test either the OFF value or the 10k value.
• How well matched do they have to be? Aion measured the stock FETs on an MXR Phase 90 and found VGS(10k) to be matched to within 0.09V, although Aion recommends matching to within 0.05V for better results.

Note that for those that can't or don't want to do their own matching, Aion sells highly-matched 2SK208-GR FETs - which are "typically matched to within ~0.01V of each other" [for VGS(10k)], which is impressive. They look to be a good value ($8 for 4), especially compared to matched NOS 2N5952's at today's prices.

• https://aionfx.com/component/2sk208-gr-matched-set/

Here's the links to read more, as Kevin go through the issues in additional detail.

• https://aionfx.com/app/files/docs/zenith_documentation.pdf
• https://aionfx.com/app/files/docs/apex_documentation.pdf

ps: same post appears on a different post on the same topic, i.e., https://forum.pedalpcb.com/threads/...-matched-quad-2n5952-jfets.16458/#post-279253