I have been testing a pile of JFETs (multiple different part numbers) with multiple different instruments to find a simple DIY method for creating matches for some JFET Beast PCB builds.
I believe that matching can be accomplished simply and at a very low cost. This thread is to show my conclusion of a rig to measure each JFET.
The journey to get to this point is a discussion for another time.
I believe that matching can be accomplished simply and at a very low cost. This thread is to show my conclusion of a rig to measure each JFET.
The journey to get to this point is a discussion for another time.
The red Harbor Freight DMM is placed in the 200 ohm measurement range. The other DMM is placed in the mA current measurement range. The red DMM is forcing 1.88 mA to measure on the 200 ohm range. It turns out that 1.88 mA is a relatively good current for characterizing a JFET to go into a JFET Beast build.
This measurement is a J111. The D-S resistance at Vgs = 0 at 1.88 mA of current is 23.7 ohms. The measurement requires about 90 seconds to settle. 2 minutes is what I use to let the measurement settle. I have plotted the settling time with a graphical sourcemeter for a large number of JFETs and 2 minutes appears to work.
whatever procedure is, I did found that - for me - every connection must be solid (soldered) excepting (of course) probes itself and JFet contacts in socket
be it Gremlins or jut my bad luck, if that's not the case, I'm paying with bunch of JFets measured doubtfully
that's why I made all my measuring jigs in permanent form .......
best possible socket for DUT, proper sockets for probe ends, relay with doubled contacts instead of manual switch
be it Gremlins or jut my bad luck, if that's not the case, I'm paying with bunch of JFets measured doubtfully
that's why I made all my measuring jigs in permanent form .......
best possible socket for DUT, proper sockets for probe ends, relay with doubled contacts instead of manual switch
Hello ZM. I agree. I am working with a designer friend to make a more solid measurement rig. That will cost $$$. I wanted this thread to show a very inexpensive and easy way to take measurements for matching.
not necessary
veroboard and IC sockets, doesn't need to be fancier
I mean, do whatever way you prefer, I was just giving my experience......... at least half of it being minor or major ookups
I have been using this jig. 5 dollars cdn shipped to my door for this matching jig by PRASI.
https://www.diyaudio.com/community/threads/another-idss-measurement-jig.353425/#post-6180935
https://www.diyaudio.com/community/threads/another-idss-measurement-jig.353425/#post-6180935
A proper design of test fixtures for small signal or power FETs requires thorough understanding of how voltage and current should be measured in the presence of unavoidable, finite contact resistances.
That may come from the test socket for the DUT, or setting switches, or otherwise.
All of the above have some resistance, which can also change during their lifetime.
Doing it properly and having PCB designed for that purpose are two completely different things.
Patrick
That may come from the test socket for the DUT, or setting switches, or otherwise.
All of the above have some resistance, which can also change during their lifetime.
Doing it properly and having PCB designed for that purpose are two completely different things.
Patrick
Hi Patrick.I appreciate your points.
The purpose of this thread is to support those builders who want to build one version or another of a mini JFET Beast.
The JFETs are USD $0.10 each. I think that most builders want a simple and cheap way to characterize, sort, match and use the JFETs for their intended Beast.
The JFETs in a Beast will likely be idling from 5mA to 20mA at somewhere between 12v and 30v. I think that my recommendation will work adequately.
My final recommendation will be to
1. Perform a preliminary sort for Idss. Place the JFETs into piles of 5mA, 6mA, 7mA, etc......
2. Measure each pile for Rds and record Rds for each JFET
3. Sort each pile by Rds
Where matched pairs are required, use the matches in the sorted piles.
The inexpensive way to measure Idss requires
1. Power supply
2. DMM with a mA range ( I found one for $20 on Amazon)
3. Breadboard and wiring
Again, the intent is for simplicity and low cost while being able to sort a pile of JFETs into usable pairs or groups.
The purpose of this thread is to support those builders who want to build one version or another of a mini JFET Beast.
The JFETs are USD $0.10 each. I think that most builders want a simple and cheap way to characterize, sort, match and use the JFETs for their intended Beast.
The JFETs in a Beast will likely be idling from 5mA to 20mA at somewhere between 12v and 30v. I think that my recommendation will work adequately.
My final recommendation will be to
1. Perform a preliminary sort for Idss. Place the JFETs into piles of 5mA, 6mA, 7mA, etc......
2. Measure each pile for Rds and record Rds for each JFET
3. Sort each pile by Rds
Where matched pairs are required, use the matches in the sorted piles.
The inexpensive way to measure Idss requires
1. Power supply
2. DMM with a mA range ( I found one for $20 on Amazon)
3. Breadboard and wiring
Again, the intent is for simplicity and low cost while being able to sort a pile of JFETs into usable pairs or groups.
I have pivoted on what to use for simple and inexpensive matching of a large pile of JFETs.
This takes me back a zillion years on how to plot the impedance vs frequency of a loudspeaker.
This rig only requires
1. 10 volt power supply
2. 10k resistor
3. Proto board with wires
1mV reading on the voltmeter equals 1 ohm D-S resistance of the JFET. Most inexpensive multimeters have a 200mV range yielding 3 significant digits for 10 ohms through 99 ohms.
The J111 that I used to evaluate the rig measures 24.89 ohms using a $7000 sourcemeter. It measures 25.04 ohms on the resistive divider rig.
The current through the JFET is nominally 1mA.
This will work for matching JFETs for mini JFET Beast builds.
This takes me back a zillion years on how to plot the impedance vs frequency of a loudspeaker.
This rig only requires
1. 10 volt power supply
2. 10k resistor
3. Proto board with wires
1mV reading on the voltmeter equals 1 ohm D-S resistance of the JFET. Most inexpensive multimeters have a 200mV range yielding 3 significant digits for 10 ohms through 99 ohms.
The J111 that I used to evaluate the rig measures 24.89 ohms using a $7000 sourcemeter. It measures 25.04 ohms on the resistive divider rig.
The current through the JFET is nominally 1mA.
This will work for matching JFETs for mini JFET Beast builds.
