Announcement

Collapse
No announcement yet.

my transistor matching adventure

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • my transistor matching adventure

    hi Everyone,

    I recently purchased 100 MJL 21194 and 100 MJL 21193. the first thing I wanted to do was match them, using a common method, the Hfe on my multi meter. I made 3 clip leads with wire so I could easily attach them to the transistor without having to bend the transistor legs to fit in the holes as they were way too close together. I figured with 100 transistors I should have no problem getting a large amount to match. These trannies are from a single lot number from on semiconductor.

    Well here are some pics of what I discovered.
    Click image for larger version

Name:	tran 1.jpg
Views:	1
Size:	92.7 KB
ID:	50892Click image for larger version

Name:	pic 2.jpg
Views:	1
Size:	92.5 KB
ID:	50893Click image for larger version

Name:	tranny 3.jpg
Views:	1
Size:	93.1 KB
ID:	50894

    The lowest Hfe value was 23 the highest was 84. There was a bell curve very noticable, in that in between the 50 and 70 range were the most transistors. Even with this there was a huge difference in gain from the lowest to the highest. So you can imagine if you have a string of 4 or 8 transistors and you had one with an Hfe of 23 and one with an Hfe of 84, how unbalanced your circuit would be.

    What was also interesting to note, was that if I held the transistor in my hand with my thumb on the back, it would heat it up a bit and the gain would go up as much as 3 points over a peroid of a few minutes.

    Now add to this that if you were building a full bipolar circuit you would have to match the 93's and the 94's so both sides of the switch interacted with the coil the same way.

    So to put this out to the group, you have all of these differences, how would you mathc the gain board to board? I know how I would do it, but I want some more ideas.

    Tom C


    experimental Kits, chargers and solar trackers

  • #2
    Tom,
    I've noticed the variation of Hfe's as well. Particularly with heat. As for me I have picked the average temp in my house, not my shed! The house usually sits at about 24 deg. So During the summer I cool, during the winter heat, till i get my ideal environment. Then I use a semiconductor analyser. I don't touch it more than a few seconds and none at all if I can avoid it. Attach the clips and then print a little sticker with the gain on it for future reference. Then I can be sure it has been tested in standard conditions.
    I did this for my first 20 MJL 21194's and got 3 groups of 4 and 2 pairs and 4 odd. I thought that was pretty good. I've just ordered another 30.
    On the board, not that i have much experience at all in this field, I would be going for symmetry, track width, length, pad size etc. The bigger the track the less of the percentage variation. Being that we match so many components if I could make a completely symmetrical board I would. Sadly my engineering skills in this area leave a lot to be desired .
    James

    Comment


    • #3
      Originally posted by Tom C View Post
      So to put this out to the group, you have all of these differences, how would you mathc the gain board to board? I know how I would do it, but I want some more ideas.

      Tom C
      Hi Tom,
      from one board to the next, the branch resistors should do the trick as long as each transistor in any given coil is w/in 2-5% ??
      Bedini Ferris Wheel Regauging Motor - Page 32
      "Match them all to about the same reading for current 2 to 5% will be fine."

      kind regards,
      Patrick

      Comment

      Working...
      X