My L212 sub amp starting sounding weak (lower output at the same sound levels for the satellites), and after having an Al electrolytic failure in one of my TEK 576 curve tracers (which use the same vintage power supply caps), I decided I would replace the caps in the JBL sub amp.
These are a real b*tch to get out without damaging the traces on your PC board, especially since the traces are thin and delaminate from the board easily with any amount of heat or force. I suggest cutting the tabs as close to the board as possible, then use a de-soldering tool like a Hakko 808 vacuum desoldering gun to remove as much solder as possible. The tabs on the outer can are the most difficult due to the heat sinking of that large outer can.
It is difficult to find direct replacements for the original parts, which were 4 x Mallory 2500 MFD 60 VDC parts in my unit. New old stock would not be that desirable as replacements due to the age of those parts. I decided to upgrade the caps with 4 x 27,000 MFD Panasonic THA 105 C industrial grade low ESR parts, which were mounted on a separate PC board. You can get various power supply PC boards from EBay vendors and some produce some very nice epoxy glass boards with heavy 2 ounce copper traces. These boards fit most of the modern prong style electrolytic caps such as the Panasonic snap in caps. These were 40 mm diam x 80 mm high, and with 1/4" stand offs just fit below the frame dimension of the original JBL pc board (see photos). I also replaced the bridge rectifier with a 40 amp 600V part to handle the additional inrush current from the large capacitors. The new caps result in a 10X increase in power supply capacitance over the original units. On the AC input side I replaced the flimsy zip cord two prong power cord with a standard fused IEC AC input module. This module allows the use of standard or audiophile removable IEC three prong power cords.
While I was doing the PS mods I decided to upgrade the amp's semiconductors as well. Although the service manual specs the output Darlington power transistors as 2n6388/2n6668 [80V, 10 amp] parts, my unit actually had 2n6044/2n6041 parts [80V, 8 amp] installed. Even though the current rating for the latter parts is lower, the safe operating area of the 6044/41 parts is considerably better so I understand why they were used instead. I ended up using the 2n6042/2n6045 100 V rated parts from ON semi (the Motorola parts are NLA and ON Semi bought a lot of the older Motoral fabs). I gain matched the output transistors with the help of my 576 curve tracer, but this is very tricky due to the extremely high gain of the Darlingtons. With very little base current (on the order of 100 mico amps) you can drive the collector current to levels that overheat the TUT and distort the results. I was actually quite impressed with the consistency of the ON Semi parts and was able to get 4 nicely matched parts out of a batch of ten easily.
I also added a bias control pot by replacing R27 (4.7 ohm) with a 20 Ohm 20 turn trimmer. This allowed me to adjust the class A bias levels to about 40-50 mA/transistor. Turns out this is about the maximum you can get for this amp because it has a strange turn on exotherm that pushes the bias current to about 3/4 amp (per transistor) when cold started, which then self regulates back to 40 or 50 mA as everything heats up. So it is best to leave this amp on all the time or you will get this high current startup everytime the amp is cold started. It also turns out my pot setting ended up at 4.5 ohms, very close to the original JBL value of 4.7 ohms. This cold start behavior is another reason to have a good safe operating area which avoids secondary breakdown of the output transistors.
Reply With Quote
Originally Posted by 4313B
