I blew out a couple output transistors on my PL400, what next?

George S.

Veteran and General Yakker
Joined
Feb 18, 2020
Messages
4,539
#61
As Perry would say...Huh??

Phase Linear Pre-Driver and Driver Selection

This post is to provide a set of guidelines for selection as the sources for original devices are drying up and substitutes will have to be sought.

First let's get the definitions clear:

Pre-Drivers are classic RCA410 location. Typically Q11 and Q12 in the PL400 and the PL700B.

Drivers are the classic PL909 location. Typically Q13-Q18 in the PL400 or Q13-Q22 in the PL700B.

Pre-Driver Requirements (less demanding than the Driver requirements):

TO-3 Package

Vce = 200V min.

Slow, Ft less than 1/2 of the output device. Or alternatively 2X it. But not the same. The larger separation, the better. This requirement affects the negative half side of each channel (typically Q12) in order to ensure stability of the quasi-PNP transistor that is formed by the local feedback combination of Q7, Q12 and Q14, 16, 18 (also Q20 and 22 for the PL700B). The positive half side of each channel can use the same Pre-Driver as the Driver as there is no such local feedback condition on the positive half.

Low current gain (this contributes to the low Ft).

Lower noise is better (hardly ever specified, determined by measurement)

Recommended devices (in order of preference, the first 2 were the only types specified by Phase Linear): RCA410, SJ2741, MJ410, MJ413, MJ423. possibly 2N6678

Driver Requirements:

TO-3 package

Vce>= 250V

Large SOA, especially ability to sustain more current in the 50msec range at 200V (160V for the PL400) is important. The more current it can withstand, the better. The 21196 is well suited in this regard in that it will safely sustain 1.5A for 50msec at 200V (this is the half sine-wave time duration for a 10Hz signal).

Ft >= 2X the predriver is ideal, the higher the better. Minimum 4MHz is acceptable, higher is better.

Lower noise is better (hardly ever specified, determined by measurement)

The more current gain, the better. The 21196 has hFE between 50-75 at 3A.

Recommended devices (in order of preference): MJ21196, MJ21194, MJ15024

What is the difference between Quasi Complementary and Full Complementary?

With fully comp circuit topology, the positive half and negative half of the signal are amplified through the final current gain sections with symmetrical circuit design through the use of NPN drivers in the plus half of the signal and PNP drivers in the minus half of the signal.

The quasi comp used NPN on the plus half, like the full comp version but the minus half had the PNP driver emulated by a combination of a small PNP transistor coupled with 2 NPN pre driver and driver instead of using PNP devices. When the Phase Linear was initially introduced, this is what the designers had available to design with so they went with it.

In the last of the PL400 and PL700 production, PL did convert to fully comp on a small number before end of life of the product.

Some benefits, amp gain is more equal on the positive and negative half of the signal. In the quasi comp version, the gain of the negative side was always higher than the positive half. The global feedback compensated for this difference however and quite effectively nulled it out.

The bias settings are now extremely close for both positive and negative half. Part of this comes from the use of precision 1% resistors in the bias resistor locations, the balance comes from the symmetrical topology.

The current limit protection circuits are also now fully symmetrical leading to a more balanced onset of protection for each of the plus and minus sides.

Of course with this PL400 Backplane board, the configuration of full comp or quasi comp is easily selected by several components, jumpers and of course the choice of output drivers.

In the full comp versions, the MJ21196 (NPN) is used in the pre driver and driver positions on the plus half and the MJ21195 (PNP) is likewise used in the pre driver and driver positions on the minus half.

The use of these outstanding drivers everywhere leads to a more robust output stage as a side benefit.
I'm going to print this out and pin it to the wall above my electronics work area. Answers a lot of ??? I've had. Thanks!
 

Gibsonian

Chief Journeyman
Joined
Aug 3, 2011
Messages
777
#64
Getting ready myself to install the latest dual mono board in my fully comp 400, hopefully done tonight so can compare Joe's older rev (not a dual mono, can't remember rev) to the latest. I think I'll be able to easily hear the difference based on my experience with my quasi comp 700, miss Cathy.
 

jbeckva

Admin/Server Dude MAA Guy
Staff member
Joined
Jun 11, 2010
Messages
6,842
Location
Powhatan, Virginia, United States
Tagline
WassupYa Mang?
#65
As Perry would say...Huh??

Phase Linear Pre-Driver and Driver Selection

This post is to provide a set of guidelines for selection as the sources for original devices are drying up and substitutes will have to be sought.

First let's get the definitions clear:

Pre-Drivers are classic RCA410 location. Typically Q11 and Q12 in the PL400 and the PL700B.

Drivers are the classic PL909 location. Typically Q13-Q18 in the PL400 or Q13-Q22 in the PL700B.

Pre-Driver Requirements (less demanding than the Driver requirements):

TO-3 Package

Vce = 200V min.

Slow, Ft less than 1/2 of the output device. Or alternatively 2X it. But not the same. The larger separation, the better. This requirement affects the negative half side of each channel (typically Q12) in order to ensure stability of the quasi-PNP transistor that is formed by the local feedback combination of Q7, Q12 and Q14, 16, 18 (also Q20 and 22 for the PL700B). The positive half side of each channel can use the same Pre-Driver as the Driver as there is no such local feedback condition on the positive half.

Low current gain (this contributes to the low Ft).

Lower noise is better (hardly ever specified, determined by measurement)

Recommended devices (in order of preference, the first 2 were the only types specified by Phase Linear): RCA410, SJ2741, MJ410, MJ413, MJ423. possibly 2N6678

Driver Requirements:

TO-3 package

Vce>= 250V

Large SOA, especially ability to sustain more current in the 50msec range at 200V (160V for the PL400) is important. The more current it can withstand, the better. The 21196 is well suited in this regard in that it will safely sustain 1.5A for 50msec at 200V (this is the half sine-wave time duration for a 10Hz signal).

Ft >= 2X the predriver is ideal, the higher the better. Minimum 4MHz is acceptable, higher is better.

Lower noise is better (hardly ever specified, determined by measurement)

The more current gain, the better. The 21196 has hFE between 50-75 at 3A.

Recommended devices (in order of preference): MJ21196, MJ21194, MJ15024

What is the difference between Quasi Complementary and Full Complementary?

With fully comp circuit topology, the positive half and negative half of the signal are amplified through the final current gain sections with symmetrical circuit design through the use of NPN drivers in the plus half of the signal and PNP drivers in the minus half of the signal.

The quasi comp used NPN on the plus half, like the full comp version but the minus half had the PNP driver emulated by a combination of a small PNP transistor coupled with 2 NPN pre driver and driver instead of using PNP devices. When the Phase Linear was initially introduced, this is what the designers had available to design with so they went with it.

In the last of the PL400 and PL700 production, PL did convert to fully comp on a small number before end of life of the product.

Some benefits, amp gain is more equal on the positive and negative half of the signal. In the quasi comp version, the gain of the negative side was always higher than the positive half. The global feedback compensated for this difference however and quite effectively nulled it out.

The bias settings are now extremely close for both positive and negative half. Part of this comes from the use of precision 1% resistors in the bias resistor locations, the balance comes from the symmetrical topology.

The current limit protection circuits are also now fully symmetrical leading to a more balanced onset of protection for each of the plus and minus sides.

Of course with this PL400 Backplane board, the configuration of full comp or quasi comp is easily selected by several components, jumpers and of course the choice of output drivers.

In the full comp versions, the MJ21196 (NPN) is used in the pre driver and driver positions on the plus half and the MJ21195 (PNP) is likewise used in the pre driver and driver positions on the minus half.

The use of these outstanding drivers everywhere leads to a more robust output stage as a side benefit.
That's some excellent info, Joe. Should probably sticky it somewhere around here... We should get you a section called "Joe's Corner" or something like that. I've always appreciated the insight, filling in the holes from my own path in 'trons.
 

rollcentre

New Around These Parts
Joined
Mar 22, 2021
Messages
48
#72
About to shake down my second amp (the '74). This WOPL-ization went by pretty smoothly. Wish me luck.

After the 3rd or 4th one I imagine it goes pretty quick.

Btw Joe, thanks for pre-bending all of the leads. That is much appreciated.

PXL_20210629_022705361.jpg
 
Last edited:

rollcentre

New Around These Parts
Joined
Mar 22, 2021
Messages
48
#73
Damn! I whiffed it again.

It bought up fine with the variac and DBT.

My DC offset # were pretty much a zero on both channels.

Bias on the left channel I dialed in about .375v but the right side is much too low, Its about .037v when the pot is in the same position as the left.

I must have missed something.
 
Last edited:

Gibsonian

Chief Journeyman
Joined
Aug 3, 2011
Messages
777
#74
Pre bent leads? I didn't know, wow. Might've got me to solder a board, actually glad I didn't know this up front. I went the easy route, but it comes tested and set for bias already. It's pop in and play.

Roll on one of your amps you gotta do the backplanes and go fully comp. You'll never know if you don't........................
 

rollcentre

New Around These Parts
Joined
Mar 22, 2021
Messages
48
#76
The bias transistor checks out. I also made sure the leads weren't switched at the control board.

I'll start going over the board with a fine toothed comb look looking for an inconsistency between the right and left side of the board. This amp was actually working, so it must have been something I did on the board, or the board wiring
 

rollcentre

New Around These Parts
Joined
Mar 22, 2021
Messages
48
#77
Visually , I'm not seeing an inconsistency, but it doesn't hurt to have another set of eyes looking, do guys see anything?

PXL_20210629_140927924.NIGHT.jpg
 

mlucitt

Veteran and General Yakker
Joined
Jun 24, 2011
Messages
3,339
Location
Jacksonville, FL
#78
Can you check all the voltages (referenced to ground) on the bottom of the Control Board. First the Left channel, then the Right channel. No inputs and no outputs connected.
 

George S.

Veteran and General Yakker
Joined
Feb 18, 2020
Messages
4,539
#80
Also, you haven't populated the complementary output portion of the control board have you? Looks like every component has been stuffed (complementary and quasi) unless I'm missing something. When I build a control board the first thing I do is fill component locations I won't be using with solder so I don't inadvertently stuff them.
 
Top