Parallel Driver Transistors

Status
Not open for further replies.
Hi all, I am after a bit of Advice in trying to upgrade a High Power Chinese Amplifier.
Just trying to pre-empt shortcomings that could result in premature failure.
As standard the board is normally kitted out with 14 pairs of 2SC5200/A1439s. Of which 12 pairs are output transistors and 2 are drivers. Recommended supply voltage is up to +/- 135 volts. I do not intend to use them at anywhere near that supply voltage and aim to use +/- 70 volts so that bridging facility could be possible. I have used MJL 21193/4 as the Output devices with MJL 4281 and MJL4302 as parallel drivers. I first thought that if I done away with one of the parallel drivers and converted the resultant spare pad into an output transistor I would increase the output stage reliability by adding an extra pair of output devices. The now single pair of 4281/4302 with their much improved soar would be adequate to drive the now 13 pairs of output devices. After reading through related articles, I am now wondering if it may be asking too much? And maybe I should restore the paralleled pair for drivers. Another question that arose is the configuration of the parallel pairs of driver transistors. They are connected in complete strapped parallel with no emitter resistors. I wondered if it would be a good idea to fit individual emitter resistors to the devices if I was going to reinstate the parallel Driver device?

The driver/s drive the output devices via 10 ohm resistors with a 100 ohm resistor between base and emitter of each output device.The drivers are driven by 2sc5171/2sa1930 via 220 ohm resistors.
 
To parallel BJT's directly is ugly and bad idea, respectfully. It happens that the thermal runaway will affect in more manner in one than in the other and as a result, one of the will be blown or at least, overloaded. So, good thermal coupling plus emitter and base ballast resistor are mandatory. Note that MOSFET's with their positive thermal coefficient may be paralleled directly as they are self balancing with no extra elements.
 
Last edited:
Hi Osvaldo, thanks for your reply. I know in output transistors there would definitely be emitter resistors for each device but wasn't 100% sure about in driver stage application. In addition I have seen application where base ballast resistors were not used. Bearing in mind the resistor values in my text, would you recommend a 220R feeding each driver individually or another value, what value would you recommend for paralleled driver emitters, 10 ohm for each emitter going to bases for paralleled output devices?
 
Emitter ballast resistor creates some negative AC and DC feedback increasing the output impedance of the device. The bigger the emitter resistor, the better will be the balance between units. So, it depends on current trough the transistor, the voltage available to be drop in them, and the requirements of the stage.
 
Hi Sakti, that was my original thinking. I could imagine that the original 2sc5200/1493s might of struggled singularly to provide enough base drive and stay within their soar, especially if the high supply voltages were used and that's why the doubled them up, but I thought 4281/4302s would be sufficient to drive singularly.
 
What was not yet mentioned, but could probably also be considered, is the concept of subdividing the 12 pairs of output transistors into two groups of 6 pairs. Then, provide a pair of driver transistors for each of such groups. This would actually boil down to lifting half of the respective base resistors of the power transistors and creating a "separate bus".
Drive each pair of driver transistors, and their respective power transistor pairs subgroup, via some equal valued base ballast resistors.
Such scenario would probably enable equal loading of the driver transistors.
I assume.
 
Yes, understood. I recently saw something similar done in an Amplifier I repaired recently. They had taken it a few steps more by driving each output device with it's own driver. 16 output devices 16 driver transistors. Not sure that for high power pa use the mods would be worth it though.
 
Hi Sakti, that was my original thinking. I could imagine that the original 2sc5200/1493s might of struggled singularly to provide enough base drive and stay within their soar, especially if the high supply voltages were used and that's why the doubled them up, but I thought 4281/4302s would be sufficient to drive singularly.

2sc5200/1493 as driver for 77VDC is enough for 4 Ohm load, although a friend tested it using four 8 Ohm speakers paralel (2 Ohm) for one hour, but not recommended.
 
....if I done away with one of the parallel drivers .... ...would increase the output stage reliability by adding an extra pair of output devices....

This is literally "a half-dozen of one and seven of the other".

If the thing might be close to blow-up, *double* the stuff in it. A 7/6 increase is hardly worth mentioning.
 
If I understand correctly: You use 12 pairs of MJL21193/94 for +/-70V. It is quite over enginered, it can handle almost any kind of load. Usual desings are useing 4-5pairs for this voltage. Drivers, and predrivers are strong or stroger than necessary. It is a bullet proof amplifier, if it has electronic short circuit protection...

Sajti
 
Hi Bigun, thanks for the advice. I only buy from main distributors, Farnell, RS etc, or from On-Semi themselves. I'm not even sure if the 2SC5200/SA1493 that come with a lot of the Chinese Amps are genuine, even though you would have thought they should be cheap enough for their manufacturers.
 
Status
Not open for further replies.