HOME PROJECTS 70WATT BIULD

AUDIO POWER MODULE 70WSP

70wsp project is a multistage discrete amplifier based on a few different designs and some work on the calculator. This amplifier is amazingly simple and the basic structure has been published for some time but this is not a copy, more an elaboration on a good design and is quite clean even with so so components. I'm going to post a couple version of this amp that I have built and tested, the first is actually the one I recommend but the others have there benefits.

The 70wsp is a single channel solid-state audio power module rated at 70watts into an 8 ohm load with less then .01% THD @ 1kHZ. The high frequency response of this amplifier is adequate but future revision may have improvements in this area but probably not. .3% THD into an 8 ohm load @ 2 watts @ 15kHZ and goes down to .1 @ 50 watts. One look at the schematic and you can see why the upper frequency is not to good I used the tip31/32 combo as pre-drivers. These device are not real fast but it’s what I had on hand at the time, you could replace these with 2sd669/2sb649 pair or even bd140/139 combo would be a great improvement. Anybody want to guess whose amps I started out building? The first discrete amp I ever built that worked well, a tribute to the design, was Rod elliot’s P3a which this amp is more or less.

This amplifiers operation is Fairly simple and easy to explain. The three stages: differential, transconductance, and current amplifier are more or less dc coupled (explained later) which I prefer, to alleviate capacitor non-linearity’s.

Differential stage


The differential stage seems to be a gift of nature in that it's the building block of so many electronic circuits and yet so simple. If you understand the basic operation of transistors then you know when current is applied to the base of bipolar junction transistors that is properly biased you get current flow equaling the transistors dc current gain times the input signal from collector to emitter or emitter to collector whatever the configuration.

With that said if you look at the schematic you see that you have a bipolar supply of 80+/- and a differential stage which is comprised of r6,r19,t1,t2 that spans the 80v supply and is forward biased to conduct about 3mA's current through r6 and then splits somewhat evenly through transistors t1,t2. T2 completes through to -v supply, while t1's collector and Potentiometer r19 make a voltage divider, which allows the difference between the 2 transistors to be fed into the next stage.

If the transistor characteristics of the differential stage (t1,t2) are closely matched then the following will happen. If you tie the emitters of two PNP transistors together, run there collectors out through 2 equal resistors attached to a -V supply and forward bias them you have a simple differential stage and any changes on the transistor bases will be amplified by the transistors hfe on the collectors. When you apply a current to the base of t1 (or t2) then an equal in magnitude yet out of phase voltage is created on the opposite transistors collector.

Voltage amp.


The voltage amp or transconductance stage is a simple bootstrapped class A amplifier. The 68uf bipolar cap takes the place of a constant current source and simplifies the design and this is why I can't really call this amp a true dc coupled amp. Here's how it works, when t5 goes negative it turns the lower half of the AB class push-pull (or B class how ever you like to think of it) output stage on with respect to the input signal. Now look at the capacitor wedged between r9 and r10 resistors, r9 is connected to the positive rail and so when the output goes negative a potential difference is created between the output and the positive rail so the capacitor charges. Now when t5 goes positive it turns on the upper half of the output stage and because there is no constant current source to feed the output stage the amp would sag with respect to the lower half of the output stage. So we throw in the bootstrap capacitor, which is now charged from the previous output from the lower half and as the output swings upward the potential difference between the output and the capacitor now discharges through r10 into the base of t7 and assists in supplying pre-driver voltage.

Current amplification.


The output stage or current amplification stage is a complimentary push-pull Sziklai pair Comprised of t3,t4,t6,t7 and a vbe multiplier r20,q2. T6 and t7 are the output transistor drivers their function is to provide the voltage stage with a high impedance load as well as providing enough current to properly power the output transistors. With Fairchild's FJL4215/4315 power BJT (bipolar junction transistor) pair (my new favorites) you'll have roughly 320 ohms at the base when powering a 4 ohm load at a peak current of 5 amps or roughly 3.5 amps RMS and the class A voltage stage would need to dissipate almost 1.5 watts to provide this power (you can figure out a transistor base impedance for a certain load by looking at it’s Hfe or DC current gain graph in the data sheet FJL4315 Data sheet I’ll explain later). This would require a substantial voltage stage so we just add the output drivers and we can then recalculate our new load on the VAS (voltage amp stage). Base drive current on the output transistors pushing 5 amps into 4 ohms (load) will be the load current divided by the transistors Hfe (dc current gain) or 80 in this case. So 5/80 or .062A or 62Ma and the resistance at the base is, drive voltage divided by the load current 20/.062 or 320 ohms roughly. Now the tip31/32 drivers graph says there hfe at 62Ma will be 95 so, .062/95=.00065 and resistance at the base 21v/.00065a=30k ohms. Now the Vas stage is only needed to push .013 watts or 13mw (21v*.00065a=.013w) There are other things to consider when calculating these figures such as heat, device leakage, and the gain curve graph given in data sheets is only 100% true if you’re your voltage out follows the graph, but for what were doing here these simple calculations will be enough. Another important note, this is only for the fjl4315, the fjl4215’s dc gain curve is different.

The vbe multiplier is used to establish a resting forward current between the two halves of the output stage. This is important because if not for some resting constant current flowing between the two output transistors there would be a sag or a dip in the output every time the signal came near or crossed the 0v output threshold. Silicon transistors require around .7 volts on there base’s in reference to there emitters to be forward biased and they also loose this .7 volts so if you have an audio signal going into the base of a transistor you would loose from 0 to.7 volts. So what you do is provide that .7 volts ahead of the output stage to compensate for the voltage drop, it’s called vbe multiplier.

Short circuit protection.


The short circuit protection is accomplished by monitoring the voltage drop across r7 and r11. When the voltage drop reaches a certain level there is sufficient voltage to forward bias either q2 or t8 depending on which phase the amplifier is in and this causes a redirection of the pre-driver base current through q2 or t8 and thus removing input from the output stage and stopping a disaster. This particular amplifier is set to redirect signals at about 9 Amps. At 2 ohms with max signal input you get about 7.5 amps out, and at .5-ohm load you get 9.2A, and at .001 ohm (more or less a short) you'll get 9.5 amps. These last test were done with DC not AC, the AC figures are close but do differ and I will provide detailed info on AC characterizes of this amp as soon as I can.

Here is a Pspice model showing simple voltages and currents around the amp to aid you while constructing you own amp. This is not the amp but is a good representation of what you should see if all goes well.

Pspice Model.










HOME