Hi, what is the SOA of these transistors , if I parrallel two of them ( 4 per channel ) how much power can I obtain in 4 ohms - 6 ohms , My friend has two nice speakers(6ohm rated at 25w rms , 5.25 inch driver ) and he wants to use them on his laptop , and he needs an amp ... he doesnt want ic ' based amps . Something around +-15 v to +- 18 v rails. Thanks
Not quite enough IMHO. With 4 ohms from ~14Vpk you're looking at >3 amps peak, and one BD139/140 is just 1.5 amps with a following wind. Better use some "proper" power transistors.
You can work out the power from the data sheet; http://www.st.com/content/ccc/resou...df/jcr:content/translations/en.CD00001225.pdf
It really depends how you use them and in what class.
It really depends how you use them and in what class.
If I apply my usual guide where maximum output power from the 2pair output stage is given by :
Pout = total Pmax / 5 = 12.5*4 / 5 = 10W
10W into 4ohms is equivalent to 6.3Vac and 8.9Vpk and would run on supply rails of ~ +-12Vdc to 15Vdc.
But I think they may struggle to meet the peak current demand of HF transients into that low a load. It may be better to try 8ohms speakers first (on +-18Vdc supplies) and see how they behave.
Pout = total Pmax / 5 = 12.5*4 / 5 = 10W
10W into 4ohms is equivalent to 6.3Vac and 8.9Vpk and would run on supply rails of ~ +-12Vdc to 15Vdc.
But I think they may struggle to meet the peak current demand of HF transients into that low a load. It may be better to try 8ohms speakers first (on +-18Vdc supplies) and see how they behave.
Last edited:
Also would I need some pre drivers for them ? I want to use an op amp and after it the BD's ( used this schematic before and it worked fine , with bigger transistors )
Hi xXBrunoXx,
Why not change to the MJE series parts (like MJE15031 and it's complement) for the outputs and use the BD139/140 as drivers. That would make up a nice amplifier for your needs, and reliable. An op amp could reliably drive those transistors, but use an NE5532 or NE5534 (x2) or equivalent. TL071 /TL072 have very low output drive, so they are out.
Why are you running the bias so high? Anything between 10 and 20 mA would work every bit as well. Your schematic didn't post, it would be nice to see what you are looking at.
-Chris
Why not change to the MJE series parts (like MJE15031 and it's complement) for the outputs and use the BD139/140 as drivers. That would make up a nice amplifier for your needs, and reliable. An op amp could reliably drive those transistors, but use an NE5532 or NE5534 (x2) or equivalent. TL071 /TL072 have very low output drive, so they are out.
Why are you running the bias so high? Anything between 10 and 20 mA would work every bit as well. Your schematic didn't post, it would be nice to see what you are looking at.
-Chris
Hi xXBrunoXx,
I think you might be asking a bit much from the BD139/140 outputs. Peak currents could kill that pair. You could try TIP102 / TIP107. Those are darlington transistors in a TO-220 case. Remember that the op amp have to supply whatever the output current is divided by the transistor beta. The other choice would be to use the BD139/140 as drivers, and something like the MJE15028 and MJE15029 (good for 8 amperes). OF course, you can also go to a more robust output as well. You need to double your bias diode string, but I would rather see a real bias Vbe multiplier in there. This drops the current requirement from the NE5532 to reasonable levels, which has got to sound better.
Keep in mind that many internet schematics do not include capacitors for stability or even reasonable bias networks. Circuits that simulate well often do not perform as expected unless realistic current levels and supply impedances are used.
-Chris
I think you might be asking a bit much from the BD139/140 outputs. Peak currents could kill that pair. You could try TIP102 / TIP107. Those are darlington transistors in a TO-220 case. Remember that the op amp have to supply whatever the output current is divided by the transistor beta. The other choice would be to use the BD139/140 as drivers, and something like the MJE15028 and MJE15029 (good for 8 amperes). OF course, you can also go to a more robust output as well. You need to double your bias diode string, but I would rather see a real bias Vbe multiplier in there. This drops the current requirement from the NE5532 to reasonable levels, which has got to sound better.
Keep in mind that many internet schematics do not include capacitors for stability or even reasonable bias networks. Circuits that simulate well often do not perform as expected unless realistic current levels and supply impedances are used.
-Chris
I have done the same schematic with M5220 ( +-25v op amp 10v slew rate ) and TIP 142 / TIP 147 with a big 20.5 - 0 - 20.5 ( 20.5 after bridge rectifier and supply caps 6800uf ) transformer 8A per rail + 6800uF. And after 11 v on output it distorts. On music i get 8-9 v before clip and it draws around 1- 1.2 A per rail ( mono ) I ve seen 1.6A and it clipped, what would be the problem ?
Also tried NE5532 on the same schematic got 12V out ( in the datasheet at 600ohm load it says 32 V p-p so 15-16v output ... what you think ?
Also tried NE5532 on the same schematic got 12V out ( in the datasheet at 600ohm load it says 32 V p-p so 15-16v output ... what you think ?
Last edited:
Well, your supply voltage represents the peak voltage your output signal can be. But, you have to subtract the drop across the outputs and the sag from the power supply. So a rough ballpark might be 18.5 volts X 0.7071 = 13.08 volts rms. The losses are guesstimates and might be higher than 2 volts. Then, the op amp might have run out of driving current and current limited the output. That's why another stage of current amplification might be needed. A CFP output would preserve more supply voltage since you are low on that.
Always make sure the circuit isn't oscillating with your oscilloscope as well.
-Chris
Always make sure the circuit isn't oscillating with your oscilloscope as well.
-Chris
why do i need another stage , if the tip142 / tip147 are darlingtons , it like bd139 - 140 and 2sc5200 - 2sc1943 ( for example ) why , do I need another pair before the darlingtones ?
If you've got those mitubishi 52xx op amps at +-18 rail or more, they may be thermally limiting themselves. They only put out about 600-700 mw, and if they burn the watts rating using up all that supply voltage they might only put out 6 ma. 6 ma x 500 gain in darlingtons is 3 amps.
If you want more current out of the op amp, use 1.3W zeners to drop the supply voltages to +-6 or so. the darlington inputs will be limited to +-1.8v. Leave the full supply voltage to run the darlingtons. You need series resistors with the zeners to keep the wattage under control. Or use TO92 voltage regulators, +-5 even.
As far this daft idea of using 6 Watt BD139/140 as output transistors, don't they have 2sc5200-2sa1943 in Eastern Europe & Asia at attractive prices? They are ~$5 here, more expensive here than serious wattage MJ4502 MJL21193/4 from On semi etc so I won't touch them. But everybody between the Atlantic and Pacific in your hemisphere seems to use them as output transistors. They must be cheap somewhere over there. I guess real On Semi power transistors are expensive over there.
If you want more current out of the op amp, use 1.3W zeners to drop the supply voltages to +-6 or so. the darlington inputs will be limited to +-1.8v. Leave the full supply voltage to run the darlingtons. You need series resistors with the zeners to keep the wattage under control. Or use TO92 voltage regulators, +-5 even.
As far this daft idea of using 6 Watt BD139/140 as output transistors, don't they have 2sc5200-2sa1943 in Eastern Europe & Asia at attractive prices? They are ~$5 here, more expensive here than serious wattage MJ4502 MJL21193/4 from On semi etc so I won't touch them. But everybody between the Atlantic and Pacific in your hemisphere seems to use them as output transistors. They must be cheap somewhere over there. I guess real On Semi power transistors are expensive over there.
Last edited:
Hi xXBrunoXx,
Remember, the equivalent DC heating power is only 0.7071 of a peak AC waveform. Also, there are losses associated with emitter resistors, emitter-base drop and power supply sag when loaded. Not to mention the voltage losses in whatever op amp you want to use to drive the output stage. All these factors gang up on your plans in order that you become depressed about a lack of power. Accept these truths and design something that will work for you.
-Chris
You don't. Either use a 2 stage normal transistor output stage (way more flexible and faster), or an output stage using darlington transistors. The down side is that you're locked into an EF type output stage. One or the other, but not one stage of output. A darlington counts as two transistors as to how you deal with them.
So do it that way, but no darlington transistors. It's time for you to sit back and think on this a bit. You have options, and one fact. A single stage output won't cut it in this application. and another, a single darlington transistor counts as two connected emitter-base (as in an emitter follower). So you have to make a choice. Darlingtons or two transistors following another pair. Alternatively you could go the CFP route, but I wouldn't recommend that to anyone starting out.
Remember, the equivalent DC heating power is only 0.7071 of a peak AC waveform. Also, there are losses associated with emitter resistors, emitter-base drop and power supply sag when loaded. Not to mention the voltage losses in whatever op amp you want to use to drive the output stage. All these factors gang up on your plans in order that you become depressed about a lack of power. Accept these truths and design something that will work for you.
-Chris
+1 on using the BD's as drivers. On your rail voltage the most cost effective outputs are the TIP 41/42. Set them up as CFP's with the TIP's to maximize what limited voltage swing you have, and the bias should track ok without mounting the bias stack to the output heat sink.
Hi wg_ski,
That's what more experienced DIYer's will do, but not starting out. KISS would seem to apply here.
-Chris 🙂
That's what more experienced DIYer's will do, but not starting out. KISS would seem to apply here.
-Chris 🙂
Don't do it this way! As the output stage usually is of the emitter follower type (even a CFP is it!), output voltage swing is, at it's best, equal to the op amp's output swing. Decreasing the op amp's supply voltages will also decrease it's output swing, thus the whole amp's swing is more or less determined by the op amp's supply. In this case it would be much (!) better to decrease the whole amp's supply voltages.
Or you use a level translator between the op amp and the output stage. I'm not sure if the TS wants it to get that complicated, though.
Best regards!
How about BC546 -556 or any other low power complementary before the TIP's ? so even if the op amp outputs a few mA it is " amplified " by the BC's and after that it goes to the tips's , I have to much gain than if bc's have 100 or so and 500 from the TIP's i get
50.000 gain ?! . only 1 mA at the base of the BC would mean 100mA at the collector ( it exceeds its output ) but it can drive the TIP's ?.
Thanks for your replies guys.
50.000 gain ?! . only 1 mA at the base of the BC would mean 100mA at the collector ( it exceeds its output ) but it can drive the TIP's ?.
Thanks for your replies guys.
- Status
- Not open for further replies.