Hm, i would like so suggest that you disconnect the outputtransistors MJE2955/3055 so you get rid of the possible CFB oscillation (even if this oscillation is unharmful 
) so youre only using BD139/140 as output transistors. This also includes removing the mysteroius 220pF cap to the left of BD140.
Now check what happens on the output. If everything works perfectly you might get a clue where the problem is
The diodes is there because in rest the amplifier shall have a small current running through the output stage so you dont get any crossover distortion, and if you have two diodes (1.4V) this will open the threshold for the transistors so you might get around 10mV over the Emitter resistors who allows a certain amount of current to run.
However, when the output devices runs hot their Vbe gets lower and by the same amount of diode voltage at their bases the voltage drop over the emitter resistors might have got very big. If this can continue the current will be so large you will blow the output.
But if the diodes are reflecting the exact same temperature as the outputs (in the CFP case, the drivers), their voltage drop vill also drop when they get hot so it will compensate the Vbe sink in the transistors.
Using a transistor (also called Vbe multiplier) to do the job the diodes are doing are MUCH better because you can trim in exactly the voltage drop and the best resting current, and if youre using for example a To126 transistor it's easy to mount it together with the outputs! (still in the CFP case on the 139/140 drivers)
I know Rod Elliot has written the drivers don't need any heatsink but i would like to say they need a heatsink. But in CFP mode, im not sure it's so good to have them on the same heatsink as the outputs. However, the drivers (139/140) would i highly recommend to put be on the same heatsink as the diodes or a Vbe multiplier (this can be a small heatsink) and put the 2955/3055 on the bigger.
An simple Vbe multiplier can be made like this. I marked it with red dots so you know where to put it. The capacitor is there so the Ac signal don't have to be pushed through the transistor. This is not 100% neccessary but many designers recommend it.
Check with a voltage-meter over the emitter resistors on the output
while turning the potentiometer (without input signal of course). Around 50mV over the resistors will make around 50mA with 1R resistors.
When you power it up, make sure the potentiometer is at maximum resistance, this is because then the current will be the lowest.
sorry for my bad english. If anyone want to correct something do so!
Good luck
) so youre only using BD139/140 as output transistors. This also includes removing the mysteroius 220pF cap to the left of BD140.Now check what happens on the output. If everything works perfectly you might get a clue where the problem is
The diodes is there because in rest the amplifier shall have a small current running through the output stage so you dont get any crossover distortion, and if you have two diodes (1.4V) this will open the threshold for the transistors so you might get around 10mV over the Emitter resistors who allows a certain amount of current to run.
However, when the output devices runs hot their Vbe gets lower and by the same amount of diode voltage at their bases the voltage drop over the emitter resistors might have got very big. If this can continue the current will be so large you will blow the output.
But if the diodes are reflecting the exact same temperature as the outputs (in the CFP case, the drivers), their voltage drop vill also drop when they get hot so it will compensate the Vbe sink in the transistors.
Using a transistor (also called Vbe multiplier) to do the job the diodes are doing are MUCH better because you can trim in exactly the voltage drop and the best resting current, and if youre using for example a To126 transistor it's easy to mount it together with the outputs! (still in the CFP case on the 139/140 drivers)
I know Rod Elliot has written the drivers don't need any heatsink but i would like to say they need a heatsink. But in CFP mode, im not sure it's so good to have them on the same heatsink as the outputs. However, the drivers (139/140) would i highly recommend to put be on the same heatsink as the diodes or a Vbe multiplier (this can be a small heatsink) and put the 2955/3055 on the bigger.
An simple Vbe multiplier can be made like this. I marked it with red dots so you know where to put it. The capacitor is there so the Ac signal don't have to be pushed through the transistor. This is not 100% neccessary but many designers recommend it.
An externally hosted image should be here but it was not working when we last tested it.
Check with a voltage-meter over the emitter resistors on the output
while turning the potentiometer (without input signal of course). Around 50mV over the resistors will make around 50mA with 1R resistors.
When you power it up, make sure the potentiometer is at maximum resistance, this is because then the current will be the lowest.
sorry for my bad english. If anyone want to correct something do so!
Good luck
I am very thankful to all of you for your valuable suggestions.Finallyl I got the output without significant distortion.Here, I am using parallel network of 5.6 mico henry and 0.5 ohm with 5 watt resitor in between the out put and load.Plz suggest me the Inductor specs (Wattage, core sige........etc).I am using a quater resistor type Inductor,which is available in the market.
with best regards
Madhu.K
with best regards
Madhu.K
I think the 5Watt 0.5 ohm resistor is going to run hot since it's maximum current will be around 3 Amps [sqrt(5/0.5)]
The inductor value isn't so critical. But it need to be thick enough to carry the current. Douglas Self uses a 2.3uH, 0.01ohm inductor in one of his designs with good results. I'm gessing 10 turn around a 25mm axle will do fine
/Fritzell
The inductor value isn't so critical. But it need to be thick enough to carry the current. Douglas Self uses a 2.3uH, 0.01ohm inductor in one of his designs with good results. I'm gessing 10 turn around a 25mm axle will do fine
/Fritzell
Enquiry
Hi Everybody ,
firstoffall I want say thanks for your valuable suggestions to my previous project.Now I am going to anlyze that Amplifier using Network analyzer.Ok.Rightnow for another purpose I wish to buit a 30-35 Watts IC based Audio power amplifier.I am requesting your suggestions.
with regards
Madhu.K
Hi Everybody ,
firstoffall I want say thanks for your valuable suggestions to my previous project.Now I am going to anlyze that Amplifier using Network analyzer.Ok.Rightnow for another purpose I wish to buit a 30-35 Watts IC based Audio power amplifier.I am requesting your suggestions.
with regards
Madhu.K
I have built an Amp with the chip TDA2020 (20W version). It was very simple and worked very good.
Here is a nice guide
http://www.electro-dan.co.uk/electronics/TDA2040.html
Here is a nice guide
http://www.electro-dan.co.uk/electronics/TDA2040.html
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.