I have this simple push pull amplifier using a sziklai setup. However, the ouput transistors are heating up fast, so far no heat felt on other components. Simulation tells me that I should be getting about 700mW dissipation on Q1 and Q2 so why the heat?
Actual Measurements:
Bias diodes - Across both is ~1.4V (as expected from simulation)
Output transistor Base - ~25V (as expected from simulation)
DC output - 100mV to 200mV. Its far from simulated value.
Actual Measurements:
Bias diodes - Across both is ~1.4V (as expected from simulation)
Output transistor Base - ~25V (as expected from simulation)
DC output - 100mV to 200mV. Its far from simulated value.
Attachments
At 700 mW you would expect 45-50 deg C case temperature with no heatsink which would feel reasonably hot. What's the bias current (calculated from voltage across R1 or R2)? You should have a Vbe multiplier instead of just 2 diodes in series if you weant to accurately set the bias current.
It's about 29-30mA
I am going to use a Vbe multiplier in the final circuit, but If I cant get it working fine with just two diodes, how much more with a Vbe multipler
Exactly - So, where are the heatsinks? If you want to "breadboard" power amplifiers, you still have to stabilise their bias current dissipation or perhaps there's even some low level oscillation which causes the excess heat. It's difficult to predict what might occur, so you would need a real oscilloscope to see what was really going on. Multisim models are probably not accurate enough to show instability and other free sim. packages are probably also weak in that area.
Last edited:
A power transistor without a heatsink is going to heat up fast, that's why we use heatsinks with them. 700mW will get any transistor hot in free air. Most packages have thermal resistances whe unmounted in the order of 30 to 100 C/W, mounting a power device on a large heatsink brings this down to maybe a couple of C/W.
Thermal resistance to free air is normally not specified for a power device as you would never use them unmounted...
Thermal resistance to free air is normally not specified for a power device as you would never use them unmounted...
So just heatsink it is, thanks to you all.
I hope this place is not like stack.
Is the use of CFP a taboo or is creating a thread about it a taboo?
I hope this place is not like stack.
1. Any CFP must be loaded with a Zobel network for stability/oscillation.
2. The base resistors on OPs should be 100 Ohms or less, especially with older ~1MHz parts.
3. The input to the drivers should be low-Z at high frequencies. Typically, that comes from the VAS compensation cap.
4. The breadboard and the lengthy wiring is an invitation to all manner of unpredictable coupling.
2. The base resistors on OPs should be 100 Ohms or less, especially with older ~1MHz parts.
3. The input to the drivers should be low-Z at high frequencies. Typically, that comes from the VAS compensation cap.
4. The breadboard and the lengthy wiring is an invitation to all manner of unpredictable coupling.
For max. homogenity:
Q6 - TIP3055
Q5 - TIP2955
But TO-3P-casings sound very relaxed, but also without high color resolution, and without contour and colors to get together, a bit flat, noisy, hung up.
Most TO-247 will sound better. Even better TO-220. They should also suffice for most power desires.
Q6 - TIP3055
Q5 - TIP2955
But TO-3P-casings sound very relaxed, but also without high color resolution, and without contour and colors to get together, a bit flat, noisy, hung up.
Most TO-247 will sound better. Even better TO-220. They should also suffice for most power desires.
But TO-3P-casings sound very relaxed, but also without high color resolution, and without contour and colors to get together, a bit flat, noisy, hung up.
Most TO-247 will sound better. Even better TO-220. They should also suffice for most power desires.
High end snake oil b u l l s h i t theory.
Most TO-247 will sound better. Even better TO-220. They should also suffice for most power desires.
High end snake oil b u l l s h i t theory.
Your circuit doesn't show any load on the output. That is, a resistor to ground or a Zobel. The circuit might be oscillating.
You should also have PSU decoupling caps on the breadboard going to ground. Make a star ground point on your breadboard. And the input should have some resistance to ground.
In a simulator having things floating will not be a problem but reality is different
Yes, your only problem at the moment is the lack of a heat sink.
But when amplifiers are built around them they often have problems that are not easy to solve, at least not without significant compromise. Resulting in a lot of off-the-reservation discussion, pseudo-science, and snake-oil sales. The kind of thing that gets moderators upset and threads closed. It’s already started here.
There is nothing inherently wrong with a CFP, and I’ll use them where appropriate - WHEN the compromises you end up making to ensure stability can be accepted. Not every amplifier needs 200+ watts, 10 ppm distortion, 100V/us slew rate, or needs to operate into a pure capacitor. If those are your goals, CFP will be more trouble than it needs to be.
aye, I indeed have the output floating, as for the oscillation, Indeed I am poor I cant afford a scope, so what I did is measure the output with the AC voltmeter, there's a fluctuating reading, even in DC.
I got them to the value such that the quiescent current on the output is low. This is a mistake, this is where a Vbe multiplier must be used.
I apologize, I can't understand this, I am guessing this is the miller cap.
Right, I made a mini antenna
I got 100uF on both rails to ground, my PSU also has decoupling caps before that, just not shown in the pic. As for the input, I indeed had it floating, I'll just ground it through resistor instead of keeping it floating if no input is intended.
Thanks for your inputs, I got something from them.