Hi jneutron
Sorry for asking the dumb question - wasn't wearing my glasses.
Thanks for the good advice though.
Hi megajocke,
Regarding the local oscillation in the output stage; I went back and checked my notes on the tests I did. The oscillation frequency varied only by a small amount with the FET amp; when I increased the value of the cap across the output, at 100nF the frequency of oscillation was around 77.3MHz - bit low for an FM transmitter!
The 1.3nF was made with a capacitance substitution box; this value gave the best peaked response in the output.
This was similar for the Sloane BJT and the Nat Semi IC based amp; but the capacitors value had a slightly more pronounced affect on the frequency of oscillation moving the oscillation frequency by 10's kHz.
What I forgot to mention was that removing the gain control resistor from the -IN terminal to ground of the FET and BJT amplifier stopped the oscillations. This put the amp into a unity buffer configuration. The amp still had a reasonable overshoot when I excited it with the impulse.
Never got a chance to try this with the IC amp.
In all cases I used a fan to keep things cool - this helped to a point. I blew the fuses on the BJT and FET amp several times and also managed to get the IC amp to shut down after what appeared to be overheating of the substrate.
My tutor thinks that the noise floor has risen in the IC amp - I don’t think it liked being made into a power oscillator.
Thanks anyhow for your advice.
Paul.
Sorry for asking the dumb question - wasn't wearing my glasses.
Thanks for the good advice though.
Hi megajocke,
Regarding the local oscillation in the output stage; I went back and checked my notes on the tests I did. The oscillation frequency varied only by a small amount with the FET amp; when I increased the value of the cap across the output, at 100nF the frequency of oscillation was around 77.3MHz - bit low for an FM transmitter!
The 1.3nF was made with a capacitance substitution box; this value gave the best peaked response in the output.
This was similar for the Sloane BJT and the Nat Semi IC based amp; but the capacitors value had a slightly more pronounced affect on the frequency of oscillation moving the oscillation frequency by 10's kHz.
What I forgot to mention was that removing the gain control resistor from the -IN terminal to ground of the FET and BJT amplifier stopped the oscillations. This put the amp into a unity buffer configuration. The amp still had a reasonable overshoot when I excited it with the impulse.
Never got a chance to try this with the IC amp.
In all cases I used a fan to keep things cool - this helped to a point. I blew the fuses on the BJT and FET amp several times and also managed to get the IC amp to shut down after what appeared to be overheating of the substrate.
My tutor thinks that the noise floor has risen in the IC amp - I don’t think it liked being made into a power oscillator.
Thanks anyhow for your advice.
Paul.
Guys I have been doing a brief look into Zobel networks because I have encountered a very strange oen in some speakers that I am upgrading.
In the diagrams I have seen the Zobel network is placed between the crossover network and the speaker driver.
I think one of the key reasons for making the speaker load appear purely resistive is to assist the crossover network to see a stable and fixed resistive load.
Xl = 2pi x FL. The impedance of a speaker driver increases with load which is pretty challenging for the crossover network.
Have I missed the point?
In the diagrams I have seen the Zobel network is placed between the crossover network and the speaker driver.
I think one of the key reasons for making the speaker load appear purely resistive is to assist the crossover network to see a stable and fixed resistive load.
Xl = 2pi x FL. The impedance of a speaker driver increases with load which is pretty challenging for the crossover network.
Have I missed the point?
Exstatic - you are correct, but that is in a different location and serving a different purpose than what these guys are discussing.
On the speaker side of the equation they can be used to normalize impedance both before and after the crossover components to help keep crossover points consistant and present a steady load on the amp - particularly important for tube gear. Some designers use them, some don't. It depends largely on the drivers, and the end goal. Some OEM crossover designs I have seen, by adding a zobel network to a tweeter or midrange circuit, they can add an adjustment pot/pad after the network to adjust the driver volume without changing the crossover point significantly (infinity).
These guys are talking about a 'zobel' type RC network on the output side of an amp (still inside the amplifier) to load the amp at or above a certain frequency to prevent the amp from going into runaway self oscillation and shortly thereafter self destruction.
On the speaker side of the equation they can be used to normalize impedance both before and after the crossover components to help keep crossover points consistant and present a steady load on the amp - particularly important for tube gear. Some designers use them, some don't. It depends largely on the drivers, and the end goal. Some OEM crossover designs I have seen, by adding a zobel network to a tweeter or midrange circuit, they can add an adjustment pot/pad after the network to adjust the driver volume without changing the crossover point significantly (infinity).
These guys are talking about a 'zobel' type RC network on the output side of an amp (still inside the amplifier) to load the amp at or above a certain frequency to prevent the amp from going into runaway self oscillation and shortly thereafter self destruction.
Hi,
If you want to do some serious work on the stability of any system (not only audio amplifier), the first thing that you have to do is to find the transfer function of the system that you want to examine on stability. Finding transfer function means also to represent it graphically with one of the available methods (for example Niquist). Then you may play with various parameters of system design (in the case of amplifier with Capacity, Inductivity, Amplification, etc.) in order to find out how your system behaves corresponding to these changes.
Different amplifier topologies will give you different transfer functions. That means that there are no universal solutions for every amplifier topology - they will vary depending on amplifier design. Therefore I will advise you to restrict yourself in advance to a certain amplifier topology.
This is an academic approach, which will demand very good knowledge of mathematics and electronics. Transfer functions of audio amplifiers might be very complicated and very difficult for calculations. It is therefore advisable to read first some book about determining transfer functions of the audio amplifiers. You will see how complex these calculations can be. But do not be scared ... there are always good approximations that can make your life easier ....
If you want to do some serious work on the stability of any system (not only audio amplifier), the first thing that you have to do is to find the transfer function of the system that you want to examine on stability. Finding transfer function means also to represent it graphically with one of the available methods (for example Niquist). Then you may play with various parameters of system design (in the case of amplifier with Capacity, Inductivity, Amplification, etc.) in order to find out how your system behaves corresponding to these changes.
Different amplifier topologies will give you different transfer functions. That means that there are no universal solutions for every amplifier topology - they will vary depending on amplifier design. Therefore I will advise you to restrict yourself in advance to a certain amplifier topology.
This is an academic approach, which will demand very good knowledge of mathematics and electronics. Transfer functions of audio amplifiers might be very complicated and very difficult for calculations. It is therefore advisable to read first some book about determining transfer functions of the audio amplifiers. You will see how complex these calculations can be. But do not be scared ... there are always good approximations that can make your life easier ....
CBRworm said:
Many thanks CBRworm for the clarification. I hadn't realised Zobel networks were used on the output of an amplifier. After rereading the previous posts that makes sense. Your comment on the PAD also just clarified something that has been puzzling me with a commercial crossover/Zobel network that I am playing around with at the moment.
Much appreciated.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.