BigDad said:
But, how can I bias the BA340 ? I inject an 1 Khz 1Vpp on the BA338 input and I connect an oscilloscope on the BA340 Output.
All I have is a noisy signal with DC offset, i try to trim the BA340 but the DC offset don't change. The only one thing happened is a little signal gain modification.
Moreover, some transistors running very hot after some trimmer turns.
Does anyone have an idea of how to bias this gear ?
starting with the DC
these circuits are AC coupled... meaning there need to be DC blocking caps between stages. Between the BA340 output and the next thing you should have a large cap (1000uF generally) that is terminated by a resistor or a transformer primary, the other end of which is tied to common (signal and power supply common, ground or zero volts ... all the same here). Note that on single-ended outputs (no transformer) I always have a resistor terminating the output side of the blocking cap to 0V so it's not hanging in space if the output is disconnected.
bias
I did not mention in that 1081 article - but I should have - is that monitoring supply current to the circuit really helps... and if you don't have a current meter on your bench supply or a multimeter that measures mA then, no problem, you can find in most BA340 implementations, a 10ohm resistor in series with the supply B+ feeding to the output amplifier module. This resistor is also acting as a fuse so if you ever fry a BA340 style output stage look for that resistor which is generally right next to or under the plug-in amplifier. These cook, sometimes without showing it, so verify that it's the correct value (usually 10 ohms) and then you can hang a volt meter across it and measure current indirectly by looking at the drop across that series R and using ohms law.
if you are dropping 100mV across a 10 ohm resistor then how much current is flowing?
100mV ÷ 10 ohms is 10mA ...
If you twist the bias up and spread the bases of the output pair too far then both output transistors conduct and they will quickly slag if you leave it that way... almost instantaneously if they are not on a heat sink... this is because they are effectively shorting the 24 volts B+ to common through that 10 ohm R just mentioned, the two outputs and their two 1.5 ohm emitter resistors... so all that stuff gets hot. If you have powered this up and had it get really hot then you should check to make sure you still have transistors that transist.
Start with the bias pot turned to the end connected to the collector of the base-spreading transistor TR5... that turns that transistor ON all the way so the voltage drop from collect to emitter is at minimum, the bases of the two output transistors are held close together and the amplifier runs cool and in class B.
I have a current limit that I can set on my bench supplies to 100mA, so first time I power up I have that dialed in just in case I put the bias pot at the wrong end or some other problem exists (belt and suspenders). Assuming I have the pot at the correct starting point the current draw should be quite low.
When I set bias I pass a tone through the amplifier and on to a distortion analyzer ... but really you don't need to do any more than look at the current draw with no signal present to get close. Turn the bias pot off the collector end, slowly, and at some point TR5 will become less conductive and the bases of TR7 and 8 will start to spread apart... to the point where both output transistors start to conduct at the same time... at this point the current draw from the supply begins to climb. A good place to leave the bias trim is right at that point right before current begins to sharply increase*.
With a distortion analyzer you can look at the distortion number and the residual distortion waveform on a scope and fine tune... you'll see that in class B you have these big spikes where both transistors are off as the output swings from one output transistor to the other. Class B is that... one or the other, never both, so there is a dead band as you cross over from one side to the other. As you spread the bases of the outputs apart the dead band gets smaller and the crossover trash distortion reduces. More base spread, beyond what you need to get rid of the crossover distortion, is taking the amplifier in to class A and current draw starts to increase, quickly, and the output devices get warm.
Having a little bit of overlap and having both transistors conducting just a little with no signal takes the amplifier just in to "class AB" ... very very low level signals are mostly operating in class A but with any significant signal you are operating class B with a well-managed crossover from P to N to P and so on...
a tip... when biasing power amplifiers ... for some reason Yamaha P2200's come to mind but whatever... this is what you are doing... you are minimizing the dead band... and if you are running tone and looking at distortion you'll see crossover distortion much better at low levels.
* as to leaving the bias trim right at that point right where current begins to sharply increase... it's the same as "tuning" an EMT plate... you increase tension until adding more tension breaks the clip and you have to start over with a new clip...
have fun
