Hello,
please help me with this design, there is a stability issue.
The DC output point (without input signal) should be 0 V with a small offset, but it's not the case, becase this point is not stable!
Initially the amplifier had an offset of about 200 mV but not stable so it changed from 100 mV to 300 mV pretty fast (so fast that you can see in a multimiter 2-3 changes at second)
I tried to add an input signal and the amplifier itself worked so good but the offset was always not stable.
After a while this offset started to drift and increased to 400mV, 500 mV, 600 mV and more, sometime it becomes negative, always with fast changes.
Every time I turned off the amplifier and then on this offset increased, even if the amplifier was enough cold. Actually there is 2 V of offset !!!
The trimmer on R4 should set the offset but it has no effect on the circuit, the offset goes for its way.
There is a stability DC issue but I don't find it.
Maybe something is wrong with the schematic?
please help me with this design, there is a stability issue.
The DC output point (without input signal) should be 0 V with a small offset, but it's not the case, becase this point is not stable!
Initially the amplifier had an offset of about 200 mV but not stable so it changed from 100 mV to 300 mV pretty fast (so fast that you can see in a multimiter 2-3 changes at second)
I tried to add an input signal and the amplifier itself worked so good but the offset was always not stable.
After a while this offset started to drift and increased to 400mV, 500 mV, 600 mV and more, sometime it becomes negative, always with fast changes.
Every time I turned off the amplifier and then on this offset increased, even if the amplifier was enough cold. Actually there is 2 V of offset !!!
The trimmer on R4 should set the offset but it has no effect on the circuit, the offset goes for its way.
There is a stability DC issue but I don't find it.
Maybe something is wrong with the schematic?
R4 sets the current gain of Q11.
Where is your DC NFB loop to the first stage ... there is none shown! No DC NFB will make it drift.
Where is your DC NFB loop to the first stage ... there is none shown! No DC NFB will make it drift.
Sudden changes in DC conditions in my experience tend to indicate parasitic oscillations. I suggest use of base stopper resistors on Q4,Q5 to begin with.
Please, let me know what do you mean with DC feedback at first stage. What should I add to schematic?
I actually made a different design, with VFA topology and never used base resistors on Q4 and Q5 with same transistors. I'll give a try but I think the problem is somewhere else.
If base stoppers on Q4,Q5 don't help then next try on Q2,3. The first places to try would be where the base sees a low impedance at RF - your C10 kinda ensures that on the input stage.
Actually a strange thing happened. I used the hot air to resoldered a transistor where current sources are located.
Well, I just power on the amplifier still hot and now the DC point is very stable at 1 mV and everithing works like a charme since 1 hour.
This thing happened yesterday but after a day the amplifer started again to be unstable.
Looks like hot air fix the issue for a time, but why?
Now the board is at room temperature so the effect of hot air should be gone, the amp is still stable and working.
But I'm sure I'll try tomorrow and again instability, because just happened.
Well, I just power on the amplifier still hot and now the DC point is very stable at 1 mV and everithing works like a charme since 1 hour.
This thing happened yesterday but after a day the amplifer started again to be unstable.
Looks like hot air fix the issue for a time, but why?
Now the board is at room temperature so the effect of hot air should be gone, the amp is still stable and working.
But I'm sure I'll try tomorrow and again instability, because just happened.
Check if there is any failed soldering particularly broken pads. Heating them cause a dilation of them recovering temporarily the well conducting path
Why have you put a 1nF cap from the output to ground?
That alone could make it unstable. You need a proper Zobel network and perhaps a series output inductor following the network.
That alone could make it unstable. You need a proper Zobel network and perhaps a series output inductor following the network.
My advice: 1k base stoppers for Q13 and Q15. And as Mooly said, remove the 1 nF cap at the output.
Classic bad connection symptoms. Could be a bad bond pad in a transistor, or a PCB/solder problem. You might be able to localize it with heat from a soldering iron and a bit of cold spray to turn the problem in and off. Also some mechanical pushing and pulling on things can be helpful. Always be suspicious of that trimmer wiper as they can produce similar symptoms.
Let us know what you find.
Let us know what you find.
The 1n cap at the output is definitively bad althoug I saw IC's with up to 0.22uF there, directly without series resistor. But it seems to be to be a thermal issue, not electrical instability.
Please ignore that cap on the output, it's not in the real PCB, just a way to simulate capacitive load behaviour.
Actually after the hot air the amplifier is working, it's very stable.
I don't see any solder problem in the PCB, neither mechanical one.
The trimmer was removed but It was ok.
I'll try tomorrow to see what happens and I'll tell you.
p.s. base resistors are a good suggestion for next design XD
Actually after the hot air the amplifier is working, it's very stable.
I don't see any solder problem in the PCB, neither mechanical one.
The trimmer was removed but It was ok.
I'll try tomorrow to see what happens and I'll tell you.
p.s. base resistors are a good suggestion for next design XD
Hello
today it started again to be unstable in its DC point.
Now I'm checking every solder joint.
In you opinion could be only a PCB issue or something related to temperature matching of some components?
today it started again to be unstable in its DC point.
Now I'm checking every solder joint.
In you opinion could be only a PCB issue or something related to temperature matching of some components?
It may be a resistor or a semiconductor with internal defective connections. When you heat the rig, the dilatation may cause a momentaneously solution to the trouble. There isn't an apparent definitively faulted piece. You will need a large patience with those kind of faults.
I need to heat the board to desoldering temperature of 260 °C to get it working every day. 50 °C or maybe 100 °C are not enough.
Than if I let cool down the board for a while, at ambient temparature (15 °C), the amplifier works anyway. But from a day to another no.
So if there was a mismatch in some currents, or unmatched components, or thermal drift for a bad design it should not work always, right?
Because I have to exclude at all a design mistake both in the schematic and in the PCB. That's my priority.
I'll have a large patience but I won't mount a new board if I'm not sure there is a semiconductor or a defective resistor.
Does the node between R15 and R16 (feedback injection point) show the same offset variations, 'following' the output offset?
If so, this drifting originates from the first stage. If not, check for this drift at the bias network.
If so, this drifting originates from the first stage. If not, check for this drift at the bias network.
One thing that would help greatly with remote debugging would be to have as many node voltages as you can get for us while not working. Especially those in the front end circuits, the 4 input transistors and the current source and sink that biases them. Annotating the schematic with those voltages would be best format.