Hello
I have an hybrid phono stage which is hybrid. First stage has a 6922 with mosfet follower. There's passive RIAA and another tube with mosfet stage.
The preamplifier is all balanced. First stage seems to work normally but there's DC variation at the four output film caps that goes to the passive RIAA. Even if the second stage has an input cap, the DC variation from the first stage affect the second stage.
1- Normally caps block DC. Why is there DC variation at each output caps of the first stage ? voltage is going from -.2 V to .2V
2- Oscillation problem ? could this cause DC at output caps ? The 4 output caps (Solens) shows the same problem
Thanks
I have an hybrid phono stage which is hybrid. First stage has a 6922 with mosfet follower. There's passive RIAA and another tube with mosfet stage.
The preamplifier is all balanced. First stage seems to work normally but there's DC variation at the four output film caps that goes to the passive RIAA. Even if the second stage has an input cap, the DC variation from the first stage affect the second stage.
1- Normally caps block DC. Why is there DC variation at each output caps of the first stage ? voltage is going from -.2 V to .2V
2- Oscillation problem ? could this cause DC at output caps ? The 4 output caps (Solens) shows the same problem
Thanks
First thing that comes to mind is that it's not DC if it varies.
I built a valve-based phono amplifier with CRCRC filtered, unregulated supply, AC heater supply and AC coupled output about a decade ago. When I tried to measure small output signals on an oscilloscope, the picture kept slowly moving up and down at random. I put the blame on slow variations of the mains voltage.
Oscillations can also cause the DC shift, if there is something at the other end of the capacitor that can behave as a rectifier, such as a semiconductor junction or a valve.
I built a valve-based phono amplifier with CRCRC filtered, unregulated supply, AC heater supply and AC coupled output about a decade ago. When I tried to measure small output signals on an oscilloscope, the picture kept slowly moving up and down at random. I put the blame on slow variations of the mains voltage.
Oscillations can also cause the DC shift, if there is something at the other end of the capacitor that can behave as a rectifier, such as a semiconductor junction or a valve.
Last edited:
Check the frequency of the variation to see if it’s power supply related. If it’s 60, 120, 50, or 100hz it’s a power supply issue.
Some deliberately bias the RIAA caps with some DC voltage.
Otherwise it could be motorboating if it varies up and down.
Check the plate supply to see if it is there also.
Otherwise it could be motorboating if it varies up and down.
Check the plate supply to see if it is there also.
It seems the problem was the TL783C regulator. Probing the V+ showed a line which was slowly bouncing up and down on the oscilloscope screen even if ripple was under 1mv.
Glad you found it. To be clear, yes SOME film caps can leak, or actually short. Foil-film iv overvoltage can punch through and short, where a metalized film will migrate away from a hole. A metalized film cap failure mode is more likely end bond failure, less common in the bigger foil film. With creativity, everything can fail.
All capacitors have DC leakage. Some almost none, others enough to worry about. In older Mylar film capacitors it was 10-100 uA per uF.
https://www.electronicdesign.com/te...whats-all-this-capacitor-leakage-stuff-anyhow
https://www.electronicdesign.com/te...whats-all-this-capacitor-leakage-stuff-anyhow
Depends on cap technology.
But the point to understand is caps have a lot of parameters, most non-linear and capacitance is not always the most significant. For instance Z5u's have the ability to short with a low voltage across them and then fuse when a little higher is applied. Thermal, humidity, vibration,...
But the point to understand is caps have a lot of parameters, most non-linear and capacitance is not always the most significant. For instance Z5u's have the ability to short with a low voltage across them and then fuse when a little higher is applied. Thermal, humidity, vibration,...
No. I built a regulated power supply from a design in Radio Electronics magazine in the early 70’s. It wasn’t stable without a Mylar film capacitor across the output. On the most sensitive range (One milliampere) of the current meter you could read the current through that capacitor.
Even a good 1000uF electrolytic has much less leakage ( nA range ) , so that is a particular case . Tube grid current is in that range , who would build capacitors useless for electronics 
10 uA per uF at 25 V means a leakage time constant of only 2.5 seconds. Amazing, I would only expect that for old paper capacitors that have absorbed moisture.
The oldest metallized polyester capacitor datasheet I have at my disposal specifies worst-case values that are decades higher: Philips 2222 352 series, 5000 seconds (100 V) or 10000 seconds (250 V...630 V) minimum, 30 000 seconds typical at 23 degrees Celsius, 16 or 33 seconds minimum, 500 seconds typical at 100 degrees Celsius, typical data from September 1977, worst-case data from December 1979.
The oldest metallized polyester capacitor datasheet I have at my disposal specifies worst-case values that are decades higher: Philips 2222 352 series, 5000 seconds (100 V) or 10000 seconds (250 V...630 V) minimum, 30 000 seconds typical at 23 degrees Celsius, 16 or 33 seconds minimum, 500 seconds typical at 100 degrees Celsius, typical data from September 1977, worst-case data from December 1979.