OT inductance + bypass capacitance = LC resonance?

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

5v333

Well-known member
Joined
Jun 30, 2013
Messages
631
Location
Gothenburg
in a triode transformer coupled output stage (pushpull in my case) - does the inductance of the transformer and the cathode bypass cap form a resonant component ?

im wndering because the lower frequencies in my tube amp are quite overwhelming and somewhat messy. when i probed one of the cathodes in the pushpull stage and sweep a sine wave , i have a resonance at about 6hz, quite wide aswell. i changed my bypass cap to 1000uF and it mght gotten a tad better but no biggy.

with about 17-20db global feedback, the lowend bandwitdth at the OT sec is about 6hz.

then i tried to short the cathodes so that the resistors went from saparate to shared, and now the resoance is more or less gone. totally different sound in the bass now. its clean and tidy, almost boring :)

 
The bump in low frequency response is almost certainly due to the onset of LF instability caused by the NFB. If you increase the value of the largest capacitor in the loop by tenfold you should see the resonant frequency drop. The cathode bypass capacitor has a small effect on loop gain so the small change you see by a large change in its value is to be expected.

Cheers

Ian
 
the bump is not seen on the output of the transformer though... this is at the cathodes.

the two interstage caps inside the loop is set at 0,42hz and 42hz. this must be okidoki? no?
 

Attachments

  • NYDEQ TUBE aMP2.pdf
    17.7 KB · Views: 7
5v333 said:
in a triode transformer coupled output stage (pushpull in my case) - does the inductance of the transformer and the cathode bypass cap form a resonant component ?
  No. The tube's internal resistance is so high that the resonance is completely damped.

  i have a resonance at about 6hz
This is the result of the NFB loop, trying to force the amp (that has a cut-off frequency of 42 Hz), to go down to 0.42Hz. The whole system meets more or less half-way, because the open-loop gain goes out of steam. You may try to reconcile the NFB loop's response  with that of the OLG.
The cathode caps add another order to the OLG's high-pass, which makes things a tad more complicated.
 
5v333 said:
the bump is not seen on the output of the transformer though... this is at the cathodes.

the two interstage caps inside the loop is set at 0,42hz and 42hz. this must be okidoki? no?

OK, I am confused because you said "I'm wondering because the lower frequencies in my tube amp are quite overwhelming and somewhat messy"

The two loop caps frequencies at 0.42Hz and 42Hz I agree are OK. However, there is a third pole caused by the transformer and there is another due to the cathode bypass.

Let's take it a step at a time. First open the loop and check if the boost is still there.

Cheers

Ian
 
alright thanks!

so what´s happening at the cathodes are just part of the whole closed loop circuit. and nothing to worry about?

yea maybe its time to open the loop again. but i think last time i checked the open loop response with the feedback resistor tied with 1.2k to ground ( measuring from the first grid to the feedback resistor ) i had a phase margin of about 80degrees in the low end.
 
5v333 said:
yea maybe its time to open the loop again. but i think last time i checked the open loop response with the feedback resistor tied with 1.2k to ground ( measuring from the first grid to the feedback resistor ) i had a phase margin of about 80degrees in the low end.
Tying caths to gnd drastically changes the gain of the input stage, so may not be indicative of the actual performance. I would think shunting them with large caps (1000uF) would me more adequate.
 
abbey i think my english didnt suffice.

i meant:

untying the feedback resistor from the cathodes of the first tubes. then add a 1.2k resistor to the feedback resistor and then ty that additional 1.2K to ground. so that the circuit - in open loop - is loaded as it would be in closed loop.
 
5v333 said:
abbey i think my english didnt suffice.

i meant:

untying the feedback resistor from the cathodes of the first tubes. then add a 1.2k resistor to the feedback resistor and then ty that additional 1.2K to ground. so that the circuit - in open loop - is loaded as it would be in closed loop.
OK.
 
i tried a couple of different ground layouts today and think i might solved the overwhelmingness and messyness. :)

ground layout is serious buissness. it has given me alot of headaches.

thanks again for helping me out with this!  :-*
 
The power supply decoupling of the first stage and possibly the next stage can come in to influence, and was once used to add some phase margin in the frequency range of a resonance.

Imbalance in the output stage could also lower the inductance seen by the output stage plate resistance, moving that RL pole higher.

Poor grounding could be allowing a form of motorboating with frequency content at the resonance, which then amplifies the disturbance due to the LF peaking.
 
actually i havnt paid attention to where decoupling caps have their ground attached. need to experiment with this.


thanks again!
 

Attachments

  • NYDEQ TUBE aMP2.pdf
    18 KB · Views: 1
Back
Top