I have various selection of o/p trannies of different vintages, some quite modern manu. Problem, on all UL o/p stage configurations I have to apply cap-resistor arrangements between anode and 40% g2 taps on each side of the o/p tranny regardless of tube type used.. ..literature suggests only arbitary values selected beween 1-2nF & 470-1500 ohms.
Okay I have one o/p tranny which in situ resonates at roughly 200Khz without suppressor components. The simple cap/ Z reactance formula seems very vague. . The trouble is I find with typical values specified it mucks up the square wave quality.
Anyone know the proper way to calculate these C/R values v.s freq ?
rich
Okay I have one o/p tranny which in situ resonates at roughly 200Khz without suppressor components. The simple cap/ Z reactance formula seems very vague. . The trouble is I find with typical values specified it mucks up the square wave quality.
Anyone know the proper way to calculate these C/R values v.s freq ?
rich
Hello Rich,
The RC network is in fact a “snubber” to dampen the hf ringing of the LC circuit formed by the primary leakage inductance and parasitic transformer capacitance. To calculate the values you need to know the values of the leakage inductance and capacitances of the transformer. Usually the ringing is sufficiently damped by the Ri of the tubes. Proper designed and applied OT’s need no snubbers at all. But when needed calculation is tricky. Best is to use a resistor as large as possible but sufficient low to dampen the ringing. The capacitor should be as small as possible but large enough to be effective. Trial and error is still the best way to go, sorry no ready made recipe for calculation the values here.
You can use a RC snubber across the ends of the transformer (across the anodes of the two PP tubes) but you will need a capacitor that withstands twice the PP voltage across the anodes then. So you will need at least a 1500V to 2KV one then. Same for the resistor.
Cheers
The RC network is in fact a “snubber” to dampen the hf ringing of the LC circuit formed by the primary leakage inductance and parasitic transformer capacitance. To calculate the values you need to know the values of the leakage inductance and capacitances of the transformer. Usually the ringing is sufficiently damped by the Ri of the tubes. Proper designed and applied OT’s need no snubbers at all. But when needed calculation is tricky. Best is to use a resistor as large as possible but sufficient low to dampen the ringing. The capacitor should be as small as possible but large enough to be effective. Trial and error is still the best way to go, sorry no ready made recipe for calculation the values here.
You can use a RC snubber across the ends of the transformer (across the anodes of the two PP tubes) but you will need a capacitor that withstands twice the PP voltage across the anodes then. So you will need at least a 1500V to 2KV one then. Same for the resistor.
Cheers
Pjotr........Yes.......... perhaps I should have known better (as a switchmode designer !!). The word "snubber" slipped from brain.
In general the âim is always to get the best LF/HF performance from any tranny, but it ain't quite so easy as you realise. I found as you suggested, use a largest value resistor as poss but I find a fairly large C as well.
It is a horrendous business dealing with the primary ends as you rightly mentioned the massive push pull voltage swing, especially dangerous when one is forcing through 20Khz at nearly a 1kV! Any scope probe at that voltage will load and disfigure the waveform and creare more ripple/overshoot. In switchmode the dv/dt is just as high and much faster but impedance much lower...... Big difference.
I have this long time habit of poking about with circuits attaching the other end of components with the soldering iron while the B+ is blazing away (500V) and keeping one hand in pocket (anyone copy that ??). Yes, I violate practically every safety ethic. Come on, there must be loads of other experienced benchers out there; Own up!!! ........ I'm still alive but shakin'.
I never use HV ceramic types on dicy HV audio stages, as they have a piezo characteristic and are quite lossy. They also sound musically "hard". The optimum solution I have found on this particular speciment is as you said "trial and error" using the conventional snubber between, A and g2 values 1K5 and SM 2nF ,with a 270R to the 43% tap. . I am content with the square wave resp at 20Khz so I will leave it at that. The overall performance is very good. NAturally any designer seeking dammned good performance will also double check that the circuit is also going to take off with difficult loads.
Perhaps it is little wonder that we don't come across more push pull problems.
My motto to anyone contemplating a push pull amp is stick to a design that have been tested with time. However with modern trannies having very high self resonance figures, (as I have) the 3dB down response happens to be at 80Khz! This fulfills fourier criteria great for true 25Khz+ audio response.
More anon,
rich
In general the âim is always to get the best LF/HF performance from any tranny, but it ain't quite so easy as you realise. I found as you suggested, use a largest value resistor as poss but I find a fairly large C as well.
It is a horrendous business dealing with the primary ends as you rightly mentioned the massive push pull voltage swing, especially dangerous when one is forcing through 20Khz at nearly a 1kV! Any scope probe at that voltage will load and disfigure the waveform and creare more ripple/overshoot. In switchmode the dv/dt is just as high and much faster but impedance much lower...... Big difference.
I have this long time habit of poking about with circuits attaching the other end of components with the soldering iron while the B+ is blazing away (500V) and keeping one hand in pocket (anyone copy that ??). Yes, I violate practically every safety ethic. Come on, there must be loads of other experienced benchers out there; Own up!!! ........ I'm still alive but shakin'.
I never use HV ceramic types on dicy HV audio stages, as they have a piezo characteristic and are quite lossy. They also sound musically "hard". The optimum solution I have found on this particular speciment is as you said "trial and error" using the conventional snubber between, A and g2 values 1K5 and SM 2nF ,with a 270R to the 43% tap. . I am content with the square wave resp at 20Khz so I will leave it at that. The overall performance is very good. NAturally any designer seeking dammned good performance will also double check that the circuit is also going to take off with difficult loads.
Perhaps it is little wonder that we don't come across more push pull problems.
My motto to anyone contemplating a push pull amp is stick to a design that have been tested with time. However with modern trannies having very high self resonance figures, (as I have) the 3dB down response happens to be at 80Khz! This fulfills fourier criteria great for true 25Khz+ audio response.
More anon,
rich- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.