Has anyone actually build the suggested zener mod and can share his listening experience ? A lot of discussions, but I am curious if anyone tried it ?
When going with my El34-PP-UL amp to triode (g3 to cathode, g2 to anode, no resistor), The amp feels soft and not very transparent/holografic anymore. Why is this ? Cant believe this is just the loss inpower....
When going with my El34-PP-UL amp to triode (g3 to cathode, g2 to anode, no resistor), The amp feels soft and not very transparent/holografic anymore. Why is this ? Cant believe this is just the loss inpower....
How about posting a schematic and parts values of your amp that does not sound as good in triode mode, as it sounds in Ultra Linear mode.
Ultra Linear mode has about 2x the gain of Triode mode.
That means that Triode mode requires the driver has to put out about 2x the voltage swing to get to the same power out as Ultra Linear. Is the driver stage straining more in that case?
Ultra Linear mode also has about 2x the output impedance of Triode mode.
So Ultra Linear mode has more variation in the resultant voltage frequency response versus Triode mode (because of loudspeakers that do not have a constant impedance versus frequency).
How does a loudspeaker sound in constant voltage drive, constant current drive, or constant power drive?
If there is no negative feedback around the output stage:
Triode mode tends to be more constant voltage
Ultra Linear mode tends to be more constant power
Pentode mode tends to be more constant current
Ultra Linear mode has about 2x the gain of Triode mode.
That means that Triode mode requires the driver has to put out about 2x the voltage swing to get to the same power out as Ultra Linear. Is the driver stage straining more in that case?
Ultra Linear mode also has about 2x the output impedance of Triode mode.
So Ultra Linear mode has more variation in the resultant voltage frequency response versus Triode mode (because of loudspeakers that do not have a constant impedance versus frequency).
How does a loudspeaker sound in constant voltage drive, constant current drive, or constant power drive?
If there is no negative feedback around the output stage:
Triode mode tends to be more constant voltage
Ultra Linear mode tends to be more constant power
Pentode mode tends to be more constant current
Blitz,
I like your Versterker. If B+ and Bias are adjusted accordingly, then KT88, EL34, 6L6, or one of my current favorites, KT66.
When comparing Triode mode to Ultra Linear mode of the same circuit, you had to turn up the signal voltage to the drivers (to make up for ~ 6dB less gain in Triode than in Ultra Linear mode).
The question is, did the drivers have a little more distortion with more signal voltage, to get the amp output to a reasonable listening volume.
That could be part of the sound difference between triode and UL modes.
I plan on experimenting with KT66 Ultra Linear again soon, and will have to compare to Triode mode then.
I like your Versterker. If B+ and Bias are adjusted accordingly, then KT88, EL34, 6L6, or one of my current favorites, KT66.
When comparing Triode mode to Ultra Linear mode of the same circuit, you had to turn up the signal voltage to the drivers (to make up for ~ 6dB less gain in Triode than in Ultra Linear mode).
The question is, did the drivers have a little more distortion with more signal voltage, to get the amp output to a reasonable listening volume.
That could be part of the sound difference between triode and UL modes.
I plan on experimenting with KT66 Ultra Linear again soon, and will have to compare to Triode mode then.
With the original 2005 comments, the outstanding query still is - why choose an approx. 30V drop??
For an 807 that was rerated for 400V triode class A use (due to Williamson operation at that idle level), then 30V equates to only a 10% increase in B+ when keeping the screen-cathode at max rated 400V.
The comments/circuit in the triode-trick link put a diode in series with the screen connection - which could be related to Dennis Grimwade's incentive for doing that - although I don't think I've come across evidence to support the presumption that screen current can go 'negative' (a requirement to reverse bias such a screen diode).
For an 807 that was rerated for 400V triode class A use (due to Williamson operation at that idle level), then 30V equates to only a 10% increase in B+ when keeping the screen-cathode at max rated 400V.
The comments/circuit in the triode-trick link put a diode in series with the screen connection - which could be related to Dennis Grimwade's incentive for doing that - although I don't think I've come across evidence to support the presumption that screen current can go 'negative' (a requirement to reverse bias such a screen diode).
Pentodes that have even the slightest kink in plate current have negative resistance (at those points along the kink).
The Plate Current goes Down as the Plate Voltage goes Up (Negative Resistance).
Look at some Pentode family of curves graphs.
That is not negative current.
I am unaware of any screen current that is ever negative.
And even if it ever could or does go negative, why would you want a disconnect diode connected to the screen?
It seems that whenever disconnects occur, there might be transients (connect, disconnect, connect . . . ). Who wants transients from the amp that did not come from the music?
Triode wired mode and Ultra Linear mode can have less screen power dissipation under large signal swings than Pentode mode. It depends on the circuit and voltages.
It is Impossible in triode wired mode for the plate voltage to be less than the screen voltage.
It is possible in classical ultra linear mode for the plate voltage to be less than the screen voltage (and far less during signal swing).
It starts out that way because there is more voltage drop from B+ in the transformer primary DCR to the plate, than there is voltage drop from B+ in the transformer primary DCR to the screen.
But it does not stop there.
If during the signal swing, the plate voltage goes down by 100V, the screen will only go down by the percentage of the UL tap. i.e. for a 43% tap, the screen will only go down 43 Volts.
You can still get a kink in UL mode (negative plate resistance).
Just take a look at some graphs of family of curves for Ultra Linear.
The Plate Current goes Down as the Plate Voltage goes Up (Negative Resistance).
Look at some Pentode family of curves graphs.
That is not negative current.
I am unaware of any screen current that is ever negative.
And even if it ever could or does go negative, why would you want a disconnect diode connected to the screen?
It seems that whenever disconnects occur, there might be transients (connect, disconnect, connect . . . ). Who wants transients from the amp that did not come from the music?
Triode wired mode and Ultra Linear mode can have less screen power dissipation under large signal swings than Pentode mode. It depends on the circuit and voltages.
It is Impossible in triode wired mode for the plate voltage to be less than the screen voltage.
It is possible in classical ultra linear mode for the plate voltage to be less than the screen voltage (and far less during signal swing).
It starts out that way because there is more voltage drop from B+ in the transformer primary DCR to the plate, than there is voltage drop from B+ in the transformer primary DCR to the screen.
But it does not stop there.
If during the signal swing, the plate voltage goes down by 100V, the screen will only go down by the percentage of the UL tap. i.e. for a 43% tap, the screen will only go down 43 Volts.
You can still get a kink in UL mode (negative plate resistance).
Just take a look at some graphs of family of curves for Ultra Linear.
With respect to 'negative' screen current, I think the latter comments in the triode-trick link try to rationalise that a negative current flow from a particular process is bad (eg. discussion on 'negative impedance') and hence putting a series diode in the circuit will stop electrons 'going the wrong way'.
Whereas there is a pre-existing flow of 'positive' current exceeds any process that imposes 'negative' current, and hence the diode is just continuously conducting for all conditions and would never operate with current reducing to zero and wanting to reverse bias that diode. So it is an incorrect appreciation of why a diode would be required in the first place, and so the diode could be removed (except for the aesthetically pleasing musicality that it seems to provide).
Whereas there is a pre-existing flow of 'positive' current exceeds any process that imposes 'negative' current, and hence the diode is just continuously conducting for all conditions and would never operate with current reducing to zero and wanting to reverse bias that diode. So it is an incorrect appreciation of why a diode would be required in the first place, and so the diode could be removed (except for the aesthetically pleasing musicality that it seems to provide).
Last edited:
I tried the Pearl "triode trick" on my amp which uses 6AR6 outputs. Can't say I heard much/any difference. If I was going to go triode mode again I'd just use the stereotypical 100 ohm resistor. That's what I started with before the Pearl trick.
John
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.