Just completed a RH807 build using 6L6GC tubes. Going through the testing phase I notice that at times the g2 is having a reverse current.
I am testing because apparently modern tubes are not of the same level of quality as those of days past - the spacing of the g2 wires is critical in power beam tubes and the design specs for g2 varies from manufacturer to manufacturer. This can lead to reverse g2 current and since KenRad speaks of not using a voltage divider for the 807 but using a proper stable g2 voltage. As a result we may be in deep water with the RH807 implementation.
As far as I gather the RH807 is highly optimized but the variations possible in the g2 current will cause imnsho a wide variation of results when using tubes that exhibit a varying amount of g2 current (even negative).
( There seems also to be a wide variation of the plate disspiation for the 6L6GC , some are as low as 19 watts, others go as high as 30 watts. )
I want to keep this amp running for a very long sime so I want to ensure that the design is "amended" so no nasties can happen to the unwary when replacement of the finals is needed.
Those of you have more knowledge than myself: Would you a) stabilise the g2 voltages using a shunt regulator and give g2 a fixed volatge (seems to be some amount of interaction when current through g2 changes or b) switch to UL?
What will the pro and cons be of each, in particular what will switching to UL do to the drive requirements (is is going to require more drive) since UL may be the simplest solution.
(the PSU has enough spare capacity for a shunt regulator so no worries there but if I can I like to stay away from solid state)
On another note: Would it be possible to drop the B+ voltage and decrease the impedance of the opt without disturbing too much of the performance? I've got the James 9123HS OPT presently sitting at 5K but would like to go to the 2.5K at 275 ~ 300V.
Thanks
AM
I am testing because apparently modern tubes are not of the same level of quality as those of days past - the spacing of the g2 wires is critical in power beam tubes and the design specs for g2 varies from manufacturer to manufacturer. This can lead to reverse g2 current and since KenRad speaks of not using a voltage divider for the 807 but using a proper stable g2 voltage. As a result we may be in deep water with the RH807 implementation.
As far as I gather the RH807 is highly optimized but the variations possible in the g2 current will cause imnsho a wide variation of results when using tubes that exhibit a varying amount of g2 current (even negative).
( There seems also to be a wide variation of the plate disspiation for the 6L6GC , some are as low as 19 watts, others go as high as 30 watts. )
I want to keep this amp running for a very long sime so I want to ensure that the design is "amended" so no nasties can happen to the unwary when replacement of the finals is needed.
Those of you have more knowledge than myself: Would you a) stabilise the g2 voltages using a shunt regulator and give g2 a fixed volatge (seems to be some amount of interaction when current through g2 changes or b) switch to UL?
What will the pro and cons be of each, in particular what will switching to UL do to the drive requirements (is is going to require more drive) since UL may be the simplest solution.
(the PSU has enough spare capacity for a shunt regulator so no worries there but if I can I like to stay away from solid state)
On another note: Would it be possible to drop the B+ voltage and decrease the impedance of the opt without disturbing too much of the performance? I've got the James 9123HS OPT presently sitting at 5K but would like to go to the 2.5K at 275 ~ 300V.
Thanks
AM
Last edited:
UL is certainly a lot easier to implement than a regulated screen supply. And that original design (pentode) doesn't exactly have a lot of decoupling between the screen and B+! I would at least add a cap to ground from the end of the 10K resistor.
You could also easily experiment with triode mode. Triode is a lot less sensitive to OPT impedance so you try changing the transformer impedance to 2.5K.
Whatever sounds best really.
You could also easily experiment with triode mode. Triode is a lot less sensitive to OPT impedance so you try changing the transformer impedance to 2.5K.
Whatever sounds best really.
Decoupling is done in the 8 W PA system in Radiocraft Libary #33 publication (5K with 8uF). But that does not take into account that the current may go negative so am loath to persue that matter.
Triode mode gives me too little output,
Thanks for the suggestion and keep them coming,
AM
That is one solution I have been navel gazing about but then started thinking of going one step further and turning it into a proper shunt regulator.
However I wonder if the missing bypass cap was deliberatedly left out to achieve something by the fluctuating g2 screen current, I don't know.
edit: one of the tube manuals mentioned that the 6L6 was designed for having an optimum working of g2 with a voltage between 250V and 300V.
AM
Last edited:
Perhaps more try it (we do not want to suck on something that has 250 - 300V on it do we? Can be very lethal) and then measure it and listen.
(PS I know of a true happening from a guy fresh out of school who asked his boss how to check if the mains sockets were connected up properly. Boss replied to stick two nails in and if it gave a tingle it was OK. The bloke did just that and died and boss went up for manslaughter. Moral of the story: never assume that the other knows what they are doing when it concerns electricity)
AM
(PS I know of a true happening from a guy fresh out of school who asked his boss how to check if the mains sockets were connected up properly. Boss replied to stick two nails in and if it gave a tingle it was OK. The bloke did just that and died and boss went up for manslaughter. Moral of the story: never assume that the other knows what they are doing when it concerns electricity)
AM
With every tube i use as final tube pentode such as : 6L6 , 6P3S , Gu50 , I use resistor and zenner at 250V strapped by MKP from +B always .
I have changed a lot of tubes and no modification occur .
Also I do not like UL sound !
When implement UL you must stay at maximum 280V ... 310V for +B with this type of pentode as the Ug2 is maximum 250...280V .
Now at the last 2 amplifier I am using smps power source and I have a distinct +250V output especially for G2 of final tube beside the +B for anodic power .
I have changed a lot of tubes and no modification occur .
Also I do not like UL sound !
When implement UL you must stay at maximum 280V ... 310V for +B with this type of pentode as the Ug2 is maximum 250...280V .
Now at the last 2 amplifier I am using smps power source and I have a distinct +250V output especially for G2 of final tube beside the +B for anodic power .
Last edited:
Tungsol 6L6GC-STR is 500V max on plate and 450V max on g2..... 30W anode dissipation and max 5 Watts screen dissipation.
From what I understand is that ultra linear can sound more clinical than the pentode mode.
AM
Again : is a personal opinion about it !
I have build , finish (and listen to ) in the last 20 years over 15 tube amplifier and my likes are :
- I like only pentode tube in SE ( and triode in PP )
- I dislike the crap sound of so call "audiophiles" SE tube like 300B
- I do not like UL whatever final tube I use over the years ( and boy I have make a lot of tests )
- In the final it is all about your years only, as everyone are different in taste about music
So this are my 2 penny thoughts .
Thanks - perhaps I did not explain properly (English is not my native language) but the clinical sound from UL is not my taste either. (might then use solid state instead)
Good to hear you like pentode in SE.
I'll try to create some shunt regulation 250 - 300 V for g2 and try that.
AM.
The RH807 design was subjected to testing with a bypass cap on the g2, with a full regulator and with just a series resistor. Kitic reported almost no difference between the different approaches.
I tried some similar approaches myself and can confirm that how you deal with the screen is not that critical.
This is attributable to the Plate to Plate feedback.
Replacing the input triode with a pentode will produce far better results than messing with the screen grid.
Shoog
I tried some similar approaches myself and can confirm that how you deal with the screen is not that critical.
This is attributable to the Plate to Plate feedback.
Replacing the input triode with a pentode will produce far better results than messing with the screen grid.
Shoog
Shoog, thanks for this very valuable information, I had not read this before.
I have several NOS ECC81 (Japanese made) and several NOS E180CC. I want to try the sound with these out before spending more money or cutting extra holes in the chassis, am trying to keep costs down (although I did splash out and spend more than I wanted on the transformers and choke: all James).
I'll start looking around for some decent shunt regulator to ensure I will not have some runaway voltage on g2. (the old tubes were made to a different standard than todays. A slight misalignment of g2 and strange things happen to the currents in g2).
Thanks to all for their inputs, I have now a way forward to resolve my concerns. AM
Or far better it is what i use : SRPP in front of pentode !
http://i27.tinypic.com/2s7jjoz.jpg
http://www.diyaudio.com/forums/tubes-valves/148638-tube-se-amplifier-6p3s-russian-tube-2.html
Last edited:
Or not !
Maybe he is in searching for what is best for him , no fuss about : just to keep open mind and try all possibilities .
OK ?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.