Thanks, for the suggestion, it'll be a good starting point for tryal of my PSE.
Will also try 480V and 45mA per tube.
You prefer the GU50 with self bias over fixed bias?
I mean single ended. Self bias means negative feedback by DC current, so you don't need to measure and adjust bias. The tube will do it for you. However, for parallel SE it would be better to have the way to adjust bias for each tube, but anyway it is better to select tubes for parallel operation.
I'm aware of what it is, but having to decide and as in my experience most tubes sound better with fixed bias, I was wondering about your suggestion.
Even with fixed bias I always add a bit of "self bias", even if it is just to measure current and stabilise the DC part.
When in parallel it is usefull to be able to adjust bias independent for each tube, as it gives more latitude on pairing the tubes.
If you have some experience on using the GU50, I'd like to profit from it, as I'm building a PSE with 2xGU50 into a LL1620/90.
PSU is about 480V (I can reduce it replacing the rectifier tube or using some stabilising circuit). I think 480V is too much for triode GU50 and to little for pentode, but as I will only be using each tube at 45mA, the grid current will not aproach the limit of 5mA. Can you help me on this?
Hi Wavebourn,
I have been playing with GU50 Triode curves for a while now, and it really looks like a 3.5k-5k transformer gives better power output with lower distortion, especially sitting at a quiescent point around 80mA at 400V. I have noticed in these forums that you typically advocate for something lower, more like 2k. What is pushing you in this direction?
A good example is attached (not my work, swiped from here).
Thanks
Attachments
One might use zener diode string with sufficient wattage to drop the voltages below 300 volts. Same as but lower dc, You may even use gas regulator tubes.
KT88, GM70, GU50. Which sounds & drives Quad ESL 57 best?
I'm a newbie with limited amplifier design knowledge. Please advise so I don't buy or have built the wrong amp for these speakers with the crazy impedance curve! Quad ESL & ESL63 Information
I'm a newbie with limited amplifier design knowledge. Please advise so I don't buy or have built the wrong amp for these speakers with the crazy impedance curve! Quad ESL & ESL63 Information
Does the speaker expect electrical damping, or sounds better with higher output impedance of the amp? I am not familiar with this speakers. Also, which tube matters less than which topology you prefer. GU-50 I would use in P-P topology; GM70 would be good for SE in A2, while KT-88 and other 6L6-likes I would prefer to see in guitar amps.
Edit: I went to the link and found that a pair of EL84 is indeed the right choice to drive this speakers. Even don't think of a "Zero feedback" fashion, with zero feedback it would sound like a crap!
Edit: I went to the link and found that a pair of EL84 is indeed the right choice to drive this speakers. Even don't think of a "Zero feedback" fashion, with zero feedback it would sound like a crap!
Last edited:
Should read " if you add positive voltage on G3 (suppressor) curves shift left", I presume.
How far do curves shift under +12v G3-bias ? what is the transconductance of G3 in a GU50 ?
and how much current is to be expected into G3 ... or is it out from G3 ... ?
Talking about G3, I wonder whether G3 can be used to adjust for symmetric bias currents; e.g. amp with 2xGU50 in PP, cathode bias, single common cathode resistor;
if tubes don't match very well, individual G3 voltage adjustment could be used to balance DC currents in the OPT ... doable ?
Last edited:
What voltage have to do with power dissipated by screen grid if anode voltage never goes below it?
Forget about pentode voltage rating for the screen grid when using it in triode mode. It is a completely different story.
Спаси́бо. Thank you.
just measured G3 current, is extremely low, maybe 0.5uA almost idependent of voltage within reasonable limits;
curves shift very little with G3 voltage, maybe 0.5mA in plate current for every 10V of G3;
300v plate & screen, 60mA plate, 470ohm cathode R;
fine tuning of plate current also works well, simple pot does the job well enough due to low G3 current.
I thought that by looking at the impedance curve at that link you would be able to determine if the Quad 57 ESLs need an amp with high or low electrical damping factor. But what is it about that EL84 tube amp (p-p or SE?) that you think would work well with the Quad ESL 57s and 63s?
Please also explain the reason (s) why low feedback amps would sound bad driving these speakers when so many otherwise amps using low feedbacksound wonderful driving other speakers.
As for required output power, my understanding is that to get what dynamics these ESLs have to offer, you need 15 wpc to get them to play at moderately loud levels. However, playing the Quad 57s much above 90db SPL can soon damage them, or if clamping protection was not added to them by users or rebuilders. Welcome - Quads Unlimited ElectroStatic Solutions - Visit
My ears won’t take much above 80db at 11 ft in my room anyway. Btw, the 57s only do down to ~ 45Hz, but I would expect that my pair of Rythmik 12" sealed servo subs are quick enough to blend well with them. Rythmik Audio • Servo subwoofer products
However, assuming that GM70 SE amps can only do 8 wpc, would a GU50 P-P amp produce 15 or 20 wpc? How would it sound with these speakers compared to the EL84, as they are both pentodes?
And are there P-P triode tube amps that might sound better? 300B P-Ps,despite very high tube replacement costs? Is the power draw and heat of GU50 P-Ps as much as 845 tube amps? Line Magnetic Audio LM-518IA integrated amplifier | Stereophile.com
I’d had hoped to be able to find someone locally who would let me hear their Quad 57s and 63s, and which might also give clues to the better amps to use with them. But I’ve had no luck at this in the NYC/Long Island area.
JacMusic inform GM70 start at 800Vdc, at this low B+ it delivery low power as under 20W.
Marco audio,
With a typical Ultra Linear output transformer, and with a typical 0 Ohm or 100 Ohm resistor from the UL screen tap to the screen:
The current from the OPT center tap to the UL tap = plate current + screen current.
The current from the UL tap to the plate tap = only plate current.
Therefore, due to the two DCRs, the screen voltage is higher than the plate voltage.
With 600V on a GU50 plate, and typical UL connections to the screen, the screen will cook!
You would have to:
Either use a separate floating screen winding (say 40%) and connect one end to a 250V supply.
Or use a follower circuit from the plate that has a large offset voltage to keep the quiescent screen voltage at 250V.
There are many examples of these two methods, but I do not know if any of them used a GU50.
Or,
Use a low voltage for both the plate and the screen (250V to the UL transformer center tap).
(probably not a good use of the GU50).
With a typical Ultra Linear output transformer, and with a typical 0 Ohm or 100 Ohm resistor from the UL screen tap to the screen:
The current from the OPT center tap to the UL tap = plate current + screen current.
The current from the UL tap to the plate tap = only plate current.
Therefore, due to the two DCRs, the screen voltage is higher than the plate voltage.
With 600V on a GU50 plate, and typical UL connections to the screen, the screen will cook!
You would have to:
Either use a separate floating screen winding (say 40%) and connect one end to a 250V supply.
Or use a follower circuit from the plate that has a large offset voltage to keep the quiescent screen voltage at 250V.
There are many examples of these two methods, but I do not know if any of them used a GU50.
Or,
Use a low voltage for both the plate and the screen (250V to the UL transformer center tap).
(probably not a good use of the GU50).
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.