The SE 1625 I just built is still doing very nicely.
Cathode bias, triode mode.
Time will tell how well it holds up to the 516V B+. With a 60VDC+ drop from the cathode resistor will bring the plate down to 456V. So far so good...but only 30mA running through the tube- plate and screen combined. 14K OPT
Nice transparent sound... better than the current production common octals I've heard in the same type of set up.
Don't forget the pre-amp/ driver stage... the pre-amp and power stage work as a team. It's amazing to simulate the stages together and see how they affect each other.
The VR tube regulating the pre-amp tube voltage supply also makes make an audible improvement.
Roger
Cathode bias, triode mode.
Time will tell how well it holds up to the 516V B+. With a 60VDC+ drop from the cathode resistor will bring the plate down to 456V. So far so good...but only 30mA running through the tube- plate and screen combined. 14K OPT
Nice transparent sound... better than the current production common octals I've heard in the same type of set up.
Don't forget the pre-amp/ driver stage... the pre-amp and power stage work as a team. It's amazing to simulate the stages together and see how they affect each other.
The VR tube regulating the pre-amp tube voltage supply also makes make an audible improvement.
Roger
Yes, that will work! Also, keep a 100R/2W screen stopper resistor directly on the screen pin.
And 1. switch only when amp is not powered, 2. use a heavy amperage switch. Regular switches cannot handle high DC, but with the two constraints above they can be used in this application.
The 807 is not particularly well suited for triode strapping either, due to screen grid voltage limitations, yet triode strapped 807 sound very nice. I never tried UL 807, but I would guess that you can safely run 807 in UL up to the voltages you would run it in triode mode. I don't have suitable UL-tapped SE iron to try it, but might have such PP iron. Will try it one day UL and report on it;
As long as plate voltages are kept in the 400V range UL should be OK. The strict 300V screen grid rating applies ONLY for tetrode mode. In triode or UL you can run it higher, provided that the screen is ALWAYS at a lower potential than the plate - which it always is in triode or UL mode. In tetrode mode, the plate can swing below the screen and then your screen simply melts if there's higher than 300V voltage on it. BTW, it will also melt if there's only 200V on it when no plate voltage is applied!
The single most important factor for running 807 is to keep the screen always lower / much lower than the plate.
Interestingly, in tetrode mode the higher the plate voltage, the lower the screen current is! That is, the screen is much happier with 700V on the plate than 400V plate. As the screen grid is the Achilles heel of the 807, anything that keeps its current down, will help tube life. It's not the voltage that kills the screen, it's the dissipation, which is kept low when the plate has high enough voltage to attract the electrons. When plate voltage is lower the screen sinks much larger quiescent current, that is, dissipates much more.
On loading: higher the load, the better control the tube has over it. As load increases, distortion lowers (up to a certain point). Also, power drops a little - 10-15% or so. 7W of great sound versus 8W of mediocre sound.
Tube datasheets usually give you a load value which gives the most power output at a given THD. However, that is not the point where the tube sounds the best. Still, almost every manufacturer sticks to those operating points because if your amp has 10-15% less power on paper than most competitors using the same tube, then it's a marketing disaster. As 95% of consumers have absolutely no hearing at all, it's the numbers that keep companies alive.
Of course not! It supposed to supersede 6L6 by anode voltage, with no anode leg (3) near the filament one (2). Otherwise, it is the same tube. I don't know how big screen voltage can be allowed under the shade of the anode with the same potential, but I believe it is nothing to worry about. In triode mode max G2 voltage for pentode mode does not apply. You can easy insert a milliampermeter between anode and screen grid and find the criterion, when power dissipated by G2 approaches it's maximum.
Does this apply to all output tubes or just 807´s?
It´s a pitty I last week ordered 5K:8 ohms trafos from Lundahl here in sweden.
But if the amp turns out good maybe I´ll ask him to make a pair of custom wound for me,they werent to expensive....
To all 6L6 siblings. They have to be driven either above destruction, or loaded as less as possible, according to their transfer characteristics.
I used them in my first Pyramid design. It was the last time I used 6L6 and siblings. I got then 40W of sound quality that satisfied me from 4 of 6P3S-E, 2 in parallel.
I used them in my first Pyramid design. It was the last time I used 6L6 and siblings. I got then 40W of sound quality that satisfied me from 4 of 6P3S-E, 2 in parallel.
I have found this to be true of most tubes.
That doesn't mean you can't get good sound from the 5K:8's.
You just can't go too high in voltage lest 2nd order distortion goes through the roof and becomes audible.
I consider myself as not nearly as knowledgeable as many on this forum.
I will present the following as a hypothesis/ theory, whichever you prefer to call it. I will let more knowledgeable people chime in and determine if it is actually fact. Once I test it and have proper data, I may have a stronger case...or find it to be inaccurate rubbish...
I have found using simulations, curves and my own ears that using higher voltage with a higher load in triode modes decrease 3rd order harmonics at the expense of increased 2nd order. Most times that I try lower impedance with lower voltage, I end up with higher 3rd order % distortion that increases as the signal (volume) decreases. I do not find this desirable...and will tolerate more 2nd order to decrease 3rd order.
As far as impedance... I use the following thought process... I do not claim this to be fact, as I am not an expert and realize I have tons to learn...but I think it is somewhere close...my math may be off, but I hope the concept is accurate.
Using quiescent current as a reference... (I know DC doesn't pass through OPT's) so its just a number that I know will be present with AC in the circuit...
Let’s say we want ½ Amp of current going to an 8ohm speaker, enough current for 2Wpk at 4Vpk, ~1Wrms
3.5K ohm= 20.92:1 turns .5A /20.92 = .023A = 2Wpk 1Wrms
5k ohm = 25:1 turns = .5A/25 = 0.02A = .02A =2W
14K ohm= 41.83:1 turns = .5A /41.83 = .012A = 2W
Not much on paper until you realize that 1Apk @ 8ohm (8Wpk, 5.65Wrms) using a 3.5k requires 47mA vs. 23mA with 14K load. This means going from 4Vpk to 8VpK @ 8ohm with a 3.5K OPT requires an additional 23mApk, whereas a 14K OPT only requires an additional 11mApk.
Less current swing=more stable and less strain on power supply regulation.
Current is what seems to shorten life. Less current is required using a higher impedance OPT. Less current also means smaller OPT’s.
The other benefit I find is when the 8 ohm speaker drops to 4 ohm during passages, the 14K OPT drops to 7K whereas the 3.5K drops to 1.75K. The 7K will result in less distortion vs. 1.75K… so the passages sounds more stable to me.
There are definite tradeoffs using this method such as-
-Higher voltage swing is required – the design of the prior stages are
more important. Otherwise higher distortion may be the end result as
higher MU tubes usually have higher distortion.
-Less power - higher voltages are required to get the same power, often
more than common octal tubes can handle
-Usually more turns of primary results in a higher resistance that pulls
down the high frequencies.
I do expect disagreements and corrections to this post... please let me know if there is a more accurate description or how it really works.
Thanks
Roger
It is not what I meant.
If you open datasheet for some 807 (or 6L6) clones, like some from Britain tubes I saw with very thorough measurements, you can see that for optimal power/distortions the tube should dissipate 50W or more on anode, that is impossible by design, so forget about optimum, and think of acceptable distortions. I've found 400V on anode / 250V on screen grid in class A is the way to use the tube in push-pull, according to my very critical demand to sound quality.
What's the point in building "just one more tube amp", if it is not of excellent quality?
If you open datasheet for some 807 (or 6L6) clones, like some from Britain tubes I saw with very thorough measurements, you can see that for optimal power/distortions the tube should dissipate 50W or more on anode, that is impossible by design, so forget about optimum, and think of acceptable distortions. I've found 400V on anode / 250V on screen grid in class A is the way to use the tube in push-pull, according to my very critical demand to sound quality.
What's the point in building "just one more tube amp", if it is not of excellent quality?
Hello Ake,
There's an operating point where you can use them:
In tetrode mode single ended, at 350Vp / 250Vg2, -18Vg1, 54mA plate, 2.5mA g2 the recommended load is 4.2K in the datasheets. This will yield 10.8W at 15%THD - which means you will get about 5 clean watts. You will get lower distortion, and bigger useful range with your 5K iron.
I have previously tried 400Vp / 300Vg2 with 5K load in SE tetrode mode. That sounded pretty good, if you try it with your Lundahl you will not be disappointed. Nevertheless, would work better with heavier load.
Janos
I agree... I will sacrifice volume and power for sound quality...to a point.
3-4 watts rms with 90db speakers in a 10'x15' room is enough volume for me.
That equates to 96db 1M away. 90db @ 8hrs requires hearing protection per OSHA... so while it may not be that loud from 8'... it is still louder than many realize.
A gas powered lawn mower at the operators ear seems to be the most common comparison for 90db.
Last edited:
It does not matter how it soldered inside, which wires/capacitors/etc it uses, if it as if disappears from the sound path.
https://www.youtube.com/watch?v=f5-OsFIt_Hk
Hmm,a little late for that now,the powertrafo is on it´s way from Poland (custom made).
Unless I wanna burn about 100v in a resistor to get it down.
Thats about 10w for 2 channels,I guess that will be quite some heat.or isnt it?
Choke input might bring it down to the correct voltage.
What voltage is the Transformer that's coming from Poland? If you have room, you can use another transformer in line with the power output to drop the voltage. A small transformer used as an autoformer might work well.
Triode strapped is still an option for a higher voltage xtfmr
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.