Your prerogative. My viewpoint is, why force a tube to do something it's completely unsuited to do (supply grid current at near zero impedance) when you can use a MOSFET like the AOT2N60 with low single digit impedance and negligible reverse transfer capacitance. Besides, at 2.3kV it's only drawing grid current above 40W.
These amps weigh about 180lbs each, but I'm able to put that weight into the power supply, with LCLC supplies on the driver, the output stage and even the filament, since I'm saving a lot of weight by using a technologically superior solution for the grid current issue. To me a win-win, perhaps not to you. That's what makes freedom so grand, isn't it?
When building HI MU tube amps they have very high internal plate impedance and the amp has a poor damping factor without lots of feedback. Best to choose speakers with lots of damping like Altec, JBL, B&C and more, another benefit is you don’t need lots of power; as the Japanese found you don’t need to use very high voltages, they work best at lower voltage, high current, the 805 can give 40 watts at 800 volts SE. Speakers designed for transistor amps are best to keep clear of as you will have boomy bass of course unless that's what you want.
Phil
Phil
No boomy bass here. Tight as a drum. Multiplying the speaker impedance using an autoformer works like a charm, with no loss of sound quality I can detect - in fact everything sounds better with the autoformer on the speaker.
I'm not the only one who has found that these tubes sound better at very high voltages and lower current. Ditto for Tubelab and Ward from Monolith. The only reason to use low voltage is fear.
I'm not the only one who has found that these tubes sound better at very high voltages and lower current. Ditto for Tubelab and Ward from Monolith. The only reason to use low voltage is fear.
Last edited by a moderator:
I just wanted to show there are other ways to use the Hi Mu triodes; I don’t mean to be critical of your amp which is a work of Art and skill.
Yes of course is just a matter of opinion; if Hi mu tubes sound better at high or low voltage. I have never made one using very high voltage so I don’t have one to compare.
If you measure a Hi Mu amp without feedback the internal impedance can be as high as 10-15 ohms that is why you need to use the Auto transformer; you must have had a lot of boom before using the Autotransformer or why would you use one?
I would have thought the main reason not to use high voltage was the enormous cost, the huge output & power transformers, and of course fear, but why use 2000 to 3000 volts when you don’t need to; I have never used more than 1000 volts, don’t need to with 100Db well damped speakers and don’t need the Auto transformers.
I have made amps using KT88’s 2a3’s 300B’s; I prefer the HI mu tubes. Another advantage is how long the Hi Mu tube last compared to KT88’s 300b’s using USA, English and Russian made tubes any way.
I made an amp for a friend using 572b’s 150 watts, he took 12Db feedback off as he loves the boomy sound with his Ushies.
Phil
Yes of course is just a matter of opinion; if Hi mu tubes sound better at high or low voltage. I have never made one using very high voltage so I don’t have one to compare.
If you measure a Hi Mu amp without feedback the internal impedance can be as high as 10-15 ohms that is why you need to use the Auto transformer; you must have had a lot of boom before using the Autotransformer or why would you use one?
I would have thought the main reason not to use high voltage was the enormous cost, the huge output & power transformers, and of course fear, but why use 2000 to 3000 volts when you don’t need to; I have never used more than 1000 volts, don’t need to with 100Db well damped speakers and don’t need the Auto transformers.
I have made amps using KT88’s 2a3’s 300B’s; I prefer the HI mu tubes. Another advantage is how long the Hi Mu tube last compared to KT88’s 300b’s using USA, English and Russian made tubes any way.
I made an amp for a friend using 572b’s 150 watts, he took 12Db feedback off as he loves the boomy sound with his Ushies.
Phil
So you put autoformer on opt secondary? 2:1? That should make tube working into higher impedance. Less power, less distortion
Exactly why my 2a3 are pumping into >12k, good bye (global) fb.
I use a Spelz Zero autoformer set to 4X impedance ratio between the amp and speaker. This gives me a speaker impedance of 13 ohms at 1000Hz, 14 ohms at 100Hz (lowest my meter goes) and 7.34ohms at the lowest impedance dip (10kHz).
Still plenty of power to crank Metallica when I want to, but Allison Krause sounds so airy and open, like a good SET should.
The OPT itself is a 12k:4 so it's already a nice stepdown.
Last edited:
I'm not offended, just pointing out the experiences of myself and others. Sorry if it came off that way.
Yes, acoustic bass like Paul Chambers was a bit loose at times, so I tried out the autoformers. What amazed me is that Everything sounded better with them in place, and I didn't lose any detail I could detect.
I figure the output impedance of my amps is more like 2-3 ohms at 2300V and 160mA due to the high stepdown of the Monolith OPTs. With a high impedance speaker I could avoid the autoformers but I love the sound of the Infinitys - those EMIM mids are hard to beat! I'm using brand new kaptan EMIM diaphragms from Apogee acoustics and they are a revelation.
I have been a ham radio operator for nearly 40 years. Many of my early transmitters used tubes, and I learned a lot from the "old timers" in my quest for electronics knowledge. The old guys always said to run the big transmitting tubes at the highest possible voltage and lowest current to get the desired power output, they would live longer. Most of these guys have left this world by now, but this used to be an accepted "truth."
My limited experience with big tube RF amps didn't disprove it, but there was not enough data to prove anything either. The big tubes are rather high impedance creatures, and their impedance must be matched to a 50 ohm antenna system. As you increase the voltage, the impedance increases, and the matching network becomes more difficult, and lossy. There will be a point where it is impossible to achieve the necessary impedance transformation with acceptable bandwidth.
There are plenty of parallels between the RF tube amp world, and the audio tube amp world. Yes, running an 833A on 2500 volts requires a much more expensive OPT than running the same tube on 800 volts, the tube may live longer, but tubes are not nearly as expensive as OPT's.
I have found with KT88's and 300B's that running the tubes at or near maximum plate voltage, while running a 5K ohm OPT instead of the customary 3K to 3.6K, will yield a much more dynamic sounding amplifier.
Each designer is free to choose the components he prefers. As a engineer recently "retired" from a 41 year career at Motorola, I always searched out newer and better components for a given task. This leads to "advancing the state of the art in two - way radio design." I first stuck a mosfet FOLLOWER in front of an output tube grid over a dozen years ago, and discovered that it truly is the right component for that job due to its high Gm. Some prefer to have a pure tube amp with technology that hasn't evolved much in the last 50 years, that's OK if that's what you want.
Modern mosfets have evolved enough to remove their other limitations for follower duty only. Would I put common source mosfet stage in a tube amp, no they are not linear enough, and posses to much capacitance which varies with the signal voltage.....even for a guitar amp.
Cathode depletion generally follows cathode current. You can get longer life out of most tubes with less current and higher voltage.
Transmitting tubes like 811 or 833 may be safer to push than something like a KT88.
I am glad I had time to talk about this with my father about this. He taught me some of the care and feeding rules of power tubes.
Also round 1990 he called the 6l6 a "Dainty little tetrode" can handed me the RCA transmitting handbook. Best advice ever.
I'm not so sure that many modern KT88's are up to the task of eating 600 volts in triode. I know that my EH KT88's started tending toward bias creep, and possible runaway over 500 volts. That was in a SSE board which uses a standard coupling cap / grid resistor / cathode bias circuit. Grid resistor was 100K.
You might be able to run more voltage if the grid circuit resistance was significantly lower.....another reason to drive the grid with a mosfet.
Modern new production tubes often have a less that perfect vacuum, or impurities that will pollute the vacuum over time, leading to grid current. This is one of the main reasons that a tube will go into runaway after a year or two of reliable service.
Another possibility is pentode operation with a lower screen voltage. Use local feedback around the KT88 to bring the plate resistance down.
In simple terms from an old ham.....cathode current kills the cathode.
I had an old transmitter that he gave me that ran a 4-250 on 1700 volts. I rewired the HV secondary in the power supply with a solid state FW bridge to give 4000 volts, and stuffed a 4-400 into it. I got about 1 KW out of that thing with the 4-400 glowing cherry red. It was still working when I gave it away several years later. It was the only thing I ever made that really scared me every time I tuned it up. Misload it a bit, and the blue ball of plasma would engulf one of the variable caps. Wiped out TV sets for half a block.....shielding, what's that. Hey I was a dumb kid with a kilowatt, what more could you want.
Lest say 10r, that is 30k, with ra 4k is better damping and match.
amazing curves
Attachments
Last edited:
I see that have less that 1% thd at 50w in the simulator
Triode / Pentode Loadline Simulator v.1.0 (20160104 www.trioda.com)
is this true?? better than 211 better than 300b, this tube work very happy in A2
All I can say is that it sounds fantastic and has been completely reliable since I built it. Haven't even had to adjust the bias since I got it dialed in.
I still haven't bothered to do the PS mods I talked about a couple pages back. Too much work when I'm already enjoying it immensely. One of these days (or years...).
Magz, I have a GM70 and a pile of 6e5p tubes I just acquired.
In triode mode, does G2 tie to plate (no gridstopper)? I have some 7788's to try as well but where does G2 tie on this type of tube(G3 is tied to plate I assume)?
I'm in unfamiliar territory when it comes to pentodes and tetrodes.
In triode mode, does G2 tie to plate (no gridstopper)? I have some 7788's to try as well but where does G2 tie on this type of tube(G3 is tied to plate I assume)?
I'm in unfamiliar territory when it comes to pentodes and tetrodes.
I use 6E5P in triode to drive GM70. Do as Magz advises, use a 100-200R carbon composition resistor from screen to anode.
I played with all the usual suspects, E180F, D3A, EF184 and so on but found I preferred the 6E5P. Of course your taste will be different from mine so be prepared to play with driver valves.
The GM70 loves grid current so if you can tolerate sand use a MOSFET source follower to drive the grid. Even if you don't intend A2 operation it helps with blocking distortion.
Merry Christmas everyone
Cheers
Matt
I played with all the usual suspects, E180F, D3A, EF184 and so on but found I preferred the 6E5P. Of course your taste will be different from mine so be prepared to play with driver valves.
The GM70 loves grid current so if you can tolerate sand use a MOSFET source follower to drive the grid. Even if you don't intend A2 operation it helps with blocking distortion.
Merry Christmas everyone
Cheers
Matt