EtherealWidow wrote:I hear that filaments ( and maybe this is just for 12ax7's) can be run off 6.3v +/- 10%. 5.7v-6.9v would be fine I imagine. Correct me if I'm way off base.
I think this correct for most tubes, +10%, -20%.
Higher shortens life, lower impacts performance.
Hey Cliff, how much higher would you call dangerous? In just about every amp I've built, I've measured my heaters at about 3.4vac to ground on each side. Would it be wise, if having custom trannies built, to have them wound around 6vac (3vac each side) on the dot, to account for higher line voltages? I imagine this would put them right at about 6.3vac out of the tranny. I've been thinking about having this done ever since I first read one of your posts where you were talking about elevated heater voltage being a problem.
Blackburn wrote:Would it be wise, if having custom trannies built, to have them wound around 6vac (3vac each side) on the dot, to account for higher line voltages?
If you are having trannies wound, you can account for the voltage flux better by having a multi-tap tranny. There is simple math here. If your optimal value is 120v:6.3v, the turns ratio is 19.05 primary to 1 secondary turn. The secondary winding will be about 20 or 21 turns, so the primary will be about 380 turns. Given how few turns there are on the secondary, it is simpler to wind the primary for 115-120-125. This should accommodate most wall supply voltages in the US. It is not a big deal for the winder to pull out a couple of extra taps if he plans for it. Assuming 20 turns secondary, one gets pulled at 364 turns, one at 380 turns, and one at 396 turns. The extra length of wire is not that significant cost-wise and it should not be too much of a challenge to get it on the bobbin and fit to the "window."
Second-guessing the voltage likely means you will never be spot on. Go measure your wall voltage once per hour for a few days and make a chart of it. You'll be unpleasantly surprised. Then wait for a really hot day when all the AC units are running all over town and see it go below 115v. On this day, you will starve the filaments if wound for 120:6.0.
ET (!?), I used a circuit out of Kevin O'Connor's TUT2 book, IIRC. Same as your hand drawn but with 2x15,000uF caps for extra smoothing. TUT 2 pages 2-60 and -61. "The choke in conjunction with the second cap gives a 20dB ripple reduction."
This is still not as good as a regulated supply.
A 6.0V regulated supply from TUT 1 Fig. 2-16, using a 5A type 338 regulator. TUT 1 was published in 1995 so the 338 has no doubt been replaced two or three times since then!
Blackburn wrote:Would it be wise, if having custom trannies built, to have them wound around 6vac (3vac each side) on the dot, to account for higher line voltages?
If you are having trannies wound, you can account for the voltage flux better by having a multi-tap tranny. There is simple math here. If your optimal value is 120v:6.3v, the turns ratio is 19.05 primary to 1 secondary turn. The secondary winding will be about 20 or 21 turns, so the primary will be about 380 turns. Given how few turns there are on the secondary, it is simpler to wind the primary for 115-120-125. This should accommodate most wall supply voltages in the US. It is not a big deal for the winder to pull out a couple of extra taps if he plans for it. Assuming 20 turns secondary, one gets pulled at 364 turns, one at 380 turns, and one at 396 turns. The extra length of wire is not that significant cost-wise and it should not be too much of a challenge to get it on the bobbin and fit to the "window."
Second-guessing the voltage likely means you will never be spot on. Go measure your wall voltage once per hour for a few days and make a chart of it. You'll be unpleasantly surprised. Then wait for a really hot day when all the AC units are running all over town and see it go below 115v. On this day, you will starve the filaments if wound for 120:6.0.
Yeah, I considered that too. Voltage is always fluctuating and it would be difficult to get it spot on, even with a multi primary. I guess it comes down to whether 3.4vac vs 3.15vac is severe enough to cause premature death of a tube or not. Since most non variac users should be facing similar circumstances due to higher line voltages, I'd assume the difference can't be that damaging, since there hasn't been much to alleviate the issue, other than using a variac. Even me, I've been using NOS or OS tubes exclusively once an amp has been tested and is in good, consistent working order and I've never had tubes die on my watch. A curious issue.
The problem is when you don't heat the heaters up enough (too low a voltage), the cathodes have to work harder for the same current levels because of the decreased efficiency (less free electrons floating around). With most preamp tubes this won't really kill them (except maybe 12AT7's in high current situations) but with power tubes you work the cathodes much harder for the same power level and eventually poison the cathodes. In the preamp tubes you lose some gm, which you may or may not notice depending on the situation, but much below 6V and some tubes aren't very happy anyways.
When you overvoltage tubes, you get the obvious overstressing of the filament itself as well as the cathode which will prematurely burn off its coating. You eventually cause the tube to become gassy which kills gm.
I've read before that it's worse to undervoltage than overvoltage and the little chart below seems to agree. While +/-10% is specified in datasheets, I treat that as an absolute maximum and aim to get as close to 6.3V as possible. I've used all sorts of tricks in the past but the best thing I've found is to know what your amp wants to see for input voltage to get the heaters right on and build a bucking transformer dealy that'll get you close enough. You'll never be spot on because the circuits we use (minus DC) are unregulated and constantly varying.
That's why I want to build some sort of switching supply that constantly keeps your special preamp tubes *exactly* where they want to be run at for optimal life expectancy.
I've seen this plot reproduced on several amp-building sites, but I can't help wondering where it originally came from, what heater/filament types it does encompass, and if it does actually apply to the type of "receiving tubes" we are dealing with.
Respected publications such as the Radiotron Designers Handbook state a heater voltage variation of up to +/- 10% is acceptable, yet this plot indicates a dramatic reduction in valve life, especially for under-voltage.
I'm particularly suspicious of the under-voltage curve, which indicates a 3-hour lifespan when a valve heater is operated at 85% of its nominal value.
I have tubes that are 40-50 years old with months of use.. maybe I am underestimating that... months of use at 45-70 vac from the line. What ya think the heaters are at 65 volts? Just over half, right?
I was amazed to hear it was a problem years ago when the internet was invented.
I have heard it's about broadcast tubes. Ya know, the ones as big as a car..
Signatures have a 255 character limit that I could abuse, but I am not Cecil B. DeMille.
Did a little digging and found out that chart came from "Audio Cyclopedia" by Howard Tramaine. Here's what Tales from the Tone Lounge says:
I would like you to study the diagram below carefully. It is from the book Audio Cyclopedia, by Howard M. Tremaine. It deals with the prospect of running the filament voltage on any tube at other than 100% rated value. The illustration shows that tubes will fail prematurely if the filament voltage is either higher or lower. For maximum tube life, the voltage must be 'right'.
I would order one of those 6,3V supplies for the bench as a reference.
I get easily confused by the noise from the 6SL7 - is it the tube or the filament circuit? Once it took me about 40 swaps until I found a quiet one for a V1 position. But half of those were used, the rest NOS. Then again, it was actually a used one that got pole position.
I recall also this was an issue with HV transmitting tubes, not the pedestrian stuff we are dealing with. Someone here did an experiment running filaments on ca. 2VAC as a way to reduce power and found the tubes lasted for days or weeks of continuous operation.
With the help of these guys (especially Martin M)....I have built two SL7 amps and I am currently working on my third.
So far, I have found no need for DC Heat. Just observe the "normal" lead dress routines for AC wires.
NOS SL7 are pretty affordable compared to most of their AX7 brethren.....but from my experience (about 10-15 tubes) I have found the NOS to be abnormally micro-phonic. Who knows where these things have been for the last 5 decades, and how they have been handled.?
Anyway.....I have had very good success with the New Production SL7 from Sovtek and Tung-Sol. The guys I have had help me "voice" the amps say they prefer the Tung-Sol over the Sovtek.
But with the pricing, it would not be too hard to spring for a few of each.....probably. I realize as well as anybody what rotten times these are for disposable cash.
Good Luck
ALRIGHT. As much fun as it is talking about the heaters alone, I still wonder, does anyone know for sure whether or not it's ok to use 6SL7's in pretty much any 12AX7 circuit?
I know that it was established that with typical 100k plate load resistor and 1.5k cathode resistor the circuit would be fine. What if that's not the case? In that link I posted with the grid leak cascode, would it be fine to substitute the 12AX7 for the 6SL7?
Passfan, in your 6SL7 Express, did you use the typical 1.5k Rk and 100k Ra?
