Good morning, afternoon and evening.
I'm new to tubes/valves so I'm starting with a simple preamp and seeing how it goes.
The circuit uses two miniature double triodes with their heaters in series. I'm running this from a 300mA constant current from an LM317. With only 47uF between bridge and regulator I'm getting almost no hum (sorry to be so subjective, but it's inaudible below the white noise - at some point I'll low pass and attempt to measure).
Is there any reason I shouldn't persist with this? Seems sensible to me as the valves take a long time to warm up, there's no surge on the bridge from cold heaters and empty 4700uF capacitors, and as the output is a cathode follower the DC output voltage rises slowly, eliminating turn-on thumps. But I've never heard of anyone using CC to drive heaters.
thanks
I'm new to tubes/valves so I'm starting with a simple preamp and seeing how it goes.
The circuit uses two miniature double triodes with their heaters in series. I'm running this from a 300mA constant current from an LM317. With only 47uF between bridge and regulator I'm getting almost no hum (sorry to be so subjective, but it's inaudible below the white noise - at some point I'll low pass and attempt to measure).
Is there any reason I shouldn't persist with this? Seems sensible to me as the valves take a long time to warm up, there's no surge on the bridge from cold heaters and empty 4700uF capacitors, and as the output is a cathode follower the DC output voltage rises slowly, eliminating turn-on thumps. But I've never heard of anyone using CC to drive heaters.
thanks
There are good reasons to use CC. The undersea telephone repeater tubes were designed to work with CC. For decades, always-on!
I’m not sure from the documentation I read, why.
I’m not sure from the documentation I read, why.
I did think that, but equally, if one heater has a lower resistance than another then a CV supply is going to make it hotter.
No go on.
In fact some tubes ( P* in european terminlogy) is designed for 300mA filament AND to be connected in series.
What tubes do you use ?
The reason it takes longer with the CCS is that the current is limited to 300mA at startup when the heater is cold. In older TV sets the tubes were connected to a low impedance source, like for example the mains supply, resulting in a high current peak when the tube is cold, and quickly settling to 300mA when the heaters are hot.
The heaters have more resistance when hot, so constant voltage drive will cause a big pulse of current that rapidly heats the filament up towards the final temperature, constant current will take a lot longer as the initial power is much less than the final power. Some filaments might not even work with constant current and get stuck in a luke-warm low-resistance mode.
Heaters in series, if properly matched, are like constant voltage, and they all heat up in step. If not properly matched the ones that heat up faster then see more voltage drop and can fail through thermal runaway.
The underseas cable uses constant current because there is no real choice - the resistance of the wire along the 100's of miles of cable is far higher than the heaters, and gets most of the power, so the heaters are in series with a large resistance (which is close to constant current already if you think about it). There may be many 1000's of volts along the cable to drive a handful of 12V heaters, setting constant current ensures the heater gets to the intended operating point.
Heaters in series, if properly matched, are like constant voltage, and they all heat up in step. If not properly matched the ones that heat up faster then see more voltage drop and can fail through thermal runaway.
The underseas cable uses constant current because there is no real choice - the resistance of the wire along the 100's of miles of cable is far higher than the heaters, and gets most of the power, so the heaters are in series with a large resistance (which is close to constant current already if you think about it). There may be many 1000's of volts along the cable to drive a handful of 12V heaters, setting constant current ensures the heater gets to the intended operating point.
voltage regulated long as the current at startup not dropping voltage.
imposing current limit is just that, longer warmup.
Hopefully without voltage drop, for too long.
Long as it stabilizes. DC is DC on heaters.
Most the benefit if any is usually lower noise.
imposing current limit is just that, longer warmup.
Hopefully without voltage drop, for too long.
Long as it stabilizes. DC is DC on heaters.
Most the benefit if any is usually lower noise.
Not arguing at all, just thinking aloud: since heaters are PTC, feeding them constant voltage makes them sort of self regulating while constant current takes much longer to stabilize.
The best of both worlds would be a constant voltage supply with a current limit a bit higher than the normal heater current.
The problem with series tubes connected in parallel is that the heater voltage is not really controlled at manufacturing or design. If you look at datasheets from different vendors for the same P-tube, you will sometimes see different Vh, but the same 300mA. You may need to match them to operate properly when connected in parallel.
About the slow startup, a current spike in the CCS could be created using a capacitor, depending, of course , of the CCS design.
About the slow startup, a current spike in the CCS could be created using a capacitor, depending, of course , of the CCS design.
A baffling problem randomly appearing on some so called "tube preamp" ,
or "hybrid" Guitar or Bass amps is that they often add a solitary 12AX7 at the preamp, half of the time only cosmetic, nothing beyond a cathode follower, fed from existing SS power amp rails so meager +40 to +70V tops.
Big problem being that 12V filaments are fed from same rails.
Using a large and hot resistor which behaves like quasi constant current source.
A problem was noticed by Techs that Marshall amps worked only with original Marshall brand tubes, and most Sovtek, but none other.
On measurement, those filaments barely reached 9VDC or less across pins 4-5
Current theory is that Russian tubes need less initial current than others, so dropping resistors were tightly calculated.
Those just reach 12Vdc, others simply never do.
Of course, when dropping 6 or 12V they pass 300 or 150mA as rated, the difference lies in the warm up thermal time constant.
or "hybrid" Guitar or Bass amps is that they often add a solitary 12AX7 at the preamp, half of the time only cosmetic, nothing beyond a cathode follower, fed from existing SS power amp rails so meager +40 to +70V tops.
Big problem being that 12V filaments are fed from same rails.
Using a large and hot resistor which behaves like quasi constant current source.
A problem was noticed by Techs that Marshall amps worked only with original Marshall brand tubes, and most Sovtek, but none other.
On measurement, those filaments barely reached 9VDC or less across pins 4-5
Current theory is that Russian tubes need less initial current than others, so dropping resistors were tightly calculated.
Those just reach 12Vdc, others simply never do.
Of course, when dropping 6 or 12V they pass 300 or 150mA as rated, the difference lies in the warm up thermal time constant.
The issue I think is that data sheet and reliability/life are based on a voltage spec and tolerance during normal operation, so that is the final arbiter imho.
OK. But that wouldn't be an issue with a "real" cc, so long as it had enough voltage headroom to force 300mA all the way from cold to 12V, surely?
CV with CC limit would be the best, agreed. But a "real" CC can be implemented with one LM317 and one 0.6W resistor, and it seems there's nothing could possibly go wrong. Said he, suddenly remembering how reliable TVs were when they were full of series connected valves...
Loved the transatlantic cable story by the way. Somewhere you really, really need reliability!
True, but you don't reason like a Corporation bean counter 🤣
A plain resistor is always cheaper and easier to mount than a Voltage Regulator which to boot needs heatsinking and in any case typically has a 35V or so voltage limit, remember they are coming from a power amp +V rail.
Minimum +42V and typically (in 300W Bass amps) between +63/70V
The objective is to spend less than winding a few extra turns around an already present core which amounts to 1 or 2 extra bucks, tops.
Here in the Marshall Valvestate 80 Guitar amp they feed 12AX7 12V filament straight from 30+30 VAC main supply windings, using two 180 ohm 7W resistors as droppers.
Current is barely acceptable 133 mA , not sure how much is needed for the initial cold filament current surge.
Yes, they do eventually stabilize, at least with Marshall labelled or Sovtek tubes.
Again, in higher power amps dropping from 63-70V DC rails problems appear 🤷
They are often mentioned in relevant Forums.
"Technical parameter" considered is co$t , not performance 🥴
Personally, since I wind my own transformers, I simply add an extra 6.3V tap to main winding and call it a day, it is actually cheaper (for me) than a 15W or 2 7W ceramic resistors but hey .... 🤷
A plain resistor is always cheaper and easier to mount than a Voltage Regulator which to boot needs heatsinking and in any case typically has a 35V or so voltage limit, remember they are coming from a power amp +V rail.
Minimum +42V and typically (in 300W Bass amps) between +63/70V
The objective is to spend less than winding a few extra turns around an already present core which amounts to 1 or 2 extra bucks, tops.
Here in the Marshall Valvestate 80 Guitar amp they feed 12AX7 12V filament straight from 30+30 VAC main supply windings, using two 180 ohm 7W resistors as droppers.
Current is barely acceptable 133 mA , not sure how much is needed for the initial cold filament current surge.
Yes, they do eventually stabilize, at least with Marshall labelled or Sovtek tubes.
Again, in higher power amps dropping from 63-70V DC rails problems appear 🤷
They are often mentioned in relevant Forums.
"Technical parameter" considered is co$t , not performance 🥴
Personally, since I wind my own transformers, I simply add an extra 6.3V tap to main winding and call it a day, it is actually cheaper (for me) than a 15W or 2 7W ceramic resistors but hey .... 🤷
Tube heaters are specified to operate specific voltages, and the currents they draw are listed; as an example, a 12AX7 works at 12.6V and draws 150mA. The typical current different tubes may draw can vary a bit. If you power the tube with a CCS, you could end up with an over or under voltage across the heater, most likely not catastrophic in amount. It certainly would not hurt to use the CCS, if you tweak the current to get to the specified voltage range. Nice thing about using a CCS is that you avoid the inrush current when lighting up a cold tube. Keep in mind that it takes a bit longer for a tube to warm up when using a CCS because a cold filament has a lower resistance than a hot one. That is just a minor, if any inconvenience, though. Since a CCS would be DC in nature, there would be no AC hum caused by it, so that's a nice fringe benny.
Strictly speaking, this is not always true. Yes, many receiving tubes were designed for a particular steady state operating voltage. However, with the advent of television and mass market design (including A5 radios), many tubes were designed for a constant steady state operating current and the voltage was just a target. In the GE data sheets, this characteristic is termed the "controlled heater warmup characteristic". Such tubes were specifically designed to operate in a series filament configuration (usually with 300mA or 600mA level). These tubes can still be used with a voltage source of appropriate value, but oftentimes their voltages are approximate.
People end up chasing there tail never reaching a definative answere.there are tubes specificlly designed for series operation.and there are audio tubes mostly designed for a specific voltage to operate. Back when tubes were designed using a ccs never came into the picture and there are millions of tubes lasting years without fancey current limiting. Your worrying about an issue that does not exist.