Tube Pre-Amplifier Power Supply Questions

[nathanpc]

Recently, out of curiosity, I've started looking into tubes (triodes to be precise). Since I've built many bipolar amplifiers, I've decided to build a tube pre-amp, just for fun and to learn more about this technology. The only problem that I have with it are the high voltages. I've build a HV power supplies for my nixie tubes, but I prefer to avoid having 100V+ floating around on my bench. So I started looking at low-voltage tubes and came across this article that talks about space charge tubes that were designed for car stereos and things like that, but sadly I wasn't able to find any of the triodes listed there on MercadoLivre (the brazilian version of eBay).

Since my search for low-voltage tubes wasn't successful, I decided to think about operating them from 30-60V since it's a region that I'm a bit more comfortable with, but then comes the problem. I don't have a power supply capable of supplying those voltages, so I thought about generating them using a simple CMOS oscillator + voltage multiplier combo, but I don't know how much current the anode requires, and I couldn't find this information in the tube datasheets probably because of the terms used in them, so my first question would be: What's the typical current of a tube operating at those voltages (maybe what to look for in the datasheets)?

Some of the tubes that I'm interested in trying (selected purely on availability and price):

• 6BA6
• 6AS6
• 6EJ7
• 6DJ8

Would any of those work with my specs? Any tips or traps for young players?

[Paul Price]

The tubes you list all require a power supply that will supply something like several tens of milliamps to do just about anything practical and usually are biased with plate voltages >100V.

The 6BA6 is a pentode, that is not used in pre-amps and usually is set to work with >100V plate voltages.
The 6DJ8 is a dual-triode that could possible be used.

The 6AS6 and 6EJ7 were designed to operate at 12-14V plate voltages.

The most popular triodes used in tube pre-amps were: 12AX7, 12AU7 and can be used to work with plate voltages around 30-50volts with the right choice of biasing and were low-noise optimized.

You should maybe search for tubes like thes3:  1T4
                                                                           3S4, 3V4
                                                                           1T5
                                                                           1U5
                                                                           1S5

These tubes were used in portable radios you could take to the beach using a  few "D" cells for their heater voltage and a larger 45-V battery to provide the optimal plate voltages. These tubes are not low-noise types and may be unsuitable for pre-amps unless you like that old radio style hiss to sooth your ears in quiet periods in tune playback.

Your best bet is to build a simple boost convertor to convert 5v or 12V to +45 or so. You can make this with even a 555 timer and N-Channel MOSFET with a VDSmax >50V and a low threshold voltage. Or youi can us a 12V to 24V transformer and a simple diode-capacitor voltage-tripler circuit to create the voltages you need.

[T3sl4co1l]

I made one of these for breadboarding,
http://seventransistorlabs.com/Images/Discrete_Tube_Supply.png
or use a UC3843 instead of all the discrete junk on the left.

The heater supply isn't necessary, but handy when working with 6V-only types when all you've got is a 12V supply.

As for selection, 6BA6 might be interesting for distortion, but remote-cutoff characteristics are generally unfavorable for non-radio use.  The rest should be interesting.  6AS6 is a dual control pentode, which means it kinda-sorta-maybe acts like a dual gate or cascoded MOSFET.

Something which has no analog in solid state is the screen grid.  In a BJT, when it goes into saturation, hFE drops, collector voltage stays low, Vbe drops a little, and the base starts storing charge (so it takes forever to turn off again).  In a tube, when plate voltage goes into saturation, plate current drops, and screen current goes up by the same amount -- cathode current remains about constant, because it's mainly controlled by grid and screen voltage, not the plate.  But now your screen grid goes into "toaster" mode.   If the screen dissipation doesn't rise too hard, this effect can be used to implement positive feedback with just one tube.

Tim

[IanB]

I prefer to avoid having 100V+ floating around on my bench

I believe in the old days people didn't tend to have such voltages floating around. They built a dedicated power supply and hard wired it to the circuit under development. Then when they operated the on switch the power supply and the circuit were powered up together, and they made sure they weren't poking around inside the circuit with their fingers when it was turned on (or when filter capacitors were charged up).

In short, you can work safely with higher supply voltages if you follow different working practices and do not have flying leads floating around from a stand alone power supply as you might do with lower voltages.

[ampdoctor]

You're working with tubes here so yes you will need to deal with higher voltages. Having said that you also won't need anything too fancy. A fairly simple linear supply will be more than adequate. They don't pull much load current so heavy regulation really isn't necessary. You might even be able to get away with a basic zener regulator. In general don't feel too intimidated by it just follow good safety precautions like religion and you'll be fine.

Tubes are much easier to work with than bjt's or opamps. Here's a few bits of info that may help.

As your first pass at valve audio, you will probably want to work with 12AX/AT/AU7 tubes. Very common and easy to work with.  I'd say to stay away from pentodes or tetrodes until you get comfortable with the particulars of tube circuits. They usually are noisy as hell, and right out of the gate you dont want to fight noise issues while you're just trying to figure out the basics.

Second thing, while they will work with lower voltages as has been mentioned, running them above 100v will give you the best results, particularly if you want good linearity.

Another thing is that these tubes used as preamps tend to prefer higher load resistances, typically on the order of 470k or above. You'll get lower noise with smaller resistances so the trick is to find the sweet spots.

last thing and you may not need to worry about this depending on your supply, is to watch your heater to cathode voltages. This is particularly important if your rails are above 2-300v and you're using cathode followers. If it's marginal you may want to float your filaments above ground a few dozen volts.  Doubt you'll be going there just yet but offering it up for completeness sake.

[calexanian]

last thing and you may not need to worry about this depending on your supply, is to watch your heater to cathode voltages. This is particularly important if your rails are above 2-300v and you're using cathode followers. If it's marginal you may want to float your filaments above ground a few dozen volts.  Doubt you'll be going there just yet but offering it up for completeness sake.

Operating the filaments at a positive voltage en ref to ground will lower the noise from the filament supply. 24 volts is a good round number. Also for just playing around 48 plate volts is enough to get any of the 12A*7 type tubes working. It will be nowhere near optimal, but enough to see them work. High voltages in tube circuits are easily manageable by adhering to good build and bench practice and being safety aware. Never get complaisant in safety. Thats when accidents happen. Building yourself a 180 volt regulated supply would be a good project and suitable for most preamp designs without fear of arcing cathode to filament in the extreme high swing on a cathode follower. I have an old 300 volt regulator that can provide 250ma as my main high voltage supply. Plenty for even small power circuits. Make up some good insulated end jumper cables and actually build circuits that are mechanical sound when dealing with high voltage. I use a plank of wood and copper tacks and solder the components to them. Makes for a cool look and secure mounting so parts are not flopping around. You can use brass too, but the local hardware store carries small copper tacks that are perfect.

[Paul Price]

Instead of biasing the filaments to 20+ volts just power the heaters/filaments with a DC voltage supply. A  7812 or LM317 can be used to create a cheap, easy to build and totally humless power supply for filament/heaters these days.

[ampdoctor]

Instead of biasing the filaments to 20+ volts just power the heaters/filaments with a DC voltage supply. A  7812 or LM317 can be used to create a cheap, easy to build and totally humless power supply for filament/heaters these days.

I've never had any issues with AC filament hum and most of my designs are 4-5 stage 100+dB instrument amplifiers. While DC supplies can be dead quiet, the can also get sideways on you and be a royal pain in the ass if you're not careful with layout and grounding. If you decide to use some current hungry power tubes, now that regulator could be required to supply as much as 4-5 amps or more depending on any number of factors. Not hard to deal with by any stretch but can get quirky for beginners. Putting a bias on the filaments is usually as easy as a 2 resistor divider and an electrolytic hanging off of the B+ rail. I love the K.I.S.S principle!

[calexanian]

Standard linear regulators have difficulty with the surge currents of tube filaments. I have seen many die as a result. Bringing the filament voltage up above cathode potential not only can reduce current induced noise (Hum) but also electrostatic induced noise.

[Paul Price]

AC transformers for heaters/filaments also cause a surge that burns out vacuum tube heaters at turnon. A solid state heater supply with a LM317 can have a slow startup by adding a single capacitor and eliminate this problem. A solid state heater supply requires these days very few components, a LM317, a 220-ohm resistor, a rectifier and filter cap, a small cap, a voltage setting pot or resistor and an electrolytic cap to bypass the setting resistor for slow startup and a small cap at the output.  The metal chassis can be used as a heatsink.

Pre-amps do not require huge tubes that require a very large current just to light them up.

AC transformer wires snaking through very low level input signal areas of your chassis can induce hum into to high impedance tube inputs.

The D.C.  heater power supply can also be floating or even biased.

The filaments/heaters have a very much lower resistance at turn on and this creates the bright flash seen at turn on as the heaters are stressed by the turn on surge.

[Mr. Coffee]

I just got some of these for a simialr project:

http://www.boiaudioworks.com/subminiature

[vk6zgo]

AC transformers for heaters/filaments also cause a surge that burns out vacuum tube heaters at turnon. A solid state heater supply with a LM317 can have a slow startup by adding a single capacitor and eliminate this problem. A solid state heater supply requires these days very few components, a LM317, a 220-ohm resistor, a rectifier and filter cap, a small cap, a voltage setting pot or resistor and an electrolytic cap to bypass the setting resistor for slow startup and a small cap at the output.  The metal chassis can be used as a heatsink.

Pre-amps do not require huge tubes that require a very large current just to light them up.

AC transformer wires snaking through very low level input signal areas of your chassis can induce hum into to high impedance tube inputs.

The D.C.  heater power supply can also be floating or even biased.

The filaments/heaters have a very much lower resistance at turn on and this creates the bright flash seen at turn on as the heaters are stressed by the turn on surge.

That said,tubes were used in the millions over tens of years without any noticeable incidence of heater burnout due to turn on surge.

The most common reason for replacement were loss of cathode emission,& heater to cathode leakage.
(Most tubes used were of the indirectly heated type)
Older tubes which you obtain may be "pulls" removed because of the latter problem.

Heater to cathode leakage can cause hum in the wanted signal,but if you did use a regulated dc heater supply,you may be able to use otherwise unuseable tubes.

Directly heated (filament) tubes had problems with uneven emission due to one end of the filament being at a different dc potential,but this is unlikely to occur with dc operated indirectly heated tubes.

People seem to be incredibly "wussy" when it comes to tube power supply voltages.
100v & above power supplies are not a hazard,providing you use a bit of care.

To add a bit of perspective,vk3ye was making tube Transmitters when he was about 12 years old!

You mainly need good filtering with tube HT supplies---voltage regulation is nice,but not essential

[Riotpack]

I built a valve preamp that required around 200V from memory - I purchased two multitap transformers (http://www.jaycar.com.au/productView.asp?ID=MM2002&form=CAT2&SUBCATID=1002#3) and ran them back to back - You can change the taps on the LV to change the output voltage.

[T3sl4co1l]

People seem to be incredibly "wussy" when it comes to tube power supply voltages.
100v & above power supplies are not a hazard,providing you use a bit of care.

'Oy mate, be a good one and touch'eze wires right quick woodja?

But really, 100s of V are just fine on the solderless breadboard.  The most I've used, I think, is 2kV between non-adjacent positions (if it's that high, I try to space them out a few positions).  No big deal.  I've done 300V or so between adjacent positions, probably more.

I won't touch an energized breadboard in the 200V+ range, but 100V and below isn't all that scary.  Watch where you're going.  Dry fingers brushing against resistor leads will tingle and alert, but nothing crazy.  Assuming you aren't standing barefoot on wet concrete of course.  Standard disclaimers, don't work with wet hands, don't work if you have a history of heart stopping, always best to work on a de-energized circuit, etc. etc...

Tim

[old gregg]

An essential addition to tube power supply (IMO) is a bleeder resistor. It will discharged cap when turn off and it's very handy when desinging circuits. I couldn't do without.  For a final product it may be removed.

I've never done breadboard tube although it seems really nice.

Keep in mind you've to stay really focus when working on live circuits and put your other hand in your pocket ! 

the annoying part of tube circuit is their enormous power consumption, making proper transformers hard to buy (unless you go "custom"). 

I think the all sucess of tube circuit is simply care (and death  )

[nathanpc]

So, I got inspired to build a very simple HV (120V most likely) power supply for my tube project. I'll probably power it from a isolation transformer (fused of course) so I don't have to dick around with switchmode power supply designs. I'll try to post an update as soon as I have something more concrete to show.

Listening to all your tips and suggestions was great. Thanks very much for sharing your experience.

[old gregg]

for a tube power supply lab, take a look here :

http://www.bartola.co.uk/valves

there's some interesting idea for such application. Current limiter and such !