How to work out heater resistor value? (And a few other submini amp questions)

[Scruffie]

Right, I am entirely tube ignorant, I have no intention of learning proper amp design as this is just to use some parts up for a project I never did before I throw in the DIY towel but I've got 4 6088 tubes I wanna use to make an incredibly quiet amp (10.5mW output power!) but I don't know how to work out the heater resistors for them.

I'm gunna be using a 9V supply and the requirements are 1.25V @ 0.02A

And on a simmilar note, the control grid voltage is -1.25V, as i'm gunna use a MAX1044 charge pump, should I use this to get that on the grid? Don't really know much about tubes, just gunna be going with standard building blocks and hoping for some kind of output (I have had these sounding before, had a 1.5V supply so I used that and just put a couple of boosters in front of one just to test them about as loud as the unplugged guitar through a 6" speaker but good enough!).

As the output power is so low, can I tie two together on the output for more? I assume that's how you end up with 100W Amps etc.

I'm even gunna try and use it through this transformer for shits & giggles as I have one around http://www.radioshack.com/product/index.jsp?productId=2103254
Entirely wrong primary impeadance (1k: but... why not. I expect little from this (and this is what I used the last time) not sure where to even get a 80k:8ohm transformer anyway, finally I have a 70ohm vintage speaker, would this produce more or less output with the incorrect matching, seem to remember reading something about this... guessing less.

[frequencycentral]

You might as well use an LM317 for 1.25v. 9V in, ground the ground, 1.25V will be at the output once it sees a load. No other doodads needed.

Yeah you can parallel more tubes for more output.

[Scruffie]

You might as well use an LM317 for 1.25v. 9V in, ground the ground, 1.25V will be at the output once it sees a load. No other doodads needed.

Yeah you can parallel more tubes for more output.

That involves buying one, would rather do it with what I have if possible.

And awesome, cheers.

[R.G.]

Right, I am entirely tube ignorant, I have no intention of learning proper amp design as this is just to use some parts up for a project I never did before I throw in the DIY towel but I've got 4 6088 tubes I wanna use to make an incredibly quiet amp (10.5mW output power!) but I don't know how to work out the heater resistors for them.

I'm gunna be using a 9V supply and the requirements are 1.25V @ 0.02A

Ohm's law to the rescue. The filaments use 20ma, and have 1.25V across them when they have 20ma through them. You plan on using 9Vdc, so the dropping resistor will have 9 - 1.25 = 7.75V across it. If your power supply is really 9.00V, not 9.4V, or whatever. So you need a resistor with 7.75V and 20ma, or R = 7.75/0.02 = 387.5 ohms. 390 ought to be fine. The resistor will then dissipate 7.75V times the 0.02A = 0.154W, so a 1/4W resistor will be OK, probably. That's per tube, obviously, if you use more than one.

And on a simmilar note, the control grid voltage is -1.25V, as i'm gunna use a MAX1044 charge pump, should I use this to get that on the grid? Don't really know much about tubes, just gunna be going with standard building blocks and hoping for some kind of output (I have had these sounding before, had a 1.5V supply so I used that and just put a couple of boosters in front of one just to test them about as loud as the unplugged guitar through a 6" speaker but good enough!).

I would insert a resistor between the cathode and ground. The resistor would drop 1.25V at the specified 650uA of plate current, or (back to you, Mr. Ohm) 1.25V/650E-6 = 1930 ohms. A 2K would work wonders. Then ground the grid through a 1M or so grid leak resistor. Again, per tube, except that all paralleled tubes can use the same grid leak. Bypass the 2K cathode resistor with a 10uF-22uF cap for more gain.

As the output power is so low, can I tie two together on the output for more?

Yes

I assume that's how you end up with 100W Amps etc.

No. You'd need 10,000 6088's to get to 100W. In this case it's only practical to get a bigger horse - that is, use a bigger tube like a 6L6. In a philosophical way, if you consider that the area of a 6L6's plate is equal to many, many times the area of the 6088's cathode/plate/etc, then a 6L6 is like thousands of 6088's in parallel. But no, it's not practical to keep adding 10mW at a time to get many watts.

I'm even gunna try and use it through this transformer for sh*ts & giggles as I have one around http://www.radioshack.com/product/index.jsp?productId=2103254
Entirely wrong primary impeadance (1k: but... why not. I expect little from this (and this is what I used the last time) not sure where to even get a 80k:8ohm transformer anyway, finally I have a 70ohm vintage speaker, would this produce more or less output with the incorrect matching, seem to remember reading something about this... guessing less.

If you don't expect much, you are likely to be happy with the results.

[Kesh]

as the filament's resistance is hugely temp dependent, i would get a 1k pot in series with the heater, set it to full resistance, measure the voltage across the heater, and turn it down SLOWLY until you get a warm tube and 1.25 volts. this caution may not be necessary, but it cannot hurt.

[R.G.]

as the filament's resistance is hugely temp dependent, i would get a 1k pot in series with the heater, set it to full resistance, measure the voltage across the heater, and turn it down SLOWLY until you get a warm tube and 1.25 volts. this caution may not be necessary, but it cannot hurt.

I'm just guessing, but I suspect that if you do this carefully, you'll get to a resistance on the pot of close to 387.5 ohms on a 9.00V source.

[Scruffie]

Right, I am entirely tube ignorant, I have no intention of learning proper amp design as this is just to use some parts up for a project I never did before I throw in the DIY towel but I've got 4 6088 tubes I wanna use to make an incredibly quiet amp (10.5mW output power!) but I don't know how to work out the heater resistors for them.

I'm gunna be using a 9V supply and the requirements are 1.25V @ 0.02A

Ohm's law to the rescue. The filaments use 20ma, and have 1.25V across them when they have 20ma through them. You plan on using 9Vdc, so the dropping resistor will have 9 - 1.25 = 7.75V across it. If your power supply is really 9.00V, not 9.4V, or whatever. So you need a resistor with 7.75V and 20ma, or R = 7.75/0.02 = 387.5 ohms. 390 ought to be fine. The resistor will then dissipate 7.75V times the 0.02A = 0.154W, so a 1/4W resistor will be OK, probably. That's per tube, obviously, if you use more than one.

Perfect explanation, thanks for even doing the maths too!
They're +/-20% Tollerance so i'm sure the .6V or so extra'll be fine.

And on a simmilar note, the control grid voltage is -1.25V, as i'm gunna use a MAX1044 charge pump, should I use this to get that on the grid? Don't really know much about tubes, just gunna be going with standard building blocks and hoping for some kind of output (I have had these sounding before, had a 1.5V supply so I used that and just put a couple of boosters in front of one just to test them about as loud as the unplugged guitar through a 6" speaker but good enough!).

I would insert a resistor between the cathode and ground. The resistor would drop 1.25V at the specified 650uA of plate current, or (back to you, Mr. Ohm) 1.25V/650E-6 = 1930 ohms. A 2K would work wonders. Then ground the grid through a 1M or so grid leak resistor. Again, per tube, except that all paralleled tubes can use the same grid leak. Bypass the 2K cathode resistor with a 10uF-22uF cap for more gain.

There is negative voltage available from the charge pump though, would there be a benefit to using it?
There's only a control grid & screen grid so I assume when you say cathode you mean the control... right?

As the output power is so low, can I tie two together on the output for more?

Yes

I assume that's how you end up with 100W Amps etc.

No. You'd need 10,000 6088's to get to 100W. In this case it's only practical to get a bigger horse - that is, use a bigger tube like a 6L6. In a philosophical way, if you consider that the area of a 6L6's plate is equal to many, many times the area of the 6088's cathode/plate/etc, then a 6L6 is like thousands of 6088's in parallel. But no, it's not practical to keep adding 10mW at a time to get many watts.

Yeah... I wasn't expecting to pull 100W out of 10.5mW tubes, I just meant getting 21mW or 31.5mW might be a good idea. I imagine 100W would be a little microphonic and noisy 

I only have 4 of the things available, so if I need to add a solid state preamp, no harm done.

I'm even gunna try and use it through this transformer for sh*ts & giggles as I have one around http://www.radioshack.com/product/index.jsp?productId=2103254
Entirely wrong primary impeadance (1k: but... why not. I expect little from this (and this is what I used the last time) not sure where to even get a 80k:8ohm transformer anyway, finally I have a 70ohm vintage speaker, would this produce more or less output with the incorrect matching, seem to remember reading something about this... guessing less.

If you don't expect much, you are likely to be happy with the results.

All I want is some output, if it's loud enough to just be a little night time amp that's not much louder than the unplugged guitar but I can run a pedal or two through it that's fine by me 

Thanks for sorting that all out  should be all I need to get something together (other than that cathode question).

[Kesh]

as the filament's resistance is hugely temp dependent, i would get a 1k pot in series with the heater, set it to full resistance, measure the voltage across the heater, and turn it down SLOWLY until you get a warm tube and 1.25 volts. this caution may not be necessary, but it cannot hurt.

I'm just guessing, but I suspect that if you do this carefully, you'll get to a resistance on the pot of close to 387.5 ohms on a 9.00V source.

Yes, if that 20mA is the actual current draw, and not some nominal value. Which is why I'd start at less than half and that and see. And you have to actually power it and not just measure the ohms when cool.

EG they rate it at 20mA because its resistance is 62.5 ohms cool (0.02 x 62.5 = 1.25), but hot its resistance triples to say 200. With a 387.5 resistor you are now giving over 3 volts.

But, like I say, who knows if that 20mA is operating or peak or what.

[R.G.]

Tube makers in general rated their filaments at nominal filament/heater voltage with a +/- 10% tolerance and the current was stated as the average of many tubes' current drain with the nominal voltage across them. That is, their specification presumed that the filaments were already hot, and when at fully settled, final hot temperature with the nominal filament voltage, would pull the specified current.

That's not to say that tubes don't/didn't vary, but tolerances were in general much tighter than with semiconductors.

Running tubes at 5% - 10% low heater voltage was a dodge used in some large systems with many tubes to lengthen the average tube life. However, the tubes were predictable enough that this worked. There are a couple of articles I've seen on this in doing tube-archeology.

The real, no-fooling, techno-geek way to do this is to use an optical pyrometer to optically measure the filament temperatures, which is what is actually being set with this. Oxide coated cathodes (in all the normal tubes we can usually get) have a specified temperature they operate at, and that's what the tube makers were trying to hit. Even then, they recognized the impossibility of having technicians have optical pyrometers, so they specified things indirectly. They designed the tubes so that when the voltage on the filaments was right, they had close expectations of the correct temperature. They then specified that voltage and the measured current at that voltage. They figured every tech could measure voltage and current, and it was the tube designer's job to get the filaments to the right temperature when the voltage and current were right.

Even today's tubes - which 6088's are not, most likely - take this approach.

As I said, you certainly can use a 1K pot and gingerly lower the resistance til you get the right current, but I'd bet that when you're done, the final voltage and current will be very close to 1.25V and 20ma, and that the resistance will be really close to 387.5 ohms. 

[Kesh]

That is, their specification presumed that the filaments were already hot, and when at fully settled, final hot temperature with the nominal filament voltage, would pull the specified current.

well then everything is all right with the world

[R.G.]

It's unusual of course, but sometimes that happens.

[PRR]

> the filament's resistance is hugely temp dependent

But we only care about the HOT resistance. That's the only way the tube will work. Yes, the filament resistance is lower cold. A large series resistor will slow-start the heating process. But the "proper" (1.5V battery) heat-up is under a second, the "slow" heat may be 3 seconds, no problem.

> insert a resistor between the cathode and ground.

6088 is a directly heated filament cathode, _not_ the "uni-potential indirectly heated" cathode of say 12AX7.

You need to treat these right.

The filament has + and - markings. Connect filament + to the + of a 1.5V battery. Filament - to battery negative, circuit common, and _grid_ return resistor. The 'cathode' has from +1.5V to zero V along its length, average about -0.7V. If grid is returned to filament negative it has -0.7V bias. The tube is designed to work this way.

> use an optical pyrometer to optically measure the filament temperatures

Well, yeah (or filamment resistance shift), in development.... but these are _hearing_aid_ tubes. The user put in a dry-cell and used it until the sound was weak. The spec-sheet shows +/-20%, but really means 1.5V new to 1.0V weak.

> can I tie two together on the output for more?

With 45V supply, use 80K load for one, 40K load for two, 20K load for four.

"Fender reverb transformer" will be close-enuff to 20K and plenty big to handle the whopping 0.042 Watts.

> use it through this transformer... Entirely wrong primary impeadance (1k

With one 6088 tube, that load gives 0.000,2 Watts (0.2mW); with four, 0.000,8 Watts.

A small 120V:3V power transformer would be a MUCH better OT.

Even with four tubes, and a pretty good (large!) guitar speaker, it will be whisper-soft. (It's for a hearing-aid... made for a "speaker" directly AT your ear.)

The power consumption of this tube is SO SO SO low, that you really should use one C-cell and five 9V batts rather than fool with charge-pumps and dropping resistors. Even with four 6088, the C-cell will keep the filaments warm for 80 hours, a LOT of playing. Five 9V will last over 100 hours. One tube, four times longer... unless you get obsessed with this sound, the batts will die of old-age not run-down. The voltages are known-good for-sure, the supply has no hum or whine. And it is portable, which may be innerestin. Get it working on batts, gig-out every night for a month or more, _then_ decide you "need" a more permanent supply.

[Scruffie]

> the filament's resistance is hugely temp dependent

But we only care about the HOT resistance. That's the only way the tube will work. Yes, the filament resistance is lower cold. A large series resistor will slow-start the heating process. But the "proper" (1.5V battery) heat-up is under a second, the "slow" heat may be 3 seconds, no problem.

> insert a resistor between the cathode and ground.

6088 is a directly heated filament cathode, _not_ the "uni-potential indirectly heated" cathode of say 12AX7.

You need to treat these right.

The filament has + and - markings. Connect filament + to the + of a 1.5V battery. Filament - to battery negative, circuit common, and _grid_ return resistor. The 'cathode' has from +1.5V to zero V along its length, average about -0.7V. If grid is returned to filament negative it has -0.7V bias. The tube is designed to work this way.

> use an optical pyrometer to optically measure the filament temperatures

Well, yeah (or filamment resistance shift), in development.... but these are _hearing_aid_ tubes. The user put in a dry-cell and used it until the sound was weak. The spec-sheet shows +/-20%, but really means 1.5V new to 1.0V weak.

> can I tie two together on the output for more?

With 45V supply, use 80K load for one, 40K load for two, 20K load for four.

"Fender reverb transformer" will be close-enuff to 20K and plenty big to handle the whopping 0.042 Watts.

> use it through this transformer... Entirely wrong primary impeadance (1k

With one 6088 tube, that load gives 0.000,2 Watts (0.2mW); with four, 0.000,8 Watts.

A small 120V:3V power transformer would be a MUCH better OT.

Even with four tubes, and a pretty good (large!) guitar speaker, it will be whisper-soft. (It's for a hearing-aid... made for a "speaker" directly AT your ear.)

The power consumption of this tube is SO SO SO low, that you really should use one C-cell and five 9V batts rather than fool with charge-pumps and dropping resistors. Even with four 6088, the C-cell will keep the filaments warm for 80 hours, a LOT of playing. Five 9V will last over 100 hours. One tube, four times longer... unless you get obsessed with this sound, the batts will die of old-age not run-down. The voltages are known-good for-sure, the supply has no hum or whine. And it is portable, which may be innerestin. Get it working on batts, gig-out every night for a month or more, _then_ decide you "need" a more permanent supply.

Okay... now i'm a little lost, as I said, I have no intention of learning all about tubes (although I don't expect it all done for me to build this, I just wanna get a simple amp down for fun) this is one of my last projects and i'd like some kind of quietish mini tube amp at the end.

How do I need to treat the 6088s right regards R.Gs advice?

I'm not gunna use batteries (other than either a 9V or 6xAA) it works off a charge pump or I sell them, i'd rather have something from them but i'm not going any further than a 9V supply working here, it stops being a 'toy' when I need to treat it specially and I got some output with a chargepump before.

Interesting regards the load though, I will consider buying one of those line transformers is it (the cheapy ones)? to use with this, as I said, I did get some output from one tube in to that tiny transformer with a transistor booster.

I expect little from this amp, at my ear would be okay, but as I said, if I can justttt about get enough power to be above unplugged electric, i'll be quite happy, this is for fun and late night playing with pedals, not anything special or i'd build another 386 amp.

[Kesh]

> the filament's resistance is hugely temp dependent

But we only care about the HOT resistance. That's the only way the tube will work. Yes, the filament resistance is lower cold. A large series resistor will slow-start the heating process. But the "proper" (1.5V battery) heat-up is under a second, the "slow" heat may be 3 seconds, no problem.

yeah, i know now. before i was wondering if the 20mA quoted was the peak, cold start, cold resistance current requirement.

Scruffie, you don't happen to have any IR LEDs? some of those have around 1.25v forward voltage.

[Scruffie]

> the filament's resistance is hugely temp dependent

But we only care about the HOT resistance. That's the only way the tube will work. Yes, the filament resistance is lower cold. A large series resistor will slow-start the heating process. But the "proper" (1.5V battery) heat-up is under a second, the "slow" heat may be 3 seconds, no problem.

yeah, i know now. before i was wondering if the 20mA quoted was the peak, cold start, cold resistance current requirement.

Scruffie, you don't happen to have any IR LEDs? some of those have around 1.25v forward voltage.

Nah, fraid not.

If you're really worried about me burning them (i'd probably just loose one anyway, as I wont do them all at once) I can always try with my adjustable supply and take it up in incriments to check 390R is enough.

[Kesh]

Not worried at all really, just something inside me rebels at regulating voltage with a single resistor.

[R.G.]

Not worried at all really, just something inside me rebels at regulating voltage with a single resistor.

Me too. Fortunately, regulation of a voltage is not what is needed.

If you think about it, we're saying much the same thing - decide what resistance you need in series with a filament to give it the voltage and current it needs. I'm just noting that the voltage and current relationship of a heated filament at thermal equilibrium is a mostly-fixed curve. Limit either voltage or current, and the other changes to match, based on the fact that the temperature of the filament changes the apparent resistance to let it settle to a fixed temperature.

I'm just observing that one doesn't need to know the curve or variation in filament resistance per power/voltage/current to do that. There's only one spot to match, and if the tube doesn't work at the point where it has the specified voltage OR current, then it's defective. That being the case, you can set either the voltage or current, the other will follow, within normal bounds.

[Scruffie]

That discussions all a bit over my head but i'll be using the 390R.

Having a look at Rick (frequency centrals) murder one schematic i'm gunna just assume I can wire it up much the same but with the new filament resistor value, comparing the 6088 & 5672 datasheets they seem much the same except for the 5672s supressor being tied to the filaments and having a few different ratings.

So i'll tie two together for the power amp (power used very loosely here) and then use another two of them wired up as triodes (as I have no idea what values I should use as a pentode, to be fair I don't know what values to use for anything except caps and the heaters, but sure it'll work) for the preamp, seems simple enough.

Murder One Schematic - http://i210.photobucket.com/albums/bb292/frequencycentral/MurderOneV6.jpg

6088 Datasheet if it's any use - http://frank.pocnet.net/sheets/138/6/6088.pdf

[iccaros]

The Resistor will work fine, these are tubes not scientific devices for medical purposes, while a resistor will change with heat but so will the heaters, in practice it does not really matter, these filaments don't get that hot, and they keep stable, hell it has not mattered on 12K5's or el84, which pull a lot more current and very in temperature much more,
What are you using for a preamp tube?
I am a fan of these tubes a preamps,
they are microphonic, so look at a 1 meg input resistance, that seams to quite them down.

[Scruffie]

The Resistor will work fine, these are tubes not scientific devices for medical purposes, while a resistor will change with heat but so will the heaters, in practice it does not really matter, these filaments don't get that hot, and they keep stable, hell it has not mattered on 12K5's or el84, which pull a lot more current and very in temperature much more,
What are you using for a preamp tube?
I am a fan of these tubes a preamps,
they are microphonic, so look at a 1 meg input resistance, that seams to quite them down.

I posted in the above post, 2 of them for output, 2 for preamp 

Yeah i've read about them being microphonic and the 1M resistance.