New Amplifier Design - "Murder One" - Submini Pentode, Low Voltage

[Renegadrian]

I think I'll end up with eight stages of voltage increase, so 16 diodes and 16 caps.

you mean 8 and 8...

[frequencycentral]

I almost wonder if you would be better of with a power transformer to supply B+, because the 317 can handle up to 60V in if I remember right.  Seems that if you had a higher input voltage (even without a dedicated power transformer) you could still get your 1.5V with the 317 but reduce the number of components.  I don't know, just seems like if you stack up 16 stages, that would be rather space consuming.  Just a thought.  Maybe you could even make it phantom powered (though I am not sure a phantom supply would provide enough current).  Are you thinking of supplying the 12VDC with a power supply or batteries or what?

16 caps and 16 diodes will take up a lot less space than a PT! I want this to be a 'safe build' anyway - PTs can be a bit intimidating. I did consider phantom, but want it to be self contained. I may end up using a 220 ohm/1 watt voltage drop resistor for each tube's heater instead of the LM317. Not planning for batteries - I'm using a 500ma/12 volt wallwart.

I think I'll end up with eight stages of voltage increase, so 16 diodes and 16 caps.

you mean 8 and 8...

No, 16 and 16 - each voltage increment uses two caps and two diodes.

[Ripthorn]

I guess I didn't read the title description well enough, says "low voltage".  I'm an idiot, so never mind me.

[frequencycentral]

I guess I didn't read the title description well enough, says "low voltage".  I'm an idiot, so never mind me.

Yup, low voltage. That's the whole point of this design for me. I want a 'Ruby' type amp in term of size and volume but entirely tube. I want it to be an easy build for pedal builders (ie, me and Adriano!) who have never dealt with high voltages. It just occured to me a while ago that the MAX1044 and these little tiny pentodes could make that possible.

[svstee]

This is an awesome idea for people like me who are a bit intimidated to try to build a tube amp from the ground up, bt want to try something a step up from a Ruby/Little Gem.
Thanks a lot man!

[DSV]

Nice project!

I've got a few suggestions for the power supply:
- maybe the MAX1044 could be replced by a NE555 as an oscillator, which is lower cost, has a higher voltage rating, and provides much more current (up to 200mA for the non-CMOS versions)
- another method could be as in here: a CD4049 in order to create the alternate voltage, and the caps and diodes for multiplication (like in here http://www.sugardas.lt/~igoramps/grass_preamp.htm )
- Finally, with a simple transformer: http://www.sugardas.lt/~igoramps/tomato_preamp_rus.htm

[frequencycentral]

Nice project!

I've got a few suggestions for the power supply:
- maybe the MAX1044 could be replced by a NE555 as an oscillator, which is lower cost, has a higher voltage rating, and provides much more current (up to 200mA for the non-CMOS versions)
- another method could be as in here: a CD4049 in order to create the alternate voltage, and the caps and diodes for multiplication (like in here http://www.sugardas.lt/~igoramps/grass_preamp.htm )
- Finally, with a simple transformer: http://www.sugardas.lt/~igoramps/tomato_preamp_rus.htm

Thanks!

This amp has a 555 PS, but the designer describes it as 'unsafe': http://www.jjs.at/electronic/class_a_subminiature.html
I don't know how to design a 555 based PS - but if you can, be my guest, I'll give it a go! The maximum voltage needs to be 90 volts.
The 4049 PS also looks interesting, i'll do some research there. i guess i just used ideas i was sure of and understood. I'm open to suggestions though.

[Zben3129]

You could try a 120v to 24v transformer to power the amp. If you wire it in reverse, then a 12vac input will give you about 60v on the secondary. You could power the amp with a 12vac wall wart and avoid having any voltages higher than 60v in the actual amp, while using a transformer supply, which I think is a better option than the MAX.

Zach

[Zben3129]

NOTE ^

The 60v you get out of the transformer would be AC, not DC, which, after rectification, would actually give to higher than 60vDC (~ 83vDC). You may want the added voltage, but if you don't you could choose a different transformer.

Determine the HT you want, and then add 5 volts to that to get your target DC voltage (HT + 5v = V_t)

Once you have target voltage, divide it by 1.414 to get your secondary AC voltage you need to be supplied by the transformer (V_sec)

Once you have your secondary AC voltage, use this ratio to determine the transformer you should use.  (120v / 'x') = (V_sec / 12v)  where x is the desired value of the transformer secondary

For example, if you wanted an 78v B+, your V_t would be 83v. This divided by 1.414 would be 60v. Plugged into the equation this would read (120v / 'x') = (60v / 12v)   which when solved would show a 24v secondary. So you would use a 120v to 24v transformer (120v is always the primary).


Zach

[frequencycentral]

I thought I'd update this thread with where I'm at right now. Here's the current configuration:



I've tried out loads of ideas, I'm pretty pleased with this one. I've decided to use a 6111 submini dual triode for the preamp, there is much more gain available than using a 5672 as a preamp. It sound nice and bright (much brighter and louder than another experiment I did which used two 6111, one as a preamp and one as a power amp). The clean sound is not as loud as I would like, but hey, I'm a dirty guy anyways. This design was one inspiration for my amp project - the designer also states that the clean sound is pretty quiet, so maybe I'm striving for something unobtanium anyway.

Each tube has its own voltage dop resistor, these are both pretty bulky but unavoidable.

It would appear that although the maximum voltage for a 5672 is 90 volts it's not possible to apply a 90 volt plate voltage via a transformer (though ok via a resistor). Here's what Jasper Oosthoek over at the AX84 forum has to say about that: "If 90 volts is the maximum for the tubes, you cannot use it with a transformer (resistor is not a problem). As soon as you get a signal in the transformer its primary AC voltage will swing around B+. This means that if the AC voltage is 30 volts peak to peak, the anode voltage of the tube will swing between 75 and 105 volts. 60 volts on the anode might be a lot safer and will enable you to have 60 volts AC P2P (30 to 90 volts)". So I'm sticking to about 60 volts. I've also added a couple of 1n4148 to limit the DC into the MAX1044 to about 10.6 volts.

For the voltage doubling stages I'm using caps rated above the voltage available at the relevant stage, this starts to become a problem over 60 volts, as 50 volt and 63 volt rated caps are small, but the 10uf 100 volt caps I ordered are huge in comparison. Maybe I'll be able to get some a little more compact.

I'm pretty sure that I'm going to go ahead and build this design as it stands, as it seems to be pretty well optimised within the design constraints.  My next amp will be a push/pull using two 5672 for the power amp running into a Hammond 125A, and a 6111 for the preamp/phase inverter. I'll again use the MAX1044 power supply at around 60 volts. I have to source a 125A in the UK before I can start on that one. I'm hoping for a little more volume so the clean tones are more meaningful - and it shall be named "Sub Atomic"!

[slacker]

This amp has a 555 PS, but the designer describes it as 'unsafe'

I think DSV was talking about using the 555 as a charge pump, not in a switchmode powersupply. You can use a 555 in a similar way to the MAX1044 to multiply or invert voltages. There's an example here.

The 4049 PS also looks interesting, i'll do some research there. i guess i just used ideas i was sure of and understood. I'm open to suggestions though.

I've tried this and it works but I couldn't get much current out of it. Using it to triple 15 volts I could only get a couple of milliamps, this is probably fine for preamp tubes with large plate resistors but I don't think it will be enough current to drive the output transformer.

Out of interest how much current does your amp draw from the B+.

[Br4d13y]

mwahahaha now it is time to take all of ricks designs and make one tube multi-effect amp with subminis

[frequencycentral]

This amp has a 555 PS, but the designer describes it as 'unsafe'

I think DSV was talking about using the 555 as a charge pump, not in a switchmode powersupply. You can use a 555 in a similar way to the MAX1044 to multiply or invert voltages. There's an example here.

The 4049 PS also looks interesting, i'll do some research there. i guess i just used ideas i was sure of and understood. I'm open to suggestions though.

I've tried this and it works but I couldn't get much current out of it. Using it to triple 15 volts I could only get a couple of milliamps, this is probably fine for preamp tubes with large plate resistors but I don't think it will be enough current to drive the output transformer.

Out of interest how much current does your amp draw from the B+.

Ian, that 555 circuit you linked to - could you tell me some more about it? What would be the voltage output, and is it just a case of adding extra stages to get more volts like to MAX? Could the zeners be replaced with something a little more common? 1n4148? I'll breadboard it, but it seem to me the parts count is going to be similar to the MAX - yeah a 555 is cheaper than a MAX though. I guess I like the MAX as its really easy to understand whats going on.

I'm not sure how much current the amp is drawing from the B+, not sure how to measure that, and my DMM is pretty basic.

[slacker]

No too sure about that circuit Rick it was just something google turned up. I simulated it using 1N4148 diodes and it doubles the voltage, I don't know if you can add extra stages.

I think the MAX 1044 or a similar dedicated chip is probably the way to go, I think I'm going to get an LT1054 to play about with.

You don't need anything fancy to measure the current draw, just put a 1 ohm resistor between the last diode from the MAX1044 and B+ and put a big cap from B+ to ground to stop the amp oscillating. Then measure the voltage drop across the 1ohm resistor and that's your current draw in milli amps.
If you haven't got any 1ohm resistors just use the smallest you've got and then use Ohms law to work out the current draw V/R = I

[svstee]

Some questions for you:
1: What is the cheapest one can find a transformer for this? Are there any other transformers that work that cost less?
2: How many watss do the 220k resistors need to be?
3: Where is the best place to get a MAX1044? I saw a couple at Futurelec, but they are so slow...
4: Can you just run this at 9v and get rid of the diodes before the MAX 044?

[kurtlives]

1. The transformer is cheap to start. Weber Speakers in the US sells them quite low.
2. You ask what the "watts" need to be a lot I see. Learn the formula.
    W = V^2/R
    Where V squared is the voltage dropped across the resistor. R is the resistance in ohms of the resistor.
   
    There are other ways to manipulate the formula but that's how I use it.
3. Most online stores sell that charge pump. Mouser, Digikey, Small Bear etc...
4. You could. Probably sound like crap though. Would be much darker and compressed from my experiences.

[frequencycentral]

I'm using 1/4 watt resistors throughout except for those specified.

Running on 9 volts will be fine, and yes those two diodes can be removed. I have to disagree with kurtlives here, it won't sound like crap, it will just be quieter, as the B+ output of the MAX will be lower. More volts = more volume. So, get the B+ back up by adding another stage of voltage increase with the MAX and it will be exactly the same.

You would also have to change the heater voltage drop resistors for 9 volts, so:

6111: 9 ohm / 1 watt (though 10 ohm would still be within tolerance)

5672: 155 ohm / 1/2 watt

[kurtlives]

Rick that was just my personal opinion and I know you understand that. I find starved plate designs really open up at 25V or so.

Why do you use resistors for getting the filament voltage where you want it. I mean sure its great if you have a constant B+. But wouldn't a regulator make more sense, then if your B+ changes your filament voltage stages constant. That is if they make regulators for the filament voltage your trying to achieve.

[frequencycentral]

Rick that was just my personal opinion and I know you understand that. I find starved plate designs really open up at 25V or so.

Why do you use resistors for getting the filament voltage where you want it. I mean sure its great if you have a constant B+. But wouldn't a regulator make more sense, then if your B+ changes your filament voltage stages constant. That is if they make regulators for the filament voltage your trying to achieve.

Chris, I think you missed something here, the output of the MAX and therefore the B+ is 60 volts, so it's not a starve plate design at all.

As for the drop resistors, just keeping it simple. And assuming a nicely regulated power supply is used. You could use a 7806 for the 6111 and a LM317 for the 5672 I guess.

[kurtlives]

I am quite aware that the B+ will be 60V with the charge pump.

At 9V like was talked about though the amp would have "starved plates" and not be as good as it could be imo.