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

[frequencycentral]

Once in a while someone comes along and says 'hey, try it this way'. That just happened, a Murder One builder Pm'ed me today, and said he'd read in the 5672 datasheet that grid 1 should be biased at -6.5 volts. He had used an extra 555 based circuit to get the -6.5 volts, and reported that it really improved the tone and clean headroom. So I had to try it!

Well I thought it should be possible to derive the bias voltage from the MAX without adding extra circuitry, so here's my quick and dirty solution:

1. I added a 10uf cap between pins 2 and 4 of the MAX1044 (+ve end at pin 2).
2. I connected pin 5 of the MAX1044 to pin 4 (grid 1) of the 5672, via a 100K resistor.
3. I added a 0.1uf cap between lug 2 of the volume control and pin 4 of the 5672.

How simple is that? The output of the MAX at pin 5 (with the cap now added between pins 2 and 4) is -5.9 volts, that's close enough for jazz. There is now a lot more control with the Gain and Volume knobs, there's a nice chimey clean sound too, but push everything up full and it still honks! Cool! 

Thanks to Aldas for this great mod!

[frequencycentral]

So here's the latest rev, with the biasing mod:

[doitle]

I'm starting to think that I might be able to build the Murder One and am in the planning stages. I have no idea how I will actually assemble it as I just plain old cannot do perf board. :/ I'm terrible. I was thinking I might be able to figure out some tag strip based construction... Not entirely sure about that. Anyhow I was looking at transformers and that Fender Reverb transformer is about 13$ on AES. They've also got a 8W SE Output Transformer for 13$. It's only 5K Ohm primary instead of the 25K of the reverb driver. I'm not sure what the output impedance is of the pentode power stage but I was wondering if that TF might be a better fit.

Another question I had was why you stopped at 80V for your B+? In simulation at least I managed to get a LT1054 pushing 12V up over 100V. Since the 6111 are rated at 150V it should be fine running even higher than you have it but I was wondering why you stopped where you did.

Also lastly what would your suggestions be for a speaker to drive? It's 8 ohm output but considering the power handling is below 1W I am wondering just how small of a speaker would be a good fit for it.

[frequencycentral]

I'm starting to think that I might be able to build the Murder One and am in the planning stages. I have no idea how I will actually assemble it as I just plain old cannot do perf board. :/ I'm terrible. I was thinking I might be able to figure out some tag strip based construction... Not entirely sure about that. Anyhow I was looking at transformers and that Fender Reverb transformer is about 13$ on AES. They've also got a 8W SE Output Transformer for 13$. It's only 5K Ohm primary instead of the 25K of the reverb driver. I'm not sure what the output impedance is of the pentode power stage but I was wondering if that TF might be a better fit.

Another question I had was why you stopped at 80V for your B+? In simulation at least I managed to get a LT1054 pushing 12V up over 100V. Since the 6111 are rated at 150V it should be fine running even higher than you have it but I was wondering why you stopped where you did.

Also lastly what would your suggestions be for a speaker to drive? It's 8 ohm output but considering the power handling is below 1W I am wondering just how small of a speaker would be a good fit for it.

I always plan my complete build from the ground up befroe I start work. With this project there are a few odd shaped parts, so you really need to be sure of how it's all going to fit together before you start your build. I'm not so sure tag strip would be the way to go - maybe you should work up a vero layout? Or get a PCB from John Lyons, it's pretty easy to add the grid 1 bias mod to the PCB (though not ideal), I'll be doing a new PCB soon, though I have an idea that I should look at biasing the 6111 cathodes first too.

The Fender reverb transformer is a good match to the 5672 pentode because the 5672's load resistance is 20K and the transformers is rated at 22.5K into an 8 ohm speaker. So you'd have an impedance mismatch were you to choose a transformer with a 5K primary.

I stopped at ~80 volts for the B+ as the maximum rating for the 5672 plate voltage is 90 volts. I've got some other submini pentodes with higher maximum plate voltages (5902, 6397, 5840), but it seems to me that any higher a plate voltage and you should be using a SMPS instead of a charge pump - the charge pump in Murder One uses 12 caps and 12 diodes but is still quite compact, to double the plate voltage you'd be doubling the number of caps and diodes, whereas a SMPS does the same job with many fewer components. The original idea behind Murder One was to see if I could build a tube amp using a charge pump for the B+. I feel that the Murder One power supply is totally safe, as opposed to the SMPS I'm using with my Sub Zero design, which can give you quite a belt if you touch the high voltage parts. Ouch.

I use my Murder One(s) through the 12" 8 Ohm speaker in my Roland Bolt 60 tube combo. I've tried it through 4x12's (I don't own one) and it sounds great too - the better the speaker the better the sound. Through cheap tinny speaker it sounds like crap. I'm planning to build a dedicated speaker cab at some point, maybe 2x12 and get some real nice speaker to put in it.

[doitle]

Well I feel a bit stupid about my voltage question now... I didn't think that the submini pentodes max voltage would be that low.

I may just break down and buy a PCB premade from someone as I dont know how much more confidence I would have in myself working with vero over perfboard.

Also I think I will look into SMPS in the future as it is a neat little concept.

I didn't realize as well that the muder one had enough oomph to power a large speaker cabinet like you've mentioned. I thought it was only .065W... At least I thought I read that somewhere in this thread maybe not. Regardless it actually sounds like a good match for me as I like to play the guitar quiet and even my dad's 5W Vibrachamp is too much for me often.

[Scruffie]

I might do a vero layout soon if that'd help although it'll probably need a bit of correcting by someone...

[Ripthorn]

I didn't realize as well that the muder one had enough oomph to power a large speaker cabinet like you've mentioned. I thought it was only .065W... At least I thought I read that somewhere in this thread maybe not. Regardless it actually sounds like a good match for me as I like to play the guitar quiet and even my dad's 5W Vibrachamp is too much for me often.

It is only .065W, but how power translates into dB is logarithmic, so .065W has a max volume 30 dB quieter than a 65W amp.  That is very significant, but if your 65W amp cranked maxes at, say, 110 dB, then your .065W amp maxes at 80 dB, which is approximately how loud a typical smoke alarm can beep (or about as loud as someone yelling).  That is why I have a 2W tube amp, about 17dB quieter than a 100W amp (of course all these figures assume into the same speaker, max efficiency of power actually put out by the speaker, etc.).  Hope that helps.

[doitle]

I have been trying to come up with a plan for my own submini amplifier similar to the murder one but also I've been trying to look at some other production amplifiers to figure out some of the aspects of design and I cannot seem to find that much in common with say a simple tube amp like an old Fender Champ and the Murder One. This is what I came up with so far and I'm sure it's mostly nonsense but hopefully in time I will learn to turn it from nonsense into something workable. One of the biggest differences I notice is that the Murder One doesn't have cathode resistors for biasing. Are you using the tubes in a zero-bias fashion? I saw in the 6112 (Similar to 6111) datasheet that it could be operated in both fashions. I have mine "configured" for cathode bias right now but the values are largely pointless as I haven't yet figured out the optimal bias point to shoot for and so do not know how to calculate the resistances. Another question is the hookup of your power pentode. It doesn't appear in the schematic to be at all similar to either a full size power pentode like the EL84 or the 6205 I was planning to use.



As for the B+ voltage I was planning to use a dual transformer system a-la the Real Mc Tube to derive a B+ of around 140 Volts. Both tubes I was thinking of using are 165V max tubes so I should be within a safe range and should have plenty of clean headroom (I think...). The two tubes planned are the 6112 and the 6205.

I'm sorry if these are stupid questions, I'm trying to learn what I can of tube design but it is slow going. When I ask my professors if they remember anything or mention that I am attempting to learn about them they almost seem to get mad at me. I have found one though who did have some interest (In making Hi-Fi Amp not Guitar Amp) but he is in Poland on vacation still.

[frequencycentral]

One of the biggest differences I notice is that the Murder One doesn't have cathode resistors for biasing. Are you using the tubes in a zero-bias fashion? I saw in the 6112 (Similar to 6111) datasheet that it could be operated in both fashions. I have mine "configured" for cathode bias right now but the values are largely pointless as I haven't yet figured out the optimal bias point to shoot for and so do not know how to calculate the resistances.

Yup, no cathode resistors. I was going for maximum gain, and found that this occured without cathode resisitors. I'm revisiting this design at the moment, and am planning to try fixed biasing of the 6111 grids at -9 volts (as per the data sheet), derived from pin 5 of the MAX. I'm still learning, and should in no way be considered an expert! I'm starting to understand biasing a lot better now, and I think fixed biasing as opposed to cathode biasing is the was to go. See this discussion thread regarding grid biasing:  http://www.diystompboxes.com/smfforum/index.php?topic=77741.0

Another question is the hookup of your power pentode. It doesn't appear in the schematic to be at all similar to either a full size power pentode like the EL84 or the 6205 I was planning to use.

The 5672 is not a standard pentode like the 6205. It is 'directly heated', which means that the cathode is internally connected to the heater, and is therefore fixed at the heater voltage which is 1.25 volts. I have some 5840 which (and some 6112), which are similar to 6205 except for the internal connection of the suppressor grid, so I'll be interested to see how your experiments pan out.

I'm sorry if these are stupid questions, I'm trying to learn what I can of tube design but it is slow going.

The questions are fine, keep asking. I'm learning too. Every time I think I've made a leap there's something else that crops up. I'm still asking 'dumb questions' elsewhere on the forum.........!


EDIT: C1 on your schematic looks like it may cut the highs too much.

[bbmonster]

Rick (or anyone),

This amp design has caught my interest, but I'm a noob to the amp-building scene--where are the speaker connections located in the schematic?

My best guess would be that it goes on the leads of the transformer, but I'd want to be sure before building

Oh yeah sorry. The fender reverb transformer has four wires coming out of it:

Two are for the primary winding - these go to the 5672's plate and B+, as shown in the schematic. I connected the 0 ohm to the plate and the 22500 ohm to B+

Two are the for the secondary winding - these go to the speaker. I connected the 0 ohm to ground and the 8 ohm to the speaker.

This was my first tube amp build too - first time I've used an output transformer too, so I'm assuming I got the poliarity of the transformer correct. Hey, it works!

I got this when I ordered a transformer from Triode Electronics. May help with that earlier post from jacobyjd. I still haven't built this amp yet, taking my time acquiring parts, which is good since I see there is a modification to the schematic.



Sorry for the size of the image, still new to photobucket, thought their resize would've have shrunk it a bit more.

[bbmonster]

Hi, I will attempt this amp soon and was reading on some safety/guidelines from various amp sites. Would a standby switch be a good thing for submini tubes? and if it is, where would it go? Also would I need a fuse somewhere? or not needed since using a wall wart and/or the amp is low power?

Also, how hot does the submini tubes get? I was wondering if lining the enclosure holes for the tubes with plastic or a grommet type thingy is a good idea.

Thanks.
-Bernard

[frequencycentral]

Hi Bernard,

I never really considered a standby switch (my VJ doesn't have one for example), if you were to add one it should break the connection between B+ and the tubes' plates, so the heaters would warm up before the plates see any voltage. With the (relatively) low voltages in this amp though, there is probably no need for a standby switch. No need for a fuse either, as the the circuit only sees 12 volts, and the ~80 volts generated by the charge pump is quite safe to touch without so much as a minor shock. The 6111 gets quite warm, as do the two voltage drop resistors, but the 5672 doesn't really get too warm at all. 

[bbmonster]

Just wondering if I got mine working properly. I perfed the MO v4 and have it connected to a Marshall MG15 (disconnected the MG amp and have wires running out the back to connect to the MO. I barely get any audio till the gain is past about halfway and the volume on max. When I do max the gain and volume then the audio is just about or a little over normal talking volume. Could the speaker I just have it connected to not be the the right ohms (or just suck)? I couldn't find out how much ohms speaker is and there is no informational label on the speaker itself.

Thank you.
-Bernard

[frequencycentral]

A good quality speaker will really help. I've been having a lot of fun with my MO recently through the 8" 16 ohm speaker of my VJ. I've been running a BM into the MO. The thing definately rawks - loud enough to get compaint of 'turn it down!' from my GF. I pretty much always run the gain and volume maxed, or nearly maxed. Think of MO as an 'everything set to 10' amp which won't damage your hearing.

I have tried MO through a crappy 12" 8 ohm speaker from an old organ, and it's really quiet. The same with cheap computer speakers. I guess the speaker it half of everything with amps.

This week I'm going to try negative biasing the preamp stages (I suspect I'll need to add cathode resistors too) to see if I can squeeze more clean volume out of it.

[bbmonster]

Thanks. I really like the sound of the amp as it is through my crappy MG. I was just making sure i built the MO correct. I haven't had the chance to play with low power amps (excluding ruby which i just use with headphones), so I'm unfamiliar with what the volume should be or the range of usefulness on the pots. Now the hard part of putting it in an enclosure and making it look nice.

Thanks again.
-Bernard

[frequencycentral]

By the way - did you check the output of the charge pump with a DMM to make sure you're getting ~80 volts? Also, try it with headphones, you be able to dial back the controls quite a bit and also get a wider range of clean tones.

As for the range of usefulness of the pots when driving speakers, the negative bias mod has improved this, and I'm hoping to improve it further with negative biased preamp mods - but the best sounds are with everything maxed or almost maxed.

[bbmonster]

By the way - did you check the output of the charge pump with a DMM to make sure you're getting ~80 volts? Also, try it with headphones, you be able to dial back the controls quite a bit and also get a wider range of clean tones.

As for the range of usefulness of the pots when driving speakers, the negative bias mod has improved this, and I'm hoping to improve it further with negative biased preamp mods - but the best sounds are with everything maxed or almost maxed.

Yep, built the charge pump first and made sure it worked, reading 77 volts. Habit I have being a programmer of "code a little, test a little".  I am using a 12v 1000ma Linksys adapter that is outputting 16v into a 12v regulator that powers the TC1044. And yes, I like using the MO with all the knobs maxed or near max.

[frequencycentral]

And yes, I like using the MO with all the knobs maxed or near max.

           

[Renegadrian]

ALL MEN PLAY ON TEN!!! 

[frequencycentral]

Got this back on the breadboard today, mostly to try negative biasing the preamp (which didn't work for me) and to try some different pentodes.

One experiment I did was to use two 5672's in parallel (but each having it's own voltage drop resistor for the heater) for the power amp - louder and ballsier. I've seen this done in 'proper' big amps, so I though I'd try it too. Well, it works with two, so why not three or four?