...and now, the "PentaDriver" 12 volt Pentode Overdrive using Submini 5672 tubes

[decc]

If you like the way it sounds I wouldn't bother. The difference isn't night and day or anything.

Note to others trying this pedal: Use the LM317, not the LM317L, for the heaters. I thought the smaller TO-92 version would be fine since it's rated at 100mA. However its max P_D is 625mW and deriving the 1.25V from a 9V battery (let alone a 12V adapter) is 775mW. I suppose I could stage two of them (9V->5V, 5V-1.25V) until I get the correct larger model.

[frequencycentral]

If you like the way it sounds I wouldn't bother. The difference isn't night and day or anything.

Note to others trying this pedal: Use the LM317, not the LM317L, for the heaters. I thought the smaller TO-92 version would be fine since it's rated at 100mA. However its max P_D is 625mW and deriving the 1.25V from a 9V battery (let alone a 12V adapter) is 775mW. I suppose I could stage two of them (9V->5V, 5V-1.25V) until I get the correct larger model.

I'd love to see some photots and hear some soundclips of you build decc!

[decc]

I'd love to see some photots and hear some soundclips of you build decc!

Well, first I have to put it in an enclosure, paint it, learn to play, etc.

So, um, did anyone else notice that the screen grids are all out of whack/mis-biased? They're supposed to be slightly negative compared to the anode in order to reflect back any electrons that get free (or something like that. I'm a bit rusty at this.)  According to The Wizard:

For normal operation the quiescent screen voltage will be less than the anode voltage. If the screen grid were to be connected to a higher voltage than the anode, then the screen current will act as the primary anode, effectively 'stealing' current from the anode. This excessive screen current can lead to overdissipation of the screen and subsequent destruction, especially if the stage is overdriven. It is possible to design stages this way, but is neither recommended nor much use for guitar.

But in my build, and in the voltages I see people posting, pin 2 is higher than pin 1. Which makes sense since they both have a 100k resistor. Since the anode resistor sets the gain, and I didn't want to mess with the turbo mod, I experimented with the screen grid resistor. A value of 390K results in V_G2 of about 6.4V and a V_A of 8.8V. (I'm running off 18V.) The screen grid is now negative to the anode and  having the idle state right at mid-rail seems correct. With the turbo on V_G2 doesn't change and the anode is at about 16V when dialed in by ear.

Doesn't sound any different though!

[frequencycentral]

So, um, did anyone else notice that the screen grids are all out of whack/mis-biased? They're supposed to be slightly negative compared to the anode in order to reflect back any electrons that get free (or something like that. I'm a bit rusty at this.)  According to The Wizard:

For normal operation the quiescent screen voltage will be less than the anode voltage. If the screen grid were to be connected to a higher voltage than the anode, then the screen current will act as the primary anode, effectively 'stealing' current from the anode. This excessive screen current can lead to overdissipation of the screen and subsequent destruction, especially if the stage is overdriven. It is possible to design stages this way, but is neither recommended nor much use for guitar.

But in my build, and in the voltages I see people posting, pin 2 is higher than pin 1. Which makes sense since they both have a 100k resistor. Since the anode resistor sets the gain, and I didn't want to mess with the turbo mod, I experimented with the screen grid resistor. A value of 390K results in V_G2 of about 6.4V and a V_A of 8.8V. (I'm running off 18V.) The screen grid is now negative to the anode and  having the idle state right at mid-rail seems correct. With the turbo on V_G2 doesn't change and the anode is at about 16V when dialed in by ear.

Doesn't sound any different though!

That's very interesting and well worth knowing. I seem to remember that while breadboarding I found that using a higher screen grid resistor also required using a higher screen grid cap to get the most gain from this circuit. One configuration I tried which worked well was 1M/10uf. Do you think the current configuration is bad for the tube in the long term? The turbo mod does set the anode voltage significantly higher than the screen voltage on the first stage (but not the second).

I'd love to see some photots and hear some soundclips of you build decc!

Well, first I have to put it in an enclosure, paint it, learn to play, etc.

[decc]

That's very interesting and well worth knowing. I seem to remember that while breadboarding I found that using a higher screen grid resistor also required using a higher screen grid cap to get the most gain from this circuit. One configuration I tried which worked well was 1M/10uf. Do you think the current configuration is bad for the tube in the long term? The turbo mod does set the anode voltage significantly higher than the screen voltage on the first stage (but not the second).

Yeah, as I understand it having the screen grid higher than the anode is going to damage the tubes. The datasheet has the typical operating point equal to that of the anode though so I guess that it doesn't have to be much lower to do its job. Too low probably has detrimental effects too.

I'm also wondering about the safety of the turbo mod. When I measured my circuit (again, running off 18V) I got 115mV across the 1k resistor. So according to the math from the AX84 quote:

V_Effective = V_A + (I_A * R_Load) = 16V + (115uA * 1.0082M) = 131V!

Now obviously that voltage isn't present, but I interpret the "effective" part to mean how much current there is. With r_a = 20k this would be 6.5mA, above the max spec of 5mA (typical 3mA). I tweaked it back down to 8.73V + (65uA * 1.0082M) = 54V -> 2.7mA. Didn't sound any different so it may be possible to accidentally tune-by-ear and end up out of spec.

I found that, at least with a single-turn pot, it's pretty difficult to hit the right spot. I would think I had it tuned and then later find either quiet notes being gated or loud ones being hard clipped as if the transistor was in the audio path. I think I'm going to see if we can get the same super-gain effect with larger anode resistors or maybe just another charge-pump stage for the first tube.

[robmdall]

decc, any update to this thread? I am getting ready to build this circuit and am very curious about your outcome.

[decc]

Sure. Let's see what I ended up with.

Screen Voltages: I think I'm correct. My build has 100k anode resistors and 390k screen resistors, although putting in a trimpot there would let you dial in something closer to ideal (maybe 330k-370k). The idea being to have the screen voltage as close to, but not over the anode voltage. I didn't try putting in new caps since the 1uF seems to work, although as I understand it smaller values will provide less gain to lower frequencies if you want to try tuning that.

2nd Stage Coupling: I was wrong about this one. What I was seeing in classic tube schematics was a cathode follower driving a Marshall/Fender tonestack. The important detail I was missing here was that M/F stack does not have a DC path to ground, while the Big Muff one does. This throws off the biasing for the tube driving it. So the cap between the anode and tone section should be there. (This only affects the DC state of the tube though. We still have an AC path for both treble and bass to ground which lowers tube #2's gain.)

Turbo Mod: I went without this just because I wanted an all-tube pedal. I know it's not part of the audio path but part of the fun was figuring out how the pentodes are supposed to work without adding in clever hacks. I think the calculations are still correct though.

Gain pot: One thing that bothered me about this pedal was that the gain knob affected volume more than actual distortion. I think in my non-turbo, 24V build the first stage has a gain of about 11x and the second of about 7x. So you get most of the distortion from the first stage (even moreso in a turbo version) and using a gain pot to control the level to the second stage doesn't really have any effect on that at the bottom half of the pot's travel. So I replaced the 8.2k stopper with a 100k one. The idea being that the voltage divider at minimum gain * second stage gain = unity, and turning up from there would start clipping. 160k would match the 7x gain better but I didn't have one on hand.

Grid Resistors & Coupling Caps:  I was putting a boost pedal in front of the tubes to see what kind of sounds I could get with a hotter signal and found that with only a light amount of pre-gain I could get the tubes to essentially go silent for a second if I hit a chord particularly hard. After some searching I came across the problem of "blocking distortion" which described the sound I was hearing. I reduced the 47n inter-stage cap to 22n and put a 100k resistor in series between the wiper of the gain pot and the grid of the second tube and the problem went away. I don't have calculations to back up these values but it didn't appear to negatively affect the tone so I'm happy.

Whew. Think that's it.

[robmdall]

Thanks for the overview. Looks like you you did some work! I have Rick's original schematic about 90% breadboarded (schedule has been rough lately). I will build the original and then work your mods into it. Again, thanks again for your input.

Obviously a huge thanks to Rick for a really cool schematic.

Rob

[decc]

I did some more experimenting with this circuit and thought I would let you guys know how higher voltages affect things. I replaced the voltage pump with a switching supply based on the MC34063A controller with an adjustable output from about 18V to 60V.

Findings:

1) This was an unusable circuit due to the noise generated by the switching supply. I used a cheap "bobbin" inductor and even though the controller was running at 90kHz it would put out a nasty hum at about 200Hz due to running in discontinuous mode. Tuning things to run in continuous mode would have helped, though I expect a toroid inductor and giant capacitor (had a 200uF 100V and it was already huge) would help more. Either way switching boost converters can be nasty beasts to tame.

2) There are diminishing returns with respect to plate voltage vs gain. I was going under the rough gain calculation of gain = gm*(R_A || R_L), and backtracking from the known resistance and measured gain I thought that the low voltage was resulting in low transconductance. However with the supply at 60V I still measured about 20x gain at maximum, not the 60x or so that would results from the 750umhos from the datasheet at full voltage. I found another equation of gain = (gm * sqrt((R_A || R_L) * r_a)) / 2. This matched what I was seeing much better and implies that transconductance doesn't fall off that much at low voltages. Since we are aiming for an overdrive pedal, not a clean one with lots of headroom, the ideal is a point where the gain is highest relative to the supply voltage. Ends up this is roughly around 24V -- so the original 12V + voltage pump was already the way to go.

[mattthegamer463]

((Thread Revival Alert))

My PentaDriver Turbo Plus enclosure so far...





I made that metal bar out of 1/4" square steel, it will go in the small holes at the top of the pedal to protect the exposed tubes.

[trad3mark]

double thread revival action!

I've ordered some subminitubes that are due to arrive any day now. They've really similar charactersitics as this, as far as i can tell. They're russian ones. I was just wondering, if you run this at 12V to the anode, what kind of gain do you get? Are we talking low levels, or serious crunch? Is it similar to, for example, the valvecaster with the 6111 tubes?

[frequencycentral]

double thread revival action!

I've ordered some subminitubes that are due to arrive any day now. They've really similar charactersitics as this, as far as i can tell. They're russian ones. I was just wondering, if you run this at 12V to the anode, what kind of gain do you get? Are we talking low levels, or serious crunch? Is it similar to, for example, the valvecaster with the 6111 tubes?

Probably worth listening to the soundclips:

I have no doubt that you guys are now sick to the teeth of me and my submini tube obsession! Anyway, I recorded a couple of new soundclips of my modded PentaDriver, which is now at the spec of the schematic on page 5 - with the turbo mod and Big Muff style tonestack.

Here's a couple of soundclips, nothing special playing wise, just trying to show what the pedal can do in terms of it's tonal variations.

Signal path:

Squier Telecaster Custom (HB) > PentaDriver > Roland Bolt 60 watt tube combo > Shure SM58 > Event Echo Gina 24 soundcard > Wavelab

This clip is with gain at half, just running through some different tone control settings: http://www.aronnelson.com/DIYFiles/up/PentaDriver_Half_Gain.mp3

This clip is with gain at full, again running through some different tone control settings: http://www.aronnelson.com/DIYFiles/up/PentaDriver_Full_Gain.mp3

[trad3mark]

beautiful! Thanks for that rick!

The tubes i'm getting, 1J29B-V have REALLY similar characteristics, so looking at these kinds of schematics is a good starting point. That and valvewizard's pages.

That is a really nice gain level though. Really really nice.... 

[mattthegamer463]

I put together a DIY Layout for the Pentadriver Turbo Plus, would someone like to verify it for me?  I've double checked it, but fresh eyes are always a good thing.

I left out the bypass footswitch and power switching, because the design was getting too complex looking and not being able to route the wires manually makes it look sloppy.

[mattthegamer463]

All done, sounds cool.





Thanks for the design, Rick.

[frequencycentral]

^^^

Nice! Very cool that you fabricated the rollbar. Any gutshots for us?

[Renegadrian]

If it sound as cool as it looks, that's KILLER! 

[mattthegamer463]

Sorry guys, I forgot I took this pic with my phone earlier.  Here you go:



The submini's are mounted to a segment of perf which is epoxied in place on top of the DC jack.  Its a tight squeeze in there. 

Also, Rick gave me the great idea of using ribbon cables in place of lots o' wires, so simple and yet so effective.  I've got two 5672's left, what should I do?

[kristoffereide]

Hello, haven´t been here a while, but as I´m having trouble with the layout here, I wonder if the IC3 is a row too high up? I have a short I can´t figure out anywhoooos...

[BluesHarry]

Sorry to dig up so old topic (golden shovel award to me), yes, the vero board layout seems to be incorrect - the pinout is pin 1 adjust, 2 vout and 3 vin, so the power should go to pin3 (not pin 2 as on vero), pin5 of the tube should be connected to either 1 or 2 (as they're both grounded) and there should be a jumper from pin 1 or 2 to ground (as on the layout shown they are not connected to ground and they should be, according to schematic.