Cherub (ECC83 --> EL84 + 6V6S)

[tux320]

(Original post at SEWATT forums: http://www.sewatt.com/node/20552)
Ever since I first saw the proverbial DIY light when building my first pedal (an MXR Distortion+), I’ve wanted to build an amp. Because of my circumstances, being 16 and so still living at home, the traditional single-ended low-wattage first build that most people start with appealed to me but at the same time, I knew that a champ just didn't offer quite the right sound I wanted. Then one day I found the schematic and a sound clip for a DR. Z Mini Z which was more to my liking.

Unfortunately I am restricted in that I don't have the tools necessary to fabricate my own chassis, nor can a currently afford them, and so I bought a pre-punched chassis from Ampmaker (good prices and great service) as well as two of his transformers. Because the chassis was punched for a noval and an octal socket I would need to modify the schematic to utilise a 6V6 output tube instead. This was achieved by changing the grid-leak to 220K, removing the grid-stopper (a 2.7k 2W) and changing the cathode resistor to 470R.

While reading up on tube theory to try and understand what changes were needed, I found various references to better ways of building a single-ended amplifier’s power supply in order to reduce hum and noise. By a happy accident I discovered that my PT had a centred tapped 6.3v heater and so I could do away with the 100R artificial centre tap and elevate the heaters by connecting them to just above the 6V6’s cathode resistor. Furthermore, I read on Merlin’s site that a 0.1uf non-polarised capacitor and a 15R resistor could be placed before the reservoir capacitor to reduce ‘hash’ noise by reducing the switching transient. In addition I was advised to add a filter after the reservoir capacitor but before the output stage to achieve an even better quality DC supply.
The final addition I made to the original schematic, which I now call Cherub, was to double up the diodes in the bridge rectifier, and add in 10n capacitors in parallel with the diodes to act as snubbing capacitors and to ensure an equal voltage share between each diode.

I have simulated both power supplies, to the best of my understanding (which is not great) in Duncan’s PSU designer and have included screenshots of the results. The maximum variation of voltage is within 10% of the original values which seems reasonable to me, though I may of course be completely wrong Very Happy .
I apologise for this turning into somewhat of an essay, but if anyone could check over the schematics and simulations, found at the download link below, or just offer any general advice it would be much appreciated.
Thank you,

Sam

The amp is now finished, as you can see from the photos below. She runs quite hot, in terms of temperature, so I've run the head with an open back thus far. I've only been able to test with my crappy 12" no-name speaker so far, so the sound has not been fantastic, but the amp does show promise. As I suspected 10 watt is very loud, but since I normally play with a large amount of fuzz I'm not that worried. Considering the gain I'm pleased with the amount of clean headroom, the cathode bypass switch being of particular use here. Unfortunately she ate the preamp valve after about a fortnight, having developed an annoying buzz on certain notes and volumes. 80-100% on the volume dial also caused squealing, even without the guitar plugged in. These problems disappeared for about ten minutes when I tried a new preamp valve, but then reappeared again. I've not yet had time to troubleshoot the build, so I haven't been able to play with it much. Designing and building the amp was an enormously satisfying process, and I can safely say that I will be building another when money and time allows, probably in about five years though, with university coming up.

Thanks to everyone who has helped me,

Sam

[Alexo]

Hey Sam, welcome to amp building! Good on you for getting an early start.

Everything looks good on your schematic, nice work.

The only things I would change would be revisions to the original schematic. At the risk of revealing my ignorance, R5, R6 and R7 all look quite superfluous to me - I would just replace that whole network, + the 1M pot, with a single 500K pot.

My guess is Z may have been just trying to cut down some signal to tame things down, via R5 and R7 and thrown in R6 to alter the taper of the pot (?). But I think that generally the fewer parts you can run, the better. 6V6's need a bit more signal to drive them than EL84's, so the extra volume reduction won't be necessary in your amp.

But then... we're darn close to being back at a Champ - a hard thing to avoid in SE builds.

I would add a 470 ohm, 1 watt resistor between B+2 and your 6V6 screen, which will make the tube last longer and give you a little bit more compression.

I would also use a 5 watt resistor on R11 instead of a 2 watt.

...if you are really concerned about hum, you could up the filter caps to whatever you want: 47uf, 100, etc., the sky is the limit with diode rectifiers.

[tux320]

Thanks for the help Alexo, it's very much appreciated. I've done away with R5 and R7 as you suggested but I’m keeping R6 for the moment so that I do not have to buy a new pot. I've incorporated the other changes you suggested as well as doing away with the needlessly overbuilt rectifier. Now I only need to finish the layout and then I can begin soldering .

Here is an updated schematic:

[ChrisM]

Why no grid stopper on the output tube?

[tux320]

When I asked about the design on ax84 I was given this piece of advice by steve25: "You could ditch the grid stopper also, as a 6v6 does not drive as easy as a EL84." If I need one then what value would you recommend. After having a look at valvewizard all I've established is that the tube's datasheet should suggest a reasonable value but I can't find where in the datasheet this information is, if indeed it is there at all. The power tube I am using is a JJ/Tesla 6V6S.

Thanks again,

Sam

[ChrisM]

I would use one personally.

1k5 10K are reasonable values (you usually see 1K5-5K1).

[tux320]

Ok, I will put back the 2k7 that I originally had there.

Thanks for your help.

[crbowman]

Looks good to me. Nice work tux!
Be sure and keep us posted once you get it built. I'm curious as to what you think.
Nothing tastes as good as vegetables grown in your own garden.
And congrats on starting a really fun, rewarding ....er......hobby.....yeah, that's what it is.....a HOBBY.....

[tux320]

I'm resurecting this thread now that I've finally got round to getting on with this build. Incorperating some ideas confirmed in this thread over at SEwatt (http://www.sewatt.com/node/21015) I have come up with a new design utilisng a parralel single-ended output stage with two transformers. More than likely I will be using the impedance mismatching trick to give a lower reflected impedence, i.e. 2.5K rather than 5K. Here are the schematics, any comments and critique are most welcome.

Thanks, Sam

(Paralleled transformers to increase power rating, as I had a spare lying around)

[tubeswell]

I have come up with a new design utilisng a parralel single-ended output stage with two transformers.

You might be better off running both output tubes in parallel to a single SE OT in terms of the signal going into the speaker. Less 'complication' with the signal. (On the other hand you may prefer it the way you have drawn it). JM2CW

[Gaz]

Have you tried that tone stack? 47p cap with a 250k pot (opposed to the Vox 47p with 1M pot) will be only adjusting super high frequencies. If you wanna reduce the Vox stack's pot to 250k, you need to increase the treble cap to about 150p to get the same response. Also a linear pot might make it kinda bright at the half way (noon) setting, but it really depends on the rest of the amp of course.

[tux320]

Thanks for the advice guys, very much appreciated.

@tubeswell
The main reason for wiring the transformers in this way was to take advantage of the spare transformer I had lying around. Right now I don't really have the disposable income for a new, higher rated transformer. For clarity, each OT is currently 5 watt. At the end of the day if it sounds bad it would be quite easy for me to pull a transformer and an output tube.

@Gaz
I haven't really experimented with tonestacks before, so this is more or less copied wholesale from a blackheart schematic I found. Thanks for the tip on the 150p cap, I'll pick one up the next time I go for parts. Originally the amp wasn't going to have a tonestack anyway hence why I made it defeatable. The inclusion was to try and make the amp more complex as I'm trying to put together a mathematical analysis of the amplifiers operation for school, a la the ax84 p1 document.

Thanks again, all the maths in the world can't beat advice from somone experienced.

Sam

[tux320]

Ok small progress update. Turns out that the 47pF tonestack capacitor was supposed to be 470pF anyway and I'd just uploaded an earlier version of the schematic by mistake. My question now is that does the effect of breaking the ground path from the mid pot to defeat the tonestack also negate the DC blocking effect of the tonestack as a whole and as a result, do I need to add a coupling capacitor in before the tonestack to keep DC off the volume pot when the tonestack is defeated? It seems to me that the DC would still be blocked even with the tonestack ground lifted however, the only amp schematic I've seen with a defeat-able tonestack was that of the Vox Night Train in which, via the use of a DPDT switch, the ground was lifted and the signal path changed so that the plate resistor now fed a new capacitor, which in turn went into the volume pot. Is this extra capacitor needed for coupling, or was it just included to compensate for the loss of the tonestack?

EDIT:
I've now resolved this issue and I've uploaded some photos of the chassis together as a mock up. I have no idea as to the spacing required between valves, but considering there must be a reasonable amount of heat I have decided to place the ECC83 in the center to try and keep the two power valves as far apart as possible. If the EL84 and 6V6S would not be too hot, I would much prefer to switch the EL84 and ECC83 as this would simplify the wiring. The distance between the mid-points of both power valves is 3", and the ECC83 is positioned exactly between them, giving a spacing of 1/2" between each valve. Any words of advice would be greatly appreciated.

Thanks,

Sam

[tux320]

After a first real day of work, here's where I am now. It made sense to me to ground the center pins of the noval sockets, not sure if they should stay that way or not though. Any recomendations of good speakers for this amp sub-£100, 12", 8 or 4 ohms? Bearing in mind I'm unlikely to ever run it completley clean

[tux320]

The amp is finished, please see the updated first post

Sam