New project : vajra 36W

[rjgtr]

The funny thing is the layout I posted isn't the exact amp I build. s2 sent me a 'kit' of the parts with his different chassis. The Vajra layout I did was for a standard 18watt chassis layout, because I thought it would be more useful to the builders here. However, I hope it was a little bit helpful to you anyway.

I really like the vibe of the Vajra, but it is a lot more like an early AC15 (EF86) mated to a Top Boost (TMB) than a normal 18watt. Still a cool variant in every way.

Wiring is always the source of most noise and hum issues in a hand wired amp, especially with a new layout. But if it were too easy it wouldn't be fun!

Keep us posted on your progress.

[clobo]

Hi again,

Just to keep in touch, I'm still fine tuning the amp but this is a slow process since real life does not leave me with much spare time.

- Overall, the amp is a keeper, certainly the most articulate and lively amp I played so far (I'm turning 56). Both channels are equally sensitive to picking attack, and very subtly pushes the amp from clean to crunch, and into either Marshall, Vox, or Fender territory.

- PP/SE
This is a funny option to try, but I don't think I'll get any usefull sounds out of it. This control works best with a single 75R-10W cathode resistor (4xEL84), or with fixed bias mode. The "current" 36W configuration with dual 150R-5W cathode resistors connected to external, and internal internal pairs (respectively) seems to mask most of the effect. Also the "previous" 36W configuration with dual cathode resistors connected to left, and right pairs (respectively) is more compatible with the control, however the whole amp becomes pretty unstable under heavy drive (has nothing to do with this control).

- TMB boost
This mod is simply a switchable cathode bypass resistor on first TMB stage (12AU7 section of the 12DW7). The pre-amp is already very clean in its boost mode, however in non-boost mode I seem to hear a subtle difference (maybe more feeback and lower impedance driving the tone stack ?)

Cathode/Fixed bias option.
Switching to fixed bias mode with 50-70% max dissipation adds a little more of everything to the amp without changing its personality. Just for records, supply jumps from 340V (cathode bias), to 380V while the amp runs much cooler. Also Bias voltage jumps from 11 to 15V allowing for more voltage swing. Over all result in even more open sound. This option is a keeper.

- Master-Volume on TMB channel.
I find this control completely defeats the designer intention, however I'm only using it to tame the hum/noise coming from the TMB channel to acceptable levels. More to come.

TMB hum and noise
With its fully open 12DW7 second stage (12AX7 section) moderate levels of hum and noise enters the PI. The final result is not acceptable in a professional setup, so I completely rebuilt the board for the 3rd time, this time strictly following Vajra published layout (pretty weird btw). I was only able to get minimal improvement, mostly because of the shorter wires around the preamp and PI sockets. I've also tried many grounding configuration without much improvement, actually I have the "standard" Ritchie 4 points (power star, driver star, TMB-ch control bus, and normal-ch star).

Next
I'm gonna rebuild a 4th version, back to basic 36W layout, but with still shorter wires around the preamp/PI sockets, and maybe use a more Fenderish layout around the 12DW7, I find a typical Black-face noise/hum level would be quite acceptable...

Waiting for suggestiions on this one...

Thanks for reading my poor english

Claude

[rjgtr]

Can you post some pics?

I know the Vajra layout I posted is a little different, but my goal was to create a layout with the same size board as a normal 18watt. It was also designed to fit in a standard 18watt chassis. My actual build was on an S2 chassis, so it deviated a little from the layout in the front panel control.

As to your hum issue, the pics might help us see something. My Vajra doesn't have a hum issue using the layout I posted. I think you might have some kind of a sympathetic issue here, either from the grounding scheme or from the filaments.

On the PP control, it is a pretty subtle change, but I like it. But I could easily live without it.

Like you, I find the fixed bias option is useful. But the cathode bias mode definitely gets you into a Voxish/18 wattish vibe better.

[clobo]

Thanks for you help Richard.

Sorry I'm not allowed to attach any file now Seems the maximum filesize for all Attachments is reached

However my current prototype is actually dead quiet, this is until I move the (temporary) master volume up. I pretty much followed your drilling guide, except I had to reduce board width by one row (my board is only 2.5 in. wide). I was lucky to find some very small axial filter caps.

BTW I just changed your own grounding pattern, for the Ritchie version I had success with in all my previous builds, however it did not improve anything this time.

Thanks again

Claude

[klingo]

Thanks Clobo for detailed review.
a new amp is trying to find a place on my "next amp" list. may be with parallel 12au7 & 12ax7 instead of a 12dw7.
don't worry with poor english, here in France we can read you 5/5 loud & clear
EDIT meaning ,i'm not able to detect any difference

[rjgtr]

Claude,

Your english is fine.

Grounding schemes are an interesting topic. A lot of people have ideas on what is best, but all have tradeoffs. Any decent scheme can be used to produce a quiet amp. I think it's all a matter of personal preference.

My layout wasn't intended to be the end all and be all of Vajra layouts, just to satify the above criteria and give people something that's been used successfully.

[clobo]

Claude,

Your english is fine.

Grounding schemes are an interesting topic. A lot of people have ideas on what is best, but all have tradeoffs. Any decent scheme can be used to produce a quiet amp. I think it's all a matter of personal preference.

My layout wasn't intended to be the end all and be all of Vajra layouts, just to satify the above criteria and give people something that's been used successfully.

Thank Richard.

I find your layout is fine, pretty inventive for getting short wires around the preamp sockets and moving critical "interfering" components away from TMB input jacks, and it still fits a standard lenght 18W board.

Back to my tests, my hum issue is not much related to layouts or grounding patterns, but in fact moving TMB signal wires around (input, tone stack, volume, master) has the greatest effect on reducing hum.

Then I read Mark D. mentioning that the original Brown Concert had a very similar split plate circuit implementation... so I rebuilt my 12DW7 section to reflect 6G12-A (see 7025, second from right). Then I also found out that using non-shielded wires actually improved hum (allows for better hum cancellation through cross-coupling between adjacent wires, signals being out of phase )

See layout model on page 2, of course I substituted all components values for Vajra
http://www.ampwares.com/schematics/concert_6g12a.pdf

At this point my amp keeps quiet until Master is turned past 7. Not perfect yet but still a major improvement.

BTW I also changed my mind about that Master-vol, this one will stay. I found out that my Les Paul could nail that Duane tone on TMB channel (everything on ten, exept Master).

More to follow

Claude

[rjgtr]

Lots of great info. Thanks Claude.

Then I also found out that using non-shielded wires actually improved hum (allows for better hum cancellation through cross-coupling between adjacent wires, signals being out of phase )

I use shielded wires sparingly. Like you, I find the shields can actually carry ground hum into the rest of the amp. IMHO, shields are best used sparingly and only when needed. A good layout and lead dress is a lot more important in eliminating hum.

Once we get the attachments issue solved I hope you can share some pics and maybe even a layout.

[clobo]

Hi,

Trying to attach file again...seems to work now.

Here's my fourth prototype. Hum reduction mostly resides around the Brown Concert "inspired" layout around the 12DW7, including cable routing (not shown). Also, I opted for a single cathode resistor in power section, for the sake of simplicity when adding thecathode/fixed bias switching. The rest is pretty basic 36W-RP style.

Thanks,

Claude

[clobo]

Hi,

For those who might still be interested, here's my previous prototype (I was unable to post). This one was pretty much inspired by Richard's own 18W Vajra.

BTW it worked very well, despite its crowded aspect. TMB components are logically placed close to the tubes, and away from inputs and controls.

Thanks Richard

Claude

[rjgtr]

I like your layouts. Very clean. The last one is very classic Marshall.

I'm still really glad you're building a Vajra. More people need to build these. Mark created a great amp.

[clobo]

Hi again,

It's been a month since my last post (unfortunately day job keeps me very busy). Remember my 12DW7 gain stage was still humming, even after many changes (layouts, and grounding patterns).

Now I finally fixed it, took me 2 majors steps into it.

1) On the original schematic, the 12DW7 actually shares the PI supply rail. This might seem like a design flaw (more than 2 consecutice gain stages), but there are other similar examples (Traynor YGM) that do work pretty well. Now back to my amp, I finally decoupled 12DWT supply (used basic 2.2K/22uF), and this got rid of all hum from the second preamp stage. No need to hide behind a Mast-Vol anymore.

2) Then there was still some (minimum) level of hum coming from the first 12DW7 stage (heard when cranking vol). This one took me a long time to fix. Finally I carefully rewired the entire 12DW7 preamp following the Fender Concert 6G12A. All input and control wires (unshielded) are now carefully twisted together along the controls, and they are also crossing the board at very precise locations (not shown in my drawing, please refer to Concert layout posted earlier). Also, there is a separate chassis ground lug close to TMB input, used as star ground for the whole 12DW7 preamp.

Not the amp is dead quiet. I now have a PP/SE control (in place of Mast-vol), but still have to implement fixed bias option. This is still my best sounding amp...

Claude

[CurtissRobin]

Congratulations on the result!
Hats off to you for the effort!

And all in all, wow!

KennyO