Tweedle Dee - 5E3-D

[Stephen1966]

Greetings

I recently built a Tweedle Dee from Charlie Wilson's sketch and thought you guys might appreciate some posts showing the results of my work.

First off, my thanks to all of you - too numerous to mention individually - who helped my thinking on the design with special mentions to Uncle Doug (a real gent. and a great educator) and Rob Robinette who seems to have spent half his life writing webpages and Randall Aiken for his immensely instructive technical whitepapers. Of course, none of this would have been possible without Charlie Wilson.

I was particularly interested in modifying the design to match the 5E3 which permits two channels with hi and lo inputs thereby maximizing the options for jumpering between channels. Another design goal was to keep the sound floor to a minimum. The grounding scheme of separate power and preamp buses was adopted from RobRob's https://robrobinette.com/RobRob_Deluxe.htm design which also gave me the wherewithal to elevate the PT CT voltage, reducing heater hum, by hooking it up to the power tube cathode resistor. As for components, I followed Randall Aiken's advice https://www.aikenamps.com/index.php/res ... -it-matter and went for the largest 'practical wattage' resistors I could lay my hands on, along with an overall increase in the number of metal film resistors. The aim was to produce an amp I could feel comfortable using much more than it was to create a purist clone of a Tweedle Dee, and so far, nothing has exploded and it works like a charm. Only one thing didn't work out, which was to include an ohm selector for the outputs but that was only because of space constraints in the chassis - two dimensional thinking on my part!

As you will see when I post the layout, it pretty much follows the 5E3 organization of components and my board was designed around the components chosen for the circuit. The circuit works for sure, but the position of the tail load resistors (sandwiched between the B+3 filter cap and the power tubes' grid leak cap and resistor network) suggests room for improvement. I think a redesign of this real estate will permit a more logical organization of the components in the signal path and allow for the (physical) installation of an ohm selector.

The tubes (I still call them valves if I am being honest) are the Sovtek 5AR4, a pair of matched EH 6V6GT's and a pair of matched and balanced Tungsol 12AX7's. Generally speaking my voltages are higher than a standard 5E3 and higher than Charlie Wilson's record of voltages from his Tweedle Dee. Everything is still within tolerable limits for the components I think, but my power tube plate voltages are on the high side so I may eventually try swapping the 6V6GT's out for 6L6's. There is very little noise at maximum gain settings: a little AC hum which I might eventually try cancelling out with a Hum Dinger pot and a little pink noise hiss, both of which aren't noticeable once I start to play. I couldn't say what the output wattage is, I would need to sell a couple of these to buy an oscilloscope but it's loud... I mean window-rattling loud.

I built the cabinet as well - a standard '57 5E3 frame with some modifications.

cab_front.jpg

cab_back.jpg

The layout was done in Illustrator. This is just the bare chassis and components.

5E3-D layout bare.pdf

I'm including a couple of photos of the top and bottom of the board as well.

board_top.jpg

board_bottom.jpg

And the assembled combo. Which will need to go into a subsequent post if anyone is interested as I seem to have reached a limit of five attachments here.

I'm very happy with the build. Nobody died and it sounds rich and authoritative with a gutsy overdrive once the volume gets around 6 on the dial. There is less volume interaction - probably because of the split cathode caps and resistors in the preamp. I'd go as far as to say there is almost none, but that is where the jumpers come in and I'm only just beginning to explore the possibilities there. The speaker is a Celestion Vintage 30. Sounds great. Real ballsy.

[xtian]

Nice work! And welcome to TAG.

[Stephen1966]

Thanks xtian. Here are some more pics of the finished build.

5E3-D-front.jpg

5E3-D-open-back.jpg

5E3-D-back.jpg

Cats + high voltages = -9 lives... Hence, the back panel's mesh.

[norburybrook]

looks great. I Built one last year and and put it in a separate head/speaker cab.

Lovely amp but it still has way to much bass unless you play on the bridge pickup all the time. perhaps the original amps bottom end is reduced by the small combo size and therefore a separate 1x12 cab is too much. I fitted a switch on the normal channel with a choice of coupling caps to reduce the bass but if I were to do it again I'd either make a small combo like the original or fit a t&B tone control.

great amps though for recording as you can high pass the low end.


and I'll echo Xtains sentiment in welcoming you to TAG.


M

[Stephen1966]

Thanks norburybrook. I was thinking about the possibilities of a separate head + speaker cab. For one thing, there would be a lot more real estate to play with, adding switches, jacks and tone stacks and the like, along the lines of the Modulus Tweedle Dee. The 5E3 combo chassis is a bit tight though and after considering a whole bunch of other mods I settled on the idea that less is probably more. The four inputs allow for a lot of different frequency variation, with the low channel inputs offering a nice smooth bass on the normal channel and a crisper one on the bright channel. Sustain is phenomenal, touch sensitivity is great and there is a clear definition of individual notes within chords, adding up to a complexity of tone. The upper partials on single notes bloom when the note is held.
I've also been trying a treble boost pedal and that thing just takes it to another level! The one thing that everybody seems to encounter though is that the overdrive doesn't really make an appearance until we are almost at the peak of the gain - starting around six on the dial. So there again, it's asking for a pedal if you don't want to blow the front row into next week.

I couldn't attest to the performance of other speaker types, but I am guessing you could find one to match the speaker cab with a little less low end.

[Charlie Wilson]

Great job Stephan1966. You may want to use JJ6v6S. I don't know how long the EH 6v6 will last.
CW

[Stephen1966]

Thanks Charlie, it's much appreciated.

I had a think about the JJ6V6S' but comparing the EH 6V6 or 6V6EH to be precise I noted that both have a rated plate dissipation of not more than 14W. I dare to say they could probably handle more but going back into the amp earlier this evening, and measuring the bias I get 39.92mA on V3 (16.6W) and 35.9mA on V4 (14.9W) which, in the case of V3, is around 120% of maximum dissipation. The plate voltage comes in 434VDC when measured to ground and 415/416VDC when measured plate to cathode. Both are within the 450V recommended plate voltage for the 6V6EH and the 500V for the JJ6V6S but those currents are too hot. It's likely the current would fry the tubes before the voltage does so now I am thinking it may be better to increase the value of the cathode resistor. Beyond trial and error I am stumped as to how to do this. I've seen values up to 330Ohm in other similar designs but if there is a way of calculating the required resistance to bring it down to the max 33.7mA or slightly less without going through a stack of components, I would be very grateful to anyone who could point me in the right direction.

Perhaps 6L6's are still a possibility?

[mojotom]

I would use the EH6V6 up to 400V, not more, and 70% max dissipation or even less (I shoot for 66-67%) and have a good amount of time available on the tubes yet they lost power after a while and need to be changed often (that’s my experience only on a a few amps for a few years). Good thing about them is they loose power and punch but do not break like a Tung Sol at those voltages.
Amps is fixed bias and at 70% you want something like 25mA @ 400V (6V6 is a 12W tube max).
I try to stay below 130*C at the hottest in the middle of the tube with a temp gun. Above I noticed a reduced lifetime.

Above 400V and up to 450V or so the JJ for me is the only choice available (modern tubes) that will work at those voltages and keep up for a couple of years or so. Some sets will go bad or have vibrations issue but a good pair can take a lot for a 6V6 (more rugged like a mini 6L6 by design rather than a 6V6 btw).

Yet unless I didn’t read your post correctly you’re really high bias wise, above 100% dissipation, so maybe check if everything is right and set the cathode resistor accordingly.

You can try JJ5881.

[Charlie Wilson]

Hello Stephan, I find tweed cathode bias amps always on the hot side, Dumble or not. The JJs can handle about 17W but I do think your plate voltage is a bit high. It seems that volume is not the goal, so you may try a 5Y3(not Sovtek) or 5V4 rectifier to bring the B+ down a bit. What power transformer did you use and do you know what the unloaded voltages are? The original Dumble amp had an unloaded B+ of 380-0-380 with an old Triad transformer. I think those old transformers sag the voltage a bit compared to the modern over spec ones. I believe this is why most builders end up with higher voltages than the original even though their transformers are 380-0-380 or even less.
CW

[Stephen1966]

Yet unless I didn’t read your post correctly you’re really high bias wise, above 100% dissipation, so maybe check if everything is right and set the cathode resistor accordingly.

Thank you both. I am working my way through your advice now and double checking the results after removing a 100K filter discharge resistor. Checking all the caps and resistors, everything comes in within single digit tolerances, all except for the power tubes' cathode cap - it should be 25uF or thereabouts but this is way out, 39.6uF! I can't see how that would augment the plate voltages and/or current but its positive terminal is, along with the cathode resistor, hooked up to the 6.3V CT and the power tubes' cathodes.

The power transformer is a Hammond 290BBEX rated at 355-0-355, the actual voltages I still need to check. I'm also going to check for continuity and double check everything against the schematic/sketch. A 5Y3 might be just the trick but given Dumble's transformer specs and voltages I have to look through everything again. Something doesn't add up, somewhere!

[Charlie Wilson]

Stephan, what is the wall AC voltage where you live(measured) and does that match the primary voltage for your power transformer?
CW

[Stephen1966]

I really appreciate the help - thank you.

Today, the wall voltage is 242V. Here in the Czech Republic the official wall voltage is 230V and in the very beginning that is what I configured the PT for. I quickly realised the discrepancy however, and corrected it by rewiring the PT to 240V primary. The wall voltage varies all the time, but the lowest it's been is 239V. If I take the amp out to Prague I expect it might drop to 230V but I don't think a 10V drop on the primary will be a problem.

[Stephen1966]

The 290BBEX power transformer is rated to deliver 355-0-355 V. It's actually coming out higher than that.

From pins 3 to 4 and from pins 5 to 6 on the Sovtek 5AR4, I have a pair of protective 1N4007 diodes. (Their polarity is correct.)

The voltage on the anode side of the diodes VAC - measured to ground:

pin 3 = 372VAC / pin 5 = 371VAC

The voltage on the cathode side of the diodes:

pin 4 = 196VAC / pin 6 = 193VAC

The high voltage leaving the rectifier at pins 2 and 8 is reflected in the line up to the standby switch

pin 2 = 434V / pin 8 = 434V and to STBY switch (measured on the terminal) = 433V

After the STBY switch with its surge protection (100k resistor in parallel with a .047uF cap, wired across the terminals) I measured the voltage again and it's about the same as you would expect. The line out voltage then goes directly to the first filter cap (B+1).

[Stephen1966]

I could possibly put a couple of resistors in series (one each on each HV line).

I'm attaching my schematic for the power section of the amp.

5E3-D Power circuit.pdf

This is the original sketch and I've changed some things now.

The primary windings of the PT are now wired for 240VAC - white and blue leads - as per the data sheet, also included.

290BBEX.pdf

R34 - the bleed resistor has been disconnected.

The ohm selector switch wasn't installed and the speaker is an 8Ohm type so is connected to the OT correspondingly.

[Stephen1966]

I explained the problem of the cathode cap and the supplier is sending a new one out to me today - Banzai Music, great service. It may not fix the problem but it will remove one more variable.