New 4-4-0 and Asteroid 60 amp porn

[passfan]

I've just finished up a 4-4-0 Express and a Komet 60 and decided to let everyone look under the hood.

The 4-4-0 is running a JAN Richardson 12av7 in V1, a Telefunken 12ax7 in V2 and a GE 12ay7 in V3. V4 is a Vintage RCA 6sn7gtb. This amp does the full on Express well,right down to the bloom, but it won't back off. Forget the clean to mean thing here folks, she won't do it.

The Astroid 60 is actually my second and I kicked myself for selling the first one so I had to go again. I used mullard and EM vintage caps and it too has Mullard, Telefunken and RCA vintage tubes. Cabinets are coming along and should be finished up within a week or two.

Enjoy them, I certainly do.

[passfan]

And the Komet. Thanks to RJ for the Edcor OT recomendation. It is smooth, clean and clear from the bottom to the top and it is a great fit so far.

[stephenl]

Any thoughts on the Fast / Gradual function? I am currently building a variant of this amp and was considering just using a fixed 50/50 split load.

[passfan]

Well it depends on how you want the amp to respond. If your looking for laid back the 82/22 is fine. If you are a precise player you may like the fast mode better. It would be interesting to hear a 50/50, but the difference is huge between fast and gradual as it sits. I believe some have built a relay type switcher to pedal drive the switch as well. Please let all know how 50/50 works out

[Colossal]

Nice build! Looks great. I bet it sounds excellent.

[amplifiednation]

Pass,
Great looking work!!

I did a variation on the 4-4-0 earlier this year. I put a choke in place of the large resistor and also added a second 6sn7 for a full quad of triodes!!

The amp sounds ballzy as all hell, it's a beast when played with my Les Paul. It's very express like, but with more bass (at least mine is). When I play it with my tele's it's just so-so.

I think that the amp needs some tweaks...it's a little harsh sounding to my ears. The first one I built I took out V1s cathode bypass cap and had the amp paired with a V30 in a combo, it sounded sweet!!! I built it very close to spec

This one i have in a headshell, paired with a matching 1x12 with a Celestion Blue. It's just not doing it for me. I have it biased around 65%, still seems too hot, but the 10k pots i have won't let me go lower. The bias with the two sides is somewhat interactive.

Sometimes I think I don't like the sound because there isn't reverb...i'm a bit of a junky when it comes to reverb, but I'm not sure...

How is yours sounding?

[passfan]

Mine sounds great with a LP and not so great with the tele. I think this amp needs to be pushed. It runs nice and crunchy up to around 1 o'clock and above 1 it starts to compress. I'm not really bassy. Must be your parallel output tubes. I don't remember what the bias pots are but I can shut mine all the way off. If I get a chance this weekend I'll work up some clips.

[Tillydog]

I've experimented with split load resistors in the 4-4-0 to get some clean headroom & tame it a bit. After a lot of experimentation, I ended up with the circuit shown below (sorry it's a bit scrappy) which I think it sounds pretty good now. (I need to update my thread on AX84 with the most recent developments.....)

The partial load resistor values are quite sensitive, as it's necessary to keep the PI reasonably balanced both clean and overdriven to keep the nice cleans and avoid harshness in distortion (according to my observations).

Note the value of the NFB resistor and the PI tail resistor, and the much larger grid-stoppers to tame the 'triode rattle' when the 6SN7 is overdriven.

Prior to these mods, my 4-4-0 was also very bassy. I cut the value of the 2nd to 3rd stage coupling cap combination (C14a & b) from 3n2 to 2n2 which helped a lot, and I also reduced the treble pot from 1 meg to 250k which did the rest, and made the tone controls much more usable IMO. I didn't notice any loss in highs going to the 220k grid stoppers on the 6SN7.

Andy

[surfsup]

The 39/150k in parallel is 31k. Why didn't you just try a 33k like on the other side? (just curious)...

[passfan]

The partial load resistor values are quite sensitive, as it's necessary to keep the PI reasonably balanced both clean and overdriven to keep the nice cleans and avoid harshness in distortion (according to my observations).



Andy

I believe this is the reason. Tolerances are just too small trying to coax some chime out of this dual triode

[passfan]

Tillydog are you british ? I just caught the 1000K which after a moment I realized was 1M. At first glance my mind registered 1000 ohm. Many thanks for the mod as I simply love the breakup out of this amp, I just miss the clean aspect. I will give these a try.

[M Fowler]

You built some nice amps

Mark

[amplifiednation]

Tillys got more experience than I do with this amp, and I've built two of them!! Mine needs some tweaks and I going to try the suggestions. For anyone who wants a taste of the express without having to attenuate, the 440 rocks. I get no clean at all though....but I have amps that do cleans quite well

Thx Andy

[Tillydog]

The 39/150k in parallel is 31k. Why didn't you just try a 33k like on the other side? (just curious)...

I didn't have a 31k and I did try 33k - I tried a lot of values, both in simulation and in real life. While the PI is clean, there's no problem - you have free choice of ratios. I played the amp for quite a while with 33k/47k and 39k/56k split loads (lowest value to B+) and while the cleans were good and the heavily distorted was OK, there was an ugly glitch between the two.

After a lot of testing, I think I tracked this down to the partial load PI being extremely unbalanced once the PI is over-driven - much more so than the standard PI. (FWIW my opinion is that the clean-to-mean transition is caused be the breakdown of the NFB loop as the output valves are driven into grid conduction. This causes the PI to be over-driven, hence its behaviour and headroom are key to the distorted sound - YMMV).

I could get a compromise of balance between clean and driven with the standard PI tail resistor (10k) - maybe 4% in each case, but the cleans had lost their 'jangle' and there was still this ugly glitch. I think this was because the lower level PI output clean became the higher one distorted (& vice versa), so there was a big change in PI behaviour at the onset of being over-driven. The only way I could find to minimise the change in PI balance between clean and driven was to increase the tail resistor and be quite picky about the split load values. From memory, the PI is balanced within about 2% clean & 4% driven, but the same output is the greater in each case, so there is only a small change between clean & over-driven behaviour of the PI.

I initially worked out the partial load ratios to try and replicate the behaviour of EL34 grids, but didn't re-calculate them after messing with the PI tail, with the result that there is a hair less drive than I intended, which means you can open the volume control up to 3:00, rather than ~1:00 If anyone feels like doing it, you could find some resistor values that give about 10% more output.

Note that the PI is now an ECC83 / 12AX7, rather than the 12AY7 in the original schematic.

You can see how I implemented it below. (The E'lytic-caps are ~10u and are between the bias voltage and ground, as I found that the bias voltage was wandering when the 6SN7 was driven to grid conduction, and generating a noise like a mosquito buzzing around your head! The big rotary switch is because I have an EL84 output stage built into the same chassis.)

Tillydog are you british ? I just caught the 1000K which after a moment I realized was 1M.

Yes, in the UK, but the 1000k is just out of habit with LTspice (which is where I grabbed the schematic from) because it recognises 1M as 1m - i.e. 0.001 ohm! It recognises 1meg OK, as long as I remember to type it.

This mod worked for me. I would be really interested to hear any feedback if anyone else tries it.

No guarantees

Andy

[passfan]

Note that the PI is now an ECC83 / 12AX7, rather than the 12AY7 in the original schematic.



Andy

Thanks for the heads up here. I made the modifications on the schematic (all but the presence and treble pots) and noticed a difference in the cleans and the overdriven sounds. I will have to try it with 12ax7's next. I noticed the drop in drive as well, which was okay in my opinion. Too much compression in the top end and I rarely strayed there because of it.