JTM45 build

I've been reading and researching for a long time now. First I looked into building a JCM800, but thought I didn't have enough time to actually built it. After that I thought I might build a 18W clone. But now I finally bought myself a JTM45 clone kit!

I would like to alter a few things and as advised by robrob's fantastic website I'm planning to post the schematics here hoping I'll get some help from you guys and girls.

In the mean time I've build a amp cradle. Below is the image as a teaser for what's to come!

 

I've got the first iteration of the altered schematic ready. It might look like a lot has changed, but it boils down to:

- Limit noise: 'preferred input network' from Blencowe.
- Split inputs to prepare for 'one wire mod'.
- Lower most capacitors to tame the bass. Cutoff frequencies around low E (82 Hz). Should reduce blocking distortion also (Blencowe).
- Arc protection for Cathode Follower (Blencowe).
- Generous use of grid stoppers (Blencowe).



Red dots mark a crossing, not a connection.
Feedback phase might be wrong. Will follow manual for that.

Looking forward to your (positive ) feedback!

 

And now the turretboard layout:



- Grid stopper moved to tubes.
- Bias circuit still a bit rough...

 

Can you give any insights into the cause of that? Cutoff frequency is still well below the low E. Bias excursion is milliseconds in both cases.

Lead spec... Can you please explain?

 

Here is the result of my homework assignment

Input: Yes, input looks a bit more like the Lead spec. My cathode bypass capacitance can be easily increased by putting another capacitor in parallel.

2nd gain stage: Here the Lead spec has a .68uF cathode bypass capacitor. This will increase the gain for the higher frequencies. I follow Bass spec here.

Tone stack: Here the Lead spec actually some more bass and middle . I follow the Bass spec here.

(output) Coupling caps PI: I have even smaller capacitors here than the Lead spec. Cutoff frequency is however lower than the Low E for all versions. The same holds for the other coupling caps I made smaller.

As for what you call soul: I can think of two ways how the lower caps can alter the 'soul' even though the cutoff frequencies are lower than the guitar frequency spectrum.
1) Lower frequencies (than the Low E) could arise from intermodulation distortion and these can 'suffer' from the lower cap values.
2) Bias excursion and recovery (Orig: 7, 29ms vs. Mine: 1, 3ms) also differ. This change is intentional however; it limits the chance of blocking distortion. I follow the advice of Blencowe on this one (and several others...). I admit, he focusses on a scientific approach and questions some common (scientifically unfounded) knowledge amongst guitarists. I guess I'm that kind a guy too, however I'm impressed at what a good set of ears can hear (e.g. Rob Chapman in the blindfold challenges).

 

I made LTSPICE models of the JTM45: one with the original cap values, the other with the new cap values.

 

Ok, I'm lost...

Below is a Bode plot of the complete amp for the original (old, normal channel) schematic, Ch. I (Normal) and Ch2. (Bright) as calculated by LTSPICE of my adjusted schematic. The smaller window on top of it is just the input stage (with formula from Radiotron Designers Handbook Rev. 4, p. 484).



Why is the 'new Ch. II' actually above all others just above 100Hz, before diving down as the frequency goes down?

Can anyone explain this?

 

An update...

First, the (bottom) graph in the previous post is wrong. I had a DC power supply with filter resistors, but no capacitors...

Second, I've been changing a lot of parts and simulating the result. Based on the advise here and parts availability I ended up quite close to the original now, apart from the input.
I've made a layout in diylc. It was a lot of work and I (of course...) found a chance to change things.
I tried to implement a multiple star grounding scheme. I think it looks very similar to Larry's Grounding Scheme (http://www.tdpri.com/threads/dc-heaters-on-jtm-style.884748/#post-8690391), apart from the fact that I have only one connection from the circuit to the chassis near the input.
I also moved a filter cap further (physically, not electronically) towards the input stages as it makes the wires to the grounding point shorter. Is there any negative side effect to this?



Legend:
Red wires: B+, Line
Black wires: ground
Green wires: grids
Yellow wires: cathodes
Blue wires: plates
Gray wires: other

Black turrets: present
Red turrets: present, but to be removed
Blue: new, to be added


Comments, tips, remarks, criticism; it's all welcome!

Ps. I couldn't upload the diy file, but if anybody wants it I can send it to you.

 

Finally did some practical work again. I mounted the chassis-mounted bias and humdinger pots. Also made a chassis connection dedicated for the line ground (if that's how it's called).
Besides that I think I have doubled my soldering experience soldering the heater wire. Getting two wires through the pre amp tube socket lugs was no easy task.



Given the nearly non-existing experience I'm pretty happy how it turned out.

 

I got the kit from ToneFactory (NL), but its very similar to MableAudio (China), Jukebox Revival (NL) and BYOC (USA?)

 

Time for my weekly update:

Slowly, but steadily, worked on the amp this week. I've mounted the face plates together with all the dials, switches and jacks. Besides that, I made a dummy load for future testing of the amp
I've also been soldering mostly around the power transformer and output transformer. I've intentionally left out the standby switch as there is no need for it. I had already ordered a new 500V cap for the supplied 475V cap to withstand the full voltage when the valves are not conducting yet.

 

I've been waiting for some parts in the mean time. They have arrived.

I mounted the components (minus 1x 100nF cap, 1x 22nF cap and some jumpers) to the turretboard. Before I start soldering I would like to ask you guys if you spot any errors. I've seen some impressive examples lately on this forum!