First question. Diodes.

Initially the kit had 1n4007 diodes. Later on, there are statements in the forum whereby the diodes to be used should be UF4007 henceforth. These are much faster than the 1n4007 and apparently better sounding.

Now in the parts list of the kit, the diode is RGP10M-E3/54. This seems to have slower recovery times than the UF4007. Steve, is there a sound quality difference? RGP10M-E3/54 is more difficult to find in the suppliers I am going to use - do I lose anything by switching to UF4007?

Thanks, Bobc.

Another question: I have 100uF 450V F&T type A capacitors. Given that we are using silicon diodes and the tube diode limit to capacitance doesn't apply, can they be used instead of the 33uF in the schematic or would that have a negative sound effect? (calculations indicate that ripple (and thus hum) will be reduced by a factor of nearly 10)

I still plan to bypass them with Mundorfs as per anniversary schematic.

(BTW: for Europeans: the best deals on Mundorf capacitors I could find were in don-audio.com. Mundorf Supreme 0.1uF at 9.22€ VAT included. I also used Mundorf MCAP EVO for the Hazen mod: 2.94€ and very small)

uFThanks!

I'll go 100uF - 1kohm - 100uF, bypassed by Mundorf supremes. These seem to be much smaller than the Jupiters and Miflex - the 0.1uF fit perfectly in their place on the board.

Also, as I had in the tool box several old unused black gates from long, long time ago, I will bypass the cathode 1000uF cap with a 100uF Black Gate. Just because.

(I will be using a Mills 12W resistor. I have calculated a little less than 6w dissipation on the resistor, and I don't want to put a big one there. I will ensure good ventilation and separation from the board.)

Edcors are very expensive to bring here to Europe (due to shipping) so I am planning to use Toroidy SE output transformers and Toroidy power transformer too. Claimed 10Hz-55kHz -3dB response!

I used the board (I have two, once I have finished the first one I will decide whether to bridge or not). Beware with the 2019 schematic - there are some lines missing. The 2011 schematic is correct and then you can figure out the anniversary mods on top of it. The boards are correct and give the right components values and positions. They also have the Hazen mods integrated. The only problem is the limited space with poly capacitors.

With the boards, I have made some changes (to save money and sometimes space):

• Changed the double .33uF+0.1uF for a 0.47uF Mundorf Supreme cap in all bypass locations.
• For the 3.3uF positions, I have used a couple of Mundorf Supreme 3.3uF that I already had.
• I have substituted the 33uF capacitors of the CRC filter for 100uF (bypassed too), hoping to minimize hum.
• I have substituted the two 20uF in series (= to 10uF) for a 15uF Mundorf MCAP 400V that I had in the toolbox. The two resistors in series do not have a task as a voltage divider between the capacitors anymore - they can be substituted by a single one working as bleeder resistor to avoid shocks next time you open the amp!
• Put a Black Gate 100uF in parallel with the 1000uF capacitor - just because I already had it in the toolbox. Bypassing a cathode bypass: let's see if it improves something.
• Used a couple of beautifully small 0.1uF Mundorf MCap EVO for the Hazen mod.

Hi, you can find it in this page, in the second post:

https://www.diyaudio.com/forums/tubes-valves/357130-quick-questions-zen-decware-...

The part that seems wrong to me in the new schematic is the capacitor bank for the input stage - which is correct IMHO in the 2011 version. Also there should be a line between the connection of the two series capacitors and the connection of the two series resistors.

The board is correct, BTW.

Hi,

Yesterday night I finished the amp. First I did some testing, starting with voltages. These were a little bit down, but the calculated currents were in line with the specified ones for each tube. This may be due to my using different transformers (Toroidy TTG-EL84SE, a toroidal SE OT model made for the EL84, but dimensioned for 30w).

Apart from this, I included all the Anniversary mods and the Hazen grid mod too (found a quite cheap place in Germany for Mundorf capacitors). In addition I went overkill with the power supply electrolytics with 100uF+100uF in the CRC filter instead of the 33uF+33uF. (Total and  absolute silence as result)

As box I took an old Hammond enclosure that I had used for a Tripath DIY amp 18 years ago. (That amp had been built to replace and improve upon a Zen84B. Biggest mistake in my audio life). I used the latest Decware circuit boards, which are truly excellent and properly used can handle all the mods. They include not only the Hazen mods but also provision for ultra linear connection - that I have not used. I used only half of the width of the enclosure - as I expected to bridge it later with another circuit if the sound was as good as I remember my old Zen84B.

I then tested the amp with a 4 ohm resistor on each loudspeaker binding posts, in the 8 ohm tap (to maximize output power)

• Frequency response was excellent IMHO, with -3dB points at 16Hz and 50kHz.

• Maximum output voltage: when the waveforms starts to change curvature (but still not truncated), 3.1v RMS, which amounts to 2.4 watt on 4 ohm.
  However, on the onset of wave "top flattening", voltage is 3.91v RMS, which gives 3.8 watt!

• Gain is 17.8 dB - quite useful.

I could not wait with more tests and hooked it to my loudspeakers. These are Gallo Strada 2, at 90dB/w/m, with 8 ohm and a comfortable impedance curve except in the extreme treble. To give them more "grunt" I used the Mapleshade stands, that keep them close to the floor to add bass by boundary reinforcement. I did not hook up the subwoofer to keep things simple.

I have owned Tripath amps, Musical Fidelity KW500 behemoths and Devialet 220 amps. The Devialet was used with the same speakers. By far the Zen is the best of them. Cavernous soundstage, with width, depth and height; and they tame the tweeter on these speakers that may have a harsh top end with lesser amps. (Tripath does not work well with them in my experience) It sounds just sweet!.

Regarding volume: I have not forced it yet but for classical music and not head banging rock and pop, it is perfectly adequate! (Boundary reinforcement may be adding another 3dB to the speakers' efficiency). I had bought a second PCB to be able to bridge it (and all parts but the transformers), but it looks like that I'll build another one in a different box for another room.

Now I left for the weekend and I will continue listening on Monday, and breaking in the amp (without breaking in it is already the best amp-speaker combination I've ever heard. With proper positioning, you are surrounded by music with just two speakers)

Thanks to Steve for making available the PCB and the design for the "sincere but broke" audiophiles!!

For the sake of completeness, the rest of the system is just a Mac running Audirvana with music from hard drive and Tidal, through Audioquest Dragonfly Cobalt, Kimber select interconnects and Mapleshade loudspeaker cables. So far, a nondescript power cable and no mains conditioning. Yes, this combo has bettered by far the imaging of the Devialet 220 I lived with for 3 years. The loudspeakers are on the floor with a mapleshade recommended 5ft distance to listener, 7 feet distance between them. No hole in the middle: rock solid center images with shape and volume. I tend to think that the formidable imaging is due to the zero-feedback and the near-omni dispersion of the gallo CDT tweeter.

A picture of the back and the front.

I substituted the two series 20uF capacitors (which equal 10uF) for a 15uF mundorf standard of the right voltage - 50% more capacitance

I replaced the 2x 33uF for two 100uF F&T - German capacitors with some prestige in Europe.

Due to space, I substituted the Mundorf supreme 0.33+0.1 in parallel for a Mundorf Supreme 0.47uF. I know it may not be equivalent from an extreme performance point of view, but close enough. These I soldered on the upper part of the board since they are small enough.

I had a couple of Mundorf supreme 3.3uF from a previous loudspeaker project - I used them to replace the 3.3uF Solens and avoid bypassing them with other supreme. I'd rather do this than expending on additional expensive caps when I already had this one!

I bypassed the 1000uF electrolytic of the cathode bypass with a 100uF BlackGate that I had from many years ago in the toolbox. Theoretically the effect of the 1000uF capacitor should affect above 2Hz - then the Black gate should impact above 20Hz

I put a 12W mills wire wound resistor. According to the calculations, the dissipation should be around 6W. Other resistors were far too big. I put it on the upper side and perforate the top plate to make sure it doesn't cook the board or the capacitors!

I used the beautiful 0.1uF Mundorf Evo for the Hazen mod. Look how small they look!

Replaced the 1N4007 original diodes for UF4007 as recommended by Steve in this board. I could not find exactly the ones referenced in the part list. I hope these are good enough.

Finally, the board seems made for the Mundorf Supreme 0.1uF - they fit exactly in place.

I used 25mm brass stands to attach the boards to the top plate - the selection of what is soldered on top was dictated by that 25mm clearance.

These pictures are prior to soldering a 470kOhm resistor replacing the two 220k in series for the bleeding of the power supply

Tomorrow when I return to that place I will run more tests with 4 and 8 ohm test resistors in the 4 and 8 ohm taps - all 4 combinations, and report on power, gain and bandwidth. I will also post FFT of all of them. I think that the 4 ohm resistor in the 8 ohm tap may be an extreme case: highest power but lowest bandwidth and highest distortion. Let's see.

Anyway for listening I am using 8 ohm speakers (which are above 8ohm until 5kHz) on the 8ohm tap.

I have spent now some time with the system. It took a lot of time finding the position of the loudspeakers, which is key to get out the most of it. In this case it is very near field: 5 feet from the speakers, with an angle of 90º (according to the instructions from Mapleshade), and 4 feet from the back wall.

The first thing is that 2.3W on 90dB 8ohm speakers is loud!!! very loud. No problem with the system volume with a normal 2v source. Working well at 50% volume (12:00). The source is a Khadas Tone 2 pro - which is excellent for the price and clearly better than the Audioquest Cobalt it replaced. Probably not enough for Mahler full blast, but clearly more than for "little girl with guitar". I plan in the medium term to use the second PCB to be able to bridge them and quadruple power. Have all the parts but the transformers.

The second thing is that, when properly set, the system "casts" an enormous soundstage, where the whole air is "energized" (in very special recordings). It sounds very 3D, with a lot of  depth and height. Very dynamic too. Much more dynamic than other multi watt amps.

Next step is to set up the turntable and a phono pre!

One month later, I opened the kit again.

In the cabin where I have the amp I have a very cheap chinese multimeter - cheap as $15 cheap ! The amp sounded fantastic but the voltages were off - much lower than expected (like 24v on the first reading point, totally off the bracket). The transformer was providing steady +-300v ac, but the DC readings were all low.

The amp was still performing flawlessly - best sound ever for me. I will come to this later.

I took a better multimeter from my home in Barcelona and rechecked the voltages with both multimeters.

With the good one, the voltages were spot on! 346-305-10.8 and 2.65. The total current drawn was 79 mA - 2.8mA in the small tubes and 36.7mA on the big tubes.

For some reason, also the oscilloscope reading were much, much clearer and better. The second harmonic on the FFT of 1kHz at 1W was -23dB down the first just after completing the amp - now it was a much better -45dB.

Also bandwidth was better on both frequency extremes. So much for break in !!

Curiously the second harmonic went much lower for higher frequencies, reaching -52dB at 15Khz.

The amp amazes me with the available power. While admittedly listening in the near field (5 feet) with the Gallo Strada on the Mapleshade stands (90dB), I have not managed to clip it even at uncomfortable levels.

I will post later some oscilloscope screenshots

Hi, I made a single input at the time, and extremely short distance between cinch/RCA and circuit; also, tightly twisted. Connected circuit ground to the mains ground too. I can hear no hum - just very faint 50hz from the subwoofer if I paste my ear to the woofer. Nothing at all from the listening seat.

I applied "brute force" to the power supply filters, by substituting the original 33uF (or 47uF) for 100+100uF mundorfs (1st and 2nd capacitor of the filter). Also bypassed with smaller pairs of mundorf Supremes, as per the anniversary mods. The big capacitors couldn't have worked with the tube rectifiers, BTW - only with the silicon diodes.

I am intrigued by your transformers - I couldn't find them in internet. Can you please give me more information? I have another circuit fully populated with Mundorfs capacitors and anniversary mods, pending only of transformers!

Hi,
ok,i just wondered what was meant in the assembly instruction with placing a jumper between ground plane and circuit ground.
The transformer website is https://www.tbtroehrentrafo.com/ and i found my model on ebay.

Hi There - I have the same question "what was meant in the assembly instruction with placing a jumper between ground plane and circuit ground."

I can't find any place to put a jumper for this connection when looking at the board or the schematic/assembly instructions?

Yes, the current schematic on the website shown in the zkit manual is correct, to my knowledge. I built mine using it and it works.
Keep us posted. Looking forward to seeing what you build.

News July 2023 - Balanced!

As I explained in the first posts, I had material for a second Zkit - except the transformers. I have not started the process for having a second Zen because I had no volume problems with the single unit, and I was perfectly happy with what I had. As I had explained, I had several highly ranked amps that were bested by the Zen - so no reason to change.

As I had a DAC with balanced outputs, and reading the opinions from Steve in the forum, I said "why not?" and got another set of transformers. To be on the safe side, and despite the power transformer was 200w, I also got another power transformer, to make sure that I had complete "dual mono".

I have connected internally the units as differential balanced - and now I can only use it as balanced. I have also a phono preamp for my turntable, but it also has a balanced output. I have drilled the casework and installed a couple of XRL inputs and cabled it accordingly.

I expected to have more power (not that I needed it: I don't party, and only on very loud periods I hear clipping) and little more. I also expected less hum: the way it is cabled, it all should cancel out (it appears on both speaker terminals). Also the linearity of the amp should improve with the opposing signal from both sides at the cathode common resistor - I think.

Results:

- Much more air and clarity.
- More bass ( for the time being, I have switched off the subwoofer since it is really not needed)
- Increased size of the soundstage. The most apparent thing is that closer instruments sound even closer; and farther instruments sound farther.
- More natural and effortless.
- Piano sounds now percussive! I had not realized it wasn't until I balanced the system. Records that sounded very good now sound very good and percussive - with more "slam". (The Zen alone has more perceived slam and dynamism than most amplifiers. I think that this is due to the lack of negative feedback. Well: with balanced, much more!)

In my opinion, for those that cannot afford a Sarah, differential balanced is the way to go. It is now more than twice better than the single unit (provided you have a balanced source)