stereo SE kt88 build ... abdellah diyaudioprojects design

[88man]

Yes, Jeff, I realize the 6N1P tube is too far back in pix. I am planning on placing the RCA jacks on the left side for a short signal path. I figured that having the rectifier closer to the caps might be better than having them closer to the PTX. I am not worried about heat as much as I am hum and noise.
You're right, I may have to submerge C1, since it's going to be determined experimentally, as I am planning trying out values from 10uF to 40uF to get a desired output of 400-420 B+. I need to get wider 2.6in x 0.75in red oak veneered wood.

Those copper ventilated socket plates look awesome! But, truthfully, I am going see how this first build turns out. I don't want to put the cart before the horse. If it sounds good, then I'll build a fancier looking build with more aesthetic chassis components - like the Derek Walton 300B SET amp.

[jmillerdoc]

Well, I got it all together last night. Haven't hooked it to a source yet. I found my B+ was for by a margin of about 35V (435V) from the PSUD2 simulation. I went ahead and changed the resistor before the driver stage on the power supply to a 30H 595DCR choke. This gets my B+ at the driver stage down to about 410V. I am going to add a resistor between the10H choke and the 80+82uf caps, gonna try about 20R to begin with and go from there to get the B+ down to about 400-410. I could have hooked it up tonight but I want to wait until I get the voltages right. Other than that all of the connections and switches work fine on the power tube cathode resistors (560+3K) and the selector for UL, triode, and pentode modes. I also had to reduce the resistor on my rectifier/filter on the filaments. I did the calculations for the proper drop and they showed that I needed about 1.3R to get the voltage down to 6.3 but with that resistance I was only getting 5v. In fact, I havent checked the B+ since I got that fixed. It should drop some now that the KT88's are conducting properly.

All in all, so far, I am very pleased with the finish. The chassis really turned out quite beautiful. It looks so much better in person. The Canary wood grain is deep and 3D in appearance. I haven't put it on a scale yet but I bet it weighs in at about 50lbs or more!

I am also in the process of hand carving a knob for the volume control out of madera and zebrawood. I will get some final pics by tomorrow (sans the knob, that will take a couple more days).

I will burn it in for a couple of weeks and tweak it a little then it will go up for sale on Audiogon, proceeds to fund my next project, the 300B MK1 mentioned earlier on this thread.

[jmillerdoc]

0

[Nikon1975]

Dear All,

I am also going to assemble this ampli soon. I have a few questions:
1) How come this schematic is so popular, while the one from Jean Hiraga is not ?
2) I see that people use various transformers, with an impedance from 2.5k to 5k. How should I make a choice on that ? (Actually this aspect is not clear to me in general)

3) I saw that the design of the power supply is kind of standard, and I found some improved version with the 10H choke splitted in two, for better common mode rejection. What do you think about ? What about a bleeder resistor ? (470k or so)
4) I also saw that in some schematics the HV+ is connected to the center tap of the 5V secondary (that I do not have). Is it ok if I use one of the 5V AC?
5) In the Hiraga design, I notice that he always put a small capacitor in parallel with a bigger one. What is the effect of this ? in theory they just sum up and adding a 0.47 to 200 will just fall in the tolerance of the big one.


Thank you for helping !!

Davide

[Nikon1975]

Dear All,

I forgot other two questions:

a) Is there any improvement in using DC for the heaters ? From what I understand it could make sense not to have AC signal running in the audio sector at all.

b)If I have a valve with a double triode, and I want to use only half, can I just leave the half I am not using disconnected ?

Thanks,

Davide

[jmillerdoc]

Davide,
You have some very excellent questions which I will try to take a stab at. Others, if you find a better explanation, please help me out here

Originally Posted by Nikon1975

1) How come this schematic is so popular, while the one from Jean Hiraga is not ?
2) I see that people use various transformers, with an impedance from 2.5k to 5k. How should I make a choice on that ? (Actually this aspect is not clear to me in general)

3) I saw that the design of the power supply is kind of standard, and I found some improved version with the 10H choke splitted in two, for better common mode rejection. What do you think about ? What about a bleeder resistor ? (470k or so)
4) I also saw that in some schematics the HV+ is connected to the center tap of the 5V secondary (that I do not have). Is it ok if I use one of the 5V AC?
5) In the Hiraga design, I notice that he always put a small capacitor in parallel with a bigger one. What is the effect of this ? in theory they just sum up and adding a 0.47 to 200 will just fall in the tolerance of the big one.

1) I can't tell you and I don't think there are any real psychics here but.....I can say that after just having completed mine.....It sounds awesome and it is very versatile. I imagine this may have a little to do with it. To be sure though, who can say for certain how many of each kind has actually been built.

2) I would stick with the original design of 5K. I went with Edcor CXSE25-8-5k with exceptional results, absolutely no complaint from their products. Me thinks a very good bargain at $80 each. Here: EDCOR - CXSE25-8-5K

3) Again, did it with the basic design above using Edcor 375-0-375 power transformer as follows: 5U4G to 15uF Solen to 100R (brings the B+ down to spec) to a 47uF Solen then to a Westinghouse vintage 10H 77R DCR choke followed by an 80uF ASC oil cap in parallel with a 82uF Solen (162uF total here) then to B+ tap. I also added a 30H 595R Hammond choke and 15uF Solen after the main B+ tap to the transformers. This filters the supply down before the driver tubes even further. I used only one filter for both sides of the driver stage where others have used x2 5H/10uF, one for each side. My voltages are spot on with this arrangement and ripple is essentially absent at the drivers and about 0.008V p-p at the transformers. The issue of common mode rejection is not addressed as well compared to split designs but it sounds perfect like this and is a lot less costly than adding all of that extra iron.

4) I am using a 5V without a CT just fine so, no it is not required. At the Diyaudioprojects.com website the schematic does not use a CT on the 5V.

5) As far as snubbing big electrolytics and even large oil caps, there is a lot to say about this on the Chipamp Gainclone thread, search gainclone and snubberized power supply, you should find plenty to read about there. Although the gainclone is a SS amp, I believe the principal apply everywhere.

And finally, your last 2 questions: I have done exactly both of these with excellent results. Whether or not it makes a huge difference in the noise level on this design, rectifying and filtering the filament supply has always made sense to me and there is much to be said about it elsewhere.

As far as using only 1/2 of the dual triode tube I did it for a variety of reasons. First for the ability to change the circuit to an SRPP 5687 driver stage later. Then I have the x2 tube sockets needed for this ready to go. Second if you are into matching tubes, it is much easier to match sections from 2 separate tubes than it is to match the triodes within a single tube, although they should be very close.

Hope this helps. See my layout at post #107 I think.

Jeff

[jmillerdoc]

Well, she is all put together and the results are simply spectacular. I have only listened to it with a pair of KT88 that were manufactured and labeled for McIntosh. Soon I will be changing these out for a pair of Svetlana 6L6-GC tubes.

I did change the grid leak resistor but not to 220K exactly. I had already ordered the 470K PRP resistors when I heard about the change. Fortunately I ordered 2 extra of all of the resistor for the amp. I simply paralleled the extras to the ones on th tag board for a value of 235K.

The sound of this amp has not yet stopped amazing me. I am all smiles when I am in front of it. I haven't had enough time to go through all of my music but there are some of my CD's I can't wait to listen to, especially the ones with vocals highlighted.

Triode is by far the silkiest and sweetest mode but the UL has tons more power and plays load well. The amp does not even clip at full power on triode mode, at least not to my ears. It just seems it wants to be pushed harder and harder....all the way to the top is sweet awesome honeylike sound.

UL is great too, a little more analytical but still warm and smooth. Clipping occurs at about 85% full power. It does not clip hard however and does not punish your curiosity with the control knob above 85%.

I did mine with the pentode option as well. Well, what is there to say really. It is inferior to the other modes but certainly is not bad by any means. It blows away my K-502 in this mode at about 25% volume. It is good indeed but when compared to the others, I will only be listening to this with some hard rock where power is helpful. About power....tons of it in this mode, I have not seen the ratings for it in pentode and am not sure what the wattage is but I would imagine more than 15 watts. It pushes my speakers way forward in this mode and is a bit more shimmering and bright. Clipping occurs at about 60-65% full power (visually on the knob). It is a bit more punishing when it clips in pentode. Where while in UL you can turn it to 100% and tolerate the distortion quite easily, you cannot do this in pentode. I haven't even tried. I like my speakers too much to abuse them like that.

As a final comment on this amps performance all I can say is....YES! Build it if you haven't. It is simply wonderful and the best money I have spent to date on an amplifier project.

In summary, while I have given this detail in multiple separate entries on this post, I will give the highlights of my build here all toghether.

OPTs - Edcor CXSE25-8-5K
Power transformer - Edcor XPWR134
Chassis - Hand-built with exotic woods including Rosewood, Madera, Canary and Dark Cedar with a 1/8" think aluminum plate for the tubes.
Connections - Vampire RCA and Binding posts.
Volume Control - Alps Blue Velvet 100K
Wiring - Silver on signal path and OFC copper on the power supply. RCA to pot and pot to driver tube is Silver shielded Silver 24ga Teflon coated wire NOS from NASA (yes spaceship wire) which is double shielded with copper mesh grounded to Main earth.
Coupling Caps - Jupiter 0.33uF Foil/Paper/Beeswax caps, original style that melt
Resistors - PRP 1watt throughout except Mills on the power tube cathodes and Xicon on the power supply (100R not in original circuit).
Bypass Caps - Black Gate STD series
Power Filter Caps - Eudya Oil (like ASC) 80uF and Solen 630V (82, 47 15x2)
Chokes - Vintage Westinghouse 10H / 77R and Hammond 30H / 595R
Filament Supply - Rectified and filtered with a 8A bridge rectifier and x2 4700uF 25V. Total filter modules #2, one for each channel (XPWR134 has x2 6.3v 3A secondaries)
Construcion - Single Chassis, symmetrical L/R layout, All point-to-point with Tagboards and eyelet standoffs, IEC filter input, Rocker switch, Blade fuse holder, Ceramic sockets, etc..
Options / Tweaks - Toggle selector for 470R to 560R cathode value on power tube. Rotary selector (non-shorting) for Triode, UL, and Pentode modes. Dual Driver tubes (in place of single tube) that uses 1 tube per channel (1/2). Modified power supply filtering (Described in post immediately preceding this one) Addition of a 1K resistor left unconnected for paralleling the 1K driver cathode resistor (with small jumper) to bring value to 500R for 6JD8 tube use in driver stage.

I think that is everything, at least that I can think of at the moment. More pics to follow shortly (wife has the camera) as well as some scope tracings if I can get my signal gen. working properly.

Again, no complaints at all with this build.......I LOVE IT.

Gonna go now......listen to it some more!

Happy Days to all,
Jeff
Lawton, OK USA

[88man]

It's always refreshing to hear good answers to good questions! Here is a dry fit layout of the SE KT88 amp. The top and bottom chassis are drilled and the oak base is made. I am going to place the first ASC cap in the CLC PSU inside the cabinet due to space and will determine which value will get the B+ to 400-420VDC. I have 10uF, 15uF, 20uF, 30uF, and 40uF ASC values from which to experiment.

I am trying to get away from another LC filter before the driver plate. Instead of an LC filter, some have mentioned the use of an RC filter (e.g. 100R + 47uF, or 250R + 10uF). I am not sure which will work better which set up will work better?... I want low noise and hum, but not at the expense of dynamics.

I've included a schematic based on the various versions introduced by a Michael Abdulla, Alexg, Porkchop, and Kegger of this SE KT88 Amp. Some have the 1kR before the KT88 and some don't... Again, I am not sure what the difference is, except I've seen more tube circuits with a 1kR grid resistor more often than not. (Oops, forgot to draw gnd).

This is an easy circuit, but the several versions of the schematic make it confusing for the first timer like myself. Here are just a few variables I've encountered on this circuit alone:

1. 330kR plate-to-plate resistor or none
2. (2) 2V LED or 1kR + 100uF cathode bias
3. Driver input filter: LC, RC, or none. 2 filters or 1 shared filter for each channel of 6N1P.
4. 1kR input grid resistor for KT88 or none
5. 1200R output grid resistor for UL or none
6. 100R output grid resistor for triode, UL, and P modes or for only triode mode.
7. 470k or 220k KT88 grid resistor to ground
8. Standby switch or none
9. Input impedence of OTP 2.5kR or 5kR
10. 1M input protection resistor or none

[88man]

Excellent detail and description, Jeff as our threads were minutes apart. My layout doesn't have as many LC networks in the PSU as your build. I hope a CLC design will be enough to suppress and filter noise? I am encouraged to hear about the great sonic results of your build.

[jmillerdoc]

That is hilarious, we posted within minutes....probably both pecking away at the same time on our keyboards. I just happened to hit submit first.

Just stuck a pair of 6L6 tubes in.......trying to get a feel for them. They are different, not better, just different.....at least for the 5 mins I have listened to them.

I also tried the 6JD8 tubes without dropping their cathode value to 500R. It works but sounds like it is in a very non-linear portion of the curve. Almost like hitting a reverb switch (though very mild). I can tell the treble is more sharp with them though, can't wait to parallel the 1K to get the value down. I will report on this tomorrow, gotta go to work now (suck).
Jeff