• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

My KT88 Williamson Amp Build

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
It was suggested that I open a thread on this beast I just built. It is a Williamson design with KT88s driving Edcor CXPP100-MS-4.2K.

The amp is actually two mono-blocks on a single chassis. The two power supplies only share the AC line. A delay circuit is used to delay start the B+ about 60 seconds after ignition of the power switch.

Tubes are Sovtek 6SN7s as drivers and Ruby KT88. I selected these tubes as "disposable" in case of catastrophic meltdown, but they have behaved flawlessly.

I have yet to make quantified measurements of distortion, power, frequency response, and phase relation, but our ears give it two thumbs up.

Stereo imaging is much better than the H/K AVR-525 that I had used for my system. I use the 525 as a preamp for now.

The outputs are set at 505 VDC at the plates and the fixed bias is set so the tube idles at 28 Watts of plate dissipation.

An externally hosted image should be here but it was not working when we last tested it.


An externally hosted image should be here but it was not working when we last tested it.


Love to hear your thoughts and suggestions.
 
congrats! One day I will get there.. still on my pre-amps;)

very nice piece. good job.

chassis is home built?

Partly. It is a Hammond chassis, but a 1/4" aluminum plate is added on top where the transformers are mounted and another 1/4" plate is mounted as an internal divider at the forward lip of the top plate for additional strength and to act as a electrical barrier between the power supply section and the audio section. The whole amp weighs in at 42 lbs.

The bias pots are mounted on another 1/4" aluminum block off of that center divider so that the pot shafts are just below the chassis top. There are three jacks that are for DVM probes to assist in setting the fixed bias. A star ground resides in the center of the power supply section under the delay board PCB.

Speaking of PCBs. There are four PCBs in the amp as you can see here:

An externally hosted image should be here but it was not working when we last tested it.


The center board is the delay circuit. A single PCB is used for the driver circuit, but the layout actually is separated down the middle so that the circuits are really independent, if that makes sense.

Under the stereo driver PCB is another aluminum plate that shields the PCB from the twisted pair filament wires. There are six short wires that connect the 6SN&GT tube sockets to the stereo driver PCB. The filament pins do not run to the PCB. There is also a blank PCB that acts as a shield from the AC switch.

The two LEDs on the front panel are as follows; one is yellow and indicates filament voltage is on and the other one is blue when the delay circuit has kicked in.

The delay circuit is a 555 timer that engages two Omron 5VDC relays (one for each channel) mounted on the delay PCB. Much to my surprise when I had those two relays top mounted they would not engage. The coil voltage was there, but they would not engage! Turns out that the coils are polarized and I was forced to mount the relays on the solder side of the board!

The front bezel is another 1/4" plate aluminum piece. All aluminum is type II black anodized.

A short story about this amp I call the Push/Pull T-Rex:

When I was in high school I was befriended by this really weird guy named Rex Reichert. He built custom tube amps and sound systems for local bands and lived in a hut on his parents property in southeast Pennsylvania. He even built equipment for Todd Rundgren. His last act was developing and marketing a board game named Nuclear Armageddon. Then he disappeared years ago and was never heard or seen from again.

Anyway, Rex taught me how to build tube amps and speaker cabinets. I built a number of tube amps while in high school, then stopped for a long, long time. Nostalgia got the better of me a few years ago and I decided to build another amp that I call the T-Rex.

Rex was a great guy, but his 200 Watt guitar heads were legendary for extraordinary bursts of flames in the middle of a performance. Unlike his amps, this one does not share the same pyrotechnic outbursts. :D

All of us that knew him miss Rex very, very much, but I am glad to provide this tribute for all to see today.
 
Very shocking. That was not the person I knew him to be at all.

When I was in high school I was befriended by this really weird guy

To me he sounds like that person you knew him to be. :)

I'm sorry, I couldn't help it.

That's a really nice amp. I was just looking at a similar schematic in the GEC amplifier book a couple of weeks ago and was thinking about building one myself. It's a nice straightforward design.

Regards,
John
 
That's a really nice amp. I was just looking at a similar schematic in the GEC amplifier book a couple of weeks ago and was thinking about building one myself. It's a nice straightforward design.


John,

Williamson topology is "classic". The big issue with the topology is instability due to phase shifts. It's good news indeed that Loren has successfully executed Williamson style, while employing moderately priced O/P "iron". This project will add to the lore of Edcor trafos being an excellent value.
 
Last edited:
Member
Joined 2005
Paid Member
Eli,

We have discussed a similar KT88/6SN7 amp on my Magnepan's on another thread- you had concerns over my OPT...Is the Edcor used here preferred over the Hammond 1650N I have?

Are you also in favor of connecting the 6SN7 plates via the UL taps here also as with the 1650N?
 
We have discussed a similar KT88/6SN7 amp on my Magnepan's on another thread- you had concerns over my OPT...Is the Edcor used here preferred over the Hammond 1650N I have?


John,

IMO, the Edcor CXPP100-MS-4.2K is preferable, if for no reason other than its 100 W. power rating. That rating implies the magnetic headroom needed, when GNFB is used, is available. The CXPP100-MS-3.3K might be a better choice with KT88 "finals".

The Hammond 1650Ns in your possession are (IMO) best mated with PP EL34s or PPP "12" W. types, like the EL84 and 6V6. The 1650N's 60 W. rating leaves little "wiggle room", when mated to KT88s. Doug Piccard's "exolinear" topology is the only thing I can think of, which allows you a 60 W. O/P using that "iron" and KT88s. How bad the freq. resp. aberations will be, I can't say. A no GNFB topology requires VERY linear O/P trafos and I have reservations about Hammond stuff used that way.

Even though both Williamson and Mullard style topologies are both "classic", I prefer Mullard. Mullard style is outright less vulnerable to instability than Williamson is. Loren is to be commended for a satisfactory execution of Williamson style, without employing extremely costly creme de la creme O/P "iron".
 
Last edited:
Greetings from France,

This seems to be a very good work, a nice KT88 Williamson push-pull.

I'm actually thinking of building my own KT88 push-pull, with parts of a defekt Music Angel XD800 amp.

I wish I can reuse the wood+metallic case, the 4 Shuguang KT88 tubes, and the 4 transformers (output + power).

For the rest I'm actually browsing Internet to find a good scheme, and wish to avoid PCB and do direct-wiring.

Could you help me chose the scheme ? What's all about this Williamson ?
What differs it from others ?

I have Tannoy Definition 500 speakers with coaxial 8 inches polypropylene speakers and 8 inches boomer.

Thank you really alot,
Regards
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.