Son of Ikezilla: an Octal Eiclone

a DIY, modified Mullard 5-20 monoblock design

Son of Ikezilla: an Octal Eiclone

Postby Thermion » Sun Apr 17, 2011 7:10 am

I came across an unfinished project stashed in my basement a month or so ago and decided to bring it to some closure. Back in 2006 I built a PPP KT88 Eiclone amp that I named Ikezilla, and a couple of forum threads describe that prototype amp. My intent was to build a pair of those and I had almost everything I needed except a proper chassis and a few odds and ends.

So I threw them together using a couple of canary wood bases, 10"x12" top plates and some Hammond Iron. The OTs are 1650T and the PTs are 278Xs. That makes 21 pounds of iron per amp. The 1650Ts have an impedance of 1.9k, so hooked the binding post to the 4 ohm taps, and plan to use 8 ohm speakers and dummy test loads. This will give me effectively 3.8k.

I just finished wiring them yesterday and powered them up last night. Everything went well, my B+ is ~550 volts, and all the other voltages are within a few percent of my design values. I wired the boards so the voltage gain tube is an octal 6SL7 and the 6SN7 phase splitter cathode is current sourced.

Next step is to get the Boonton fired up and see what these will do. My plan is to bias the KT88s real hot. The target is 100 Watts at 1k <1%THD. I also want to get a quad of the new KT120s and see what they will do. There is enough capability in the 278X that you could bias the KT120s at 80-90 ma.

JT
Thermion
 
Posts: 181
Joined: Mon Jul 07, 2003 6:04 pm
Location: Fisher, IL

Postby Shannon Parks » Sun Apr 17, 2011 7:24 am

This will be fun! Still not convinced we won't burn up a test load. (666)

Shannon
User avatar
Shannon Parks
Site Admin
 
Posts: 3764
Joined: Tue Mar 18, 2003 5:40 pm
Location: Poulsbo, Washington

Postby dcgillespie » Sun Apr 17, 2011 8:01 am

I'm confused. So, you changed the reflected impedance because you are only using two output tubes per channel now (i.e. "son" title)? I'm assuming you are using the UL taps on the transformer as well.

If so, and your power supply will maintain 550 volts B+ at full power, then you should just be able to develop 100 watts RMS into your load. Producing < 1% THD @ 1 kHz at that power level should pose little problem.

Dave
dcgillespie
KT88
 
Posts: 399
Joined: Sun Mar 08, 2009 5:34 am
Location: Ball Ground, GA

Postby Thermion » Sun Apr 17, 2011 2:13 pm

Dave, you're not confused. You nailed it. I scrapped PPP for PP but I'm using the same OT and the UL taps. The downside is that it won't be too happy on a 3 or 4 ohm speaker. The up side is my recently acquired B&W 802s are a pretty easy 8 ohm load. Will post pics in a day or two.

JT
Thermion
 
Posts: 181
Joined: Mon Jul 07, 2003 6:04 pm
Location: Fisher, IL

Postby dcgillespie » Sun Apr 17, 2011 6:55 pm

Thanks for the clarification! There's another very big upside as well. Using the transformer as you plan to do will increase its usable power bandwidth and efficiency notably. All of that will be very good for power and distortion, and just make your final results even better.

Dave
dcgillespie
KT88
 
Posts: 399
Joined: Sun Mar 08, 2009 5:34 am
Location: Ball Ground, GA

Postby Shannon Parks » Mon Apr 18, 2011 5:44 am

Initial testing was 80W @ 1% THD @ 1kHz with the KT88s. Having issues with my function generator that kinda slowed testing down, and didn't do any sweeps with my software (just did it all by hand). I think a quad of KT120s are gonna be ordered. Trying to keep the B+ from sagging.

Man - those 1650Ts are huge!

Shannon
User avatar
Shannon Parks
Site Admin
 
Posts: 3764
Joined: Tue Mar 18, 2003 5:40 pm
Location: Poulsbo, Washington

Postby dhuebert » Mon Apr 18, 2011 8:08 am

Why don't you guys want to do PPP? My BFA makes 150 Watts easy into 8 ohms all day long. The power transformer does tend to run at 50C but everything else is cool. So far I have used up 2 sets of KT88s, one set of Sovtek and one set of JJ. Bass players complain that the BFA seems down in power so I take it home and measure it. Last time it was down to 50 watts. The Kt88s were completely exhausted! I'm happy with the JJs, they lasted longer than the Sovteks and retained their bias settings for the life of the tube, which the Sovteks did not. Anyways: 100 watts from PPP KT88s is easy work and will run cool and last for years. My $0.02 worth. Carry on.

Don
User avatar
dhuebert
KT88
 
Posts: 820
Joined: Thu May 01, 2003 9:26 am
Location: Winnipeg Manitoba Canada

Postby Thermion » Mon Apr 18, 2011 6:02 pm

Why don't you guys want to do PPP? My BFA makes 150 Watts easy into 8 ohms all day long.


Don, I agree with you. PPP is the way to go for tube power. My PPP prototype did 150 watts easy. At this point I just wanted to finish a pair of Eiclones, dust off the rust, and get back into the game. I had almost everything I needed collecting dust so this is what I came up with. Actually, I think they are gonna be pretty sweet amps.

PPP takes everything to a new level in terms of weight, size and structural stability. I still want to do it, but have some things yet to figure out.

JT
Thermion
 
Posts: 181
Joined: Mon Jul 07, 2003 6:04 pm
Location: Fisher, IL

Postby dcgillespie » Mon Apr 18, 2011 7:37 pm

How much was the B+ sagging? With a 3800 Ohm load, it will require ~ 550 vdc at full power to the output tubes to produce 100 watts RMS into a secondary load. To have decent passive regulation at the required current level, a voltage doubler supply is almost a must.

Dave
dcgillespie
KT88
 
Posts: 399
Joined: Sun Mar 08, 2009 5:34 am
Location: Ball Ground, GA

Dialing the SOI in.

Postby Thermion » Sun May 01, 2011 3:46 pm

Shannon and I went to a vintage vinyl sale yesterday morning, and afterwards decided to get some in some "tube time." So we went into my basement, threw on a newly acquired disc on the platter, and started tuning my new Ike amps.

The open loop gain measured 42.7 db, and we applied 18.3 db of global feedback. We then tweaked the cap values and step circuit along with the AC coupling cap to ensure stable performance. Our test amp had a Philco (Sylvania) 6SN7 in the LTP phase splitter, and the GC stage was a Phillips JAN 6SL7. The power tubes were a pair of JJ KT88s. They were biased to 64 ma each with a B+ of 537 volts. Voltages on the LTP and GC were 440 and 249 respectively.

Here is some preliminary data for an output of 1 watt:
Freq, OPV, IPV, %THD, Gain
20, 2.70, 0.17, 0.083, 24.0
1k, 2.82, 0.17, 0.077, 24.4
20k, 2.72, 0.17, 0.152, 24.1
100k, 1.24, 0.17, 0.62, 17.3

And for 10 watt output:
Freq, OPV, IPV, %THD, Gain
20, 8.56, 0.54, 0.175, 24.0
1k, 8.93, 0.54, 0.054, 24.4
20k, 8.64, 0.54, 0.735, 24.1
100k, 3.97, 0.54, 2.00, 17.3

And for 1% THD distortions levels as a function of frequency:
Freq, OPV, IPV, %THD, Power,
20, 12.5, 0.78, 0.99, 19.5
1k, 25.7, 1.55, 0.98, 82.3
20k, 10.2, 0.63, 0.99, 13

So, the 1%THD at 1K power is 82.3 watts. This is probably a good level for these amps with KT-88 outputs. The next step will be listening tests. I recently received a new quad of Tungsol KT-120s to try in these amps and see what they can do. Pictures to follow.

JT
Thermion
 
Posts: 181
Joined: Mon Jul 07, 2003 6:04 pm
Location: Fisher, IL

Postby dcgillespie » Sun May 01, 2011 6:27 pm

I guess I'm thinking that more power should have been developed before 1% THD was reached at 1 kHz. I have routinely used UL PP 6550s with Acrosound TO-330 transformers (3800 ohms, 40% UL taps), and produce 75 watts RMS from 20 Hz to 20 kHz with no more than 1% THD throughout that range, with 20 db NFB. At 1 kHz at this power level, THD is typically well under .2% This requires 485 vdc at full power to the output transformer. Granted, my supply is actively regulated, which certainly helps produce the low distortion results, and the Acrosound is a premium transformer, but I would think that the Hammond could produce better results than you are getting.

Many years ago, I developed a "production" amplifier I use for parties, PA needs, and just to knock down walls when needed. It uses the original Hammond 1650T before they added the UL taps. This amplifier uses PPP 6L6GCs with fixed bias, and an un-regulated conventional full wave SS supply of about 490 volts quiescent (~460 vdc @ full boil) for the plates and screens in (obviously) pentode mode. With only 10 db of NFB, it will produce 112 watts RMS @ .8% THD @ 1kHz.

The Hammond is a tricky transformer to work with however, requiring different FB networks for each output configuration -- at least mine did. And, it clearly performed best when configured for an 8 ohm output, as 20 kHz power output dropped an additional 20% at 4 ohms over what it already dropped to at 8 ohms (93 watts @ 2.90%). For production work however, it has been a real workhorse.

Although I've never worked with one, the newer version of the 1650T with UL taps is (I believe) purported to be a much improved transformer over the non-UL design, so I hope you can coax a few more low distortion watts through that puppy!

Dave
dcgillespie
KT88
 
Posts: 399
Joined: Sun Mar 08, 2009 5:34 am
Location: Ball Ground, GA

Postby Thermion » Mon May 02, 2011 6:03 am

so I hope you can coax a few more low distortion watts through that puppy!


Yes, I'm hoping so too. I'm using a 6SL7 instead of a 12AX7 which may be part of the problem. Also we could increase the global feedback a few db. More thorough sweeps need to be done across the frequency range and plotted.

JT
Thermion
 
Posts: 181
Joined: Mon Jul 07, 2003 6:04 pm
Location: Fisher, IL

Amp Pictures

Postby Thermion » Mon May 02, 2011 6:07 am

Here are some pictures of the project.
Image
Image
Image
Image
Image
Image

The bases are canary wood, and the top plate is finished in black hammertone.

JT
Thermion
 
Posts: 181
Joined: Mon Jul 07, 2003 6:04 pm
Location: Fisher, IL

Postby dcgillespie » Mon May 02, 2011 6:46 am

Very nice looking amps indeed!! The workmanship is first class for sure.

A couple more thoughts:

1. I've never been able to coax Sovtek tubes to produce the same power or distortion levels that GE 6550s will produce. This could be a significant factor.

2. It appears that your supply is rated for 200 ma, whereas the amplifier will typically pull over 300 ma in the territory you are trying to operate in. That means that B+ drop could be significant.

3. It appears you have your 1650T set up for a 4 ohm output, yet are taking FB from the 8 ohm tap. As I mentioned earlier, unless these transformers have changed significantly, your feedback will be much more effective when taken from the actual secondary configuration driving the load. To that point, the production amplifier I mentioned has a small mini switch installed to toggle back and forth between a 4 and 8 ohm FB network, so that the proper network can be in place as the load arrangement is changed to suit the requirements.

Dave
dcgillespie
KT88
 
Posts: 399
Joined: Sun Mar 08, 2009 5:34 am
Location: Ball Ground, GA

Postby Shannon Parks » Mon May 02, 2011 7:15 am

dcgillespie wrote:3. It appears you have your 1650T set up for a 4 ohm output, yet are taking FB from the 8 ohm tap.


Keen eye, Dave. These pictures must be from a couple weekends ago before Thermion rewired it. We are indeed taking the feedback from the 4 ohm tap (ie the same tap as wired for the speakers). In this config, we are treating the 1650T as a 3800 ohms to 8 ohms transformer (measured 4000 ohms). :))

Shannon
User avatar
Shannon Parks
Site Admin
 
Posts: 3764
Joined: Tue Mar 18, 2003 5:40 pm
Location: Poulsbo, Washington

Next

Return to eiclone

Who is online

Users browsing this forum: No registered users and 9 guests

cron