Some Ikezilla Questions

a DIY, modified Mullard 5-20 monoblock design

Some Ikezilla Questions

Postby kt88pppamp » Sat May 14, 2011 10:22 am

Does the Ikezilla KT88 quad amplifier take feedback off of the 8 ohm tap off of the 1650N or the 16?

Are the values you used for C7, R8, C8, and R49 the same as here for Ikezilla?

http://www.diytube.com/phpBB2/viewtopic.php?t=891

Should I go with a 12BH7 instead of the 6SN7?

What were the final THD readings?
User avatar
kt88pppamp
KT88
 
Posts: 530
Joined: Tue Sep 26, 2006 12:11 am
Location: Rochester, NY

Some Ikezilla Answers

Postby Thermion » Sun May 15, 2011 12:34 pm

Does the Ikezilla KT88 quad amplifier take feedback off of the 8 ohm tap off of the 1650N or the 16?


In the previous Ikezilla PPP prototype, the binding posts were connected to the 4 ohm taps on the Hammond 1650T and the feedback was then taken from the 8 ohm tap. I had the binding posts on the 4 ohm taps because I was planning to use the amps with my Apogee Centaur speakers which present a nominal 4 ohm load.

In the current Son-of-Ikezilla PP project, I am using the same Hammond 1650T outputs with the binding posts connected to the 4 ohm taps. I am using these amplifiers with a pair of B&W 802 speakers which present a nominal 8 ohm load. Configured in this way, I am effectively doubling the primary impedance of the Hammonds from 1.9k to 3.8k. This is a better load for a pair of KT-88s than the 1.9k would be. The feedback is taken off the binding post (effective 8 ohm tap). So the feedback is taken off the same tap as the speaker connection.

Are the values you used for C7, R8, C8, and R49 the same as here for Ikezilla?


For the current Son-of-Ikezilla project, the final feedback values are, C7=680pF, R8=2.2k, C8=330pf, R49=3.32k.

In the PPP Ikezilla prototype, I used a RevA board that I modified to include the step circuit and LTP current source. I looked and did not find my notes from when Shannon and I tweaked the feedback values. Inspection of the prototype did not give a clear answer. If you want these values, I can take the time to look further. However, I am not sure we optimized these values at that time. I have learned a lot since then.

Should I go with a 12BH7 instead of the 6SN7?


I don't have any personal experience with the 12BH7 tube. I have used 6SN7s and 6CG7s in several projects with very good results. The 12BH7 looks like it would also be an excellent choice in the LTP. I think Shannon is planning to use this tube in a project.

What were the final THD readings?


I haven't done any further sweeping except to confirm the previous results after the feedback values were installed. (82 watts, 1 khz, 1%thd) I am currently using these amps and have heard nothing but very nice, transparent sound at all sound levels and music types over the past week. Today, I did the final component change replacing the Solen 0.22uF coupling caps with Multicap RTX 0.22s. ( Just arrived yesterday from Handmade Electronics.) These amps seem to be a near perfect match to the B&Ws.

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

Postby kt88pppamp » Sun May 15, 2011 1:09 pm

Are the stock IKE values on the most recent revision a good place to start when tweaking the FB for PPP?

I have access to a scope, function generator, and a spectrum analyzer. I don't think the function generator has a sweep capability, but I have to look.
User avatar
kt88pppamp
KT88
 
Posts: 530
Joined: Tue Sep 26, 2006 12:11 am
Location: Rochester, NY

Postby Thermion » Sun May 15, 2011 3:37 pm

Are the stock IKE values on the most recent revision a good place to start when tweaking the FB for PPP?


The stock values will at least get you in the ballpark. The values depend on several factors including the choice of tubes and output transformers. Tuning an amp can get complex. You have the tools to tweak the feedback pretty effectively. First, disconnect the feedback completely and measure the open loop gain of the amp with a resistive load on the output and a low level sine wave at 1khz on the input. Adjust the input until you get around 1 watt output. For any of the Ike based amps, the open loop gain should be around 40db. Use a 20k or larger pot and put it in place of R8. Adjust the pot until you have ~20 db of feedback (reduction in gain). Remove the pot and measure the value. Look for available R8 resistor values that will keep you within 18-20 db of overall feedback and install.

Next, change to a square wave input at and observe the output response on your oscilloscope. Normally, you would expect to see ringing on the square wave response of the output. Try different values of C7 (or a variable cap) in parallel with R8 until the ringing is reduced, but not eliminated. Measure the frequency of the ringing from the OScope display. Use a value of 3.3k for R49 and calculate a value of C8 that will attenuate the signal below the ringing frequency. Install and look at the square wave response. If you have a nicely rounded corner on your square wave with little or no ringing. You are probably done, although there could still be problems with highly reactive speaker loads.

If you use the current source on the LTP and have AC coupling between stages, you will need to look at the low frequency stability. With a low level sine wave, sweep the frequency below 20 hz and look for resonances. Adjust C1 to solve any problems. In my Son-of Ike amp, I used a value of 0.68uf for C1( AC cap between the GC and LTP stages).

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

Postby kt88pppamp » Sun May 15, 2011 4:59 pm

Have you heard of those resistance-capacitance decade boxes? They are expensive but RIT provides them to us for doing our lab work. Since I am doing my amp through their EE department, they are providing me with some pretty nifty gadgets.

A decade box would prove much better than a pot for this purpose, cause you simply can just dial in the right value and wallah!

I use them all the time for transistor and op-amp based projects, yes I am mostly a solid state wiz, but for living room hi-fi, tubes rule. Anything professional, you pretty much have to go solid state and digital for the headroom.
User avatar
kt88pppamp
KT88
 
Posts: 530
Joined: Tue Sep 26, 2006 12:11 am
Location: Rochester, NY

Postby kt88pppamp » Sat Feb 18, 2012 6:00 pm

OK, its been a while but I am still working on my PPP KT88 amp. I just passed to point where I monitored the square wave response on the oscilloscope. With no capacitor across the feedback resistor, I can see no ringing. The square wave is almost perfectly square, only with a very small "rampoff" effect.

Is this a miracle or am I missing something?

The driver stage I am using is a 12AY7 in SRPP and a 12BH7 differential phase splitter.

I performed my testing with an 8 ohm wirewound resistor.
User avatar
kt88pppamp
KT88
 
Posts: 530
Joined: Tue Sep 26, 2006 12:11 am
Location: Rochester, NY

Postby Shannon Parks » Sun Feb 19, 2012 5:43 am

kt88seamp wrote: I just passed to point where I monitored the square wave response on the oscilloscope.


With a 10kHz square wave? Also, what is your open loop gain and your closed loop gain?

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

Postby kt88pppamp » Sun Feb 19, 2012 3:46 pm

I placed a 47pF cap across the 10k ohm feedback resistor. That took care of the ringing. I don't know if the 3.3k ohm resistor and capacitor in parallel with the 1MEG resistor, between the upfront stage and phase splitter, is necessary.

Image

The waveform looks pretty clean. Adding capacitors to the 3.3k network makes the ringing worse.

What do you think?
User avatar
kt88pppamp
KT88
 
Posts: 530
Joined: Tue Sep 26, 2006 12:11 am
Location: Rochester, NY


Return to eiclone

Who is online

Users browsing this forum: No registered users and 13 guests

cron