I want to build an ultralinear KT88 amp and have narrowed my choice down to 2 designs:
http://www.drtube.com/schematics/grant/g60ams.gif
or
Integrated5050
The reason for choosing the Grant is that a friend has a similar amp (same topology, KT88s and different part values + both small valves are ECC82) that sounds really excellent, so other than the output iron it's a known entity.
However Patrick Turner is held in high regard and seems to be very straight talking and really know his stuff. So maybe his amp could be even better?
Why is the gain only ~11 though? Seems a strange choice!
My other question is about transformers...
Research indicates that Sowter and Lundahl are my best local(ish) options. I can get Hammond for about half the price but these seems to be less well regarded. Any other suggestions and how to choose between these two?
The Lundahls seem to be very small physically for the power ratings. Is this because they are C-cores? Is the C-core really an advantage?
Sorry if these Qs have been asked before but I've sifted a lot of threads and not really found what I'm looking for.
General: looking for 35 to 60W to drive 87db/W speakers with ~6 Ohm impedance.
Thanks,
Dave
http://www.drtube.com/schematics/grant/g60ams.gif
or
Integrated5050
The reason for choosing the Grant is that a friend has a similar amp (same topology, KT88s and different part values + both small valves are ECC82) that sounds really excellent, so other than the output iron it's a known entity.
However Patrick Turner is held in high regard and seems to be very straight talking and really know his stuff. So maybe his amp could be even better?
Why is the gain only ~11 though? Seems a strange choice!
My other question is about transformers...
Research indicates that Sowter and Lundahl are my best local(ish) options. I can get Hammond for about half the price but these seems to be less well regarded. Any other suggestions and how to choose between these two?
The Lundahls seem to be very small physically for the power ratings. Is this because they are C-cores? Is the C-core really an advantage?
Sorry if these Qs have been asked before but I've sifted a lot of threads and not really found what I'm looking for.
General: looking for 35 to 60W to drive 87db/W speakers with ~6 Ohm impedance.
Thanks,
Dave
well, I can't offer any comments on the circuits you posted....but...
Here are a couple of additional designs from this forum (in case you haven't seen them):
http://www.diyaudio.com/forums/tubes-valves/156699-mullard-5-20-kt88-pp-blocks.html
http://www.diyaudio.com/forums/tubes-valves/161702-opus-5-0-modern-mullard.html
Both are Mullard based PP KT88 UL circuits with a few modern improvements, and lots of commentary about the designs. Putting a CCS in the tail of the LTP makes a lot of sense. Looks like the Turner 5050 has one, as well as the two Mullard designs I just posted. The Opus uses mosfet followers for additional drive capability.
Here are a couple of additional designs from this forum (in case you haven't seen them):
http://www.diyaudio.com/forums/tubes-valves/156699-mullard-5-20-kt88-pp-blocks.html
http://www.diyaudio.com/forums/tubes-valves/161702-opus-5-0-modern-mullard.html
Both are Mullard based PP KT88 UL circuits with a few modern improvements, and lots of commentary about the designs. Putting a CCS in the tail of the LTP makes a lot of sense. Looks like the Turner 5050 has one, as well as the two Mullard designs I just posted. The Opus uses mosfet followers for additional drive capability.
Last edited:
I recently priced a set of James output transformers since I travel to Taiwan occasionally and could buy them direct and roll them home in my suitcase, saving the $100+ shipping cost. But the 100W P-P models suitable for the KT88 P-P Mullard amp mentioned in boywonder's post that I'm building, would end up costing a whopping US $375 each.
O'Netics transformers are much more reasonable, though of course they don't have the snazzy James box covers, but would otherwise be just as suitable.
I wish I could buy snazzy box covers for transformers...
..Todd
O'Netics transformers are much more reasonable, though of course they don't have the snazzy James box covers, but would otherwise be just as suitable.
I wish I could buy snazzy box covers for transformers...
..Todd
Dave,
The Grant employs Williamson topology, with the unusual selection of a 12AX7/ECC83 as the voltage amplifier/phase splitter. "Normal" for Williamson is a 6SN7 or an electrically equivalent type in both small signal "holes". The 12AU7/ECC82 is decidedly inferior to the 6SN7. The "Achilles Heel" of Williamson topology is low freq. instability. Absolutely top notch "iron" is a must in a Williamson implementation. Mullard style circuitry has far fewer stability issues.
Shipping costs for "iron" can eat you alive. If good stuff can be sourced close to home, that's where you buy. The Sowter Type U070 is close to perfect, in an "off the shelf" item. Contact Brian Sowter (support@sowter.co.uk) about ordering along the U070 lines, with a 3.3 KOhm primary impedance. The 4.5 KOhm primary of the U070 is certainly usable, but you'll need about 525 V. of B+. A 3.3 KOhm primary allows a 475 V. B+ rail to work well. It's a matter of step down ratio.
You should have little trouble in executing either Williamson or Mullard style circuitry with Sowter's excellent product, provided reasonable care is taken.
The Grant employs Williamson topology, with the unusual selection of a 12AX7/ECC83 as the voltage amplifier/phase splitter. "Normal" for Williamson is a 6SN7 or an electrically equivalent type in both small signal "holes". The 12AU7/ECC82 is decidedly inferior to the 6SN7. The "Achilles Heel" of Williamson topology is low freq. instability. Absolutely top notch "iron" is a must in a Williamson implementation. Mullard style circuitry has far fewer stability issues.
Shipping costs for "iron" can eat you alive. If good stuff can be sourced close to home, that's where you buy. The Sowter Type U070 is close to perfect, in an "off the shelf" item. Contact Brian Sowter (support@sowter.co.uk) about ordering along the U070 lines, with a 3.3 KOhm primary impedance. The 4.5 KOhm primary of the U070 is certainly usable, but you'll need about 525 V. of B+. A 3.3 KOhm primary allows a 475 V. B+ rail to work well. It's a matter of step down ratio.
You should have little trouble in executing either Williamson or Mullard style circuitry with Sowter's excellent product, provided reasonable care is taken.
Hi nic6paul,
O'Netics is Bud Purvine's company. (username BudP here.) He does high-end custom transformers for DIY'ers, but his bread and butter is supplying amplifier manufacturers with transformers. He's in Seattle I believe. And he does not have a website for his business, so I guess his hands are plenty full without a website.
And thanks for the tips about snazzy transformer covers. Love it!
..Todd
Last edited:
Yep, I've heard good things about Electraprint OPT's too. Though I understand he (Jack?) does not provide more than one secondary winding. I would want 4 + 8 ohm taps for any project I build. I have a multitude of various (DIY) speaker designs that I would use the amp with. Is he perhaps more accommodating than his reputation portrays?
..Todd
Last edited:
IMHO another interesting article that is worth reading http://www.next-tube.com/articles/Veen2/Veen2EN.pdf
Regarding trafos Choose a Transformer in Audio
MagneQuest cobalt core trafos are the most exotic ones AFAIK. Lundahl seems to be a good ballpark.
Thanks for all the replies!!
Another source is guiding me towards the GEC 30W KT88 self biased circuit:
http://www.jacmusic.com/KT88/SCHEMATIC3.gif
but using 2X GZ34 rectifiers per monoblock and low value caps - this is claimed to give potent bass but still sound very fast and alive.
The other claim is that 30W with the soft clipping effect of self bias will sound almost as loud as 50W of fixed bias.
Eli,
I spoke with Brian earlier, he seemd to think most people building KT88 PP used 4.5K to 6.6K primaries. 3.3K seems very low. What current would you bias at for 375V B+ and 3.3K?
Understand your comments on 6SN7 versus ECC82, there are some mixed opinions and some think there is not much in it sound/performance wise.
One thing I don't understand is your comment "The "Achilles Heel" of Williamson topology is low freq. instability. Absolutely top notch "iron" is a must in a Williamson implementation." Surely big iron is needed for LF stability, not neccessarily high performance i.e. high bandwidth iron?
Another source is guiding me towards the GEC 30W KT88 self biased circuit:
http://www.jacmusic.com/KT88/SCHEMATIC3.gif
but using 2X GZ34 rectifiers per monoblock and low value caps - this is claimed to give potent bass but still sound very fast and alive.
The other claim is that 30W with the soft clipping effect of self bias will sound almost as loud as 50W of fixed bias.
Eli,
I spoke with Brian earlier, he seemd to think most people building KT88 PP used 4.5K to 6.6K primaries. 3.3K seems very low. What current would you bias at for 375V B+ and 3.3K?
Understand your comments on 6SN7 versus ECC82, there are some mixed opinions and some think there is not much in it sound/performance wise.
One thing I don't understand is your comment "The "Achilles Heel" of Williamson topology is low freq. instability. Absolutely top notch "iron" is a must in a Williamson implementation." Surely big iron is needed for LF stability, not neccessarily high performance i.e. high bandwidth iron?
Nice articles. Many thanks.
Although when Mr Van Der Veen describes negaitve feedback as "nasty" it does make me wonder if he is really being objective.
Pick up a copy of "Valve Amplifiers" by Morgan Jones and read thru the section that deals with distortion.
jeff
Jack at Electra-Print will make your transformers with whatever secondaries you want. 
Every tube power amp I have made in the last 10 or so years has used Electra-Print iron.
I have had a few of his single secondary 6-ohm transformers, and let me tell you, the increased bass because of the ideal coupling is amazing. I am a believer!
The question is do you want unused copper in the core with a 'normal' multi-tap (Which is not ideal)
Or do you want fantastic coupling and bandwidth to a single secondary? (this is 'less loss' in my opinion)
Regardless what you decide, Jack makes a simply fantastic product.
Every tube power amp I have made in the last 10 or so years has used Electra-Print iron.
I have had a few of his single secondary 6-ohm transformers, and let me tell you, the increased bass because of the ideal coupling is amazing. I am a believer!
The question is do you want unused copper in the core with a 'normal' multi-tap (Which is not ideal)
Or do you want fantastic coupling and bandwidth to a single secondary? (this is 'less loss' in my opinion)
Regardless what you decide, Jack makes a simply fantastic product.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.