PP KT88 Mono blocks?

[TubeMack]

Here is a quick qestion. Do you find yourself running out of power with your St-70?

I should have said, your sweet st-70

Yes. It doesn't quite get there. Also, I don't get the graceful clipping I hear about. It's nice up to a point then suddenly sounds very harsh?

 

[ray_moth]

Harshness at clipping is not unusual with NFB, especially with pentodes, even if they are in UL.

 

[waltube]

Hi,

in attach you can find a schematic of a monoblock that I have published on Audioreview magazine in Italy.
It can be configured in UL. triode and Pentode.

input:
http://www.multitask.it/diy/input.JPG

out:
http://www.multitask.it/diy/out.JPG

It has a Sowter 3575 in as splitter in input.
The outp. trafo sa a single wiring at 3 ohm , the primary is 800 ohm a-a

It reaches 200w in Pentode, 170 in UL and about 100 in triode, rms; more impulsive.
The filter capacitor are 4 pcs of Safco 330uF /500 Vdc, total is 1320 uF at 450 vdc

With open loop the response is very good; I decide to use a switch for 6 dB of FB.

Ciao

Walter

 

[athos56]

Now thats a project

 

[Bandersnatch]

hey-Hey!!!,
I had first took on a similar project( 5 pairs of EL34 on an 800 Ohm a-a OPT. I could not devise a simple driver circuit capable of driving the multi-parallel grid resistors well enough( without putting another power-capable stage in there ). What are you using for grid circuit value? Driver plate load and tube type?
cheers,
Douglas

 

[waltube]

hi,
some additional info:
Input stage
R1, R19= 150R Holco H4 R3, R20= 1.000 ohm Holco H2
R4, R2= 1.500 ohm Holco H2 R5, R6= 5.100 ohm Holco H4
R7, R9= 100.000 ohm Resista WK4 R8,R10= 22.000ohm / 4w resista WK2
R11, R12= 1.000 ohm Holco H4
R13, R14= 220.000 ohm Holco H4 R15, R16= 47.000 ohm / 4 w Resista WK2
R17, R18= 2.200 ohm / 4w Resista WK2
C1, C2= 0,22 uF / 630Vdc polypropilene
C3, C4, C9, C10, C11, C12, C13, C14= 0,68 uF / 630 vdc polypropilene
C5, C6, C7, C8= 10 uF /500 Vdc assiali
C15, C16 = 220 uF/ 25vdc low Esr
TR1= Sowter 3575

This is the pair of amp:
http://www.multitask.it/diy/amps.JPG

I will try to post the tests.

Ciao

Walter

 

[TubeMack]

Wow, looking at the schematics I find that amp may be a little over my head! I'm having trouble even following it. It certainly has no shortage of drive ability. Very nice!

 

[Bandersnatch]

Wow, looking at the schematics I find that amp may be a little over my head! I'm having trouble even following it. It certainly has no shortage of drive ability. Very nice!

hey-Hey!!!,
There are missing parts, so don't feel bad...

I'd qualify it as having no shortage of drive requirement; even if one goes to twice the maximum grid resistance value, that power stage is presenting each phase of the driver with a 25k AC load( at best ).
cheers,
Douglas

 

[waltube]

Hi

this is a brief summary of a CCL.
CCL is a test we made at Audoreview for a limit caratheristic of amp conntected with different load.
It is automatic (therei is a home-made program) with an Audio Precision One who drive the amp with impulsive signals:
load 16 ohm: 114,93 watt
Carico 8 ohm: 181,1048 watt
Carico 5,3 ohm: 192,3417 watt
Carico 4 ohm: 185,7374 watt
Carico 3,2 ohm: 174,4531 watt
Carico 2,7 ohm: 162,269 watt
Carico 2,3 ohm: 153,6463 watt

Ths is in UL; it very good because the power is increasing from 16 to 4 ohm and at 2,3 ohm is very high.
I will change something on main trafo; will be bigger with 10% more of Vdc.

Ciao

Walter

 

[rknize]

Originally posted by TubeMack
Yes. It doesn't quite get there. Also, I don't get the graceful clipping I hear about. It's nice up to a point then suddenly sounds very harsh?

Yes, my (VTA) ST-70 is similar. You are getting some of the soft clipping when the output stage is occasionally being stressed to drive your speakers, but short of driving the whole amp into clipping. As Ray mentioned, once the amp starts hitting the rails consistently (as seen on a scope), the NFB seems to be what pushes the amp over the edge. The feedback tries to compensate for the clipping by driving the amp even harder into clipping. The sound from your typical transistor amps (well mine anyway) seem to crumble much earlier than this...maybe because they have even more NFB. My ST-70 can get as loud or louder at 35 WPC and still sound good than my reasonable-quality-but-old 70 WPC transistor amp can. Those are 70 "honest watts"...it's more like a 120WPC amp by today's measuring standards.

I didn't really experience this until I built a Tubelab SE with only 2WPC. I could drive it into clipping (as seen on the scope) and it still sounded pretty good. Only when it was driving well into clipping would it break-up like the Dyanco. The NFB factor really didn't occur to me until George mentioned it in one of my threads.

Years ago I was in your boat trying to build a 100WPC tube amp and I collected most of the iron and other parts to do it. I've since changed my approach and instead got a set of efficient speakers. Now the Dynaco is way more power than I could ever need unless I wanted to hear my system clearly from a block away. The tubes reserved for that big amp are now being used in a small SE amp.

 

[athos56]

Years ago I was in your boat trying to build a 100WPC tube amp and I collected most of the iron and other parts to do it. I've since changed my approach and instead got a set of efficient speakers. Now the Dynaco is way more power than I could ever need unless I wanted to hear my system clearly from a block away. The tubes reserved for that big amp are now being used in a small SE amp.

This is what I did, built a pair of speakers using high efficiency speakers out of pro drivers (BMS compression driver followed by a B&C 8" mid and 8" RCF woofer) My 8 watt Kt-88 SET never goes above 1/4 volume, I'm thinking that a Tubelab 45 will be plenty. If I ran my ~16 watt Moskito at full output I think I could play a small venue . Those drivers are rated for more than 100 watts but I think I'd bring my shoddy apt building down in the process. More efficient speakers are always an option, might even be cheaper

 

[TubeMack]

Later on I would like do a small complete SET system, But as mentioned, iI want mono block this time around to have the flexibility to run Maggies as well as some other tough combos.

No one other than Waltube has thrown out any schematics up to now, so let me put this out there!

What if I took the same (easy) route I did with the ST 70, and base the amps very loosely on Dynaco Mark 3's?

I could use the "Posiden" Driver PCB whish really is nothing like the Mark 3 driver. Upgrade the power supply using the SDS PCB, Pehaps using the SS rectification option. And finally ugrade to some custom Electra Print Trannies? All this would go into larger custom chassis.

Would this completely knock me out of CJ / ARC territory?


I'd also like some comments regarding the posiden PCB. It a modified (EICO) classic Mullard LTP circuit with a ccs in the tail.
Here the PDF with the schematic and notes:

http://www.diytube.com/unidriver/poseidon.pdf

 

[rknize]

80-100W is achievable with KT88s/6550s and friends in your "standard" Williamson-decended PP designs, which are all over the Internet. You can either push a pair to their pure-pentode limits or back off a bit and go with parallel UL pairs for each channel (this was the direction I was going). 8 tubes just for the output stage plus the needed iron and everything else turns real expensive real fast. If you want to go triode, you are talking even more tubes. I'll dig around to find the schematic I was using as a starting point.

Take a look though the "Dusty Files" and the Schematic Index at Triode's site for lots of vintage designs:

http://www.triodeel.com/schindex.htm

Obviously you can do better with what is available today...I just wanted to make the case that maybe it's not worth the effort.

Russ

 

[Eli Duttman]

The PROBLEM with Williamson topology is stability. The O/P "iron" MUST be "super-duper". It's way too easy for the Williamson style circuit to oscillate, instead of amplify. Can you say phase shifts?

The Mullard style circuit I suggested early in this thread is a lot less vulnerable to stability trouble. A very modest amount of global NFB is necessary, when UL PP KT88s are driven by a 6GK5/12AT7 small signal combo. THD will be tolerable open loop. GNFB is used more for damping factor considerations, than any other reason.

BTW, the KT88's control grid is tough enough to tolerate positive current. Augmenting a LTP with source followers DC coupled to the KT88s, as described in Tubelab's PowerDrive and MOSFET Follies allows Class "AB2" operation and some extra power O/P.



edit: added AB2 stuff

 

[TubeMack]

Uh oh my thread is dying

At this point i'm pretty set on a modified Mullard LPT design as suggested by Eli. This is what was used in my first build so I know the circuit. I'll follow Eli's advice and skip the PPP and go with with a deuce setup to give me the best chance of success.

Now
Do I go P2P, or take the easy route of using the Mullard derived Posiden PCB?

Did anyone get a chance to check out the schematic for it that I posted above? I'm very curious about any thought on the IC choice for the CCS in the tail?

I know some don't like the basic choice of triodes for this PCB but other options are possible with this board.

 

[boywonder]

Have you looked at Triode Dick's Bill and Mono Bill?

http://translate.google.com/transla...h?q=triode+dick&hl=en&rls=com.microsoft:en-us

You'll have to click on the "projects" link, as the URL as posted here doesn't quite get you there.

Using the google translator is interesting........

This looks like a PP UL circuit with a CCS for the KT88's, although I don't know what the output power is. There is a stereo version (Bill) and a mono version (Mono Bill) on the site with fairly good documentation if you can get past the translation

 

[Bandersnatch]

Uh oh my thread is dying

At this point i'm pretty set on a modified Mullard LPT design as suggested by Eli. This is what was used in my first build so I know the circuit. I'll follow Eli's advice and skip the PPP and go with with a deuce setup to give me the best chance of success.

Now
Do I go P2P, or take the easy route of using the Mullard derived Posiden PCB?

Did anyone get a chance to check out the schematic for it that I posted above? I'm very curious about any thought on the IC choice for the CCS in the tail?

I know some don't like the basic choice of triodes for this PCB but other options are possible with this board.

hey-Hey!!!,
I'd go for a cascode FET or 10M45 for the CCS. I'd also do the bias network differently; the 100k grid resistors should just go to the wiper of a pot so you can adjust the idle current of each tube separately and w/o disturbing the other.

No way I could endorse the 12AU7. If you rig for 6CG7 you can also try 6H30( drop in; the CCS takes care of the operating point).
cheers,
Douglas

 

[waltube]

Hi Tubemack,

you can use my schematics with only one quartett of EL34/channel; you can get about 80 watts. It will be easier.
The input circuit will be the same.
I suggest to use the Sowter 3575; in this way you have a perfect splitter.
- here you can see the two output of 3575
http://www.multitask.it/diy/3575-2.jpg

-here there is a spectrum of both signals
http://www.multitask.it/diy/3575-3.jpg
it seem to be one line but are two , due to the perfect symmetry ; the distortion is absent.

I look on the schematic of Poseidon and looks not so good.
It has a low pass filter between the first stage and the input of the splitter; I can't understand why. It is no normal.


Ciao


Walter

 

[Eli Duttman]

Uh oh my thread is dying

At this point i'm pretty set on a modified Mullard LPT design as suggested by Eli. This is what was used in my first build so I know the circuit. I'll follow Eli's advice and skip the PPP and go with with a deuce setup to give me the best chance of success.

Now
Do I go P2P, or take the easy route of using the Mullard derived Posiden PCB?

Did anyone get a chance to check out the schematic for it that I posted above? I'm very curious about any thought on the IC choice for the CCS in the tail?

I know some don't like the basic choice of triodes for this PCB but other options are possible with this board.

Well executed P2P can have a performance edge. It also allows total flexibility in selecting parts.

There are 2 major issues in circuits, like Mullard's, that employ GNFB: O/P trafo core satuation and slew limiting. You deal with core saturation by rolling infrasonic noise off at the amp's I/P with a 1 pole filter set for a F₃ in the 15-17 Hz. range. Resistance to slew limiting is obtained by employing high gm small signal types. The 6GK5 is a very slick package of high mu, high gm, and low R_P. The 12AT7 is quite good in those areas too. The icing on the cake with the 'T7 is an open loop distortion spectrum extremely well suited to driving PP "finals".

BTW, the 'T7 sounds GOOD with 200-220 V. on the plate and an I_B of 3 mA.