• 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.

Dynaco VTA-inspired KT88 PP AMP project

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Dear Friends,

Inspired by the discussions around the possible schematics to use my father came up with the idea to combine the driver stage from the VTA ST70 AMP using 6SN7 tubes with the output of Mark III based on KT88 for more power. The draft of the schematics is attached.

Power transformed will be of toroid type giving 250 Wt output. B+ voltage before the choke will be 504 V.
I have some questions regarding the schematics:
1. Is this in general fine or have I missed something?
2. I want a OT with wide frequency response. The original schematics shows a 4.3K transformer. I found the LL1679-PP 9k7, 4k5, 2k6, giving 45, 105, 188 Watt transformer. Will it suite the needs? Or can you recommend something else?
3. There will be 2 chokes (1 per channel power supply). The Mk3 has this choke: 1.75 Henries, 200ma. DC current, 62 ohms, 400 vdc (max) 10% Tolerance. are these parameters enough? I saw in other schematics that chokes of 4-8 Henries are used.
4. Various schematics advice different capacity of C13 and C15 - from 22 to 47 uF. The more the better?
5. Did I correctly draw up the NFB? It is taken from the VTA 70 schematics.
6. Any suggestions on the capacitor types? I can take made in China or some European brands?
7. Is there any software to imitate and check how would this project actually work and whether some changes are needed?

Thank you in beforehand for your help
 

Attachments

  • Dynaco style 6SN7 KT88 PP with CCS.jpg
    Dynaco style 6SN7 KT88 PP with CCS.jpg
    247.5 KB · Views: 1,249
Last edited:
Your common 11.2R output tube cathode resistor seems odd. I would thinks you want an even resistor value=(ie: 10R) for easy calculation. Also if you separated it into two 10R (or 1R), you can read bias for each tube.

Unless you want to match the tubes yourself, adding an AC balance resistor may be an good idea (add an trim pot in series with R27.).

I would use 100uf+ for C13 and C15.

GNFB seems fine to me.
 
Considering that the output stage will need approx -45V bias, then in rough terms, driver voltage swing per side will need to be 45V/1.4 = 32V in order to drive the output stage to full power. With a 4.3K primary, I expect full power output at the speaker terminals will be around 55 to 60 watts, if using a stout 490V B+.

The phase inverter/driver stage will have gain I expect of somewhere between 6 and 7 (let's say 6.5) per side (meaning between each inverter output and ground). Voltage amp will have gain before feedback of say 12. So if you apply say 14 dB feedback (which is a voltage gain reduction in first stage of about 5x), then input sensitivity of the amp will be approximately 32V/(6.5 * 12/5) = 1.64V. Which is just about right, IMO.

Therefore a 6SN7 voltage amp will offer plenty of gain IF you keep feedback to a reasonable amount (around 14 to 15 dB).

Negative feedback topology looks correct. Final FB resistor size is easiest determined on test. Final FB capacitor size is output transformer dependent, and must be determined on test.

===

Things I'd add or consider changing:

  • 100Ω resistor between each screen pin and its respective UL tap lead.
  • Consider direct coupling the voltage gain stage to the inverter. That will remove the time constant (high pass filter) of that R/C coupling between the gain and inverter stages, and make it easier to apply feedback while keeping the amp stable. However DC coupling those stages will require raising the cathode voltage on the inverter stage to something like 100V to 120V, so the biasing circuit/cathode circuit of the inverter will need to change to conform.
  • Add current sense resistors in the cathode circuits of each output tube so you can test for DC balance of each tube independently.
  • Add a way to bias/balance each output tube separately (maybe already accounted for, we don't know for sure since you haven't shown the bias circuit). It can take the form of individual bias adjustments for each tube as you've eluded to, or it can be a bias/balance scheme for the pair of tubes.
  • By the book, the 150K grid return resistor on the output stage is too large for fixed bias. I'd follow the spec sheet max value listed: 100K. But if you want to be able to sub in a 6550 or KT120, you should make it no larger than 51K.
  • Going with a 51K grid return resistor on the output stage solves another lurking problem: If you keep the R/C coupling between first and second stage, the time constants between the two stages' R/C filters are too close together. If you move them farther apart you will have an easier time obtaining good LF stability when feedback is applied. Using a 51K grid return resistor with a 0.33 uF coupling cap as shown will give you a ~9.5 Hz 'cutoff' frequency, which will separate it from the cutoff frequency of the first stage R/C filter by about 2.7 octaves, which is much better separation than currently shown on your schematic, which is only one octave.
  • You may need some additional HF compensation after feedback is applied. Look for example at the Dynaco Mk. III schematic at the small feedback cap from one of the UL primary taps back to the junction in the cathode circuit of the first stage. You may also need/want to add an HF step circuit. These tuning circuits and values chosen are best determined on test. HF tuning in a feedback amp that includes the output transformer in the loop is output transformer specific.
 
Last edited:
I'm looking at a similar project

But using a little less B+...probably end up between 450-460vdc using Edcor for power and LL1682 OPT...5.5K/5. Will also be dual mono, so I've got pairs of both. The first 6SN7 duties will be split into 2 and handled by a pair of 6J5's.The schematic shown is the one given with VTA's 6SN7 replacement board for the ST70, and should be able to accommodate several different output tubes, given proper B+. The pic I'm posting should help with bias scheme.

Screen-2019-02-06_06-43-05.jpg
 
kward thank you for this vast input, which I will definitely take into consideration. Would you recommend the LL1679-PP 4,5k OT for this schematics?

John Ess, thank you.
Yes, that is one of the versions for the bias scheme I have available, which also comes with one of the VTA boars versions for the ST 70. I hope and as kward calculated that it will be able to work with KT88. The LL1682 does not seem to fir for me as I am targeting 60 Wt per channel.

Francois G, ok, I will once again recheck the schematics to be sure that everything fits.

Thank you all for your input!
 
May I offer some suggestions, based on constructing KT88 amps. Put a balancing network in the plates of the phase splitter. When a sine wave is applied to your amplifier it can be adjusted to supply equal drive to the KT88's with the aid of a CRO. Also a balancing network in the cathodes of the KT88s al la Williamson is useful. Perhaps adjust the cathode resistor(s) to drop the current to 140 or so mA. It prolongs the life of your output tubes.
Lastly consider triode operation; it sounds more musical.
 
Dear Friends,

Any suggestions on the OT? I am planning to buy them in Europe. Could you recommend some models for >60 Wt output? Besides the 2 ones mentioned in this thread?

Are there any particular requirements to the choke? My father said that we could make ones by ourselves (and I am sure that he can do it)? How many Henries would you recommend?
 
peteki,

I believe that matched plate loads on a CCS cathode coupled splitter will by definition give equal amplitude and opposite phase output signals.

That is one of the characteristics of using a real CCS for a LTP phase splitter instead of using a "resistor pseudo current source".

As to being able to adjust the plate loads on the splitter, that would make up for the not so perfect balance of the Output Stage Tubes.
 
I do not need a lot of power with them, I'm thinking of keeping the + B around 450Vdc, so the source would be more economical besides being more secure with the current Kt's. I believe something between 35 and 45W, I want the least distortion possible.

Thank you for your help!
You might look at their boards...the 6SN7 board is what's being discussed here and I'm taking a similar approach and will be using 450-460vdc. tubes4hifi home page
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.