The "iron" is intended for EL34s and they are the obvious choice. However, PPP 6П14П-EB (6p14p-ev), AKA EL84M, "finals" will be OK too.
As for small signal circuitry, the tried and true Mullard style is a definite possibility. Mullard style is not "fussy" about O/P "iron" and rarely has stability issues.
As for small signal circuitry, the tried and true Mullard style is a definite possibility. Mullard style is not "fussy" about O/P "iron" and rarely has stability issues.
Small caveat: Sonic Frontiers, which was very early on taken over by Foreign Inwestors... operated under the odd premise that Premium parts made for a premium amp.
History clearly proved that wasn't even close to being true.
Assemblage amp worked, but it was hardly a great contraption.. by any yardstick.
History clearly proved that wasn't even close to being true.
Assemblage amp worked, but it was hardly a great contraption.. by any yardstick.
The transformers were made by Hammond, or ATC Frost, i bought extras from Hammond many moons ago. Built with Caddocks, MIT RTX and Blackgates it was actually rather nice. Compared to Audio Research and Krell mono blocks at the time it is dynamically limited, but quite nice otherwise
Last edited:
I probably bought the surplus Sonic Frontiers OTs at the same time as you did ticknpop. Could you share what you did with your OT's?
You are a great source, ticknpop. Please let me know your insight into these builds, and I shall appreciate the Assemblage ST40 schematic.
Does anyone have the characteristics for these Sonic Frontiers Assemblage ST 40 transformers?
Does the "iron" have ultra-linear taps on the primary side? Based on the 40 WPC claim, I'm guessing no UL taps.
Some digging showed SF used a 12AX7 section, a 12AU7, and 2X EL34s per channel, for a total of 7 signal path tubes. That easily could have been Mullard topology, but other options did exist. As the seminal 5-20 shows, Mullard style topology is voltage amplifier DC coupled to a long tailed pair (LTP) phase splitter that's cap. coupled to PP "finals". I'm not thrilled by the small signal types Mullard chose, which were (IMO) sales, rather than technically, based. The EF86 and 12AX7/ECC83 exhibit low gm, which is a distinct negative. EICO did a better job with a 'X7 section voltage amplifier and 6SN7 LTP, but I'd like to see gm emphasized a good deal more.
A constant current source (CCS) loaded 6922 section voltage amplifier and an ECC99 LTP should provide plenty of open loop gain, while exhibiting resistance to HF error correction signal induced slew limiting. Constant current sink (CCS) load the LTP's tail to force symmetry between the 2 sides.
Assuming full pentode mode "finals" are employed, regulate O/P tube g2 B+ for max. open loop linearity. Let the GNFB loop deal mostly with damping factor and, secondarily, distortion.
My intention is to squeeze every last drop of performance out of the O/P "iron" possible. A multi-tapped negative voltage multiplier will "feed" the LTP tail CCSes and provide "fixed" bias for the O/P tubes.
Some digging showed SF used a 12AX7 section, a 12AU7, and 2X EL34s per channel, for a total of 7 signal path tubes. That easily could have been Mullard topology, but other options did exist. As the seminal 5-20 shows, Mullard style topology is voltage amplifier DC coupled to a long tailed pair (LTP) phase splitter that's cap. coupled to PP "finals". I'm not thrilled by the small signal types Mullard chose, which were (IMO) sales, rather than technically, based. The EF86 and 12AX7/ECC83 exhibit low gm, which is a distinct negative. EICO did a better job with a 'X7 section voltage amplifier and 6SN7 LTP, but I'd like to see gm emphasized a good deal more.
A constant current source (CCS) loaded 6922 section voltage amplifier and an ECC99 LTP should provide plenty of open loop gain, while exhibiting resistance to HF error correction signal induced slew limiting. Constant current sink (CCS) load the LTP's tail to force symmetry between the 2 sides.
Assuming full pentode mode "finals" are employed, regulate O/P tube g2 B+ for max. open loop linearity. Let the GNFB loop deal mostly with damping factor and, secondarily, distortion.
My intention is to squeeze every last drop of performance out of the O/P "iron" possible. A multi-tapped negative voltage multiplier will "feed" the LTP tail CCSes and provide "fixed" bias for the O/P tubes.
Attachments
Help with Sonic Frontiers outut transfomer
I have two pairs of Sonic Frontiers OTs and don't remember the details. It has printed on the transformer: "Sonic Frontiers, SFOT-04, HEK94". They seem to be single output (6 ohm?) and not Ultralinear, i.e. like a triode OT with only two primaries with center tap. Could you help me identify and use these guys? They are quite large, almost the size of Dynaco MkIII OTs and definitely larger than Dynaco ST 70 OTs
I have two pairs of Sonic Frontiers OTs and don't remember the details. It has printed on the transformer: "Sonic Frontiers, SFOT-04, HEK94". They seem to be single output (6 ohm?) and not Ultralinear, i.e. like a triode OT with only two primaries with center tap. Could you help me identify and use these guys? They are quite large, almost the size of Dynaco MkIII OTs and definitely larger than Dynaco ST 70 OTs
Just a reminder it's against the forum rules to start multiple threads on the same topic.Why is low gm a "distinct negative"? I don't recall the 5-20 having a big rise in HF distortion, which would be a sign of running out of slew rate. The EL34 is an easy valve to drive, even in UL mode. A 6SN7 LTP needs a CCS tail, but the high mu ECC83 can manage with just a resistor tail.Eli Duttman said:
In addition to "running out of steam", low gm/high RP renders the setup vulnerable to HF error correction signal induced slew limiting. High gm allows for good swinging of volts, in the time domain.
BTW, EICO did not use either a CCS or a negative rail in combination with 6SN7 LTPs. Refer to the HF89 schematic.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.