This thread is about designing a vacuum tube front end for the Sony VFET DIY amp. There are already a number of sand based front ends, but I thought a vacuum option would be nice, too!
Design targets:
(This post is a followup to the discussion that started in the "Scourge, Bulwark, Marauder, Dreadnought front end cards for DIY VFET amp" thread, and relevant posts from this thread will be moved here.)
Design targets:
- Should be relatively easy to build so the average DIY VFET builder can cope with it.
- Should fit in the DIY VFET chassis. If possible, the chassis should not need any modifications (like drilling more holes).
- Avoid dangerously high voltages -- this is for rookies!
- The front end needs to provide high input impedance and low output impedance to drive the VFET/SIT output stage.
- Single ended input signals with up to ±2 Vpeak (4 V peak-to-peak).
- Voltage gain of approximately 7-8.
- Should not impose its own distortion over the distortion profile of the VFET output stage (except maybe if the harmonics of the front end are opposite phase than those of the VFET output, resulting in cancellation of harmonics).
- Must avoid any possibility to zap the VFET in the output stage, even in case of a (high-voltage) failure. The VFETs are almost unobtanium!
(This post is a followup to the discussion that started in the "Scourge, Bulwark, Marauder, Dreadnought front end cards for DIY VFET amp" thread, and relevant posts from this thread will be moved here.)
Last edited:
I have no experience with vacuum tubes and no idea what's best for them. Consult with deeply experienced tube experts, who might be easier to find in
https://www.diyaudio.com/forums/tubes-valves/
https://www.diyaudio.com/forums/tubes-valves/
If I were to add a vacuum tube input stage to my VFET amp, it would include a separate transformer for the plate and filament voltages. The Antek 50VA models include both plate and filament windings. Both sets need to be regulated for best performance.
Edited to add:
Adding a tube input stage is definitely going to be a challenge. I would probably get a separate top plate for the amp, and mount the transformer and tubes to the topside, with the voltage regulators on the underside. Keep the amp clean looking, with a vintage flair.
Edited to add:
Adding a tube input stage is definitely going to be a challenge. I would probably get a separate top plate for the amp, and mount the transformer and tubes to the topside, with the voltage regulators on the underside. Keep the amp clean looking, with a vintage flair.
Last edited:
The VFET amp already includes a HF rejection filter PCB, designed by Nelson Pass. His design document (link) presents the circuit schematic of the filter, plus the exact part numbers of all components. That ought to be plenty enough to get you started on a detailed analysis of the filter. Then you can decide for yourself whether additional "dedicated PS filtering" would be (a) beneficial; or (b) a waste of time.
Pete Millet mad a preamp (el escorpion) that uses a step-up converter to get the B+ voltage.
I copied this approach in order to make a preamp for Mofo (Thread "Build this Mofo" post #2468 ff)
For the VFET we could use the 36V B+ to run two tube heaters (IL861RFT) in series. The step-up unit needs to be redesigned.
Alternatively we could use an external SMPS with a lowish voltage of 100VDC for B+.
Thanks for these hints! Looking at Pete Millets step-up converter, it looks like this board might be a bit large to fit into the VFET kit chassis. Maybe there are other options? What did you use in your own amp?
Also, one could consider adding a second SMPS brick. For instance, a low-power SMPS brick with 48 VDC could be stacked on top of the existing 36 VDC, resulting 84 VDC. There are a number of tubes that would work well with this B+ voltage. However, a second brick would mess with the "KISS" appraoch and would require an extra hole in the chassis.
Finally, low voltage tubes might be an option. The ECC86 was designed for low voltages. The 6H30Pi works well at low voltages.
Also, one could consider adding a second SMPS brick. For instance, a low-power SMPS brick with 48 VDC could be stacked on top of the existing 36 VDC, resulting 84 VDC. There are a number of tubes that would work well with this B+ voltage. However, a second brick would mess with the "KISS" appraoch and would require an extra hole in the chassis.
Finally, low voltage tubes might be an option. The ECC86 was designed for low voltages. The 6H30Pi works well at low voltages.
Perhaps the 6-pin jack which matches the plug on the Mean Well supply, is one member of a series of parts from the jack manufacturer. Perhaps there are other members of this series, having more than 6 pins. Perhaps one of them is a perfect mechanical fit to the chassis holes for the 6-pin jack, but provides (let's just fantasize) 9 pins. Now you can bring in (+36, GND) and also (+160, GND) and also (AC6.3, \AC6.3) without modifying the chassis. You just have a teeny little "supply concentrator" box behind the amp. This little box has 3 input cables (MeanWell_In, HV_DC_for_tubes, AC_for_heaters) and 1 output cable. The output cable has got the correct 9-pin plug of course.
I don't know whether this chain of imagined possibilities, happens to exist on the earth, or not. But it might be worth a bit of investigation if you despise the idea of drilling extra holes in your Museum Piece chassis.
_
I don't know whether this chain of imagined possibilities, happens to exist on the earth, or not. But it might be worth a bit of investigation if you despise the idea of drilling extra holes in your Museum Piece chassis.
_
I looked at the MeanWell DC boosters. There are some options with a 48 VDC output that could be "stacked" onto the existing 36 VDC. However, non of these converters will take 36 VDC as input.
I was not aware the PSU connector has 6 pins!
How about a little adapter cable with suitable connectors for multiple external bricks to achieve this:
Pin 1: GND
Pin 2: 36 VDC for the output stage
Pin 3: 144 VDC for B+ (just as an example, three external 48 VDC bricks in series)*
Pin 4: 5.0 or 6.3 VDC for the tube heaters
Pin 5: nothing
Pin 6: nothing
As far as I can tell, 4 pins should be enough. Pins 5 and 6 could be used for dual-mono supplies. I don't think we'd need more than 6 pins.
*This raises the question how much DC would be considered safe, and under which circumstances.
I was not aware the PSU connector has 6 pins!
How about a little adapter cable with suitable connectors for multiple external bricks to achieve this:
Pin 1: GND
Pin 2: 36 VDC for the output stage
Pin 3: 144 VDC for B+ (just as an example, three external 48 VDC bricks in series)*
Pin 4: 5.0 or 6.3 VDC for the tube heaters
Pin 5: nothing
Pin 6: nothing
As far as I can tell, 4 pins should be enough. Pins 5 and 6 could be used for dual-mono supplies. I don't think we'd need more than 6 pins.
*This raises the question how much DC would be considered safe, and under which circumstances.
Nelson Pass's design document attached to post #1 of the other thread, shows six wires emerging from the power connector (jack) inside the chassis. Study the figures and decide whether you disagree. That document was reviewed carefully by several people chosen by Nelson, I doubt the final version got this particular detail wrong. But hey, it's not impossible. Just unlikely.
The Meanwell SMPS brick that I received with my parts kit has a 4-pin DIN connector, which also includes an outer shield. On the DC inlet side, there are six pins, but two of them are reserved for the shield GND, so only four pins remain to be assigned by the user.
Mark's idea of looking for a 6-pin DIN version that fits in the same mounting hole may be something to look at. It will be a different part than what is supplied with the kit.
Mark's idea of looking for a 6-pin DIN version that fits in the same mounting hole may be something to look at. It will be a different part than what is supplied with the kit.
I am looking at this, page 11. The power connector has four pins + shield/chassis. The first two pins are connected using two wires to the input filter, where the two wires are joined together at "VSUPP +". The other two pins are wired and joined in the same way to "VSUPP -". This means the connector of the 36 V SMPS brick has 2 x POS and 2 x NEG pins/wires, which are joined inside the VFET amp chassis.
It would be possible to join the pins/wires at the 36 V brick outside the amp, and thereby free up two of the chassis connector pins for something else (like B+ and heater supplies).
However, depending on the B+ voltage involved, this solution might defeat the idea of avoiding high voltages that may potentially be dangerous. What would be a reasonable max. DC voltage that would be considered as "not dangerous" in this situation?
simple use the bare wire ,no connector at all or fly connector
Sure, that "works". And I would even allow myself to think about this if it were just for me. However, my idea is to come up with a design that should work well for the average VFET kit builder.
Last edited: