DIY Sony VFET Builders thread

[Soundhappy]

..and hopefully, also reduce the speaker turn on thump.

Vfet SE don't have any thump on psu turn on, is a total silence in the speakers.
Very liitle and very short time thump on turn off only ~ 1V no big deal for full range drivers.
Anyway is a hobby have fun

 

[TungstenAudio]

Likewise, my VFET amp has never had a turn-on thump. The slight turn-off thump went away after I added 1000 uF of capacitance to the power rails on the output stage cards.

 

[Soundhappy]

It would also allow for significantly larger electrolytic rail caps on the board
of one wanted to do that and avoid the SMPS from going into hiccup mode.

This circuit is different
current rise slowly so high probability SMPS hiccup mode can't detect large capacitors.
You can try 2200 uF or more without any modifications ?
Best regards

 

[TungstenAudio]

Yes, I think 2200 uF per side would probably be Ok, but I haven't tried it yet. That would add up to 5400 uF total. Since there is a 0.12 mH inductor in series with this capacitance, it will slow down the charging current a little.

The main thing slowing down the amp turn on are the 1k + 4.7k resistors in series with the 1000 uF cap C2 across the opto-isolator in the output stage.

 

[TungstenAudio]

No, as I have mentioned, the extra 1000 uF for each card were soldered directly to V+ and G.

 

[Soundhappy]

@ TungstenAudio

Ok could you show us two soldering points on the pcb please for more clarity and avoid errors ?
Thanks in advance

 

[TungstenAudio]

I used the same soldering points that are used for the 16ga wires on the edge of the card. One of the G eyelets should still be open. The connection to V+ can be made on top of the wire connection. It depends on how one chose to connect the power wires during their assembly.

 

[Soundhappy]

Thanks TungstenAudio I understand better now.

 

[grataku]

I’ll post the Gerbers and anyone can make or organize their own GB as I don’t have the time for that. I have tested the circuit, now modified for 36v operation. Works well with a 6 second delay using a 1M resistor on the RC timer with two 10uF 25v X7R capacitors in series to achieve 5uF and 50v.

I have to correct a mistake I made in saying that the circuit would use the switch to switch a low current flow. It would indeed still switch the main current flow, but upon contact, there would be a 30ohm NTC in series to significantly reduce the arcing, and hopefully, also reduce the speaker turn on thump. It would also allow for significantly larger electrolytic rail caps on the board of one wanted to do that and avoid the SMPS from going into hiccup mode.

It's fine I totally get the time constraint thing. I understood the functionality to be different in that the panel switch would be used to tun off a solid state device that would do the actual switching. I'll look at the circuit again.

 

[Nelson Pass]

You can add capacitance to the supply until the Meanwells won't start. Actually,
the Meanwells will pulse charge the caps up as long as the circuit itself doesn't
start to draw current, so you can play with the capacitor on the constant current
source to delay that further if needed.

 

[grataku]

The M2(x?) has very tube like turn on, since the Vfet is much more tube like musically speaking, it would only make sense to add that feel to it.

 

[Soundhappy]

You can add capacitance to the supply until the Meanwells won't start. Actually,
the Meanwells will pulse charge the caps up as long as the circuit itself doesn't
start to draw current, so you can play with the capacitor on the constant current
source to delay that further if needed.

Thank you Mr. Pass
Filtering board have free space around 1K uF so can be stuffed with twice size bigger capacitor
If needed CCS caps as well..I do try this easy modyfications after run of batch3 in the store..

 

[nicoch58]

frugly !

 

[xrk971]

It's fine I totally get the time constraint thing. I understood the functionality to be different in that the panel switch would be used to tun off a solid state device that would do the actual switching. I'll look at the circuit again.

There is another board with an SCR in addition to the MOSFET. This one has a remote turn on/off with a small signal switch. It would be a little more complicated but Jhofland has made this too.

 

[xrk971]

I just installed the Jhofland designed inrush limiter in my VFET amp. It resides between the power switch and CLC filter board. It’s powered by the 36v from the switch. Ground is referenced from the CLC filter board. It has about a 6 second period of flowing current through two 8D-20 NTCs in series. The cold resistance is about 20ohms. Upon turn on, they heat up to about 65C in 6 seconds. They cool down to about 32C in normal operation when the low RDson (6.5mOhm) MOSFET kicks in (the green LED glows to show that the MOSFET is conducting). The voltage drop across the inrush limiter is about 50mV or so. The inrush limiter allowed me install a 4700uF 63v Nichicon cap right on the +V and GND pins of the power rails. The caps work just fine with no hiccup. There is no turn on or turn off thump anymore. I would imagine that the addition of the larger rail caps will have to reduce the level of crosstalk. It might have some other benefits too - perhaps more bass authority?

I’ll let this burn in another day before buttoning it up.

 

[xrk971]

Here’s a closeup of the inrush limiter in action:

As you can see, it’s quite compact.

 

[nicoch58]

X fast as always

 

[ClaudeG]

Hi X,

Interesting as ever...

2 questions:
- Did you notice any sonic differences?
- Did you try to start your amp with its now 10400uF WITHOUT your inrush limiter? Could well be it still starts...

Many thanks

Claude

 

[xrk971]

I never dared try powering a pair of 4700uF caps without the inrush limiter. I will have to try it later by bypssing it with a jumper.

With my plain ears, I could faintly hear some crosstalk by pressing my ear to one channel while driving the other one in stock mode. With the 4700uF in place, I could not hear anything, not even the slightest hint. I think with an FFT measurement with -130dB noise floor it will reveal what the crosstalk is easily.

 

[xrk971]

I decided to button up the amp. It's ready.

Installation of the CLC filter and Inrush Limiter on the floor panel prior to final assembly:

Detail of grounding wires for the indivusal heatsinks to the floor panel:

Detail of Inrush Limiter, I got rid of the extra activation connector by jumpering the input voltage to the activation pin and tyingthe ground to a standoff bolt:

Detail of CLC filter with the heatsink grounding attached to the standoff nuts. I am using brass M3 standoffs with washers rather than the stock M4 bolts and nuts. It was too difficult to install the CLC while it was soldered to the amps, and to keep the screws, standoffs collars, and nuts all aligned without falling off:

Detail of pin crimped connectors on underside of amp PCBs, showing how strain is reduced on the wires as the wires are silicone insulated soft wire and the crimp pins prevent solder from wicking up:

Interior closeup showing wiring assemblies:

Deatail of top view of the amp:

Detail of rear panel with all connections while sitting at main amp postion:

Buttoned up amp sitting in main position ready for listening: