PDA

View Full Version : The 40th Anniversary Sony VFET Amplifier from Pass Labs



Odd
02-05-2014, 01:58 PM
At a special press briefing held during the 2014 International CES

http://blog.sony.com/press/sony-demonstrates-pass-laboratories-40th-anniversary-commemorative-vfet-amplifier/

61475

1audiohack
02-06-2014, 07:23 AM
Those are super cool looking. Ill bet they don't sound bad either. :)

I want 'em!

Odd
02-06-2014, 08:48 AM
Those are super cool looking. Ill bet they don't sound bad either. :)

I want 'em!



So do I.
I read somewhere they were not for sale, we have to wait and see.

BurntIsland
02-06-2014, 09:02 AM
As mentioned in our report from the 2013 Burning Amp Festival, the veteran designer showed his "simplest amp," created with “two transistors, a [Jensen] transformer and wire—the fewest parts he'd been able to use to make it work.” The transistors weren’t identified back then but it turned out that they were Sony Vertical Field Effect Transistors or VFETs, a device Sony had developed 1972 to mimic the transfer function of a 300B triode tube. Sony and Yamaha both made amplifiers using this unique device, but discontinued them in the early 1980s when lower-priced power MOSFETs became available soon after.

Nelson’s Burning Amp used two complementary VFETs, part numbers 2SK82 and 2SJ28, the only ones Nelson had to hand, but he then set to work finding more. Eventually, he uncovered a stash in Singapore and bought all of them, enough to make the single pair of “Sony VFET Amplifier 40 Year Commemorative Edition” amplifiers at CES. This uses 24 pairs of VFETs as complementary followers and front-end/driver circuitry based on a Pass Labs’ XP topology. Output power is specified as 250W into 8 ohms—it sounded like more! (http://www.stereophile.com/content/pass-labs-pays-tribute-sony)

gferrell
02-06-2014, 11:24 AM
Sort of reminds me of the old Ashly FET-2000 amps.

61476

Mr. Widget
02-06-2014, 08:04 PM
So do I.
I read somewhere they were not for sale, we have to wait and see.There will be no more... Nelson found the last NOS supply of V-FETs to build them. He built the amps as a tribute and presented them to Sony as a gift.

Widget

Dave_72
03-25-2014, 12:11 PM
There will be no more... Nelson found the last NOS supply of V-FETs to build them. He built the amps as a tribute and presented them to Sony as a gift.

Widget

That sucks...

Cyclotronguy
03-25-2014, 04:55 PM
Not quite accurate Mr. Widget! Only the last known NOS stash of Sony Vfet's.

To beg a line from Kief; Gerber files and metal "Are only a shout away, a shout away, just a shout away!" So maybe a couple more channels.

And yes the sounded really magical in the Sony CES room on the Sony speakers. Best I've ever heard the Sony speakers sound.

Cyclotronguy

SEAWOLF97
03-25-2014, 06:18 PM
My daily amp , connected to ESS Heil AMT bookshelves, is a Technics SU-G90 integrated (130 wpc). It is marked "FET direct coupling" and "new class A". Not sure what type of FET it is.

Running digi files from the computer , it is a great combo (prolly not worth much tho).

Cyclotronguy
03-26-2014, 12:04 PM
If you like it, regardless, then I would think it's worth quite a lot!
The beauty of this hobby, is that whatever puts a smile on your face, works.

Cyclotronguy

Allanvh5150
03-26-2014, 12:23 PM
Interesting. V FETS are still readily available. I use them all the time.

Allan.

Ducatista47
11-12-2014, 11:23 PM
Interesting. V FETS are still readily available. I use them all the time.

Allan.
Are those power VFETs, or low power pieces like most JFETs are? Many preamplifier stages use JFETs, but power JFETs of recent manufacture were limited to, I think, two manufacturers, one being now defunct.

Only the defunct ones (Semisouth) were high voltage units, of interest to electrostatic listeners like myself. When a suitable high voltage amp can be built, no transformers are necessary to drive electrostatic speakers and headphones, eliminating significant distortion. The only common way to build such an amp is to use tubes as the high voltage devices. The available solid state units bootstrap lower spec transistors; not a great way to go. They end up either not producing enough voltage swing to make the high end headphones sound as designed, or operating the transistors out of their design parameters. This is why the only tubes I still use are in a Stax amp from Japan. It still didn't cut it, but a friend modified it for me with higher voltage tubes; he replaced dual triodes with single triode bottles with much higher voltage capability. Some rewiring was also necessary, and it all works beautifully now.

I could mention that while MOSFETs can make outstanding output devices for audio, JFETs have some remarkable advantages over them. Nelson Pass has all the good ones left. I am lucky to have two of his First Watt power amps equipted with them.

I would have thought that power VFETs would be big news. What sort of pieces are they?

LowPhreak
11-13-2014, 01:25 AM
"At a special press briefing held during the 2014 International CES, Nelson Pass announces a few people have won the lottery, or have more money than brains..."


Good luck.

1audiohack
11-13-2014, 06:05 AM
Hi Ducatista;

Just curious, how high is the output voltage of the modified Stax amp you speek of?

Thank you,

Barry.

PS. I have no idea why there is a frowny face on the top of this post and I can't get it off.

Allanvh5150
11-13-2014, 10:14 AM
Although the classic 2SJ56 and 2SK176 and there smaller counterparts are not made now, they are still widely available from various sources. Also, the newer and higher powered 2SJ201 and 2SK1530 are in production and no problem at all to get hold of.

Allan.

Ducatista47
11-13-2014, 04:32 PM
Hi Ducatista;

Just curious, how high is the output voltage of the modified Stax amp you speak of?

Thank you,

Barry.

This is a somewhat complicated field. The short answer is about 960 volts peak to peak - about 340 volts RMS. If you would like to know what this actually means in this application, I have put this page together to give a feel for the requirements of driving electrostatic headphones. Stax of Japan is the premier company in this field. Read on if you are curious.

Firstly, check out this two page thread. That site if full of misinformation and BS, but the contributors "Kevin Gilmore" and "Spritzer" are the gold standard for information about electrostatic headphones and their ancillary equipment. They are the world leading experts outside of the Stax inner sanctum, which seems to closely follow what they say in any case.

http://www.head-fi.org/t/582518/electrostatic-amplifiers-voltage-ratings

Here are two posts from it.

From Kevin Gilmore:

Electrostatic Amplifiers Voltage ratings



I get asked about this fairly often, there are a number of different ways of measuring the voltages
so here in one place is all the amps i know about, and all are rated the same. The voltage is
peak to peak voltage measured from one stator to the other stator.

For solid state amplifiers, the output voltage is usually very close to the power supply rails.
For OTL tube amplifiers, the Rp of the output tube(s) usually takes about 60 volts off of the
maximum voltage swing, for Stax made amplifiers the negative power supply is increased to
compensate for this. These are theoretical maximums, in practice the voltage swing will be less.

There is of course more to these amplifiers than just the voltage ratings, Slew rate into a
capactive load is also very important, i will be measuring some of these amplifiers and will
add slew rate at a later time.


2400 volts peak to peak (+/-600 power supplies)
koss esp950

2300 volts peak to peak (+/-575 power supplies)
KG’s new Silicon Carbide jfet based space heater.

2000 volts peak to peak (+/-500 power supplies)
Stax T2, DIY T2, KGSSHV (ixys parts)

1800 volts peak to peak (+/-450 power supplies)
KGSSHV (sanyo parts)
craig has said his new electra amp does 1800 volts, power supply unknown
Ergo Jecklin

1600 volts peak to peak (+/-400 power supplies)
BH, BHSE, singlepower ES1,ESX (after being modified to remove the 100V offset)

1400 volts peak to peak (+/-350 power supplies)
KGSS,Stax T1,T1s,T1W,313,323,717,727,sra-14s,srm-1 mk2,srm-monitor
woo audio wes

1300 volts peak to peak
sra-10s/12s

1200 volts peak to peak (+/-300 power supplies, and +600 capacitively coupled)

KG tube #1,aristaeus,HEV70,HEV90,RSA A10, Stax srm212,srmXh,srm300,Exstata
Egmont,Cirolan,SRM-1,srm-3,woo audio ges,srm310

1100 volts peak to peak
sra-3s
srm-001 (when the batterys are at full power)

From Spritzer:

Well peak-to-peak has become a bit of an industry standard in the DIY world with only Stax using RMS. Still it is easy enough to convert, RMS*1.41*2 to get the P-P value. The SRM-300(310) is listed at 350VRMS so that would be roughly 1000V P-P. Stax always post conservative specs plus these are measured, not simply derived from the rail voltages (in this case +600V regulated).

My amp is not directly listed there, but Stax publishes these specs:


SRM-007tII SPECIFICATIONS:


Type
:Vacuum tube output stage Low noise dual FET input Class A operation, Pure balance DC amplifier configuration Earspeaker driver unit


Vacuum Tube
:4 (6FQ7 / 6CG7) 6S4 in my modified example


Frequency Response
DC to 100,000Hz +0,-1.5dB w/ SR-007 or SR-404


High Harmonic Distortion
:Max. 0.02% (with 1kHz, 100V r.m.s. )


Gain
:54dB (x500)


Input Terminal
:3 ( XLR x 1, RCA x 2) with 1 RCA parallel out


Rated Input Level
:200mV / 100V Outputs


Max. Input Level
:30V r.m.s. at min. volume


Input Impedance
:50Kohm, 50kohm x 2 during balance


Max. Output Voltage
:340V r.m.s. (1kHz)


Standard Bias Voltage
:PRO 580V x 2


Power Voltage Frequency
:117V (50Hz / 60Hz)


Power Consumption
:55W


Temperature & Humidity
:0 to 35ºC / 90% max. without condensation


Dimensions
:7.7 (W) x 4.1 (H) x 16.5 (D) in


Weight
:10.4lbs


Color
:Silver or Black



I am not claiming that my example is higher voltage, I don't know, but rather that it yields the stated voltages without drama or just plain crapping out. Calculated (RMS X 1.41 X 340) it comes out to 959 volts p-p. Spritzer said it sounds like a KGSS, which has a 1400V peak to peak, so I don't wonder why I am happy with the sound. Bear in mind that electrostatic drivers don't need much current, but the amps have to be beasts because of how the capacitive load varies with frequency. Again, from Dr Gilmore:

Virtually 100% capactive which means that the reactance varies linearly with frequency. Starts at literally meghoms at DC, and goes all the way down to under 18k ohms at 20khz. Moisture in the air has some impact on this.

To put this all in perspective, Stax headphones can be put into three groups, in decreasing difficulty to drive: the SR-007 series, the current top of the line SR-009, both still in production, and everything else. The first two are large, round and expensive and may be the World's most comfortable headphones to wear. My SR-007Mk2 is exceedingly difficult to drive in a high fidelity manner. The modifications to my amp put it into the "not the best but works adequately" category. The best amp currently for sale is the BHSE - the Blue Hawaii Special Edition - by Justin at Headamp. I have heard it and it is superb. I cannot afford one.

Reading the entire thread will impart more information about the demands placed on these amps. FYI, Kevin Gilmore is an engineer who works for the math department at Northwestern University, so he is in the Chicago area, where I met him in 2010 at a huge international headphone event. My audio buddy and I brought some gear and participated. I listened to my phones on his DIY T2 amp and I have seen the light. And the heat. He told me it would literally fry an egg. The original T2 bankrupted the Stax corporation. Spritzer is the online name of a gentleman in Iceland, who I finally met there too. We were already friends from online exchanges and some equipment going back and forth. He modified my Stax amp. Talking to these men about Stax is like talking to Nelson Pass about solid state power amps. Take anything you hear to the bank.

Hope this answers your question.

PS The bias voltage for Stax headphones was changed some years ago. If you apply the high bias to the older units it will destroy them. Be careful.

Perhaps a little glossary is in order. The PRO 580V X 2 reference indicates two 580V sockets. The 580V (high) bias refers to the voltage biasing the headphones, not the amp circuits. The legacy non pro phones used about half that. KGSSHV is a Kevin Gilmore Solid State High Voltage (amplifier).

If you think the voltages are high in these amps, there is a fellow that built some using mercury vapor rectifier tubes. It gives you a lovely blue glow and something like 5000 volts. Goes to show what it can take to go all the way with this quest. Images posted below, pictured with a Stax SR-007. He obviously knows what he is doing. Here is a quote from him about two top amps:

" I'm usually perplexed when people talk about "power" in relation to electrostatic headphones and amplifiers.

The headphone is a very high impedance load consisting almost exclusively of capacitive reactance. It requires a large voltage swing and enough current in the amplifier output stage to maintain slew rate and frequency response, but it consumes next to no true power no matter how much you "throw" at it.

Electrostatic amplifiers are voltage, not power, amps. Conceptually they are more like an ultra high gain preamp rather than a traditional speaker power amp, which unlike a voltage amp, is designed to transfer both current and voltage into a low impedance load. The only signifigant "power" involved in an electrostatic amp is the dissapation of the output devices and their respective loads.

Rather than an undefined or mislabeled discussion of an amps output "power", it might be better to discuss the amp in terms of it's ability to drive a load. Driving ability, reduced to basics, is a product of gain(usually about 60dB), output voltage swing (usually about 600Vrms), and output stage current (usually 15-20mA to ensure slew rate at frequencies of interest.) I'm passingly familiar with the designs of both amps, and as far as I know, both easily meet these requirements.

63589635906359163592