JimL

Remember that increasing the output standing current means the output stage has to dissipate more power, which means more heat output. So you need to have enough heatsink in order to keep the output transistors and heatsink MOSFETs cool enough with the increased current. IIRC the onboard heatsink version KGSSHV is not designed for higher current but the off-board heatsink version can tolerate more current if the heatsinks are big enough.

So, there are a couple possibilities. The first is that you have reached the limits of the amp, and the second is that you've reached the limits of the headphones. If it is the second, then there is no solution to the problem other than changing to different headphones. If it is the first, it depends on whether you are hearing distortion due to the amp running out of voltage, or current. The KGSSHV uses +/-450 volt or +/-500 volt power supplies, so its voltage limits are higher than the BHSE (which uses +/-400 volt PS), and close to or the same as the DIY T2 - in fact, there is less than 1 dB difference between the max voltage output of the KGSSHV with 450 volt PS and the DIY T2. The alternative is that you are hearing distortion due to current limiting, which could be alleviated by increasing the standing current in the KGSSHV. However, a more important point is, that if you are listening regularly at "deafening volumes", you are likely irreparably damaging your hearing. One sign of this is that if you hear ringing in your ears after a listening session, that is, as someone colorfully put it, "the scream of a dying acoustic nerve." You can replace headphones and amplifiers. You cannot replace your ears. Once your hearing is gone, it is gone.

The 580V voltage is for the headphones - this is sometimes called biasing the headphones. When you change the resistors on the output stage you have to adjust the currents in the output stage so that the output sits at zero volts with respect to ground - this is referred to as biasing the output stage. These are two separate things that are not related to each other except that they are sometimes referred to with the same term - "bias."

Hmm. I think you need to do a bit more math and some simple measurements. 1) The advantage of a cascode circuit is it has a high gain. The disadvantage is it has a high output impedance - roughly, the plate resistance. With a plate resistance of 140 kilohms into a MOSFET buffer with 100 pf input capacitance, that produces a -3 dB roll-off point of approximately 11-12 kHz, as Kung points out. Since there is no feedback to extend the frequency response, the amplifier will roll-off above that point at 6 dB/octave. You should be able to easily measure this using a sine wave generator, a good DMM, and a 100 pf cap as a dummy load. 2) Your notion that a 5.1k output resistor degrades the performance of the amp in controlling the diaphragm is questionable. Most electrostatic headphones approximate a 100 pf capacitor. The impedance of the headphone is therefore approximately 80 kilohms at 20 kHz. The traditional way of calculating the ability of an amp to control the transducer is the damping factor, which it the ratio of the impedance of the transducer divided by the output impedance of the amplifier. In this case, 80 kilohms divided by 5 kilohms is 16, which is perfectly adequate. Another way to look at it is to ask what the roll-off frequency of a 100 pf headphone is when driven by a amp with 5 kilohms output impedance - the answer is over 300 kHz, which is definitely sufficient.

Wow, way cool!

Horses for courses. As John Broskie of TubeCAD points out, series regulators are well suited for class AB power amplifiers, where the current demand on the power supply can vary from tens of mA to several amps, whereas shunt regulators are best suited for class A amplifiers where the variation in current demand is relatively limited. Arguments about which “sounds better” approach religious discussions, although it is interesting that more designers claim that shunt regulators sound “better” than the reverse – for example, Broskie, the late Allen Wright of Vacuum State Electronics, Richard Marsh, the late John Camille, designer of the Kyrie 211 SET amplifier, Lynn Olson, designer of the Amity, Raven and Karna, Kevin Carter of K&K Audio, and Frank Cooter. I think Kevin Gilmore is one of the few designers who has gone on record saying that series regulators are as good or better. Incidentally, Richard Marsh wrote an excellent basic article on power supplies for Audio Amateur magazine (1983, issue 3). Religious arguments aside, there are pluses and minuses to both approaches. In favor of series regulators, they can have excellent specifications in terms of noise, output impedance, input rejection, etc. For example, the low voltage regulators designed by Walt Jung have a lower output impedance than 0.2” (yes, one fifth of an inch) of 18 gauge wire. Of course, to achieve that performance you would have to incorporate the regulator circuit on the amplifier board, as any intermediate wiring would negatively affect the impedance. Series regulators are frequently more efficient than shunt regulators as they only require a few mA extra current draw for regulation. In favor of shunt regulators, they have a constant current draw from the raw power supply, can both source and sink current, and have a relatively simple path to ground. By using a constant current source in series between the raw power supply and the shunt, this limits the signal currents to the regulator and amplifier area, whereas the varying demands of a series regulator means that some simplified form of the signal current has to circulate between the raw power supply and the series regulator. The argument that a shunt regulator can both source and sink current is thought by some to be a major reason for its sonic superiority. However, this can be abrogated to a significant degree by placing a bleed resistor after a series regulator to bring it up to 50% of maximum load, as first suggested by Richard Marsh. This allows a series regulator to also source or sink current up to the level of its bleed current, but also decreases the efficiency advantage of a series regulator. For a phono preamp, which runs in class A, you definitely want low noise. The GRLV definitely qualifies for that. K&K Audio sells a shunt regulator kit which it claims is suitable for phono preamps, and says it is based on the Salas regulator, so I assume that it is low noise also, though without suitable measuring equipment it is impossible to know how they rank in terms of noise.

Has anyone seen the specs? They claim a maximum output of up to 15 watts/channel and a total power consumption of 80 watts. By comparison, the BHSE draws 180 watts, the DIY T2 over 200. Subtract 40 watts of filament power for four EL34 tubes, and that leaves 40 watts. By comparison, the SRM-T1 draws 45 watts, minus its filament power, it draws 37 watts. Now, the theoretical maximum efficiency of a class A push-pull amp is 50%, the theoretical maximum efficiency of a class B push-pul amp is 67%. 15 watts per channel times two channels is 30 watts. With a power consumption of 40 watts that works out to 75%. Ladies and gentlemen, this is not only beyond state of the art. It is beyond state of the possible.

Filament voltage of 5.7v should be fine. David Berning used 5.7VAC filaments in his EA-230 power amp, which was rated as Stereophile Class B back in the early 1980s. I used it to drive my Quad ESLs for many years.

Beautiful work Michael! 12 VDC is perfectly acceptable for offset (remember Stax in their tube amps specifies +/-15V), but if you look at the schematic, the resistors that set the quiescent point are identical between bottom (output) and top (CCS) tube, so I believe the offset depends on the match between the bottom and top tube. Matched octets! Actually, probably 2 sets of matched quads would be fine.

Sorry, but Cavilli wasn't some big company, it was basically one guy, Alex Cavilli, and it's no longer in business.

Speaking of reliability, I was at CanJam NYC this year, they had a HiFiMan table and the Shangri-La Sr. amp was on display - but it wasn't working. Quoth the HFM rep, "Well, it's been shipped around a lot." Uh-huh.

That could certainly be true, but I think that's giving way too much credit to the builder.

And their T2 build pic looks like they stole it from a Kerry build.

I dunno. According to the XE currency converter website, 12k rubles = $211

So, how much did the Canorum V3 cost you, and, if they're so good, why are you selling them?

The TO220 version of the DN2540 is rated at 15 watts dissipation, but I assume that is with adequate heat-sinking. I would bet that the bare TO220 without heatsink will only tolerate a couple watts, so I assume you are mounting all devices on a heatsink.

Decreasing the gain can reduce noise/distortion - however, that is not a problem with this design. It can also increase the risk of the amp oscillating. In case you don't know this, oscillation is NOT a good thing. Further, decreasing the feedback resistors means more of the signal current goes to driving the feedback resistors rather than driving the headphones, which is the whole point of the exercise. In case you can't take the hint from spritzer, here's the bottom line. Unless you think the designer didn't know what the heck he was doing (in which case, why are you building his design?), I would ignore the suggested changes, and build the amp as designed. And this comes from someone who has done a fair amount of circuit modifications.

Apparently according to a post on HF, the Jr. will list at $4k for the headphones alone.

I asked about transformers and the Woo guy said no, but he may have thought I was only asking about output transformers.

I can confirm that the Shangri-La was not working. Someone at the HFM table said it's been shipped around a lot and I just said, "uh-huh," in my best John Wick imitation. the Shangri-La Jr. was working, but the person at the table didn't even know what output tubes it was using, and the HFM literature was just pictures with no text. Also, they had a second HFM stat headphone sitting on the table unplugged and when I started to plug it in to the Jr. amp a person at the table stopped me and said, only one headphone at a time, even though the Jr. amp had two output jacks. OK, whatever.

So, a little more on the new Woo amp. One 6SN7 per side, two 300B outputs, "proprietary output technology." The Woo person refused to say any more about that, except he said the GES was output plate resistor and the WES was output chokes - common knowledge, and implied that this wasn't either of those. My BS sensors were on at that point. I asked about how they got enough gain and he was not forthcoming about that either. Forgot to ask if they are still using passive power supplies. Also, didn't notice at the time, but although it looks like the amp is the upper chassis, the headphone jack looks like it is coming out of the lower chassis, where the power button is. Don't know what that is about.

Will be at CanJam tomorrow AM, but leaving to meet friends at around 1 PM.

Not really. If the parts are fake and you solder them in, what you save in money you lose in time with a broken circuit. And then, you have to spend more time and money removing the bad parts and buying and installing real parts. Not worth it IMHO.

A look on the VInylEngine cartridge database failed to turn up any Apogee. Perhaps a look at that website will refresh your memory: https://www.vinylengine.com/cartridge_database.php

Perhaps you're thinking of one of the Adcom HO MC cartridges. They had a line of them for a few years, got good reviews in Stereophile by Anthony Cordesman IIRC.