|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
64.122.72.31
In Reply to: RE: 0.5A? posted by Lew on January 25, 2014 at 12:49:09
So this value looks like it will be safe so that the tubes don't roast if the B+ is higher than I am expecting.
Follow Ups:
Si tells us that the meter range is from zero to 20A (see above). Does that cause you to re-think? Anyway, the Ampere reading shown when the tubes are balanced would be a good place to start, yes?
I don't think 20A is right. I bet its more like 2.0 amps. A 20A scale would not allow for accurate reading of the bias.
In a Futterman the tubes banks are in series, so you can read the current through the entire amplifier with a single reading.
Dear Ralph,
It is a 20 amps meter, for confirmation kindly see the picture and advise further, please.
Thanks for the info.....can we see more pic of the amp??....not the OTL we see everyday................looks vary cool.......
It would not have a reading of '05' marked on it, would it?
At any rate we can be confident that there are no currents anywhere in the output section that approach such a value, but there would be some that would be 1/10th as much, so I think it reasonable to surmise that, despite the markings, this is a 2Amp meter.
What values have you been reading on it so far?
I think the markings on the meter face in between the first and second numerals in each case are meant to serve as decimal points. Thus O^5 is meant to be taken as 0.5A, etc., as Ralph says. But your confusion is quite understandable; it's just that 20A makes no sense.
Not to mention the fact that even half of 20A, 10A would tax most circuit breakers. Good point.
Interesting you mention that Lew. I used to have Silvaweld OTL Referernce monoblocks (4 x 6C33C per channel...huge suckers) that I had to start sequentially so they didn't blow the circuit breaker too often. The first turn on ALWAYS tripped the breaker and the second time was ok (guess the power supply was already partially juiced).
Still one of the most spookily real sounds I have ever gotten from amplifiers, made my Acoustats sing with joy...but oh the HEAT!!! Rasied the T in my room by 10°C in about 2 hours! Couldn't live with'em.
At one time, I had an all Silvaweld system, OTL monos, SWC1000 preamp and the SWH650 phono. Now I have just the phono because I went to integrated amps. The KR VA350i and NAT Symbiosis I have now have elements of that spooky realism of the OTLs but not quite. The one thing that bothered me about them was that the sound while super transparent and real in some ways was also somehow incoherent in the imaging and stageing and I think it had to do with what I observed on the bias and DC offset meters during music swings.
The needles "pulled to one side of scale, indicating that the matching between tubes was off DYANMICALLY and the bigger the swing the worse it got. Perhaps that is what is happening with all of the Totem-pole type OTL but without meters you don't see it. My SET amps always sound rock stable once setup and that was something the OTL never sounded like...it was nervous sounding because the image and soundstage were not stable and it appears the amp was also not stable...at least into an electrostatic speaker load.
I regret selling the SWC1000 preamp...that was a real reference product.
Brad, I think you are doing what all of us audiophiles are guilty of at one time or another. You've heard something in your system with the Silvaweld OTLs and you've made an observation of the meter swinging, and you are putting the two together to draw a conclusion ("OTLs cannot drive ESLs") that is not justified by any actual data. I don't know about the Silvaweld OTLs, but you cannot find a better way to drive an ESL than with an OTL, except if you can bypass the audio transformer and direct-drive the ESL with an OTL. Over a 35-year period, I have owned nothing but ESLs and driven them with either a Futterman-type amplifier or a Circlotron (Atma-sphere) amplifier. OTLs are particularly stable with an ESL load, vs any other common amplifier topology, solid state or tubes. (I've tried them all, but that's not the point. The point is based on the physics of the situation.) This is not to say that you personally are "wrong" in preferring your particular SE amplifier to your particular pair of Silvaweld OTLs.
As to heat, I have run six 6C33Cs per side in my Atma monoblocks, so I know what you are talking about. Now I run four 7241s per side. 7241s get even hotter, but it seems that fewer tubes trumps using slightly cooler tubes; the 7241 amplifiers do not drive me out of the room in summer.
Hi Lew,
I am quite aware that done properly direct drive OTL is the best way to drive ESLs. However, I have found that they are not all completely stable into those kinds of loads.
For example, I built a pair of Transcendent Sound Beast monoblocks and after completing the build and having it double checked by an EE friend, I hooked them up to my Acoustats and at anything beyond a whisper the DC offset started swinging back and forth wildly and the Whub whub whub sound from the speakers, indicating that the amps were oscillating and unstable into the Spectra 2200s load.
The Silvawelds were not actually unstable but it seems clear to me that they were not remaining well balanced under dynamic conditions, which definitely would increase distortion and it was audible.
but those two instances do sound a bit odd. I wonder why that would happen; I know of no special problem with the two speakers you mention. But I don't know whether each of the two amplifiers is a totally kosher Futterman type, either. I think Bruce Rozenbitt adds some tricks to the standard formula for a Futterman, and Mr Park and his colleagues were a bit unorthodox in their Silvaweld designs, if my SWH550 phono stage is any criterion. (Not even Mr Park has a SWH550 schematic; I traced it painfully with my meter.) Maybe Ralph can comment. Even with loads that were very unsuitable for it, my old Futtermans never malfunctioned. It could sound bad into low Z loads, but it tolerated them in terms of stability. Circlotron will drive anything, but again low Z speakers are not the best match. However, the circlotron just seems to lose power into low Z, whereas the Futterman loses sound quality as well, suggesting distortion goes up. Did you consult with Rozenblitt, re your experience with his amplifier?
As you may recall, you helped me out when I was deciding on purchase of the SWH550. I've enjoyed it very much. Thx.
Well the Spectra 2200 go from 2 ohms in the bass to about 16 ohms in the mids back down to 2 ohms in the highs. Couple that with the high capacitance and the fact that "The Beasts" and the Silvawelds both have about 10db of negative feedback and maybe you have a recipe for oscillation.
At this point, if I ever get electrostats again I will direct drive them with something like an Acoustat Servo Charge amp (there is a shop in Holland that makes a pretty nice derivative of this amp), or a circlotron OTL (Atma-sphere or perhaps Joule Electra).
Is the Joule a bona fide circlotron? There is not much info available. It seems they use a single-ended input stage of some kind.
Is Joule still in business?
To the best of my knowledge, yes it is a circlotron... at least the output stage.
I revised the RIAA circuit inside my SWH550 quite extensively, while conserving intact the very nice tube-rectified, choke-loaded PS and voltage regulation. (I did nothing to change the values for RIAA; I changed only the input and output sides of the audio circuit.) Recently, I have been using a Aesthetix Janus full-function preamplifier as the source for my second system, using Beveridge speakers. The Janus was a gift from the estate of my close friend who is now suffering from dementia, very sadly. It's obviously a great piece, but I had been vaguely dissatisfied with phono. Last night I hooked up my Silvaweld to one of the high level inputs on the Janus linestage section and proved myself correct. The Silvaweld just blows away the Janus phono (using input from a Grace Ruby cartridge on a Dynavector tonearm). Holy cow! I was up until 2 AM listening.
In my main system, the Atma MP1 sounds noticeably superior to the Silvaweld. So, the Janus has a way to go. At some pint, I will mess with it.
Re this system; did you give up on the Klyne phono stage in the end? I wouldn't mind hearing more about this. Also more comments about Beveridge vs Soundlabs.....
Hi Lew,
Remind me again, what did you change in the circuitry of your Silvaweld?
First, I sort of disconnected the high-gain FET-based input section, because I only wanted it for MM and MI cartridges. So the input goes right to the grid of the first tube. In the 550, he used two SRPP topologies, one at the input that adds gain for the RIAA, based on the two halves of a 12AY7/6072. I did not want to change that topology at the input, because it would change the output impedance of that gain stage and I would have had to re-design the RIAA network, which is above my job description. This first SRPP had a resistor from cathode to ground bypassed by an electrolytic. I removed both parts and replaced them with a pair of Schottky diodes in series, ground to cathode. This biases that tube at about -1.7V. Then I changed the topology of the second SRPP so it is now a plate-loaded common cathode stage. (In other words, I take the output off the plate of the lower tube, instead of the cathode of the upper tube.) Here too, I replaced the cathode resistor and bypass cap with a Schottky. The change from SRPP to plate-loaded gives a higher output impedance but it sounds better. Since this stage feeds a CF, it is no problem to raise its output Z a bit. Thus the plate of the lower tube is now direct coupled to the CF output stage. There I disconnected the feedback loop between the cathode of the CF and wherever it went and instead installed a solid state constant current source between cathode of CF and ground, using a kit supplied by K&K Audio. I also replaced the tube used for CF from (I think it was) a 12AY7 to ECC99, which has a much lower Rp, and gives a lower output Z for the whole circuit. (ECC99 is completely compatible with 12AY7 socket.) I am using Mundorf silver/oil output coupling capacitors. Gosh, it's goooood.
I am wondering if my circuit is really the same as yours because I have only 3 tubes in the whole circuit (not counting the power supply of course). If the Input for each channel is SRPP then that would be 2 out of 3 tubes used already. The last tube, I am assuming after the RIAA network, would have to serve both channels and could not be a second SRPP + cathode follower, which would require at least 2 more tubes than my phonostage has (2 tubes more for the second SRPP one per channel). I suspected that my input was SRPP then the network and then I thought either a CF out but I didn't think that would be enough gain or a plate loaded common cathode (but then the output Z would be high) for additional gain.
Mine also has the JFET inputs for MC, which I use even with MM and it sounds better that way.
First, each of the tubes is a dual section triode. The SRPP is constructed using one half of a triode for the bottom section and the other half of the same tube for the top section. It does not require two discrete tubes to construct the SRPP.
IIRC, my unit has 5 tubes. Two tubes PER CHANNEL are used for the input and output voltage amplifier functions I described. That makes 4 tubes. Then, each half of the 5th dual section triode is used as a CF, one half for each channel.
I've traced the entire circuit of the SWH550. You may recall that I was unable to get a schematic from any source, including Mr Park. So I finally and painfully figured it out myself. I now have a schematic of the original circuit. In the course of my search, I also looked for info on the SWH650, which is what you have, I think. No luck there either. I am rather curious regarding differences. Does your unit have a separate outboard PS chassis? Mine has the power trannie, a tube rectifier, and the input choke and large capacitor all mounted outboard. The transformer is impressively massive. Then on the main chassis, the left hand half has voltage regulators and CCSs, all constructed from tubes, plus the DC filament supply. The audio section is confined to the right hand half of the PCB. My unit has an attenuator mounted in the center of the front panel. It's pretty extreme, so I wondered how the 650 would be any more elaborate. (I assume the the 650 is in some way a step up from the 550, only because of the model number progression.) I did nothing to the PS except to upgrade some capacitors from electrolytic to polypropylene film and to add some fancy film type bypass capacitors, just empirically.
Ah so mine is different because it has only 3 tubes so its not possible to have two SRPP circuits per channel like yours has then the fifth tube making the cathode follower.
Does the 650 have a separate PS chassis, like the 550?
Hi Lew,
No, the power supply and circuit board are in the same full size chassis. The power supply has a pretty large power transformer, a big choke and it uses a 6C19PI power triode for the regulation. There are a couple of other tubes (rectifier and a voltage reference I think) in the circuit as well. I think a big part of the good sound is the power supply, which seems well done.
The circuit has a couple of JFETS for the MC input then there are three tubes in a circuit that I do not know how it works. Coupling caps are all Auricaps, which I find to be a very good cap.
I had assumed that the input was a cascode to lower noise and the output I was unsure of. The RIAA is passive.
Seems like the two power supplies are analogous but perhaps the 550 PS is more space-consuming, since the outboard chassis is needed. The RIAA circuits must be different, as you suggest, based on the difference in the number of tubes, if nothing else. What tube types are used in yours?
Mine uses a pair of 12AT7s at the input (one per channel), then 12AY7s for the second gain stage after the RIAA, then it had another single 12AY7, each half of which was used for the CF for one channel. I replaced that latter tube with an ECC99, blessed as a good idea by Mr Park.
If you can send to my private email or post some photos, that would be cool, and I can do the same.
All three are 12AY7s.
As I think I mentioned, my 550 is switchable for "MM" and "MC" levels of gain. There is a toggle on the front panel. In MC mode, the input signal goes to a JFET which just adds gain and then feeds the aforementioned RIAA circuit, which is all tubes. In MM mode, there is a relay that shunts the JFET to ground, thus bypassing it. Does the 650 also have provision for MC cartridges, using JFETs, or have we been over this ground, already?
Yes, but I will tell you again. THere are JFETs that switch in and out from a front panel switch...just like yours I think. I found that even with a high output MM that the MC input sounds better.
(1) I apologize. You did clearly tell about the JFETs before I asked my stupid question about whether it can handle MC cartridges.
(2) You don't mean to tell me that you use the MC inputs for MM and other high output cartridges, do you? Most likely that would result in dramatic overload of the RIAA circuit. I must be mistaking your intended meaning.
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: