How many watts per channel do you run with?

I know I'm running BELOW 2-Ohms.... 1.8 static... probably down to 1 or less at decent volume.

The Apogee Scintilla ribbons IIRC were 1R nominal, and notorious amp killers, so yes it's possible, and with lightstar the rail voltage is withing 5 volts over the head of the output transistors at all times, so 17 amp paralleled transistors can squirt 34 amps.

Damn! I had typed out a nice long reply. However, apparently I had not logged in and lost everything I had typed.

Don't worry about raining on my parade. I have a strong ego.

Let me make this a bit clearer. While the Moscode 401HR use six output per channel, they are not the same MOSFETs that were used in the New York Audio Labs Moscode 600. The older devices are capable of higher current output. 600 was rated at 300 W per channel into 8 ohms and 500 W per channel into 4 ohms. Harvey Rosenberg drove both an electric drill and a blender with the output of the Moscode 600. The amps are stable into 1 ohm.

The Martin Logan Prodigy is a nominal 4 ohms speaker. At one point I asked George Kaye, the designer of both amplifiers, what I could expect out of a single 600 bridged. He he said that the amp would deliver 55 volts/channel at 20 Hz and about 1100 to 1200 watts. That was of course into an 8 ohm speaker. I was planning to use autoformers to double the nominal impedance of the Prodigy to 8 ohms. As you know, a common ground amp that is bridged, when looking at an 8 ohm speaker, essentially sees a 4 ohm load.

Since one channel of each amplifier is connected directly to the woofer section of the prodigy's (2 10 inch drivers each) with a nominal impedance of 4 ohms and the other channel of each amplifier is connected directly to the electrostats with a nominal impedance of 4 ohms dropping to 2 ohms at 20,000 Hz, I think the 500 W per channel or 1000 W per amplifier is well within the stated capabilities of the amp. I can see how my original post could have been confusing, indicating that the entire panel was a 2 ohm load, which it is not.

When I add each individual 200 W subwoofer amp to the two Moscode's, I think I still come up with 2.4 kilowatts.

Yes it does sound amazing!!!!!!!!! I am constantly tweaking.

Roger Sanders is spot on with his wire recommendations. When I first tried the Belden in place of my Cardas hexlink golden 5-C I noticed very nice additional ambient recovery, but some loss in dynamics and high-frequency extension. After five days of burn-in on my DUO-TECH the Belden had retained all of its wonderful qualities plus had acquired the dynamics and high-frequency extension. I am now planning to open the prodigy and run Belden from the back of the binding posts up to the crossover for the electrostatic section.

Another amazing tweak was the recently acquired Blue Circle Audio BC 68, designed specifically for Martin Logan electrostats.

Voices are so palpable it is absolutely spooky. It is without a doubt a midrange to die for.

I got lost early. At 55 v/ch. into 8 R yields a little over 378 watts/ch (Watts = V²/R)

......When I add each individual 200 W subwoofer amp to the two Moscode's, I think I still come up with 2.4 kilowatts.......

I hesitated to post in the first place, but did mainly for the benefit of anyone reading the OP and thinking they can just double their 4R rating to determine what their amp'll do at 2R.

Nice amp, and nice ML's.......<i don't see an emoticon for 'green with envy'>

Bill D and Rich.

I think you can help me out on this. Obviously I am not an engineer, so electroncis theory and the math for it are not my stong suite.

But... George said the amp would deliver 55 volts at 20 Hz. Now I am not entirely sure what that means. The amps have 90 volt rails. When bridged the speaker would "see" 110 volts at 20 Hz because the speaker terminals are not connected to ground but rather the two positive terminals.

It is often heard said that bridging gives one four times the output (assuming no power supply sag or limit). An EE friend of mine pointed out that that is somewhat just semantics.

Bridging, as it is generally done and my friend calls "series" bridging, does not give one four times the output of the stereo amp, while it does give all the output of the stereo amp to essentially on channel.

To put it another way and in round numbers. Usually an amps output is stated into 8 ohms. I have heard it said that a 50 watt per channel stereo amp will do 200 watts bridged into mono. That does sound like 4 times the output but it is not.

Firstly the amp can produce 100 watts with ea channel producing half of that output. So, a 200 watt mono amp is really just twice the output of the stereo amp, but again it is not. The stereo amp will, theoretically double its output into 4 ohms (a few with huge power supplies actually will). So our 50w/ch stereo amp will do 100w/ch into 4 ohms and when we add those channels together we get a total output of 200 watts from the amp. When we hook up an 8 ohm speaker to the bridged mono amp it is effectively seeing a 4 ohm load and therefore a bridged 50w/channel stereo amp can deliver no more power than the stereo amp unbridge. It would of course be flowing more current with the 8 ohm speaker while bridged, just as it would in stereo with two 4 ohm speakers.

The power supply and output devices of the Moscode 600 are capable of producing 1100 to 1200 watts, whether bridged or not. George did build some bridged Moscodes for a recording studio, because that is what they wanted. It was not his feeling that the amps sound as good bridged as they do in vertical biamp application.

I, of course, had to learn for myself. I obtained a DB Systems DB-7 bridging/phase inversion box and bridged and ran the Prodigy full range. It DID NOT sound as good. George was right again. It was not the DB systems unity gain stage screwing up the sound because it is still in the chain being used for phase inversion only.

The bottom line on my power is that I THINK I have a lot. Wether I have what I thought or not is moot. I do have a lot of headroom, more power than I ever use, especially with my nearfield listening and am really, really happy with the sound.

I have some Russian NOS (1979 and 1982) 6H30p-DR in the inputs and some Ediswan NOS in the driver positions. It is really sweet and juicy like these nutz..

OK Show me the math. So, I can become unconfused. You know that the rails are 90 volts, the MOSFET output devices are QN113 (8310) or K176 (Japan B1), I think Toshiba. There are 6 devices per channel, theoretically how much current can these devices deliver?

Thanks for the help.

Bruce

.........George said the amp would deliver 55 volts at 20 Hz. Now I am not entirely sure what that means. The amps have 90 volt rails.

The rail voltage may be 90V, but when loaded the PS can't sustain it....what he's saying is that into a load his amp will sustain a 55V from the 90V rail.

If the power supply is a water pump, it's pushing a stream of water 90ft into the air. Switch from a needle hole to 1/4 inch pipe and now it can push 50ft.

...... When bridged the speaker would "see" 110 volts at 20 Hz because the speaker terminals are not connected to ground but rather the two positive terminals.

It is often heard said that bridging gives one four times the output (assuming no power supply sag or limit). An EE friend of mine pointed out that that is somewhat just semantics.

There's alot of semantics when discussing bridging, but you've got the gist of it right.

.....Bridging, as it is generally done and my friend calls "series" bridging, does not give one four times the output of the stereo amp......When we hook up an 8 ohm speaker to the bridged mono amp it is effectively seeing a 4 ohm load and therefore a bridged 50w/channel stereo amp can deliver no more power than the stereo amp unbridge.....

Here's the semantics part. The amp can't see, and the load doesn't mysteriously change. The load is the load and it's strapped across the hot side of both channels. Since the amp is now being asked to deliver twice the voltage, it needs to provide twice the current. This is effectively the same as halving the impedance, so in a sense it does become a question of semantics.

..........You know that the rails are 90 volts, the MOSFET output devices are QN113 (8310) or K176 (Japan B1), I think Toshiba. There are 6 devices per channel, theoretically how much current can these devices deliver?
.....

I can't find datasheets for these devices, but if the designer said it'll sustain 55V that's 378W into 8R. By way of comparison the M-1.0t i'm playing with will do 400W into 8R (56V) so it's rms is close.

With dynamic 20ms pulses it'll do 928W into 2R, and as long as the PS can provide the current, those MOSFET devices should handle at least that much dissipation.

Rich,

Thanks for taking the time to go through all my babble.

can't find datasheets for these devices, but if the designer said it'll sustain 55V that's 378W into 8R. By way of comparison the M-1.0t i'm playing with will do 400W into 8R (56V) so it's rms is close.

So, if I understand this, the claimed spec of 300 watts per channel 20 to 20k, both channel driven into an 8 ohm load is more than met. And the amp probably will deliver 500 watts or more per channel into a 4 ohm load.

With dynamic 20ms pulses it'll do 928W into 2R, and as long as the PS can provide the current, those MOSFET devices should handle at least that much dissipation

So, your amp more than doubles its output into a 2 ohm load, at least for 20ms. Since, I believe, there is very little content in the 20kHz range and my speaker approach 2 ohms as high frequency reproduction goes up to 20kHz, the amp would be seldom asked to drive such a low resistive load (I understand that resistance and impedance are not the same, but not why).

That being the case, very little content up where the speaker impedance is very low, why do many amps have such a time with electrostats? Is it the capcitive load they present or a combination of that and the low impedance at higher frequencies?

Thanks for you time and expertice.

Bruce

The speakers are rated 4 ohms, that doesn't mean that they don't dip down lower that that. For example, the Original Amazings were rated at 4 ohms. During certain transients, they would dip down as low as 1 ohm which not many amps are designed to handle a load that low. Electrostats are the same way, rated at 4 ohms but can dip lower. Both of the speakers also have a low sensitivity as well.

It just occured to me that there is another question that I did not ask the designer, or anyone else.

I have made the assumption that George meant that the amp will deliver 55 volts at 20 Hz was into an 8 ohm load, based on the question I asked him. He did not state at what load it delivered 55 volts.

Would a theoretic amp or my amp or any amp deliver a different voltage at half the load, just as it would flow more current (within its design limits) at a lower resistive load?

Thanks for the response.

Bruce

If the first watt sucks, why bother!?!

.....Since, I believe, there is very little content in the 20kHz range and my speaker approach 2 ohms as high frequency reproduction goes up to 20kHz, the amp would be seldom asked to drive such a low resistive load (I understand that resistance and impedance are not the same, but not why).

I haven't seen a freq/impedance plot for MLs, so I don't know at what freq it gets to 2R, but that's the gist of it...as freq rises the impedance shrinks. In DC circuits there's a resistance to current flow, when you introduce an AC signal now you're talking impedance, and it'll vary with freq. Inductors are wound coils of wire....as freq increases the collapsing magnetic waves oppose the current flow more and more, so impedance rises. Capacitors are two plates separated by a dielectric......as freq decreases the period of the signal approaches the time constant of the capacitor, and impedance increases. It kind of 'fills up' and starts blocking signal instead of passing.

.....That being the case, very little content up where the speaker impedance is very low, why do many amps have such a time with electrostats? Is it the capcitive load they present or a combination of that and the low impedance at higher frequencies?.....

I don't know all the reasons, but:

Electrostatics are basically a huge capacitor...as freq increases impedance decreases, presenting a harder load to drive, as you've said.

In addition, the reactive (mostly capacitive) load means the voltage and current will be not inphase, as in DC. With an output device rated at 15Amps, that spec is actually best case, with little or no voltage across it...basically the transistor is fully 'on'. In a purely resistive circuit at zero crossing the transistor will have the rail voltage across it, and not be able to pass much current, but won't need to. In a purely capacitive circuit current leads voltage by 90 degrees.

If you imagine an AC signal oscillating between 100V rails, in the resistive case max current will be demanded at max voltage, which is okay since that's withing the safe operating area of the output device (say 15 amps). Take the reactive case and (90 degrees is a very extreme example, 25 is more likely for a crossover, but) now max current is demanded 90 degrees sooner, or at 50V. With 50V across it, the device is no longer able to tolerate the current demand, so it's harder to drive a reactive load.

Another aspect of electrostatics makes it even harder. A typical cone speaker is a current device......run more current through it's coil and it produces a stronger magnetic field, which more strongly opposes the magnet surrounding the coil, which more forcefully moves the cone.

An electrosatatic is a voltage device....introduce a higher voltage to the membrane and it is repelled more strongly by the electrostatic field surrounding it, which moves the membrane more forcefully.

This means that a given amp may play typical speakers to very loud levels with 50V. As the drivers impedance changes, it'll have to source more or less curent to do that, yes. Now take the electrostatic. To get to the same loudness level it may have to provide 70V. Add to that the fact that electrostatics are typically less effecient than cones and it may be 80V (these numbers are from a hat btw). Compounded with the reactive nature of the load, you're taking an already higher demand for current and just making it worse.

...Would a theoretic amp or my amp or any amp deliver a different voltage at half the load, just as it would flow more current (within its design limits) at a lower resistive load?....

The ideal amp would produce a constant voltage regardless of load. The lighstar and Sunfire's evidently do this, with their tracking-down converter. When the load halves they double the current to sustain the voltage.

It also gets down to power supply design. Some go for a tightly regulated supply and massive capacitor banks with lower rail voltage. Some go for looser regulation and higher rails.

I like this idea because it can provide higher voltage for transients, and doesn't lead to an 80lb monoblock.

With the Carver M-series, they are loosely regulated, The M-1.0t has a 100V rail, but that sags to 55V at 8R (still very good) and dynamically it'll provide nearly 70V. Halve the load and it'll do 42V and 60V, which is not theoretically ideal, but is still very good at more than 900W into 4R

F1nut,

Well great! I am really happy the first watt sucks. That's a good thing, right?

Bruce

Rich,

Thanks for both those explanations. It really does help.

I had a basic understanding of caps and wire inductors. That makes a lot more sense why electrostats are hard to drive and what people mean when they say that estats are voltage not current driven speakers.

I appreaciate your time.

Bruce

TFM 55 on a pair of Apogee Stages.
Previously used a pair of modified Dynaco MKIII`s approx 60 watts a side, Bryston 4B 250 a side,Adcom 555 MKI and MKII 200
a side. and various others pre Apogee