Edge laser amps

[Prune]

Huh? Is this snakeoil or not? From their website:
http://www.edgeamp.com/technology.html

The NL series of amplifiers utilizes Edge's proprietary Laser Optical Bias Circuitry. This new circuit incorporates a 630 nm wavelength laser in each channel operating directly on the silicon substrate of the bias transistors. The laser bias circuit in turn is part of a servo feedback network. As the signal rises, the output of the laser increases, and as the signal decreases the laser output falls. This keeps the bias at the optimum operating point allowing a tenfold increase in gain. The additional gain is achieved while maintaining signal integrity (linearity) -- at light speed -- a first for solid state amplification. The result is the clear musical nature of the NL Series as observed by well-informed reviewers and buyers

Anyone have more info?

 

[Nysan]

??

Sounds like BS to me ....
Why should an optimal stable bias result in a tenfold increase in gain?
Im assuming they are talking about Vbias for the output stage ..

/Dave

 

[Nelson Pass]

Sounds like they are using an optical isolator to adjust bias.
I refer you to my patent pages at www.passlabs.com

 

[ingvar ahlberg]

Vactrols are among the parts You always have to replace in old guitarramps. 630 nM is visible red light so it seems to be just what mr Pass says, an ldr.

 

[cowanrg]

i live in boulder, and have talked with the guys from edge... they say its really more wow factor than anything, but basically, a laser is fixed on an open transistor, and it changes the characteristics of it. supposedly the actual power transistors are NOT being laser-baised.

 

[sss]

this "Laser Optical Bias Circuitry" ment to confuse regular people that know nothing in electronics so they will think its doing something usefull because its "laser" wow !! cool

 

[Prune]

That's what I thought, too.
On the other hand, reviewers seem to like them, even at the 75 grand price (or because of it ...)
It's the sort of stuff that makes me think of blind tests again

 

[cowanrg]

they do something. im not good enough with electronics to know the explaination of that they did, but it does do something. it changes the properties of the transistors, thats all i know.

 

[sam9]

Here's a guess. If they aim a laser at a transistor, the transistor is going to be warmer than it would be if no laser were aimed at it. If the intensity of the laser depends on a voltage (but which voltage??), then by manipulating the the temperature of the transistor they are are manuipulatiing the conductance. So I supose it could indeed be a way of controlling the bias. Perhaps it gets you better (i.e., quicker) thermal tracking than the various schemes involving mounting the Vbe multiplier on output devices, heat sinks or what not.

Clever, but are more mundane schemes really so lacking that this makes a difference?

 

[Eva]

When light is applied to a reverse biased PN diode juction, leakage current increases [the more light applied, the more leakage current]

The same happens for C-B junctions of NPN and PNP transistors : When light is applied to the juction, its leakage current increases and thus base current and colector current increase

All bipolar transistors and diodes are subject to this light-dependent leakage phenomena, but obviously, all devices except phototransistors come into an opaque case that doesn't allow external light to reach the junctions

Also, an optocoupler is nothing but a LED diode and a bipolar transistor placed in the same case in such a manner that LED light is allowed to reach the junction of the transistor

So at the end, all that laser ******** may be nothing but an exotic optocoupler. Also, optocouplers doesn't work at light-speed since bipolar transistors and LED diodes are much slower [see datasheets]

 

[sam9]

On reconsideration, the idea has a bit more merit that I thought. (Assuming I'm guessing right about the general function.) Such a scheme might well provide the close thermal tracking you can get in CFB output section, but while using an EF design. It might also make physical layout of an EF section more flexible by removing the requirement to place all output devices (per channel) on a single heatsink.

All this is idele speculation, of course.

 

[sss]

So at the end, all that laser ******** may be nothing but an exotic optocoupler. Also, optocouplers doesn't work at light-speed since bipolar transistors and LED diodes are much slower [see datasheets]

correct
these days its not so hard to make an amp with stable bias circuitry

 

[Eva]

I forgot to mention that the use of an optocoupler to ser the bias of a class AB output stage and operate it in 'pseudo class-A' was patented long time ago

I don't remeber where I saw the patent, I think it was in Nelson Pass site about 3-4 years ago

 

[Nelson Pass]

My patent did not distinguish between amounts of bias or
particularly a mode of operation. It simply provided feedback to
the bias voltage generator about the common mode current in
the output stage, and an optoisolator was an ideal way to
avoid having the output stage load the bias circuit.

 

[Christer]

The suggestions that the original text might refer to an opto-coupler
sounds like a reasonable assumption. However, at least AFAIK,
there are no optocouplers using a laser LED, so unless I am wrong
here, something is still fishy, or just marketing crap.

Edit: Well, there could of course be a combo of a laser diode and
a photo transistor, not on the same chip, but why do that is
beyond me. Maybe marketing again?

 

[Eva]

Take a 2N3055, drill a hole on top its case and stick a laser diode on it

Long time ago, when phototransistors were very expensive and hard to find, people used to drill holes in the cases of 2N3055....

 

[Christer]

Take a 2N3055, drill a hole on top its case and stick a laser diode on it

Long time ago, when phototransistors were very expensive and hard to find, people used to drill holes in the cases of 2N3055....

Well, that's certainly creative engineering, but I thought
semiconductors very much preferred to be in airtight packages.