LM1875 as transconductance amp for Fostex

[m@]

I've searched for info on this to no avail, but excuse me if it's been discussed ad nasuem.

I'm planning on building a MLTL for a set of Fostex full range (from 4.5 - 8 inches, undecided).

I was thinking about using an LM1875 in 'current source mode' to drive it a la Nelson Pass.

It's my understanding that to do so you put the speaker in the feedback loop with a resistor (znom/10) to ground.

My question is, given the wide bandwidth this would allow, will there be stability issues with doing this? Is this an oscillation magnet or otherwise going to be a nightmare?

I was going to put the chip in the speaker cabinet somehow to make an active speaker.

Has anyone done this successfully? Any hints you can share?

Any/all feedback is apprecaited. Cheers.

[Igla]

Well, here is my version of mix-mode amp. I think it will drive Fostex speakers just fine. You can tune it with R1, R2 and R3 for your needs.
Don't forget-R2 is dissipating some power, so calculate the wattage you'll need.
Changing the resistors you can set the sound you like.For example-changing
R3 from 4,7k to 1k and the mids and hights
starts to approach ordinary GC sound.
You must try for yourself and find the right combination.
I've tried different values for the resistors and found the
best results with values:
R1=220 to 330 ohm
R2=0.22 to 0.33ohm
R3=1k to 4.7k

[m@]

Thanks for the feedback.

Are you suggesting that a fostex would need some dampning, instead of using a 'pure' current source? Or are you fine tuning the sound in general with that network?

Have you noticed any stability issues doing this?

[planet10]

A combination of reading a review of the Lizard Wizard (or something like that) -- a 3875 transconductance amp) -- and getting a chance to listen to one of Duo's tranconductance amps has me very interested in this.

Duo's amp is ~15w, has an output impedance > 1M ohm, and was not designed with speakers in mind (it was designed to drive a CRT), We tried it on 3 different Fostex... on the FE108eS in Aiko it was bloated in the bass. Not near enuff control. With the Fonkens it sounded really good but just a bit fat in the bottom (some might like that). Driving the µFonken just plonked on top of the Fonkens it was superb (but Duo reports that the same with the speaker ssitting on the desk either side of his computer monitor it wasn't so good)

He promises a variable Z transconductance amp designed with audio in mind... you turn a knob to dial the amp in for your speaker & room.

dave

[m@]

Interesting. My fostex experience is pretty limited (limited to smaller drivers) so I couldn't imagine having too much bass. I guess it will be necessary to adjust depending upon the driver and enclosure.

So anyone around have thoughts on stability?

Every gainclone I've built osciallated as some point. One was input/output coupling and another was a misplaced zobel, so they were both manageable. I am hoping that opening up the bandwidth doesn't make it worse.

For the record, I've cooked at least 6 LM's and almost every line level opamp project I've put together went crazy. I seem to be a magnet for out of control opamps.

[Igla]

I had no problem with stability. Like I said, you can adjust the sound to your leaking. This amp suppose to go well with fullrange drivers with lightweight membranes helping them in the bottom region. Check Pass Labs site - there should be an article about amps driving fullrange drivers.
Of course, ordinary loudspeakers might sound bloomy with such amp because of the damping factor.
Regards;

[gootee]

Another name for a transconductance amplifier is "voltage-controlled current source", commonly abbreviated as VCCS. You can find some of the various usable circuit topologies and design considerations by googling for something like 'VCCS schematic OR circuit OR design'.

[Archwn]

Hi All,

May I ask if it is the same as buffer? i.e. gain = 1 (more or less)

[BWRX]

The gain of a transconductance amp is expressed in A/V. If you have a transconductance amp with a gain of 1 A/V then I guess technically you have a transconductance buffer, although buffer most commonly refers to a voltage buffer.

[KSTR]

Originally posted by m@
I was thinking about using an LM1875 in 'current source mode' to drive it a la Nelson Pass.

It's my understanding that to do so you put the speaker in the feedback loop with a resistor (znom/10) to ground.

This would be a true transconductance amp, Zout=infinity. It requires stabilization means, to counter the inductive effect of the voice coil. You must make the speaker load look resistive for HF, and you must obey the minimum gain requirement for the chip (or use noise gain compensation for lower gains).

Details see:
http://apex.cirrus.com/en/pubs/appNote/Apex_AN13U_C.pdf

- Klaus