Luxman Lv-105

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
luxman lv-103

i also need schematic for the above amp, i had an lv105 which i fixed by taking good mosfets from this amp, the amps used the same trannies in the output stage-2sk405/2sj115...now all four mosfet output trannies are bad....for the lv103, i intend to replace the mosfet outputs with darlington bipolars byt modding the output drivers and bias circuits, but i need the schematics in order to proceed...


calling on netlist, please help!!!!

thanks a lot in advance....
 
Administrator
Joined 2004
Paid Member
Hey these sound pretty good actually!

You can modify the bias circuit to use IRFP fets (140 / 9140). They have low impedance gate drive, so you don't run into the IRFP140 gate charge issues as much. I would rather do that than modify for bipolar output, although it is not that hard to change. It's just that it would be a completely different animal then.
When replaceing the fets, check the gate resistors. If they open, you will blow the new parts for sure.

-Chris
 
Luxman manuals in Sweden

Hi, just wanted to tell you all, i bought two manuals for my Luxman L-114A (service & owners manual) from a guy called Sven Eiman in Sweden. He has a website were he sells manuals and spare parts for Luxman. I payed 110 sek (about 10 €) for both, and they were original in pristine condition. I think he delivers abroad too.

/Andreas
 
anatech said:
Hey these sound pretty good actually!

You can modify the bias circuit to use IRFP fets (140 / 9140). They have low impedance gate drive, so you don't run into the IRFP140 gate charge issues as much. I would rather do that than modify for bipolar output, although it is not that hard to change. It's just that it would be a completely different animal then.
When replaceing the fets, check the gate resistors. If they open, you will blow the new parts for sure.

-Chris

The mod you describe would only work for the 103 for sure (rail voltages!). You may find that IRFP240 and IRFP9140 are FAR better pairs than 140/9140 or 240/9240, assuming the rail voltage is below +-45V at all times. In fact, out of all IRFP parts, 240/9140 are the closest you get to true complementaries, rivaling even the actual lateral complementaries. On average, the P/N ch differences of this pair will be about the same as toleraces between any two parts of the same polarity. In contrast, IRFP140/9140 and 240/9240 differ between 30 and 50% in gain (gm). IRFP240/9140 differ below 5% and typically 3%, plus the capacitances are much better balanced between P and N.

I have also successfully repaired LV103 and LV105 with 2SJ162 and 2SK1058 LMOS, with no mods to the driver. The gate capacitances are still within limits for the driver.
 
Administrator
Joined 2004
Paid Member
Hi ilimzn,
Correct on the 240 / 9240 's for the LV-105. I have easy access to the IRF parts in Canada, my good supplier for Japanese parts closed up shop due to economic reasons. These are from my Adcom stock ;) .

I think my main point was to illustrate the driver circuit in the Luxman units was capable of driving the gates of the IRF parts ( the worst to drive as far as I'm concerned).

-Chris
 
anatech said:
Hi ilimzn,
Correct on the 240 / 9240 's for the LV-105. I have easy access to the IRF parts in Canada, my good supplier for Japanese parts closed up shop due to economic reasons. These are from my Adcom stock ;) .

I think my main point was to illustrate the driver circuit in the Luxman units was capable of driving the gates of the IRF parts ( the worst to drive as far as I'm concerned).

-Chris

Well, they are not THAT bad, unless you need to drive more than one pair. But, the IRFPs that the data says are complements, really are not, and that is usually the culprit for their reputation for being worse sounding (to put it mildly).
For ove 95V rail to rail, the best I have managed is IRFP9240 for the P part and IRF 340 for the N part. On average they differ about 7% on gm, and are pretty well matched capacitance-wise. In contrast, IRFP240 and 9240 differ by as much as 50% on gm, hardly what one would call complementary.
My reasons for using them were originally that I could get 5 IRFP parts for one 2SJ/2SK. Since then, having invested a lot of time into the complementarity issue, I have concluded that the myth of their inferior sound is just that - a myth. Yet, people still take IRFPxxx and IRFP9xxx to be complements if the xxx is the same number. The only thing they are complements on is the maximum Vds, and nothing much more than that. Once you get past it, they reveal themselves to be unbeatable price/performance wise, and not just because their price is low.
 
Administrator
Joined 2004
Paid Member
Hi ilimzn,
I'll believe that 100%. Even the International Rectifier shows them as complimentary pairs. Adcom used them that way (and many others). This is why I have them, for warranty replacement. I'll have to try your suggestion - thanks for the info.

My reason for calling them tough to drive properly is their gate drive characteristic. The gate abruptly changes charge rates as you increase the gate source voltage. The standard TO-92 type transistor, or worse, a vacuum tube will not properly drive these. You need a lower impedance drive circuit and more feedback to compensate. The lateral fets have a much more linear gate characteristic. That means you can make a design mistake (low drive current) and it will not sound horrible.

It would be interesting to swap the fets in a Counterpoint SA-100 and see how it sounds. It just has to sound way better.

-Chris
 
Chris,

The situation is not that bad because it largely happens only under speciffic circumstances. In particular:
- Vgs close to Vgs@Idmax for the part - typically 5-6V on VMOS.
- Low Vds - typically below 10V for VMOS, this is when capacitance nonlinearity becomes very evident
- Common source connection

What this sums up as is, Vd ~~ Vg, in a typical amplifier this only happens when the output voltage and output current are close to maximum simultaneously. That being said, LMOS has a different gate geometry, but it still suffers from the same basic problem, just less - and not even a whole lot less. In fact, for rail voltages below +-45V, one can often use one pair of IRFP VMOS instead of two pairs of LMOS, at which point LMOS loses it's advantage on better gate charge characteristics. On higher rail voltages, in domestic use it is less likely you will encounter operating conditions as above, when gate charge problems become evident on VMOS (and LMOS, just less intensive).

The above is one good argument or DC coupling the tube stage to the MOS output (though it does generate a different set of problems to deal with downstream). Unless very large coupling caps are used (and this can be a problem becvause these often need to be high voltage), the gate charge step tends to reflect on the voltage of the coupling caps, in particular, for small coupling caps, you get a 'rectification' effectsimilar to what happens when the driving voltage on the SS side is clamped. For a while (which the amp needs to spend amplifying signals well below clipping), the bias current is suboptimal, until the coupling caps recharge. The gate charge step generates a less pronounced version of the effect well before clipping, and also produces dual 'notch' type distoprtion, which means high harmonic content (4th and over) - compounding the possible low bias problem.

As you say, the only real way around it is having a lot more drive current than raw gate capacitance figures suggest, by lowering gate capacitances and gate charge storage effects, and/or using larger tubes or an extra SS driver stage for the MOSFETs.
 
Administrator
Joined 2004
Paid Member
Hi ilimzn,

I will agree and add that most mosfet output stages are common source. On top of this, most designs deliver inadequate drive current to the gates. I've often said that you need a bipolar power output stage to properly drive gate charge in the vertical type fets. This is to avoid the double notch type distortion. Obviously you can use a fet driver as well but I am just making a point. The "power" stage I refer to is enough to dissipate the standing power from the quiescent current required.

An intermediate stage between the coupling cap in a tube stage and the output fet stage would solve those problems and bring a higher degree of safety to the design. The aim of the driver would be to present a more constant impedance to the VAS stage and thus linearize the output stage. I did this with a bipolar output with very positive results (driven by a tube stage).

I like using Counterpoint as a model because it falls short in many areas and illustrates well the problems with this topology. It's also an easy design to understand for those who are learning.

-Chris
 
Sorry for hi-jacking this thread, but I cant believe I've found some people that know about the LV-105. I've been trying for the past 8 or 9 years to get my fathers LV-105 repaired as I damaged it while playing my music a little loud, but have never been able to find a place that will fix it. The amplifier comes on and the tubes light up and what not but it's supposed to "click" when it's ready, but it never does. MOSFET's possibly?? I brought it to a place to have it looked at when it first went, and they said it was some overload chip that died and they could not get parts. Does anyone know of a place that will service LUX or know how to fix it?? Thanks, and look forward to hearing from you.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.