Luxman L100 pointers needed

Hi.
I am working on a Luxman L100, which has weird behavior. It sometimes works, other times one or both channels will output some 40v to 50v DC without the protection kicking in. I think it is a gnd issue, but haven't seen any bad wires, traces or solder joints. When it works, the sound is good, idle current and DC offset is within spec. It will sometimes work for half a day at a time. Temperature does not seem to influence it much, as the weird behavior sometimes appears from cold start.
Other times it will play a while before the error occurs.

I would be happy if someone could give some pointers on where to look first, as this amp isn't exactly the most fun to poke around in.

/Jørgen
 
Administrator
Joined 2004
Paid Member
Hi,
I was authorized Lauman warranty and worked on those models.

Basics. If you have a DC offset issue it is generally in the voltage amplifier section. If an output shorts it will blow a fuse and / or emitter resistors plus other stuff. The short will drive the output to that supply rail. The voltage amp section sees this and turns on the other output as hard as it can. That means uncontrolled current flow through all the outputs on that channel. Replace all and the driver transistors for certain on that channel. You would probably also lose some resistors that may look fine by sight.

Failure of the protection circuit to operate is a totally different failure. Also, with high current events to the speaker, the relay contacts can melt and short. Larger amplifiers and Marantz can do this (lots of current available). Otherwise the relay contact spring return will disconnect the load once you de-energize the relay coil. In this case I would expect turn-on and off pops if the amplifier was operating normally.

-Chris
 
Hi Max
Thanks for the tip. I assume you mean the isolation disks between the heatsink and transistors, since it works sometimes. Last it was working, I checked the voltage between the emitters on all output pairs, and that was ok. I will replace the isolation disks.
/Jørgen
 
Hi Chris
Thanks for the pointers.
The startup delay is working, no pops.
It just doesn't disconnect the relay when the fault occurs, so I get close to full rail voltage on the output terminals. Also, the fault appears randomly on either channel, or both.
So my first thoughts where that it might be a problem with proper gnd, or at least something that could affect either channel.
I will get some cooling spray and try on the small signal transistors in the voltage amp section
/Jørgen
 
Administrator
Joined 2004
Paid Member
Hi Largen,
You may be losing the negative regulated supply, check terminals C09s and C10s on the power supply PCB, the pass transistor is Q902 (2SC1161). I have seen the 2SC1161 transistors go intermittently open in all kinds of circuits.

So you do have the proper time delay before the relay closes, so we might be dealing with a leaky transistor.. Check Q910 (2SA562) and Q909 (2SC945). Since you're there, check the values of the high value resistors like R943 and R936. Also check the base - emitter voltages of the two transistors in the fault condition. They should both be off. Shorting R937 (22K) will shut off Q909 and should open the relay immediately.

My manual PDF is pretty rough, so my designations may be incorrect.

-Chris
 
Hi Chris

I got to take a closer look today.
The voltages on regulated supply where not symmetrical nor stable. So I adjusted them to ±58v using the two trimmers on the edges.
I do not know what those voltages should be, but I'm guessing I'm in the ballpark.
The voltages are not stable and fluctuate between 54 and 62 volts, mostly on the negative side. I pulled the PSU pcb and checked the two power transistors, they seem ok. The 2sc1161 has a hfe of ~62.
I checked the two other TO220 transistors while in circuit. One of them reports over 600 in hfe which seems odd?!
Also the two resistors you mentioned have been replaced. I have also resoldered all connectors as some of them where bad.
There is no doubt that it has been repaired before. All polarized caps have been replaced on the pcb. Btw, there seems to be an error on the pcb. The 330k resistor and the parallel diode have switched places, but electrically it is correct when following the traces. As for the fluctuating voltages, I have rarely come across bad zeners, so I am inclined to replace the small signal transistors. All diodes seem okay, so do the resistors.
/Jørgen
 
Administrator
Joined 2004
Paid Member
Okay, Q904 and Q906 (2SA733) are a differential pair and form the error amp for the negative supply. D910 (WZ240) is the reference for that regulator. D906 provides some kind of timing for the regulator, so the network of R906, R908 and C904 also needs to be checked. Use your scope on D910, it should be clean - no noise and definitely shouldn't vary in voltage. Expect approximately 24 VDC. The regulated output voltages should be 55 VDC (from the circuit description).

Q904, Q906 and Q903, Q905 should be matched pairs for hFE. That will minimize drift. It's too bad they are too far apart to mount together (so temperature tracks). You may be able to get Q904 and Q906 together.

If the beta of Q508 (2SB537) drops or is loaded, it could drag that reference voltage down or make it noisy.

The hFE of >600 is very odd unless the part that was installed is a darlington transistor. Using Darlington transistors for Q507 and 508 wouldn't be a bad thing I don't think. That would reduce the draw on the reference supply and also current variations. Overall, good for regulation. If it was mine, I might do that along with some other big improvements.

Zener diodes can go bad and be unstable / noisy. Not too common, but it wouldn't surprise me if they got a run of bad diodes. It wouldn't hurt to install a pair of 24V 1 watt zener diodes in there, like 1N4749A. If they were 0.5 watt, then use 1N5252B. You want to have the zener current in the right range for best performance.

-Chris
 
Hi.

Thank you. I decided to put it back together for a test after resoldering, and to check D910. D910 is clean, and at 24.32 volts.
My (cheap hantek dso2d15) scope does show random hf noise, about 4mV, but it also does so, or close to it, when the probe is shorted. This is in AC mode. DC mode shows clean and stable voltage. My old analog Philips pm3055 gives the same measurements.

I cannot set the voltage to ±55 volts.
The positive side will go as low as 56.5V,
while the negative side won't go below -58.0v

The amplifiers offset has become much more stable. It was low also before but was jumping som 20-30mV sometimes, that is now gone. One channel it really steady at 2.6 mV, the other jumps all the time between -1 and +3 mV.

No fault so far.

I checked R911/912/913/914, they are all within 10% of their nominal value.

Measuring D908/D910 in series, I get 48.41V.
Same measurement on D907/D909 gives 49.38 volts. D907 is around 24.8 volt. That is 1 volt in difference for reference between pos/neg side, which probably explains why I can't go lower.

I think I have some 24 volt zeners in a bin somewhere, but the are probably bzx types ½ watts. But seeing the two zeners on each side in series gives 10 volts over each of the 2.2k resistors, which equals around 4.55mA.
Multiplying this by 24 volts we get approx 0.11W. So I think these ½watts will do nicely.

Now I need to check if the relay protection is working again. After that, the mute circuit need checking

Thanks for all the good tips.

I will leave it on for the rest of the day to see if the fault comes back
/Jørgen
 
Administrator
Joined 2004
Paid Member
I checked R911/912/913/914, they are all within 10% of their nominal value.
Okay, not good enough. Replace with 1% metal film, that will also make it more stable with temperature. If your new zeners are a touch lower in voltage, that will also help. These were supposed to be adjustable below and above the target voltage. Feel free to change resistor values so the controls might end up mid turn. Check that D905 and D906 are not forward biased.

Yes, let it "cook", an hour or so is sufficient.
 
I am wondering if that voltage of 55 volts is correct. Doing the math with the feedback resistors 18k + 10k trimmer + 4k7...

Yes, schematic says R911/912 are 5k6, but mine are 4k7, and seem to be original.

Anyway, assuming the trimmer is set in center, we get 23k to ground and 9.7k to the output. The reference is 48v, so we get
(48V/23k) x 9k7 + 48V = 68.2V.

Minimum voltage would then be
(48V/28k) x 4k7 + 48V = 56.06V.

My zener voltages are a bit on the high end, so the math actually adds up, matching my measurements family well.

Also wondering why they bleed so much voltage in the regulator. I have around 83V on the emitter of 2sc1161.

Also, the voltage for the outputs are at around 57 to 58v.

And I would assume the input stage would run on higher voltage.

Looking at the vas stage transistors, these are rated for 120v/130v, which
would be ok for 55v, but not anymore.

So seen as a whole, I find it to be a odd
circuit, or is it just me?

If I have done some calculation errors or forgotten to take something into account, please point it out.

/Jørgen
 
Last edited:
Administrator
Joined 2004
Paid Member
Hi Jørgen,
The service manual is very clear. 75 VDC input, 55 VDC output. They have indicated a 55 VDC regulated output voltage in more than one place. I would say they used a lower rated voltage transistor than I would have. I would install higher rated parts (always go by C-E breakdown).

Normally I design with trimmers centered, but don't assume others do. AS for the math, it isn't adding up for me. Your error amp is a diff pair, that means your reference on one base should equal the feedback (divider) on the other base. So 48 VDC on each with very low gain to make up 55 VDC (or anything for that matter). Now, if R912 is actually 5K6 (not 56K) it makes a lot more sense. I sure wish they marked voltages on the schematic!

Anyway, I am pretty certain the regulated outputs should be 55 VDC and no higher. It is not uncommon for the voltage amp stage to run at a slightly higher, regulated voltage in the better amplifiers.

-Chris
 
Hi Chris. You are correct. I missed the connection between the two pins on the trimmer. That's why it doesn't add up. I simulated it in LTSpice. This gives me the same voltages as I am measuring. The voltage cannot come down to 55v using the values in my circuit, or the ones from the schematic. So I am going to change it at little. I changed the top zeners to 15v, so ref is now 39v. I'm going to remove the connection between the two trimmer pins so the base is connected to the trimmer center pin only. Also adding a pull-up from out to that center pin, in case it slips.
That should give me plenty of room to adjust to the correct voltage .

Regarding the nonworking muting circuit, I checked that there is voltage on the relay.
There is, or rather, C05S has 12 volts, which go to the muting relay. Haven't checked anything other that that yet.

/Jørgen
 
A few notes, the supply is now working - ish.
I have checked all resistors and transistors,
and it's adjustable to 55.2v
In LTSpice the same circuit lets me adjust from 45 to 60 volt with my change. Think I will put my scope on it to see if it oscillating.
Perhaps the small ceramic caps have gone bad. All of a sudden, one of the poweramps put out rail voltage. But pressing a little against the pcb made it go away. So probably needs resoldering

/Jørgen
 
Administrator
Joined 2004
Paid Member
Hi Lampie,
No, there aren't. No current sources in there. Look at the schematic.

Not only that, but while those diodes can go open, most I see are fine and I have original spares. They occasionally can go intermittent open. I do watch for this, but there are other components that have their own failure modes.
 
Hi. Yes, I know about those diodes.

Anyway, the amplifier is probably being retired,
or maybe repurposed into a poweramp.
I send the owner some pictures after
removing the front to look at the muting circuit. The relay is upside down, and glued into place, using wires to the pcb.
The touch button itself has no connection to anything at all, and I don't know how it is supposed to look. I can sometimes trigger it by touching the input on the pcb directly, but it doesn't stay on. Other times it does stay on, but won't go off again when touching the input a second time. Voltage on D903 is good with 11.8v , but there is only about 9.5V on the relay when on. According to the schematic, there should always be around 12v on one of the relays coil pins, but there is not. The connectors on those pcb's have been cut off, and all wires are soldered into place. Whoever did this last "repair" made a terrible mess out of it.
Screws are missing everywhere, and there is quite some rust here and there, that I could not see before.
The headphone jack is broken, so is one of the fuse holders on the bottom. It simply has too many faults to continue to try and fix it.
The input selector has a crack...
The list goes on.
I will se what the owner has to say about it.

Thank you for all your good advice.

/Jørgen
 
Administrator
Joined 2004
Paid Member
Hi Jørgen,
lol!
Welcome to my life. Everything is hacked these days and I hate it!
Yes, assess the unit, find out what the owner knows. I'm working on a 2500 in a similar state, it's a disaster. 4 shops have been into it. I fixed the original fault, not dealing with all the previous horrible work. These are worth too much to not repair.