ROTEL amp sounds dull

[Chris Babineau]

Emitter resistors as far as I know are always some fraction of an Ohm i.e. .47 .22. Most ohm meters don't do a good job measuring a fraction of an ohm. But if a basic meter is all you have you could lift one end of the resistor in question and compare the value to some of the other emitter resistors (also with an end lifted) and if the discolored one is different than it should be replaced. The associated transistor may have been stressed and possibly should be checked as well.

An emitter resistor is like a current sense and helps balance out the load with parallel power transistors. Some call this "current hogging".

 

[homemodder]

The resistors are not wirewounds, with the inductance of wirewounds the amp would oscillate, they should have been of a higher power rating, they do not drift because of time but because of high power usage into difficult loads, speakers can dip into 2 ohm load at some frequencies. They are only rated to 2 watt which is a low for the power rating.

The weak link in rotel is always the preamp section, the amp itself is pretty good.

 

[carlmart]

I adjusted the bias and what a difference!! It's sounds so much better! The only thing is that one of the emittor resistors looks a bit black. I checked it and they all seem to be 0 ohm on my multimeter but on they should be 0,22 Ohm K (that's 220 Ohm?). So are they all burnt? On them is written: "MPC70 0.22 Ohm K fil 2502".

That's a 2watt wirewound resistor indeed. Definitely a bit low on wattage if the load is difficult. Going up to 5watt would be safer.

Whether it's bad to be wirewound or not may be arguable. I think a non-ww resistor could sound better, and nowadays you can find them pretty affordable.

If one resistor is black it may be dirt or maybe burnt. Replace it. If you fill courageous go to a lower resistor, like 0.1 or o.15: it may also improve the audio quality.

Also change the bias trimpot and put a multi-turn type. Take note what value the Rotel one is set for, and set that on the multi-turn to start testing.

Also the bias seems to drift. When I set it at 10 mV it 'll be 20 or so in a while and next time I switch it's different as wel, though it won't get past 25 mV or so. Left and right are different (like 5-10mV difference). Is that a problem?

Can you tell me how the Rotel keeps bias stable? That diode/transistor should be in contact with the heatsink or an output transistor. It should be physically in contact with that. Being that alright, a drift, until temp stabilizes, is normal.

And about the caps: I ve seen lots of brands on the internet. What would be a good choice in my amp? What about these: Nichicon Muse KG 10000uf 63v at 90 euro? I couldn't find any at digikey btw.

Digikey does not sell Muse, AFAIK. They have Panasonic and Philips, if I remember well, which are also good. If you live in Europe and want to change capacitors buy Siemens cans, which are very good. Rifa are also very good.

But about the dullnes you report, I would look that somewhere else, perhaps in the preamp section. I couldn't find an schematic for that amp, to see where to improve.

 

[SQLGuy]

Mouser carries Nichicon, including the KG and KZ series. No minimum order and pretty good prices. But, they don't carry the 10000uF 63V in this series. The highest they seem to have available is 6800uF at 63V, and these are non-stocked.

As for the emitter resistors' wattage, at 200W continuous into 4 Ohms (does the Rotel do that?), they'd be dissipating 1.22W each. I think 2W is safe, unless this amp is stable into 2 Ohms and doubles all the way down.

 

[homemodder]

The resistors are metalplate cement resitors, if you google on how they are manufactured youll notice they have nothing to do with wirewounds, thats what mpc stands for. Wirewounds have higher inductance which will cause this amp to oscilate as it uses a triple darlington output and most probably does not even have a output coil. I have repaired some models where a repair shop substituted these resistors for plain wirewounds, and then didnt bother to put the amp on a scope, disasterous. Wirewounds are fine to use here if high quality non inductive brand like mills is used. Use 5 watt resitors.

If the caps are changed to Muse types, I think you going to be dissapointed in the sound. The Muse are not bad but compaired to the slitfoils they dissapoint. For same or better quality youll have to go up to siemens and now made by epcos, sikorels, they cost a little fortune, but maby you can get them somewhere at affordable price.

 

[Bartt]

Thnx for all the advice! I replaced the emittor resitors for a 5W version MPC-71 except for 2 (could only get 10 of them). I also replaced the trimpots with multi-turn types. The sound is better and more stable in time. The bias doesn't drift that much anymore and left and right channel are more or less the same.

I still have the feeling that something elso is (also) wrong. The Rotel has been repared (under warrenty) twice because of some damage related to my 18' Volt subwoofer I connected to the "speaker remote" terminal (couldn't handle it when I switched it on at medium volume...). Lots of joints have been resoldered and probably some parts have been replaced. What else could be wrong and how can I test it?

Maybe I should check the zobel resitor too. Is that the one near the speaker terminal?

How can I check if the output transistors are okay?

 

[homemodder]

Hi Barrt, good to see that the bias is more stable, I would however try to get the other 2 resistors replaced as well with same types. Check the zobel resistor and if possible if you could take a photo of the output transistors but in a way so I could see their markings. Im not afraid that there might be a blown one, but as you say that the amp has been repaired before Im wondering if original parts were used. Some repair shops put in fakes, usually the bias transistor is also mounted on the this heatsink and I would like to see how it was done and if the original transistor is still in place.

 

[Bartt]

I tried to make a picture of the output transistors, but it's a bit difficult to get a good shot with everyting on it. Left and right channel are not exactly the same:

3x B817 D 5F3 and 3x D1047 D 5E3
3x B817 D 5F3 and 3x D1047 D 5F1 (!)

Is that bad?

Do you know where can I find the zobel resistor?

 

[Bartt]

hope you can see the picture now ...

 

[Lumba Ogir]

Bartt,
5F1 is batch designation, I would not worry too much about different numbers.

 

[homemodder]

Those output transistors are good, they are factory originals. Your pic is good, one can clearly see. The one thing I do notice from that pic is that the bias transistor is not mounted upside down on to the output transistor but at a 90 degree angle. Im not sure on this model but on most its always mounted in line. The reason is that when mounted in line it has larger surface contact with the output transistor which is important for thermal tracking and correct biasing. As it is the metal tab on the back of d600 has very little contact on that d1047. Have a look at the attached pick in this post, its a different model but about all rotels that I have seen are mounted like this. The transistors in question are the ones with the orange wires. After that adjust the bias a little higher than factory spec with the amp warmed up, around 15 mv is good.

Before replacing the zobel resitor test to see if they are ok, before they fail the factory ones tend to turn brown as if burnt, if they are measuring good and look ok leave them be.

After this the amp should be performing close to factory spec, with only maybe the age of the capacitors having any real effect.

 

[homemodder]

Here is a pic.

 

[Lumba Ogir]

homemodder,
how are you?

Those output transistors are good, they are factory originals.

It does matter, however.

 

[homemodder]

Hi Lumba

Im good, yourself ??


Bart
Drop me a email through link on my profile and Ill point you to a online shop where you can purchase those powersupply caps if you thinking of replacing them, without denting your bank account too much. The manufacturer is the supplier to rotel for those caps. I assume it uses for 4 10000uf caps so it would cost around 55 euros to replace all with same and go up to double the price for superior quality ones.

The zobel resitor is r669 and r670.

 

[Bartt]

I checked the zobel resistors and they seem to be fine (+/- 11 ohm). I couldn't mount the bias transistor upside down, because it would have less contact than before. So I turned it straight up (see picture). The wires do make contact though with the small heat sinks. Is that a problem? The're not gonna melt are they?

The heatsink of the output transistors seems to be a bit hotter than before 41 C° (left) and 45 C°(right). The small heatsinks in between are quite a lot hotter, around 73 C°

I've set the bias at 15mV. The left channel got back to 11 later and doesn't deviate more than 10% since. The right channel increased to 18mV and when cooling down for a while to 13mV. The strange thing is that the right channel seems to increasing or decrease during the measuring, while the left channel doesn't: is that normal?

 

[homemodder]

MMM maybe we shouldnt turn the bias up that much, i was expecting the temp rise on the outputs but not so big a diffrence on the drivers. Maybe keep it around 12mv max. I sure hope when you say contact you dont mean electrical contact, that could be disaster. I see space is tight there. Something is up with that right channel, but i cannot think what is is. It will vary a bit with voltage and power fluctuations but im not so sure its should be changing like that all the time. In later models rotel improved the vbe multiplier for use with triples, ill have a look and let you know if something can be done without messing with the circuit, maybe a darlington transitor as the temp sense.