Yamaha EMX640 Limiter light on ch2 stuck on

[DrGonz78]

So I have a problem with this Yamaha emx640 where the limiter light on channel 2 is stuck on. I presumed at first that there could be a fault in the power amp like a bad output transistor. I have checked the amp out and there are no faulty transistors or any components related to the limiter circuit that could be causing such an issue. The amp turns on and the relay kicks over. There is no DC on the output, both channels, before or after the relay kicks on. I have stable -/+15v power rails. Channel 1's limiter light is functioning as it should too. I am going to go clean out all the output jacks right now but I don't feel that will solve the problem. The schematic is all over the place and is really hard to follow. I have included a link to the schematic and pages 20 and 24-27 have the parts for the where the limiter is located. I have to admit that I need some guidance on where to start troubleshooting at this point.

https://elektrotanya.com/yamaha_emx-.../download.html

[J M Fahey]

Sorry Dr Gonz, did my best and tried until I got a headache but that schematic was chopped into a completely unreadable mess.

I bet the original Yamaha one is good to excellent, maybe somebody can get it or worst case straight from Yamaha, the Elektrotanya version is useless.

That said, if amp works well then leave as is.

[DrGonz78]

Yeah the schematic is all chopped up and yes it is a headache. I think I will compare voltages between channel 1 limiter and the stuck on channel 2.

[Enzo]

I have the factory file, let us try here:


O found it easy enough to trace the circuit, yes the LEDs are on one page then driven from elsewhere.

[DrGonz78]

Hey thanks Enzo!!

[J M Fahey]

Ok, now it has some logic.
Itīs their version of a DDT limiter, it detects when amp starts clipping.

It checks 2 points which normally are at same or close voltage ... with amp idle or working within the linear area that is

Normally compares input signal to feedback signal ... which "should" be the same, thatīs NFBīs job.

Now when amp clips, they are NOT the same any more (original is pure sinewave, output is flat topped) and NFB becomes *crazy* trying to correct output, now you suddenly have "large" pulses >700mV peak where before you had, say, a couple mV error correction signals.

That >700mV peak signal triggers a transistor which triggers another .... until one triggers some kind of VCA and if necessary, another turns the LIMIT Led on.

I donīt care too much about a stuck Led by itself, thatīs why I suggested earlier "if amp works anyway, leave as is", but of course if one channel limiter is stuck ON, then that channel will put out very little power compared to the other or even be muted, so first of all reconfirm whether both channels work the same or not, mainly to reduce the troubleshooted area.

As a side comment, this is a true Rube Goldberg design, using 24 parts where they could have used 4, "just because they can".
FWIW my own IOC limiter, doing exactly the same function, uses 1 transistor, and the associated limiter uses 1 LDR and 1 resistor , if I want it to be "polite" (PA or Keyboard) and a poorly biased and uncompensated FET if I want it to be "nasty" (Guitar/Bass).
As a side note, I have one EMX640 in my boneyard, maybe Iīll repair it now Enzo was so kind as to supply a working schematic, all I had before was the Elektrotanya one which is unusable.

Not blaming Elektrotanya itself which is a goldmine and has saved my bacon thousands of times.

OK, so rolling up imaginary sleeves:

Analyzing just one channel, they are practicallybtwins and even sort of share part numbers b(R152 in one becomes R252 in the other and so on):

1) power amp is inverting so very low input impedance, needs to be driven by an Op Amp.
Input differential pair is Q205 (base grounded through R210) and Q206, base gets NFB through R215 and input signal from IC101 through R209 and C203. So far so good. Input impedance is R209 so 680 ohm.

2) output appears at collectors into mirror transistors Q209/Q210 , both points are "one diode" above -V rail but difference between them is a couple mV at most.

Now when amp clips you have >700mV and Q208 is triggered, passing current.

3) which triggers two PNP "Digital transistors" :
a) Q226 through R250 > Q227 > R252 > Pin15 at CN101/901 > Limit LED on the Preampboard.

too sleepy to continue now , 5AM here, woke up just for the typical "old Guy midnight leak" , will tomorrow follow the other trigger chain which will lead to the real Limiter/VCA.
Which I donīt want to even *see* now or Iīll lose sleep, already lost time searching and reading about DTA114 "digital transistors" and their complementaries.
Oh well.

[Jazz P Bass]

DTA114. Digital Transistor.

A PNP transistor that has two internal bias resistors.

DTA114E-D.pdf

[DrGonz78]

Yup the Led's are on page 20 and are fed from pin 15 of the ribbon cable. I had gone through and traced out where things go on the CRAP version of the schematic and thanks to Enzo I won't have as much of a headache now. I went to all associated transistors in the circuit to check them out. At that point I could not find any transistors or resistors that were bad. Now with the better version of the schematic I will go through the circuit and check voltages and compare the good side to the bad, looking for discrepancies. I did see those DTA-114E digital transistors and it took a second or third look at the schematic to realize those resistors were integrated into the transistors. The diode in both DT-114E's are good and perhaps I can measure the resistance later tonight. I will compare the good side's to the bad throughout to narrow things down.

[J M Fahey]

Yes, thatīs the waybto go.
Do not pull parts to measure outs9ide, sloww as a snail, undecisive and potential for damage is high, start by measuring voltages and comparing.

If one Led is ON then the coresponding transistor(s) are passing current, so they will either be forward biased , in which case we must find what causes that with no actual clipping present or one is shorted.
In both cases we will find some different voltage explaining that.
Only then you search for a bad part (in a very limited area) which could cause that.
But you already know this

Earlier I had written this, no sure why it had disappeared from my screen, I leave it because it may still be relevant:

........................................................ .....................



Ok, going through "morning" coffee and following the actual limiter trigger/signal chain.

VERY Rube Goldberg style, as in "why make it simple when you can complicate it?"

Instead of just using the detection transistor Q208 and activating an: LDR/FET/CA3080/whatever to do the actual VCA, they used a "differential VCA" which is relatively common in Synthesizers ... maybe this part of the circuit was designed by a Yamaha keyboard guy.
At least the (very poor by the way) Digital Reverb is based on a downgraded SPX90 type chip so it looks like different areas from the Yamaha think tank "collaborated" here.
That might explain its elaborate clumsiness.

The chain is: Q208 > R245 > Q225 > IC102 > Q204 > Q203 > ..... now that I remember, lost mains for a few minutes. Oh well.

[Enzo]

When I see something like this, the first thing I determine is whether the function of the light is being triggered, or if the light circuit itself is stuck.

In any case, even if you don't understand it, the LED is generally turned on by some part, which in turn was turned on by some part, and work your way back through the schematic. Especially if you have a working channel to compare to.

[DrGonz78]

So the voltage readings throughout the limiter circuit led me to finding Q208 shorted from Emitter to Collector. Also the DTC114es and DTA114es pin outs were a pain to figure out at first and they are pin1= Emitter pin2 = Collector pin3 = Base. Found only one data sheet that showed the correct info for these particular transistors. Confusing part is that there are EBC versions out there and most data sheets available are those ones.

Anyhow take a look at the voltage charts... After taking down voltages I went and compared Q208 readings to Q108. At that point I tested Q208 and discovered the short from emitter to collector. I checked a bunch of other voltages and this appears to be my culprit. A 2sc2240 transistor. Now time to see if I have a replacement in the shop or something that might work as a replacement.

[DrGonz78]

Must be my lucky day or something. Third and fourth transistor I picked out of old stash bin were both 2sc2240 gr transistors! I love when that happens.

[DrGonz78]

Well I removed the transistor but it tests good. I am thinking there was a solder bridge between emitter and collector solder pads. The short is not there any more. I will test the unit again but for tonight I am off to dream of fixing amps.

[J M Fahey]

When you are so fuzzy minded that you grab the soldering iron by the tip and instead of yelling look at it vaguelly wondering where the BBQ smell might be coming from, THEN itīs a good idea to go get some sleep and leave Tech problems for tomorrow

Just as a side comment, not sure if it worries you, but Q108 having exact same voltage at base and emitter , plus having way more positive voltage at collector is fine.

Itīs not a gain stage , itīs only triggered when amp clips, in which case significant voltage *does* appear and it starts pulling current.

[DrGonz78]

I put it all back together and it's fixed, limiter light not stuck on. The collector on q208 reads 12v just like q108. Thanks for all the support from Juan and Enzo, especially for a better copy of the schematic! Now let's get this contraption out here!!