To prevent oscillation of "common" Rotel circuit, I would recommend:
1. add another C609 on Q615 C-B terminals.
2. add C613 which usually is not installed.
Those above can be added even without looking into scope.
Now, if still oscillates, 15pF in parallel NFB R637.
If still oscillates add C611 (which I never saw installed in any Rotel model). As Rotel uses other type of correction - with C607 to the ground.
If C611 added, no need for C607. Only before removing C607, C611 must be installed in safe size, e.g. 100p, and now oscilloscope needed to see how looks like square wave front overshoot. Lower this value till will be acceptable wave form with not too much overshoot and amp still stable.
1. add another C609 on Q615 C-B terminals.
2. add C613 which usually is not installed.
Those above can be added even without looking into scope.
Now, if still oscillates, 15pF in parallel NFB R637.
If still oscillates add C611 (which I never saw installed in any Rotel model). As Rotel uses other type of correction - with C607 to the ground.
If C611 added, no need for C607. Only before removing C607, C611 must be installed in safe size, e.g. 100p, and now oscilloscope needed to see how looks like square wave front overshoot. Lower this value till will be acceptable wave form with not too much overshoot and amp still stable.
I saw a lot Rotels burned just for one simple reason - not enough prevention from oscillation, e.g. zobel is missing, plus missing named above measures... So common scenario to burn Rotel with smoke: turn on your amp with connected speakers and volume on normal listening position (9:00), then set Input selector on Phono, then try to play with turntable wires disconnect and connect by one wire and try different order. Most likely your turntable grounding will be fed by just one of RCA cables to avoid ground loop, therefore upon connecting the "other one" first you will connect only signal wire without shield and then amp goes into oscillation with smoke. This was tested. Similarly you can get with CD wires if only one shielding is connected to the ground which is rare case.
Yes, that is more or less what I tried to no avail.
I repaired the dead channel and tried again to load with 100nF (this time under strict current limited lab supply control) - of course oscillation as before.
Put in the capacitors you mention - still oscillations, although possibly at different frequencies.
Mind you, my upgrade already includes the removal of the distorting heavy lag (R623 || C607) loading of the VAS, and my VAS4 (EF-VAS-EF) replacement modules already contains a 100pF Miller cap.
The only thing that stopped the channel's instability with the 100nF load - was replacing the 0.22R 2W output resistor with a coil.
Ok, who (apart from someone like me) would load an amp like that in real life? Chris mentions 'stupid' speaker wires - but 100nF?
Most I have seen in 'super-duper' braided cables is 5-10nF, but I probably stand to be corrected on this.
One note is that you should be careful to willy-nilly put in C613 bias transistor bypass caps in a RA-820AX.
I once tried to add 1uF film caps (as I have done safely in numerous RA-931 upgrades), and the amp went unstable. Not fatally, but it heated up quickly.
There is most probably a reason why Rotel only put in a 0.1uF in the 820AX's Left channel - and none in the Right.
Finally, of course the missing Zobels should be soldered in their places on the pcb.
Your observations of reasons for Rotel amp destruction is very interesting and informative indeed! Thank you.
Per
Very true.
To install choke (coil and R in parallel) and Zobel network does not cause much hassle as holes drilled for those components on PCB in all Rotel amps, only components never installed.
Luckily later models like RA-921, RA-931, RA-01, 02 etc doesn't show instability in normal conditions. Perhaps PCB better traced...
Well, when compensating an amplifier you would normally feed it a square wave and connect various reactive loads. You look for ringing at the edges, or all out oscillation. You really need to know where you are. You should do this when installing different output transistors than were designed, drivers too.
Another thing I have seen is some amplifiers need a capacitor from the input of the diff pair (signal side) to audio common.
Yes, some speaker wire does have enough capacitance to cause instability in a marginal amplifier. We've had all kinds of "bright ideas" here in North America for speaker wire, and that includes using RG-8u coaxial cable. I have even seen some speaker crossovers with some capacitance directly across the input terminals.
Another thing I have seen is some amplifiers need a capacitor from the input of the diff pair (signal side) to audio common.
Yes, some speaker wire does have enough capacitance to cause instability in a marginal amplifier. We've had all kinds of "bright ideas" here in North America for speaker wire, and that includes using RG-8u coaxial cable. I have even seen some speaker crossovers with some capacitance directly across the input terminals.
Hello all diyAudio readers of this thread.
With the Silly Season
rapidly approaching I thought that a little brain teaser could maybe be helpful to counter all the silliness.
This is the schematic of a RA-980BX power stage. There is a serious flaw in the drawing - can you find it?
(I only posted the drawing of the right channel, but the left schematic has the same mistake so that wouldn't have helped you.)
(Ok, one hint: IF this schematic had been correct, the amp would immediately have blown.
)
First prize for finding it is - nothing.
Second prize is double that.
Per
With the Silly Season
rapidly approaching I thought that a little brain teaser could maybe be helpful to counter all the silliness.This is the schematic of a RA-980BX power stage. There is a serious flaw in the drawing - can you find it?
(I only posted the drawing of the right channel, but the left schematic has the same mistake so that wouldn't have helped you.)
(Ok, one hint: IF this schematic had been correct, the amp would immediately have blown.
)First prize for finding it is - nothing.
Second prize is double that.
Per
Lol!
Yes, caught it also.
Had it actually been connected that way, new output and driver time + more. Resistors also and possibly all the way back to Q614, Q616.
Drawing errors are pretty common. Component designations can also be wrong as well as values. They used to proof read manuals, I don't think they do anymore. Many are drawn in a way that is very difficult to follow on top of that.
Nice catch. Knowing how it is supposed to go, your brain just fills it in. I can see where a new technician could be caught by that.
Yes, caught it also.
Had it actually been connected that way, new output and driver time + more. Resistors also and possibly all the way back to Q614, Q616.
Drawing errors are pretty common. Component designations can also be wrong as well as values. They used to proof read manuals, I don't think they do anymore. Many are drawn in a way that is very difficult to follow on top of that.
Nice catch. Knowing how it is supposed to go, your brain just fills it in. I can see where a new technician could be caught by that.
Yes, R628 (not Q628) will pull all the driver and power transistors full ON and a spectacular amount of current will quickly destroy them before any fuse blows.
The forward connection from the lower VAS Q616 should of course go from its collector just like the top Q614 (and it fortunately does that in the amp).
But the proofreading was apparently done the morning after the company party, just look at the interesting direct speaker connections:
Anyway all, have a very Merry Christmas
and a Happy New Year
!!
Best,
Per
The forward connection from the lower VAS Q616 should of course go from its collector just like the top Q614 (and it fortunately does that in the amp).
But the proofreading was apparently done the morning after the company party, just look at the interesting direct speaker connections:
Anyway all, have a very Merry Christmas
and a Happy New Year !!Best,
Per
Hey Per!
Lol!
Now that you mention it. Really stupid errors that are fairly common in audio. There likely is an service bulletin on these errors, not included with the service manual when you buy an original (back then).
I especially like the fuses in-line with the speaker connections. Excellent. After engineering an amp to have low distortion, let's put a distortion causing element in line with the speakers.
As a tech, we don't look at things outside of the area we need to in normal work. We don't assess the entire design unless we have a good reason to since time = money. But the company who designs and builds the equipment ought to at least be expected to produce accurate documentation. There is no excuse for this.
Merry Christmas, and a very happy New Year!
-Chris
Lol!
Now that you mention it. Really stupid errors that are fairly common in audio. There likely is an service bulletin on these errors, not included with the service manual when you buy an original (back then).
I especially like the fuses in-line with the speaker connections. Excellent. After engineering an amp to have low distortion, let's put a distortion causing element in line with the speakers.
As a tech, we don't look at things outside of the area we need to in normal work. We don't assess the entire design unless we have a good reason to since time = money. But the company who designs and builds the equipment ought to at least be expected to produce accurate documentation. There is no excuse for this.
Merry Christmas, and a very happy New Year!
-Chris
Hello and a good day to all, I also have a Rotel 971MK2, a bit modified, I m working on improving the pre-amp part now, replaced caps with solid ones, and I m on the op amps part, currently I added a BB2604 in the tone part also, stock it has one 2114D, maybe will order 2 OPA1612 and be done 
), what got my attention upon checking the RA-02 and RA-06 schematics was some resistors on preamp and volume pot part:
RA971- R501-502 is 150R
RA02/06 same resistors are 470R
Also RA971 R505-506 is 680R in RA02/04 it's 470R
Last but not least R507-508 is 3.9K in RA971 but 2.7K in RA02/06
Why this difference? does it improve something? more attenuation? and other bias for opamp?
Thank you all for your opinions
Criss
), what got my attention upon checking the RA-02 and RA-06 schematics was some resistors on preamp and volume pot part:RA971- R501-502 is 150R
RA02/06 same resistors are 470R
Also RA971 R505-506 is 680R in RA02/04 it's 470R
Last but not least R507-508 is 3.9K in RA971 but 2.7K in RA02/06
Why this difference? does it improve something? more attenuation? and other bias for opamp?
Thank you all for your opinions
Criss
Was the OPA2604 one you already had or one you have recently bought as they are long out of production? Reason for asking is that I suspect most for sale today are fakes.
That opamp was one of my favourites years ago along with the single OPA604.
I had it from another Rotel, I also tested LME49720 but the RF interference problem was beyond bearable, they sounded nice tho, but they where harsh, 2 x 2604 sound is way warmer and it has something relaxing to it,
Well I should test some OPA1642, maybe both the 1612 and 1642
Good
that is as genuine as it gets. Don't risk buying any today, they are pretty much guaranteed fakes. I agree. The OPA2604 did have a bit of magic to it sonically when used in the right applications.
Forgot to add the schematic where these resistors are placed:Hello and a good day to all, I also have a Rotel 971MK2, a bit modified, I m working on improving the pre-amp part now, replaced caps with solid ones, and I m on the op amps part, currently I added a BB2604 in the tone part also, stock it has one 2114D, maybe will order 2 OPA1612 and be done
RA971- R501-502 is 150R
RA02/06 same resistors are 470R
Also RA971 R505-506 is 680R in RA02/04 it's 470R
Last but not least R507-508 is 3.9K in RA971 but 2.7K in RA02/06
Why this difference? does it improve something? more attenuation? and other bias for opamp?
Thank you all for your opinions
Criss
RA971- R501-502 is 150R
RA02/06 same resistors are 470R
Also RA971 R505-506 is 680R in RA02/04 it's 470R
Last but not least R507-508 is 3.9K in RA971 but 2.7K in RA02/06
Why this difference? does it improve something? more attenuation? and other bias for opamp?
Hi Criss,
R505 and R507 determine the gain of the opamp: Vo/Vi = 1 + 3900/680 = 6.7
In RA-02 this calculates as: Vo/Vi = 1 + 2700/470 = 6.7, ie. the same gain.
R501 does cause a bit of overall gain reduction with R503, but very small: Vo/Vi = 100000/(150 + 100000) = 0.999
The RA-02's 470R gives slightly more attenuation: Vo/Vi = 100000/(470 + 100000) = 0.995 Difference not audible.
Per
RA02/06 same resistors are 470R
Also RA971 R505-506 is 680R in RA02/04 it's 470R
Last but not least R507-508 is 3.9K in RA971 but 2.7K in RA02/06
Why this difference? does it improve something? more attenuation? and other bias for opamp?
Hi Criss,
R505 and R507 determine the gain of the opamp: Vo/Vi = 1 + 3900/680 = 6.7
In RA-02 this calculates as: Vo/Vi = 1 + 2700/470 = 6.7, ie. the same gain.
R501 does cause a bit of overall gain reduction with R503, but very small: Vo/Vi = 100000/(150 + 100000) = 0.999
The RA-02's 470R gives slightly more attenuation: Vo/Vi = 100000/(470 + 100000) = 0.995 Difference not audible.
Per
Thank you for your help and inside with this, interesting abordation on that part by Rotel, so in essence they kept the same gain, it puzzels me why they went with 470r/2.7K instead of the usual 150R/3.9K
Also, I was wondering, got my hands on some OPA2228 today, but they need at least a gain of 4 or 5 to be stable as far as I undersand they would be stable in this circuit?
I know many swear that op-amps dont change the total ballance of sound, but swaping op-amps in this 971, got me thinking otherwise, I swapped from Signetics 5532, to aTI5534, to AD823, I even have 2 pieces OPA-627 genuine (if someone wants some), LME49720NA was the best overall but the RF problem was beyond fixable, but the OP-amps are another thread, my main mods where swapping most transistors and implementing the light version of your mods, will post some pictures tomorrow, but all I can say is that its beyond all my other amps as far as sound quality and the mere pleasant feeling it gives
Thank you for all your time and help
Also, I was wondering, got my hands on some OPA2228 today, but they need at least a gain of 4 or 5 to be stable as far as I undersand they would be stable in this circuit?
I know many swear that op-amps dont change the total ballance of sound, but swaping op-amps in this 971, got me thinking otherwise, I swapped from Signetics 5532, to aTI5534, to AD823, I even have 2 pieces OPA-627 genuine (if someone wants some), LME49720NA was the best overall but the RF problem was beyond fixable, but the OP-amps are another thread, my main mods where swapping most transistors and implementing the light version of your mods, will post some pictures tomorrow, but all I can say is that its beyond all my other amps as far as sound quality and the mere pleasant feeling it gives
Thank you for all your time and help
Last edited: