It's supposed to help the transistors switching over across the zero point. It's possible the effect is more noticeable at low levels? I don't know. I usually install a 1 uF film cap for that purpose, but the value is based more on what I see used in commercial amplifiers than testing.
Given that a manufacturer will not install parts that are unnecessary, there must be a good reason why it is there in most designs. That's my thought process. At some point I will do some testing on this. That's after my RTX arrives.
-Chris
Given that a manufacturer will not install parts that are unnecessary, there must be a good reason why it is there in most designs. That's my thought process. At some point I will do some testing on this. That's after my RTX arrives.
-Chris
I have had a further look into the Vbe multiplier circuit.
Having found no effect of introducing a capacitor bypass around it, I looked at the quiescent current setting, as it has been proposed that the standard 4mV bias setting (giving an Iq of 18mA in the output transistor pair) is really too low.
And there could be some truth in that. The enclosed pic shows the distortion RTA of an upgraded Rotel. Even though the standard 4mV setting only gives distortion peaks a few dB over the respectable -105dB noise floor, increasing the Iq to 24mA actually reduces the dominant 2nd harmonic and thus the overall THD - albeit not by much - 0.0039 to 0.0033.
So, what if we increase the Iq even more to try to drive the output stage further into class A?
Well, there is no free lunch.
The list below shows the corresponding figures for:
bias mV setting - quiescent current - power dissipation - heatsink temperature and distortion:
4mV - 18mA - 1.4W - 30C - 0.0039%
5mV - 24mA - 1.7W - 32C - 0.0033%
6mV - 27mA - 2.1W - 35C - 0.0032%
7mV - 32mA - 2.4W - 38C - 0.0031%
8mV - 36mA - 2.8W - 42C - 0.0034%
Over 36mA the 5th to 9th harmonic distortion peaks start to rise sharply, indicating that something else than crossover distortion is afoot. And, the small internal Rotel heatsink of course gets increasingly warmer - maybe not alarmingly, but clearly there is no reason to run it hotter than needed.
So, as a first compromise I decided to set the (cold) bias to 6mV, only to find that after about one hour of operation it had slipped back to just over 5mV. I guess that this means that Curtis' original thermal stability design is basically sound, which is good news. And I can definitely live with the result - and the sound!!
Having found no effect of introducing a capacitor bypass around it, I looked at the quiescent current setting, as it has been proposed that the standard 4mV bias setting (giving an Iq of 18mA in the output transistor pair) is really too low.
And there could be some truth in that. The enclosed pic shows the distortion RTA of an upgraded Rotel. Even though the standard 4mV setting only gives distortion peaks a few dB over the respectable -105dB noise floor, increasing the Iq to 24mA actually reduces the dominant 2nd harmonic and thus the overall THD - albeit not by much - 0.0039 to 0.0033.
So, what if we increase the Iq even more to try to drive the output stage further into class A?
Well, there is no free lunch.
The list below shows the corresponding figures for:
bias mV setting - quiescent current - power dissipation - heatsink temperature and distortion:
4mV - 18mA - 1.4W - 30C - 0.0039%
5mV - 24mA - 1.7W - 32C - 0.0033%
6mV - 27mA - 2.1W - 35C - 0.0032%
7mV - 32mA - 2.4W - 38C - 0.0031%
8mV - 36mA - 2.8W - 42C - 0.0034%
Over 36mA the 5th to 9th harmonic distortion peaks start to rise sharply, indicating that something else than crossover distortion is afoot. And, the small internal Rotel heatsink of course gets increasingly warmer - maybe not alarmingly, but clearly there is no reason to run it hotter than needed.
So, as a first compromise I decided to set the (cold) bias to 6mV, only to find that after about one hour of operation it had slipped back to just over 5mV. I guess that this means that Curtis' original thermal stability design is basically sound, which is good news. And I can definitely live with the result - and the sound!!
Attachments
Hi guys recently i bought a Rotel RA-840BX2 and did some servicing/cleaning.Two diodes was to fragile and was damaged.What will be the best replacement for regulator diode for D901? The servicing manual doesn't show much detail of the part. I checked on RS components there are few different types of 1N4148 diodes.
Hi Lawrence, sorry for not replying sooner.
You must have done some heavy cleaning to break a diode on the PCB! That said, D901 resides in the DC power supply which can run quite hot. I would check the solder side of the PCB for scorching and copper track/solder joint integrity.
Any 1N4148 replacement will do.
You must have done some heavy cleaning to break a diode on the PCB! That said, D901 resides in the DC power supply which can run quite hot. I would check the solder side of the PCB for scorching and copper track/solder joint integrity.
Any 1N4148 replacement will do.
.Hello all. I have just obtained a Rotel RA-840BX2 in very good cosmetic condition but unfortunately it blows fuses on power on, there is a short to earth, which on initial inspection seems to be from the twin transformers. If this is the case would anyone know the va rating and secondary voltage requirements of replacement toroidals? Thanks.
no tested them all, but just noticed q613 d600k is placed backwards on to the heatsink, but there is still a short somewhere
Note that Q614 is supposed to be "backwards" ie. E-C-B, but not Q613.
But really, that should not cause any blown secondary AC (?) fuses, just maybe a strange output stage offset.
I still think that one of the output drivers or power trannies have gone to its maker. Try isolating the power stage from the supply to see if the V+ and V- rails are ok.
But really, that should not cause any blown secondary AC (?) fuses, just maybe a strange output stage offset.
I still think that one of the output drivers or power trannies have gone to its maker. Try isolating the power stage from the supply to see if the V+ and V- rails are ok.
16 ohms is a strange value. But if the short disappears and the fuses stop blowing when you isolate the heatsink the problem probably lies there. The heatsink should not conduct any current.
What I would suggest: First with power off
1) Measure the resistance of the trannies' secondary windings. (Should be <1ohm)
- and then with power on, heatsink not touching chassis:
2) Measure the voltage at speaker outputs (Should be in the mV range)
3) Measure the voltage between speaker+ output and heatsink (Should be 0V)
- if it is a high voltage, power off and
4) Measure the resistance between each power transistor collector to the heat sink. (should be infinite or high Mohms)
If R642 (33k) is down to 4ohm, the parallel 5A speaker protection fuse must have been blown and the output stage suffered a huge current load. But even so, 4ohms is also a strange value..
OR - your multimeter or test leads have a fault. We all tend to trust our equipment readouts more than our own judgement.
Just my 2p,
Per
What I would suggest: First with power off
1) Measure the resistance of the trannies' secondary windings. (Should be <1ohm)
- and then with power on, heatsink not touching chassis:
2) Measure the voltage at speaker outputs (Should be in the mV range)
3) Measure the voltage between speaker+ output and heatsink (Should be 0V)
- if it is a high voltage, power off and
4) Measure the resistance between each power transistor collector to the heat sink. (should be infinite or high Mohms)
If R642 (33k) is down to 4ohm, the parallel 5A speaker protection fuse must have been blown and the output stage suffered a huge current load. But even so, 4ohms is also a strange value..
OR - your multimeter or test leads have a fault. We all tend to trust our equipment readouts more than our own judgement.
Just my 2p,
Per
I guess from your silence that the issue is still not resolved?
One thought: Is it really a short that make the fuses blow - or is it amp output oscillation? I have just had a RA-931 in that blew its secondary fuses so I powered it with +/-20V from two lab power supplies with current limiters.
All trannies and passives had measured ok, but as soon as I increased the bias past +/-1.1V to turn on the output stage the Lch kept going into a violent ~1MHz oscillation.
A 100pF Miller capacitor from VAS collector to base stopped it.
So, if you don't have access to two lab supplies and a scope, it may be an idea to just try adding two 100pF NPO caps to Q615/Q616 C-B and crossing a few fingers? It can't do any damage.