SUMO Power Amp

Status
Not open for further replies.
Jump to Latest
#23
Federico,

The Nine Plus schematic doesn't look like an original drawing. Is it a copy of an original, or perhaps taken from the amplifier itself? I'm just curious about where the schematic came from and how sure you are that it is accurate.

Thanks!
 
#26
There is no need of temp. feedback
I think.

On the main heatsink we can obviously
see the 10 output device, while
the drivers Q11, Q14 are on a small
heatsink on the board. Q12, Q13
are on the board without heat sink
since they work only during peaks with load
<4 ohm.

Thermal stability is inerent to the adopted topology.
Look at the US patent 4229706.

However the output devices are not heavily biased,
say about 8W ( total 8x10x2= 160 W) at steady. That is a current of something less 200mA per device.


The amp become very hot, however, and there is a fan
to take count of that.

thank for the question Gert

Federico
 
#28
The amp .works always in class A if the load is
no smaller than 4 ohms. When this happens
it switchs to A-B thanks to the aid of Q13
and Q12. However its class A is not
classic ( pure ) class A as you can read on
in the forum:

Amplifiers > Pass Labs > SE ClassA with low dissipation @ idle - is this it ?
(old thread)
message #26 and following.

the nine + is a great amp. I love it, I love they both
since i have two identical amp. in passive
vertical biamp

bye

Federico
 
#30
Millwood, here is my attempt at a basic explanation.

The gain path itself is actually pretty straightforward. The dual opamp at left converts an unbalanced input signal to a balanced output signal and provides all the voltage gain for the circuit. Each phase of this balanced output signal is then buffered by a two-stage emitter follower, consisting of driver Q11/Q14 and its associated bank of 2SD555 output transistors, to drive the speaker load.

Overall feedback is taken from each output terminal back to the inverting input of the appropriate opamp. The 0.1 uF blocking caps in the feedback network result in 100% feedback at DC, so the dual opamp also acts to servo the output at 0VDC. So far, this is all pretty normal practice.

Now for the bias scheme. The amplifier contains four separate but interdependent current loops. Loop 1 goes from the positive terminal of the top 45V floating supply through the top bank of output transistors and back to the negative terminal of the bottom 45V floating power supply. Loop 2 flows from the positive terminal of the bottom 45V power supply through the bottom bank of output transistors and back to the negative terminal of the top 45V power supply.

Loops 3 and 4 go from the +55V front end power supply through driver transistors Q11 and Q14 respectively, flowing both through their respective emitter resistors R9/R30 and through the bases and emitter resistors of their respective output transistor banks, finally returning to the -20V power supply through their respective adjustable resistances R7-R14 and R29-R34.

The current in loops 3 and 4 controls the idle current in the output stage by defining both the base current in the output stage and the output stage bias voltage across driver emitter resistors R9/R30. The total current in each loop at no signal is held constant by Ohm's Law, since at no signal there is a constant 20V across R7-R14 and R29-R34. Regulation of these loop currents also prevents thermal runaway in the output stage by limiting the amount of base current available to fuel it.

The current-setting resistors in loops 3 and 4 pull down to a -20V supply, so they will continue to conduct current even when the output terminals swing negative to nearly the same level, again due to Ohm's Law. This means that output stage base current (and hence collector current) will continue to flow at full swing, albeit within limits and with the help of Q12 and Q13 as Federico has pointed out above. While this doesn't necessarily translate to linear operation, it at least avoids problems associated with switchoff in the output stage.

I don't know your background, but if you found the Bongiorno patent unclear and want to dig deeper into the techncal aspects of this circuit, I'd suggest doing a web search on Kirchoff's current law and the gain relationships (Ie=Ic+Ib, etc.) for bipolar transistors. Some understanding of these concepts may help shed more light on what is going on here.
 
#31
Joe Berry said:
Millwood, here is my attempt at a basic explanation.
Click to expand...


thanks, Joe, I will go through it later in greater details but thanks a million for getting me started.

BTW, I have been waiting for people to comment on the use of LF353/412 in the amp. People commented about how wonderful the 9+ sounded but I also heard a ton of how woeful those op-amps are.

What gives?
 
#32
millwood said:



thanks, Joe, I will go through it later in greater details but thanks a million for getting me started.

BTW, I have been waiting for people to comment on the use of LF353/412 in the amp. People commented about how wonderful the 9+ sounded but I also heard a ton of how woeful those op-amps are.

What gives?
Click to expand...
Those opamps do stink but at the time it was built they were the 'hot' units.
 
#34
millwood said:
I have been waiting for people to comment on the use of LF353/412 in the amp. People commented about how wonderful the 9+ sounded but I also heard a ton of how woeful those op-amps are. What gives?
Click to expand...

A general approach to improving opamp sonics is to maintain their output stages in class A operation. Usually this is done either by means of an external pulldown resistor or current source, or by ensuring that the opamp output stage doesn't have to deliver more current to a load than it can deliver in class A.

It looks to me like the Sumo topology uses both of these tricks. As long as driver transistors Q11 and Q14 are conducting, R7+R14 and R29+R34 would act like external pulldown resistors for the two opamp sections. And since the driver+output section only needs a few mA of drive, the output stage of each opamp probably stays in class A under all operating conditions.

I can't comment from experience on specific opamp choices, but I have read glowing reports of amplifiers that use them, such as the Monarchy class A amp. So it seems that, if used properly, they can work very well. Of course, if you DIY, you can substitute your own discrete front end if you prefer.
 
#38
Op-Amps in the NINE:

James Bongiorno participates in the Yahoo group "SAE-Talk"
and has clearly stated:

"Regarding the dual opamp, that is a SELECTED LF353. An off the shelf part will NOT work. Do not mess with this. The original units had specially made LF353's that were laser trimmed to withstand
higher supply voltages without shutting down."
 
#39
computeruser said:
Op-Amps in the NINE:

James Bongiorno participates in the Yahoo group "SAE-Talk"
and has clearly stated:

"Regarding the dual opamp, that is a SELECTED LF353. An off the shelf part will NOT work. Do not mess with this. The original units had specially made LF353's that were laser trimmed to withstand
higher supply voltages without shutting down."
Click to expand...


Hi,
You could use a OPA2604 here but James B. vigorously dislikes Burr-Brown opamps..........😎
 
Status
Not open for further replies.
Top Bottom