Amplifier Assymetrical Rail Voltages

Paul C

Paul C

Super Member
OK, this is a little strange to me...

I have two power amplifers here that have assymetrical rail voltages.

Both are Onkyo's, very similar but for voltages, voltage ratings on some caps, output devices. Same circuit topology, same boards, layout, etc.


M-5140, rated 105 w/ch at .09% distortion, output devices are STK4044V chips, rails measure +61.2 / - 59.6

M-5150, rated 150 w/ch at .09% distortion, output devices are STK4048V chips, rails measure +72.0 / - 70.6


There are a pair of diodes on the center tap of the transformer, wired parallel, one one way, the other reverse, before it goes to chassis ground. I have not seen this anywhere else. The outer leads of the transformer go to the bridge rectifer, then to the filter capacitors like any other power supply. See page 7 of the attached service manual.

I suppose these diodes are responsible for the offset voltages. But my question is... why?

Onkyo calls this their "Delta Power Supply"... which, other than those two diodes, is pretty normal.

Something to do with the Sanyo ouput chips?

Idle offset voltage on the speaker terminals is a very low 1.8 mv and 2.3 mv... very low. There is no idle current to set, no adjustments. That is automatic internal to the Sanyo chips.

Comments?

Paul
 

Attachments

Register to hide this ad
Thats strange..I've see similar on the primary side off of the (mains)ground pin to chassis ground (ground loop breaker).
 
Yes. outputs are common (chassis) ground.


But Onkyo must be doing something right here... these amps are rated at very low distortion (.09%).
 
Are they both North American model ?

International models with voltage switch and/or Asia/Europeen market could be different.

Just a guess

Thanks
 
any assymetrical current draw will swing the rails in that case. Doesn't really make much sense to me. Primary side yes, as a DC blocker, but secondary, what am I missing?
 
Paul C said:
I suppose these diodes are responsible for the offset voltages. But my question is... why?
Click to expand...

Hi

You can without any problem short this back to back diodes that are connected from the transformer middle point to ground , and measure again.

If anything become correct you must have a shorted diode (inspect).

If the imbalance persist , or the transformer has a inherent imbalance or one rail has more power consummation than the other and the voltage go down by virtue of the ohmic resistance of the secondary... :smoke:
 
So, I spent some time looking at the schematic. For the life of me, I can't figure out what they would put paralleled diodes in the common leg of the power supply. The diodes called for there are 3 amp/200PIV - pretty regular diodes. I doubt they have anything to do with the offset voltage after the filter caps, though, as that's set by the xformer winding count.

Frankly, the voltages on the opposite rails are close enough that they're within 10% of each other. Without doing any math, its my feeling that this is within regulation limits for most SS amps. I wouldn't worry about it too much.

Cheers,

David
 
inductor said:
the voltages are only as accurate as the transformer secondaries are wound, they are reasonably accurate, also if there is slightly more load on one rail this will reduce it a little, there is no reason to 'design it in' to have unequal supplies as the o\p would clip unequally
Click to expand...

Sounds more like this to me. Since it's only 1.6v difference. Doesn't take much load to drop a volt or two.
 
Tube_Dude said:
Hi

You can without any problem short this back to back diodes that are connected from the transformer middle point to ground , and measure again.

If anything become correct you must have a shorted diode (inspect).

If the imbalance persist , or the transformer has a inherent imbalance or one rail has more power consummation than the other and the voltage go down by virtue of the ohmic resistance of the secondary... :smoke:
Click to expand...


No, the voltage imbalance is apparently designed in, and the assymetrical voltages are indicated on the schematic. :headscrat


Remember, both windings drive both rails via the bridge. So even if the windings were assymetrical, it would not matter, both rails should measure the same.

So, I am back to the diodes, which I don't see in any other power supply, just Onkyo's "Delta Power Supply". "Delta" in calculus means "a change or difference between mathematical values".
 
Paul C said:
No, the voltage imbalance is apparently designed in, and the assymetrical voltages are indicated on the schematic. :headscrat


Remember, both windings drive both rails via the bridge. So even if the windings were assymetrical, it would not matter, both rails should measure the same.

So, I am back to the diodes, which I don't see in any other power supply, just Onkyo's "Delta Power Supply". "Delta" in calculus means "a change or difference between mathematical values".
Click to expand...

Here is my opinion about all this...

There is two points in your initial question.

The first point is the assymetrical voltage. I agree with you, this is designed in the original circuit. The reason for this? You see the over current protection circuit (Q502, Q503 and related components) is not symetrical, this circuit is trigered only by positive swing current to the speaker load. To be effective everytime, the voltage at the emmiter of Q503 must be higher than the maximum output voltage anytime, and particularly at positive output clipping. The clipping voltage at the output of STK4044V is limited by the driver and predriver stages within the STK4044V, and not only by the voltage at pin 14. It is a wise decision from Onkyo to maintain the collector of Q502, and incidently the emitter of Q503, at a higher voltage than the maximum output voltage. Otherwise Q502 could be reversed biased and you would loose your over current protection.

To acheive this, I think the windings in the transformer are not identical. If you disconnect the two red wires at the output of this transfo, you should measure different voltages for the upper winding and lower winding. If there was no connection between the center tap (which is not exactly in the center in my opinion) of this transfo and the common point between the two filter caps (C903 / C904), if both filter caps were identical, if the load on both sides of the supply were the same, the supply would be symmetrical. There is many "if"...

The second point, the diodes D902 / D903:

As you know, the tolerance in electrolytic capacitors is very large, usually +80% / - 20%. Imagine that C903 is much larger than C904. At turn on, the initial current from the transformer will be very large and if there was no connection point between the center tap of the transfo and the common point between the two caps, the charging current (I) would be the same in both caps, and the charging voltage (I=C dV/dt) would be much smaller across C903 than across C904. As you see, something must be done to help both capacitors to charge at the same voltage even if they have different capacitance.

This point is important not only at turn on, but each time a very large current is going to the speaker load one supply rail can discharge faster than the other if there is great discrepancy between the two caps.

In almost all power supplies there is a direct connection between the center tap from the transfo and the common point between the two filter caps. Better than nothing but not perfect, because the secondary windings have some resistance and you can still have unbalance between the two supply rails if the two caps are very different.

In our example, if C903 is much larger than C904 the voltage at the common point between the caps will be higher than the voltage at the center tap on the transfo because the voltage across C903 will be smaller than the voltage across C904. D903 will be forward biased and the upper part of the transformer will work harder to charge C903 and help both capacitors to charge at the same voltage. There is two current loops from the transformer. The first one, from the total winding across D901 and both C903 and C904, and the second one from the upper winding across D901 and C903 and D903.

For me, this is the most interesting aspect of this amp.

By the way, when you connect both speakers (A + B), they are in series, not in parallel with this amp. The frequency response of one speaker will be affected by the impedance curve of the other speaker. Not the best thing to do!
 
Ahhh! I think you have it... the protection circuit. Onkyos have pretty complex protection.

People can gripe about the use of the big Sanyo output chips, but they work very well. After turn on, no load, no signal, both channels quickly settled down to about 2 mv offset voltage within a minute or two.

Charging the power supply caps... hmmmm... I see what you mean.

I knew about the speaker series wiring. Other Onkyos are like that. But later units had the A and B speakers in parallel. For that exact reason I don't use two pr of speakers with those older Onkyos.
 
Paul C said:
Remember, both windings drive both rails via the bridge. So even if the windings were assymetrical, it would not matter, both rails should measure the same.

So, I am back to the diodes, which I don't see in any other power supply, just Onkyo's "Delta Power Supply". "Delta" in calculus means "a change or difference between mathematical values".
Click to expand...


No, it will only be true if there is no middle point connection from the transformer.

About the back to back diodes , for me that was a marketing gimmick, much publicized at the time , but that Onkyo left to fall in the new generation of amps...
 
You must log in or register to reply here.
Back
Top Bottom