Why 4 Ohm Loads Stress Your Amplifier

[sanyofreak]

Consider your bow shot. Your generous analysis has huge errors...sorry.
An ideal amplifier, like the Levinson ML 2 (25w RMS), will deliver twice the current if the impedance is halved, and twice again if halved again. So this amp is 25W into 8 ohm, 50W into 4, 100W into 2, 200W into 1....etc. The ML2 is also a nice example because it is class A. The same current flows across the outputs with or without a load. This enables the amp to drive short circuits with no damage.

I said before,
I'll say again for my sake, It is EXTRAORDINARY to find a SS amp that perfectly doubles its output with each halving of the load. If You find one, It will be very expensive.(Thats what I said) I've quickly checked out the Levinson ML 2, And as we all would expect, by anyone's stretch that I know of - this is an extraordinary amp. I dont know what the MSRP was on this amp, but when it is revealed I know I will agree with my perception - very expensive. It's also a 25WPC amp that appears to weigh in at well over 50 pounds - and thats a monoblock folks - ya need two fer stereo, You got'em, and I'll definitely consider You on the Extreme "High end" as far as price goes. You'll also likely have a pair of amps that weigh in at more than a G-22000(docked), but is (only) 25WPC at 8 ohms. See what I'm gettin at ?? Perfect doubling IS a herculean feat, and in no way would I consider an amp that does not acheive this spec to have any corners cut, on that premise alone.

The number one item that makes the ML 2's specification possible - is this.....It incorporates a fully regulated power supply, trust me, the perfect doubling feat simply cannot and will not be done on this earth without that feature. (in a class AB amp) Even if my dual 500VA transformers and outrageous filtering were behind a 10WPC amp, without regulation, it still 'prolly will not double its power with a halving of load. The expense of a regulated supply that can deliver the power necessary for an audio amp is definitely not only beyond ANY receiver ever built, but also beyond most seperates.
I agree the Concept is a wonderful piece of gear, but with a rating of 165W/Ch into 8 ohms, would not perfect doubling be more like 330 W/Ch into 4 ohms ?? 250 is more than half the watts over its 8 ohm rating, not a far cry, but a cry from the ideal, and this is excellent performance still.
Now here's a real tangent - in an amp like the Levinson ML 2, can anyone guess what the amps Dynamic headroom spec is ???
I've spoken of this before.
But to agree with EW, this is still a worthwhile and informative discussion, and 240's points are definitely something to consider with multiple 8 ohm speakers or driving a single pair of 4 ohm rated speakers.

Other points still are valid, remember, its the actual Junction inside the BJT's case that does all the magic, and all the smoke if she blows. Every power transistor has a spec on how well the case itself is thermally "coupled" to the Silicon junction. Then, the case has to be able to conduct that heat to the sink itself, (use your grease or goop)and finally the sink to "air" thermal efficiency. Somewhere out there is a member Hotgas, who can help us with all of this type of physics, but no matter, If that junction cant transfer its dissipation to the case efficiently enough - death, so in some cases fan cooling will also not be enough protection,because the junction cant shed its heat to the case/heatsink/air, quickly enough. Like EW says, if Your listening too loud, in a nutshell.
Great exchange and please, no offense to anyone here, and I will say that the ML 2 and the Concept 16.5 are both awesome products.

 

[Lefty]

Well having read through these many posts I think I have come to agree with those that have identified output device power dissipation limits (SOA) as the cause of the 4 ohm vs 8 ohms load questions. I don't think the increased output device current value is that big a deal as much as the power dissipation rating of the output devices. Here is my thinking, cutting to the chase so to speak:

Most class AB amps and receivers utilize fixed (but not regulated) DC power rails. These might be say +/- 50vdc. Just using AC RMS continuous value calculations one can show that running at a continuous 50 watts output to a 8 ohm load we would have:

20 vac output, 2.5 amps output and the output transistors would have to dissipate 75 watts of heat ((50v-20v) x 2.5A)

At 4 ohm load with the same 50 watts output:

14.14 VAC, 3.5 amps and heat dissipation is 125 watts ((50v-14.14) x 3.5 A)

So just changing from a 8 ohm load to a 4 ohm load caused the heat dissipation to increase by 50 watts, the same amount as is going to the load!

Now if the manufactures had put in a means to lower the DC rail voltages, say by having and switching the transformer secondary winding taps for running 8 or 4 ohm speakers, one could easily drive 4 ohm loads as easily as 8 ohm loads. I think that some of the models did use automatic DC rail voltage switching as a means to improve efficiency, Class G ?

So I've come to believe, as others have stated, that device power dissipation, heat sink capacity and such is the real limit on the ability of any given amp to be able to handle 4 ohm loads as safely as they handle 8 ohm loads, because we are usually stuck with a fixed rail DC voltage buss and we much keep the output devices within their safe operating area, SOA. Output device current values don't seem a real limit to me as most power output transistors were rated at 10 amps max or more. It's the power dissipation increasing with a lowering load resistance that is the weak link...

Come to think of it, from my old ham radio days I recall with power output tubes the single most important specification was the tube's max power dissipation rating, everything else you could design around or to, but you couldn't improve on it's power dissipation without more forced air cooling.

Anyway, any comments. Have we beat this horse enough

Lefty

 

[240 Volts]

SanyoFreak, agree with pretty much all you say about amplifiers doubling their output into 4 ohms and doubling again into 2 ohms.

I'm going to play Devil's Advocate here ('cos I'm like that sometimes ) :

We all agree that the ideal amplifier would have perfect voltage regulation and would double it's output power into 4 ohms (assuming that money was no object). :yes:

So if you had the choice between 2 otherwise identical amplifiers, "A" and "B", that had the following power outputs, which would you choose?
Amplifier "A" : 100 watts into 4 ohms, 50 watts into 8 ohms.
Amplifier "B" : 100 watts into 4 ohms, 80 watts into 8 ohms.

Would anyone really choose the ideal amplifier "A" over the less than perfectly regulated amplifier "B"? :scratch2: I thought not!!

- Richard B.

 

[240 Volts]

...... But, if you take a look at this thread:

http://audiokarma.org/forums/showthread.php?t=50369&highlight=SPEC

You'll see where EW put his dummy loads on the first one of mine he rebuilt and at 8 ohms it didn't clip until 340 watts per channel!

He then when on to say that 400 wpc at 4 ohms would not surprise him.

So, is all this good enough to get that SX-450?????

I read that thread the first time around - the SPEC 2 sure is a damn fine amplifier. Never see any here in the UK :tears: , but if one became available..... Hell, I might even give up BOTH my SX-450's for one!!

I think I'm going to claim the prize myself - found the following :

It certainly implies that the output is the same into 4 ohms as it is into 8 ohms. What it makes very clear is that it is the power dissipation / heating that would have been the problem had it not been for the speaker impedance selector switch.

Regarding the measurements that EW got for your SPEC 2 :
Yes, sounds like 400 watts into 4 ohms would have been possible for a short time with the switch in the 8 ohm position, but I don't think you'd be able to do it for long without melting something.
EW said that the supply voltage was 93V in the 8 ohm position - wonder what the voltage is with the 4 ohm setting? For 250 watts into 4 ohms you'd need 45V peak output - taking transistor saturation voltages into account you'd be struggling to get 250 watts if the supply voltage was halved by means of a series / parallel changeover arrangement.

- Richard B.

 

[crooner]

It's all in the numbers game. The reason why they are not rated at 4 ohms is the THD figures slightly higher for said load. I am sure the distortion is still well within the audible limits. Still, Pioneer chose not to publish them for marketing reasons.

 

[crooner]

It should be noted that Leonard Feldman of Audio Magazine measured 460 watts of short-term power per channel for the SX-1980. This is the equivalent of 2.3 dB of IHF Dynamic Headroom. (Audio, Sept 1978, page 76).

 

[crooner]

Great post sanyofreak!! :thmbsp:

I said before,
I'll say again for my sake, It is EXTRAORDINARY to find a SS amp that perfectly doubles its output with each halving of the load. If You find one, It will be very expensive.(Thats what I said) I've quickly checked out the Levinson ML 2, And as we all would expect, by anyone's stretch that I know of - this is an extraordinary amp. I dont know what the MSRP was on this amp, but when it is revealed I know I will agree with my perception - very expensive. It's also a 25WPC amp that appears to weigh in at well over 50 pounds - and thats a monoblock folks - ya need two fer stereo, You got'em, and I'll definitely consider You on the Extreme "High end" as far as price goes. You'll also likely have a pair of amps that weigh in at more than a G-22000(docked), but is (only) 25WPC at 8 ohms. See what I'm gettin at ?? Perfect doubling IS a herculean feat, and in no way would I consider an amp that does not acheive this spec to have any corners cut, on that premise alone.

The number one item that makes the ML 2's specification possible - is this.....It incorporates a fully regulated power supply, trust me, the perfect doubling feat simply cannot and will not be done on this earth without that feature. (in a class AB amp) Even if my dual 500VA transformers and outrageous filtering were behind a 10WPC amp, without regulation, it still 'prolly will not double its power with a halving of load. The expense of a regulated supply that can deliver the power necessary for an audio amp is definitely not only beyond ANY receiver ever built, but also beyond most seperates.
I agree the Concept is a wonderful piece of gear, but with a rating of 165W/Ch into 8 ohms, would not perfect doubling be more like 330 W/Ch into 4 ohms ?? 250 is more than half the watts over its 8 ohm rating, not a far cry, but a cry from the ideal, and this is excellent performance still.
Now here's a real tangent - in an amp like the Levinson ML 2, can anyone guess what the amps Dynamic headroom spec is ???
I've spoken of this before.
But to agree with EW, this is still a worthwhile and informative discussion, and 240's points are definitely something to consider with multiple 8 ohm speakers or driving a single pair of 4 ohm rated speakers.

Other points still are valid, remember, its the actual Junction inside the BJT's case that does all the magic, and all the smoke if she blows. Every power transistor has a spec on how well the case itself is thermally "coupled" to the Silicon junction. Then, the case has to be able to conduct that heat to the sink itself, (use your grease or goop)and finally the sink to "air" thermal efficiency. Somewhere out there is a member Hotgas, who can help us with all of this type of physics, but no matter, If that junction cant transfer its dissipation to the case efficiently enough - death, so in some cases fan cooling will also not be enough protection,because the junction cant shed its heat to the case/heatsink/air, quickly enough. Like EW says, if Your listening too loud, in a nutshell.
Great exchange and please, no offense to anyone here, and I will say that the ML 2 and the Concept 16.5 are both awesome products.

 

[crooner]

The Technics SA-1000 actually has independent B+ rails for 4 and 8 ohm loads. When it first powers on, it measures the approximate load attached to it and connects the proper supply. Very clever...

Now if the manufactures had put in a means to lower the DC rail voltages, say by having and switching the transformer secondary winding taps for running 8 or 4 ohm speakers, one could easily drive 4 ohm loads as easily as 8 ohm loads. I think that some of the models did use automatic DC rail voltage switching as a means to improve efficiency, Class G ? Lefty

 

[EchoWars]

EW said that the supply voltage was 93V in the 8 ohm position - wonder what the voltage is with the 4 ohm setting?

Doesn't drop that much...6 or 7 volts IIRC. Almost doesn't seem worth the trouble.

 

[thinker10]

Some of the Bryston do double up at 4 ohms.

I don't know of many that does.

Even most Pro Amp don't.

Thank You

 

[rwortman]

Any heat issues arising from operation at lower impedances would then be a funcion of the output device itself not the heatsinks attached to it. The device will be drawing much more current and if limiting is not employed it will self-destruct no matter how large the heatsinks may be...

This may be true but only if the heatsinks are large enough. Not enough or no heatsink will always derate the device. Once you have enough sink and air to it you are dealing with the thermal transfer characteristics of the transistor package itself. Someone above said that it's not the heat that kills the transistor, it's the current. Seems to me that is the same thing. Heat is how current kills transistors. We can apply enough voltage to break down the junctions but hooking up a lower impedance pair of speakers doesn't apply more voltage to the transistor. It just draws more current which can cause the transistor to fail............by frying it with heat.

 

[ron-c]

Good discussion. A couple of quick points though. The output stage of an amp can be optimized for 2, 4, 8 or many other impedances. Amps that claim to double power are in fact designed for 2 or 4 ohms and have very low power into higher impedances which means they weigh a lot and clip and eat tweeters on 8 ohm speakers.
Yes switching the voltage rails will allow an amp to work into 4 or 8 ohms. Ohms law again. The trick is to keep distortion low at both voltages and the higher cost of a two voltage power transformer. Power supply cap size and most importantly recharge rate is also key at higher frequencies.
If you want the lowest distortion with the least amount of negative feedback the output stage will have sweet spot design wise. This usually does not fall at a useful impedance however. It may be like 2.7 ohms in the MC402. One solution is the autoformer which will match the 2.7 to 2, 4, or 8 Ohms but the drawback is cost, weight and size.
Todays modern transistors are much improved in both Safe Operating Area and the mounting area to the heatsinks. This of course improves things heat wise. Having a failure in an amp's output stage has actually become a rare event compared to where we were a decade ago.

Thanks,
Ron-C

 

[Urizen]

Bob Carver states that his original Sunfire will push 300 watts in 8 ohms and will double to 600 watts at 4 ohms and 1200 watts into 2.

Downtracking he calls it. I use a Sunfire in my main system.

How does he do it? :scratch2:

 

[crooner]

Bob Carver states that his original Sunfire will push 300 watts in 8 ohms and will double to 600 watts at 4 ohms and 1200 watts into 2.

Downtracking he calls it. I use a Sunfire in my main system.

How does he do it? :scratch2:

It's probably only FTC rated at 8 ohms. Preconditioning at 2 ohms would certainly smoke it!

Carver, BTW, was originally opposed to the FTC rules in the mid 1970s. He's pretty much a believer of peak power for short periods of time.

 

[crooner]

Ron-C, I was going to mention autoformers but you beat me to it!

I've noticed nowadays manufacturers don't pay enough attention to FTC preconditioning when rating amplifiers. Stereophool has reviewed many recent units that shut down or overheat during the preconditioning period.

Perhaps it's because the FTC is not enforcing the rule like in the 1970's

Pioneer's SPEC 2 and SX-1980 appeared when those rules were strictily enforced, hence their conservative ratings.

I am sure they would rate even better if measured today...

 

[Soran]

also the toshiba 7150 is a 2x150 watt receiver but gives 2x230 watt into 4 ohm.

 

[ron-c]

We have new rules on heat and are only allowed a 20 degree C raise from room temp. by the CE rules. So most units will shut off and turn on while preconditioning. Unless they have a fan of course which means fan noise. If a manufacturer rates an amp at 8, 4 and 2 ohms it must pass on all impedances. This is why you see the cryptic 'stable at 4 or 2 ohms' on many spec sheets. What does that mean? It means it will not catch fire and has nothing to do with power output.

Thanks,
Ron-C

 

[biggal76]

In an amp rated for only 8 ohms or greater, why couldn't you hook a 4 ohm resistor in series with your 4 ohm speaker. Then you could safely run a 4 ohm speaker off of your amp, right? If so, what would that change about the sound produced? Would it reduce the efficiency? By how much? Also, if this is a good solution for people paranoid about overheating their amps (like the people who were thinking of hooking up fans to their units), what type of resistor (besides it being 4 ohm) would you need?

Also, my SX1250 will not work when both A and B speaker selectors are pushed even though I only have speakers connected to A. Are the A,B,C connected in series in the big SX receivers or has mine been altered?

 

[ron-c]

You can add a 4 ohm resistor but it had better be big because it will dissipate 1/2 the power. If you have speakers on both A and B you will hear sound as they are run in series so as not to lower the impedance to the amp.

Thanks,
Ron-C

 

[240 Volts]

......Also, my SX1250 will not work when both A and B speaker selectors are pushed even though I only have speakers connected to A. Are the A,B,C connected in series in the big SX receivers or has mine been altered?

You can add a 4 ohm resistor but it had better be big because it will dissipate 1/2 the power. If you have speakers on both A and B you will hear sound as they are run in series so as not to lower the impedance to the amp.

Pioneer has definately built different "versions" of the SX-1250 - this is what I found looking at mine :

Two of my 3 three SX-1250's are as follows :
The speaker selector buttons operate completely independently, i.e. pressing in a button does not release any of the others, so it's possible to have speakers A+B+C all selected at once. The A, B, C outputs are all in parallel, however there is only output to the speakers if one or two sets are selected at once - selecting all three at once cuts off output to all of them.
There is a warning next to the speaker outputs which states :
"CAUTION --------- SPEAKER IMPEDANCE
A, B, C ------------4 OHMS OR MORE / SPEAKER
A+B, A+C, B+C ---- 8 OHMS OR MORE / SPEAKER"

The third of my SX-1250's is completely different :
Pressing any of the speaker selection buttons releases any of the other two which are already selected - so it "normally" only runs one set at once.
However, two sets of speakers can be selected at once by pressing both buttons simultaneously - these outputs are then connected in parallel.
The warning next to the speaker outputs states :
"CAUTION --------- SPEAKER IMPEDANCE
A, B, C ------------8 OHMS OR MORE / SPEAKER"

So my experience is that selecting multiple speakers on the SX-1250 is done in parallel and not in series.

Finally, if you add a 4 ohm resistor in series with your 4 ohm speakers then not only will it reduce the volume level but it will also completely screw up the SX-1250's damping factor - it will fall from 30 to only 1. :tears:

- Richard B.