Carver Crimson 275 Measurements

[Blackdog]

I read that response, and had a look at the posted schematic. All I can say is bull shit.
The magic resistor in question is a 0.1 ohm in series with the secondary (which is grounded) and the output terminal.
It is also connected to a 270 ohm resistor that is part of the feedback network. The junction of these two resistors is the negative output terminal.
Not only would that 0.1 resistor make power readings worse, with the junction of the two resistors mentioned forming a voltage divider, there is only a very small sampling of signal "floating".
That resistor can only make things worse for measuring. It should be better if bypassed.
None of this dealers explanation makes sense.

I would like to see a comparison test result with the negative terminal grounded and floating as specified in that response.

 

[mhardy6647]

No, assuming it does work as advertised, something I'm seriously sceptical of, the peculiar grounding only affects measurements, as loudspeakers are floating devices so would connect as the amp requires. Anybody buying this and using it for the purpose of driving loudspeakers won't be affected.

S

Good point. I was taking Dr. Carver's rebuttal in its own context (i.e., at face value) -- in my defense.

 

[sergeauckland]

It doesn't short out the main feedback circuit. If that were the case (a short from output to input) there would be almost no gain at all and the amp would behave very badly indeed. If it bypassed the feedback, the gain would be excessive and Paul would surely have noticed. It's just another compensation factor that, if I understand correctly, shouldn't affect the amps behavior into a fixed load.

I read that response, and had a look at the posted schematic. All I can say is bull shit.
The magic resistor in question is a 0.1 ohm in series with the secondary (which is grounded) and the output terminal.
It is also connected to a 270 ohm resistor that is part of the feedback network. The junction of these two resistors is the negative output terminal.
Not only would that 0.1 resistor make power readings worse, with the junction of the two resistors mentioned forming a voltage divider, there is only a very small sampling of signal "floating".
That resistor can only make things worse for measuring. It should be better if bypassed.
None of this dealers explanation makes sense.

I would like to see a comparison test result with the negative terminal grounded and floating as specified in that response.

I haven't seen a circuit diagram so was thinking that this resistor was creating some current feedback but I agree that a resistor of only 0.1ohms is unlikely to do much. In which case, bull shit is more likely. Nevertheless, if the manufacturer recommends that the amp be measured with a floating load, then that's how it should be measured in the first instance. Then, if by some miracle, it meets its spec, but not with a grounded load, we have something to look into.

S

 

[amirm]

I would like to see a comparison test result with the negative terminal grounded and floating as specified in that response.

What measurement do you want for this?

 

[Zackthedog]

I read that response, and had a look at the posted schematic. All I can say is bull shit.
The magic resistor in question is a 0.1 ohm in series with the secondary (which is grounded) and the output terminal.
It is also connected to a 270 ohm resistor that is part of the feedback network. The junction of these two resistors is the negative output terminal.
Not only would that 0.1 resistor make power readings worse, with the junction of the two resistors mentioned forming a voltage divider, there is only a very small sampling of signal "floating".
That resistor can only make things worse for measuring. It should be better if bypassed.
None of this dealers explanation makes sense.

I would like to see a comparison test result with the negative terminal grounded and floating as specified in that response.

It's current feedback, right? The Pye HF-25 used the same arrangement, with a pot across a .27 ohm resistor as "variable damping." In fact a number of tube amps at one point offered this feature.

 

[Thermionics]

Can we see a shot of the innards of the amp?

I'll need to dig one up - I might have one from a while back. It's not going to win any neatness awards.

 

[Zackthedog]

I haven't seen a circuit diagram so was thinking that this resistor was creating some current feedback but I agree that a resistor of only 0.1ohms is unlikely to do much. In which case, bull shit is more likely. Nevertheless, if the manufacturer recommends that the amp be measured with a floating load, then that's how it should be measured in the first instance. Then, if by some miracle, it meets its spec, but not with a grounded load, we have something to look into.

S

No, I think you're right. It doesn't take much resistance to do that. Here's the Pye HF-25 (you have to scroll down):

100 Amplifiers – Part 3 , 1955 – 59 | Lilienthal Engineering

What I don't get is that the schematic shows the zobel network across the outputs as grounded on the negative side. So isn't the negative transformer secondary already grounded, *before* the current feedback resistor? Or is the schematic wrong? Or am I missing something?

 

[atmasphere]

Blocking?

No. They round the corners when clipping. So if the clipping is not severe you may not even know its happening.

 

[Blackdog]

What measurement do you want for this?

What measurement do you want for this?

Sorry, should have power comparison.

 

[Thermionics]

Can we see a shot of the innards of the amp?

 

[SIY]

No. They round the corners when clipping. So if the clipping is not severe you may not even know its happening.

TBH, I have never seen this in a tube amp not designed as a distortion box (i.e., for musical instruments). Every hifi tube amp I've ever measured shows flat-top clipping, since the limitation is reached when the output stage starts to draw grid current. And of course, blocking follows shortly thereafter. AB2 amps will similarly clip pretty sharply, but they're rarities.

Disclaimer: I haven't measured clipping in your OTL circlotrons, so is that clipping behavior different? Can you show some examples for power amps?

 

[TriodeLuvr]

Not to go all Perry Mason on this*, but:

If the amps are difficult to test (because of some magic grounding scheme that's not intuitive to someone who does this stuff for a living) -- would that not mean that it would be difficult to use for an ordinary dope-smokin' schlub off the mean streets of Seattle?

Would that not mean such, Dr. Carver?


___________________________
* The Raymond Burr version, that is.

View attachment 178860

The home user would ordinarily not experience the problem. The grounding issue is only present if a grounded appliance is attached to the negative speaker terminal. A number of amplifiers in the past have used a similar topology relative to the speaker output. In any event, Paul has used a testing method that entirely circumvents this issue, and I assume the new recipient of the unit will do the same.

 

[TriodeLuvr]

TBH, I have never seen this in a tube amp not designed as a distortion box (i.e., for musical instruments). Every hifi tube amp I've ever measured shows flat-top clipping, since the limitation is reached when the output stage starts to draw grid current. And of course, blocking follows shortly thereafter. AB2 amps will similarly clip pretty sharply, but they're rarities.

Disclaimer: I haven't measured clipping in your OTL circlotrons, so is that clipping behavior different? Can you show some examples for power amps?

The tube amp typically reaches 1% distortion well before clipping. This means that if a solid state amp and a tube amp are rated for the same output power at 1%, the tube amp will deliver more usable power on musical peaks.

 

[atmasphere]

TBH, I have never seen this in a tube amp not designed as a distortion box (i.e., for musical instruments). Every hifi tube amp I've ever measured shows flat-top clipping, since the limitation is reached when the output stage starts to draw grid current. And of course, blocking follows shortly thereafter. AB2 amps will similarly clip pretty sharply, but they're rarities.

Disclaimer: I haven't measured clipping in your OTL circlotrons, so is that clipping behavior different? Can you show some examples for power amps?

I've seen it plenty of times. Any tube amp has this behavior (assuming we're talking about the same thing IOW not blocking distortion). If the amp employs feedback the onset of sharp flat top clipping will be more pronounced. Our OTLs are no different in that regard although the driver is designed to be able to run the power tubes with grid current; class A2 if the load impedance is high enough, class AB2 if it isn't.

The traditional explanation for this behavior is 'space charge' due to an electron cloud forming around the plate structure of the power tube near clipping. A pentode will therefore have a sharper clipping characteristic than a power triode.

 

[MakeMineVinyl]

Is that four vintage oil capacitors at the top?

 

[Blackdog]

What I don't get is that the schematic shows the zobel network across the outputs as grounded on the negative side. So isn't the negative transformer secondary already grounded, *before* the current feedback resistor? Or is the schematic wrong? Or am I missing something?

That strikes me as odd as well. If you ground the negative terminal with the test equipment only a very small portion cfb is grounded. It should make no difference to power.
If you look at the Pye circuit there is a pot for variable cfb. One side is grounded. I doubt tbat made much difference in power.

The other thing I can't get my head around is the 4.7M resistor on the AC inlet going to ground.
Last time I saw this was an amp from the 50's.
Not quite legal today, but maybe more of Carver's voodoo.

 

[SIY]

The tube amp typically reaches 1% distortion well before clipping. This means that if a solid state amp and a tube amp are rated for the same output power at 1%, the tube amp will deliver more usable power on musical peaks.

1% is not even noticeable on a scope trace, much less be anything like clipping. Here's a plot picked from a random review on the first page results at Stereophile:

source https://www.stereophile.com/content/doshi-audio-evolution-monoblock-power-amplifier-measurements

Here's another off that page:

source: https://www.stereophile.com/content...cbloc-monostereo-power-amplifier-measurements

I'm not seeing that soft clipping in either example. It goes through the roof just as sharply as any other kind of amp.

 

[atmasphere]

Total FWIW dept.: the Edcor EM04906 is not a catalog item. Further, all the Edcor parts sans end bells have purple bobbins- they look distinctly different. So the transformer in question does appear to be a custom unit. FWIW.

 

[atmasphere]

1% is not even noticeable on a scope trace, much less be anything like clipping. Here's a plot picked from a random review on the first page results at Stereophile:

View attachment 178880
source https://www.stereophile.com/content/doshi-audio-evolution-monoblock-power-amplifier-measurements

Here's another off that page:
View attachment 178881
source: https://www.stereophile.com/content...cbloc-monostereo-power-amplifier-measurements

I'm not seeing that soft clipping in either example. It goes through the roof just as sharply as any other kind of amp.

Its easier to see if you observe a sine wave and push it to clipping. You'll see the sharp edges of a solid state amp make the sine wave look like someone clipped the top off with a pair of scissiors. A tube amp will round those corners- you have to push it pretty hard to get the corners as sharp as a solid state does.

The image is what I'm talking about. If you just look at the distortion curve you won't see this.

 

[Zackthedog]

That strikes me as odd as well. If you ground the negative terminal with the test equipment only a very small portion cfb is grounded. It should make no difference to power.
If you look at the Pye circuit there is a pot for variable cfb. One side is grounded. I doubt tbat made much difference in power.

Yes, though I suppose you could argue that varying the damping factor could affect the LF response. Would I be wrong?