G.Kleinschmidt said:
Hi Kleinschmidt,
Of course, RC-filtering does improve the PSSR, that is, wrt the main power supply. I did say you can do BOTH, didn't I? But that's not my point. If one askes: how can I improve the PSRR, you don't answer: just take perfect power supply (only Mike is allowed to say things like that).
forr and I were discussing the PSSR wrt to the power rail of the VAS (and current mirror) itself.
Cheers,
estuart said:
Well for Pete's sake, the whole purpose of R-C filtering the rail’s for the low power circuitry is because the power supply isn’t perfect. Any yes, you did say that one can both improve the VAS compensation scheme and R-C filter the rails. I agree with you, but so what?
My first post on this topic was to simply state that, from what I have read of his work on the topic, D. Self’s tendency to stick with simple miller compensation of the VAS revolves around the fact that R-C filtering the low-power supply rails generally provides perfectly adequate PSRR, with little or any measurable benefit gained from further complicating the VAS compensation scheme. That’s all!
Cheers,
Glen
G.Kleinschmidt said:
forr said:
Hi Glen,
let's go back to the beginning. I just gave an answer to forr's question, see above. The subject deals with alternate compensating schemes and definitely not with trivial things like improving the power supply lines. Besides, it makes a big difference in terms of cost, space and elegance, whether to use a big 1mF cap (as Self suggests) or a small 100pF cap. For the rest, I don't disagree with you on the effectiveness of a RC filter. That’s all!
Cheers,
PS: As a matter of fact, D. Self has referred to Ribner & Copeland, but omitted any details.
Hi all,
Thanks for your comments.
Both scrutinizing every power supply injection mechanism and examining every solution to cure this evil are of interest for greater general knowledge about amplifiers.
It can sometimes help to have more than one technique at hand to enhance performances.
Sometimes we may not want to have brute RC filtering at all, in the absence of an output relay, for example, so the trick of adding a grounded cap of same value as the Miller one in the other branch of the current mirror is wellcome
Or, on the contrary, ultra-perfectionnists may want to combine all the possibles tricks to get performances far better than those really needed.
What about decoupling the collector of the emitter follower at the input of the VAS ?
Is there any interest to use a more sophisticated current mirror ? I am thinking of an other emitter follower, loaded by a resistor, to supply current to the base of each transistor of the current mirror (scheme used in the second stage of Bob Cordell's Siliconix amp)
Thanks for your comments.
Both scrutinizing every power supply injection mechanism and examining every solution to cure this evil are of interest for greater general knowledge about amplifiers.
It can sometimes help to have more than one technique at hand to enhance performances.
Sometimes we may not want to have brute RC filtering at all, in the absence of an output relay, for example, so the trick of adding a grounded cap of same value as the Miller one in the other branch of the current mirror is wellcome
Or, on the contrary, ultra-perfectionnists may want to combine all the possibles tricks to get performances far better than those really needed.
What about decoupling the collector of the emitter follower at the input of the VAS ?
Is there any interest to use a more sophisticated current mirror ? I am thinking of an other emitter follower, loaded by a resistor, to supply current to the base of each transistor of the current mirror (scheme used in the second stage of Bob Cordell's Siliconix amp)
Ayre Acoustics MX-R fuse blown
I just read my latest copy of Stereophile and saw the review of the Ayre MX-R. First and foremost, this is a VERY nice amplifier, both in construction and performance. Although I am not a fan of zero-feedback amplifiers, the distortion achieved in this amplifier is very impressive given there is no NFB (on the order of 0.05% into a 4 ohm load at 30 Watts, where it appears to be highest).
During tests into a 2-ohm load this amplifier failed. Since we have discussed protection circuits and power transistor robustness at length earlier in this thread, I thought it would be interesting here to discuss what might have happened. Perhaps Charlie will join in and shed some light on it.
Most of us have at one time or another had an amp blow up on us, and we often know that it can be difficult to unravel which bad event happened first, as there is often collateral damage.
It is also particularly interesting that this amplifier uses the ThermalTrak bipolar power transistors.
The amp put out 360W into 8 ohms, 640W into 4 ohms, then broke at 775W into 2 ohms. Two 10-amp rail fuses blew and when John Atkinson replaced them they blew again.
Charlie mentioned in his manufacturers' comments that the 10 amp fuse is under-sized for a test into 2 ohms, and that normally a 15 amp fuse would be used for such a test. Of course, a blown fuse should not do permanent damage, but Charlie pointed out that the amplifier was mistakenly shipped with a driver protection fuse that was too big. Thus a driver transistor was apparently destroyed.
Here's what provoked my curiosity. At 775 watts into 2 ohms, the peak load voltage is 56V and the peak load current is 28A. If I did my math right, the average rail current should correspond to 0.5 * 0.63 * 28 A = 8.8 A for a Class AB output stage (this should be true for either a single-ended design or a balanced design like the Ayre). This is less than the value of the 10 Amp rail fuse. Of course, fuses have tolerances. But a Buss 3AG 10-amp fuse is rated to withstand 135% of its rating for quite a while - on the order of a half hour, I think. So this makes me wonder if it was the fuse that went first and the collapse took the driver with it, or if a different sequence of events may have actually occurred.
For example, the amp is running pretty hot under these test conditions, probably dissipating about 400W of heat. The heat sink over-temperature cut-out on this amplifier is set at 70C. Could there be something going on with the use of the ThermalTrack transistors for bias stability control? I'd guess Charlie has at least four pairs of output devices on each side of the bridge. Might there be an issue concerning how you choose which output transistor's tracking diode (or combiantion of several) to use for bias setting? Could there have been some kind of a thermal runaway situation? I wonder how big Charlie's emitter ballast resistors are.
Or could there not be enough output transistor SOA at a heat sink temperature of 70C. The transistor spec seems to say that there should be enough if four pairs are being used. At an output transistor beta of 50 and 28 amps into the load, the base current could have been as high as 0.56A; could the driver transistor have run out of SOA?
I'd like to hear some thoughts on what may have happened here.
Again, this really appears to be a beautiful amplifier, and it is unlikely that something like this would ever happen in a consumer's environment.
Cheers,
Bob
I just read my latest copy of Stereophile and saw the review of the Ayre MX-R. First and foremost, this is a VERY nice amplifier, both in construction and performance. Although I am not a fan of zero-feedback amplifiers, the distortion achieved in this amplifier is very impressive given there is no NFB (on the order of 0.05% into a 4 ohm load at 30 Watts, where it appears to be highest).
During tests into a 2-ohm load this amplifier failed. Since we have discussed protection circuits and power transistor robustness at length earlier in this thread, I thought it would be interesting here to discuss what might have happened. Perhaps Charlie will join in and shed some light on it.
Most of us have at one time or another had an amp blow up on us, and we often know that it can be difficult to unravel which bad event happened first, as there is often collateral damage.
It is also particularly interesting that this amplifier uses the ThermalTrak bipolar power transistors.
The amp put out 360W into 8 ohms, 640W into 4 ohms, then broke at 775W into 2 ohms. Two 10-amp rail fuses blew and when John Atkinson replaced them they blew again.
Charlie mentioned in his manufacturers' comments that the 10 amp fuse is under-sized for a test into 2 ohms, and that normally a 15 amp fuse would be used for such a test. Of course, a blown fuse should not do permanent damage, but Charlie pointed out that the amplifier was mistakenly shipped with a driver protection fuse that was too big. Thus a driver transistor was apparently destroyed.
Here's what provoked my curiosity. At 775 watts into 2 ohms, the peak load voltage is 56V and the peak load current is 28A. If I did my math right, the average rail current should correspond to 0.5 * 0.63 * 28 A = 8.8 A for a Class AB output stage (this should be true for either a single-ended design or a balanced design like the Ayre). This is less than the value of the 10 Amp rail fuse. Of course, fuses have tolerances. But a Buss 3AG 10-amp fuse is rated to withstand 135% of its rating for quite a while - on the order of a half hour, I think. So this makes me wonder if it was the fuse that went first and the collapse took the driver with it, or if a different sequence of events may have actually occurred.
For example, the amp is running pretty hot under these test conditions, probably dissipating about 400W of heat. The heat sink over-temperature cut-out on this amplifier is set at 70C. Could there be something going on with the use of the ThermalTrack transistors for bias stability control? I'd guess Charlie has at least four pairs of output devices on each side of the bridge. Might there be an issue concerning how you choose which output transistor's tracking diode (or combiantion of several) to use for bias setting? Could there have been some kind of a thermal runaway situation? I wonder how big Charlie's emitter ballast resistors are.
Or could there not be enough output transistor SOA at a heat sink temperature of 70C. The transistor spec seems to say that there should be enough if four pairs are being used. At an output transistor beta of 50 and 28 amps into the load, the base current could have been as high as 0.56A; could the driver transistor have run out of SOA?
I'd like to hear some thoughts on what may have happened here.
Again, this really appears to be a beautiful amplifier, and it is unlikely that something like this would ever happen in a consumer's environment.
Cheers,
Bob
Re: Ayre Acoustics MX-R fuse blown
This is admittedly unwarranted speculation on my part, not knowing the extent of damage beyond a probable failed driver.
Most probably fuses blew **after** a catastrophic failure, be it a shorted power output transistor or shorted/open driver, even a bias circuit open failure or pcb track blown.
With more data from visual inspection and semicondutor checking, a more educated guess could be made.
Rodolfo
Bob Cordell said:
This is admittedly unwarranted speculation on my part, not knowing the extent of damage beyond a probable failed driver.
Most probably fuses blew **after** a catastrophic failure, be it a shorted power output transistor or shorted/open driver, even a bias circuit open failure or pcb track blown.
With more data from visual inspection and semicondutor checking, a more educated guess could be made.
Rodolfo
G.Kleinschmidt said:
I am not really sure about that. I ask this in another post with respect to op amps.
RC Filters on Op Amp supply pins - PSRR Post #16
---------------------------------------------------------------------------
I have a quick question for people in the know. Does adding an RC network on the supply pin of an Op Amp boost it's PSRR ?
I know it would attenuate the junk going into the pin, but does it actually improve it's PSRR ?
What if the op amp was bootstrapped also ? RC filter in a loop sort of thing...
I was thinking of an RC filter with a 3db point set at the point where the PSRR/loop gain falls with frequency.
Plus PSRR is reduced by gain also.
Hi Bob,
speculating on that failure and your comment that 4pair NJL should be able to cope when bridged.
160W into 2ohm equating to 320W into 4ohm @ 60degree phase angle can be sustained @Tc<=75degC when MJL4302 are used as active devices. (the highest power/highest SOAR plastic packaged devices in my speadsheet).
388W into 1ohm causes problems (=776W into 2ohms).
Tc<=40degC lets the devices go just outside the 1second SOA with a phase angle of 60degrees.
With TC @ <=75degC, the devices go just past half way between 1second SOA and 100mS SOA. when the power is pushed to 500W at the same phase the devices are just outside the 100mS SOA.
A bridged 2ohm load is asking too much.
A resistive 2 ohm test load should be easy except for maximum current.
That is, I suspect, the clue.
At the 28Apk the drivers will be pushing out in excess of 550mA.
I will guess the driver(s) gave out first, particularly if they were bolted to the same sink or had inadequate separate sinks.
speculating on that failure and your comment that 4pair NJL should be able to cope when bridged.
160W into 2ohm equating to 320W into 4ohm @ 60degree phase angle can be sustained @Tc<=75degC when MJL4302 are used as active devices. (the highest power/highest SOAR plastic packaged devices in my speadsheet).
388W into 1ohm causes problems (=776W into 2ohms).
Tc<=40degC lets the devices go just outside the 1second SOA with a phase angle of 60degrees.
With TC @ <=75degC, the devices go just past half way between 1second SOA and 100mS SOA. when the power is pushed to 500W at the same phase the devices are just outside the 100mS SOA.
A bridged 2ohm load is asking too much.
A resistive 2 ohm test load should be easy except for maximum current.
That is, I suspect, the clue.
At the 28Apk the drivers will be pushing out in excess of 550mA.
I will guess the driver(s) gave out first, particularly if they were bolted to the same sink or had inadequate separate sinks.
Since the amplifier popped driving a resistive load, we can
rule out phase angle issues.
An AP performing a power sweep works quite rapidly, requiring
less than a second to perform an analysis at a particular
point, and it apparently died on completion of the 775 data point.
Whatever the cause, the failure was the result of quite short
term conditions.
rule out phase angle issues.
An AP performing a power sweep works quite rapidly, requiring
less than a second to perform an analysis at a particular
point, and it apparently died on completion of the 775 data point.
Whatever the cause, the failure was the result of quite short
term conditions.
Nelson Pass said:Since the amplifier popped driving a resistive load, we can
rule out phase angle issues.
An AP performing a power sweep works quite rapidly, requiring
less than a second to perform an analysis at a particular
point, and it apparently died on completion of the 775 data point.
Whatever the cause, the failure was the result of quite short
term conditions.
Hi Nelson,
This is a good point - it had not occurred to me that the AP would be completing the test so quickly (I don't own an AP, so such tests are longer in duration for me). This could also mean that my speculation that the amplifier was hot at the time was wrong. On the other hand, it might be possible that JA was doing the clip test in some kind of a manual mode, I suppose.
On the other hand, I wonder what the spacing between points taken by the AP was? If it was one Watt, does that mean that the AP would have ramped the power up over a 775 second period?
Bob
john curl said:
Yep, you're right. We're definitely nosey.
I agree completely that we shouldn't hold it against Charlie if he doen't tell us why it broke, or any other details for that matter.
Actually, I think he had a problem with a non-isothermal heat sinking structure, and that in the next rev he should mill the amplifier out of a solid copper billet, then gold plate it so it looks nice. Just kidding
.Cheers,
Bob
Opamp PSRR
http://www.diyaudio.com/forums/showthread.php?s=&threadid=97726&highlight=
I also curious about this. I ask the question here :
http://www.diyaudio.com/forums/showthread.php?s=&threadid=97726&highlight=
Bob Cordell said:
I hold Hansen in high regard - If I didn't have such easy access
to my own product, I would probably buy a pair of his.
And I don't think this is being nosy. Atkinson's taking out the
amp provides an interesting data point. We all have an interest
in comparing real failure rates against the "probable" failure rates
based on part specs.
Bob Cordell said:
Of course it costs very little to repair such damage, and I am
nothing if not discrete (pun).
