here are some ideas on loudspeaker nonlinearity
with current and voltage drive
http://www.gedlee.com/nonlinearity.htm
with current and voltage drive
http://www.gedlee.com/nonlinearity.htm
I just wondering. Since PRR and Steven has told us the way to measure/calculate damping factor, is there any PRACTICAL way to measure other properties of:
-slew rate
-THD
-S/N ratio
With ordinary hobbyist equipment (voltmeter, scope, dummy load)?
There is some measurement device out there, like NeutrikA2, but it is very very expensive, and maybe only factory can buy it.
-slew rate
-THD
-S/N ratio
With ordinary hobbyist equipment (voltmeter, scope, dummy load)?
There is some measurement device out there, like NeutrikA2, but it is very very expensive, and maybe only factory can buy it.
Slew rate: there are several ways.
Feed a sine and increase the frequency and voltage until it becomes more of a triangle. The slope of the triangle is the slew rate. Or for a go/no-go test: feed 1KHz at just below clipping, switch to 20KHz, see if it is more sine or triangle.
In real life, if it is pretty-sine at full output up to 6KHz, it will pass music very nicely. Full power sine at 20KHz would burn your ears and tweeters, it really does not happen in anything you want to listen to.
NOTE: full power at high frequency will burn-out some amplifiers in a few seconds. Don't come crying to me.
THD: good THD measurements are very fussy. You can get a useful "Total Garbage" measurement by comparing input to output. You need to set the two gains so the signal cancels. If the amplifier is non-inverting, you need an inverter somewhere. Good 2-channel scopes can do this internally. Trim the gain for minimum 1KHz wobble, crank up the gain, and look at the residual. At high and low frequency you probably can't null the signal without an adjustable phase-shift network.
S/N: define your specification (there must be a zillion ways to specify S/N) and use an AC milliVolt meter.
The sound card test programs do simplify S/N and THD, though you will probably have to build pads to lower the input and output levels below 2V, maybe more like 0.2V for some cards. Always run a "straight wire" test so you know how good/bad your sound card is. Some are pretty good, some are about useless for solid-state work. You can not tell by specs: my "32-bit" onboard sound is missing the top bit and noise starts at the 15th bit, so the usable range is more like 14 bits or 84dB S/N and the test program I'm using always shows high THD. But I have done some very fine measurements even on older cards.
Feed a sine and increase the frequency and voltage until it becomes more of a triangle. The slope of the triangle is the slew rate. Or for a go/no-go test: feed 1KHz at just below clipping, switch to 20KHz, see if it is more sine or triangle.
In real life, if it is pretty-sine at full output up to 6KHz, it will pass music very nicely. Full power sine at 20KHz would burn your ears and tweeters, it really does not happen in anything you want to listen to.
NOTE: full power at high frequency will burn-out some amplifiers in a few seconds. Don't come crying to me.
THD: good THD measurements are very fussy. You can get a useful "Total Garbage" measurement by comparing input to output. You need to set the two gains so the signal cancels. If the amplifier is non-inverting, you need an inverter somewhere. Good 2-channel scopes can do this internally. Trim the gain for minimum 1KHz wobble, crank up the gain, and look at the residual. At high and low frequency you probably can't null the signal without an adjustable phase-shift network.
S/N: define your specification (there must be a zillion ways to specify S/N) and use an AC milliVolt meter.
The sound card test programs do simplify S/N and THD, though you will probably have to build pads to lower the input and output levels below 2V, maybe more like 0.2V for some cards. Always run a "straight wire" test so you know how good/bad your sound card is. Some are pretty good, some are about useless for solid-state work. You can not tell by specs: my "32-bit" onboard sound is missing the top bit and noise starts at the 15th bit, so the usable range is more like 14 bits or 84dB S/N and the test program I'm using always shows high THD. But I have done some very fine measurements even on older cards.
Back to damping factor etc.
I don't really know a whole lot about this kind of stuff but I get the gut feeling that if your amplifier had an output impedance of ZERO i.e. the damping factor by definition was INFINITE, the *actual* damping of the speaker would be 1/Qts or whatever 1/Qt? it may be when in the box That sounds a bit ordinary, doesn't it?
I don't really know a whole lot about this kind of stuff but I get the gut feeling that if your amplifier had an output impedance of ZERO i.e. the damping factor by definition was INFINITE, the *actual* damping of the speaker would be 1/Qts or whatever 1/Qt? it may be when in the box That sounds a bit ordinary, doesn't it?
Congrats for this conclusion !
It was already mentioned many times but not many people seem to listen.
IMO a DF of 100 is sufficient. I think it is more important that an amp's output impedance is independant of frequency (not fully achievable), output power and load impedance.
Regards
Charles
High HF NFB Can Mean Ringing Into Reactive Loads..........
Originally posted by phase_accurate
Congrats for this conclusion !
It was already mentioned many times but not many people seem to listen.
IMO a DF of 100 is sufficient. I think it is more important that an amp's output impedance is independant of frequency (not fully achievable), output power and load impedance.
I reckon that it is not mandatory for an extreme or constant damping factor across the full audio frequency range, and can actually be a hinderance.
Good damping factor is required at low frequencies and 100 or 200 is plenty.
High frequencies do not really require damping at all because the crossover networks in a typical decent quality loudspeaker provide local electrical damping of the mid and hf drivers.
Reduced NFB at higher frequencies can imply increased distortion and lowered damping factor, but this does not have to be sonically damaging, and in practice is a benefit.
I am familiar with an amplifier that does exactly this and it sounds great, so good in fact that a bunch of these are going into a seriously high end international mastering studio.
The figures are as follows.........
Harmonic Distortion into 8 Ohms
1W 100 Hz 0.01%
1W 1 KHz 0.01%
1W 10 KHz 0.05%
10W 100 Hz 0.01%
10W 1 KHz 0.01%
10W 10 KHz 0.05%
100W 100 Hz 0.015%
100W 1 KHz 0.02%
100W 10 KHz 0.20%
Intermodulation Distortion
50 Hz and 400 Hz 0.01%
50 Hz and 4 KHz 0.03%
Damping figures are not given, but I understand that they are fairly high (300 maybe) at low frequencies, and reducing at higher frequencies.
These figures don't look spectacular, but I think not worse than say Nelsons Aleph stuff, and better on paper than a heap of other stuff.
What the figures don't describe is the clearness, relaxedness and fluidness in the mid and higher frequencies, and the higher frequency distortion is not audibly detracting/distracting in any way.
This technique also implies better stability and less reactive load dependence, and these qualities indeed allow sonics detail and blackness eons better than all SS NFB amplifiers I have previously heard.
High NFB at high frequencies implies very high frequency open loop bandwidth if the amplifier is to remain stable into reactive loads.
High NFB at high frequencies can also cause SS sonic harshness and forcefulness.
High NFB at high frequencies is not electrically required, so enjoy the dynamic sonic benefits of lower NFB in the mids and highs I say.
Eric.
Originally posted by phase_accurate
Congrats for this conclusion !
It was already mentioned many times but not many people seem to listen.
IMO a DF of 100 is sufficient. I think it is more important that an amp's output impedance is independant of frequency (not fully achievable), output power and load impedance.
I reckon that it is not mandatory for an extreme or constant damping factor across the full audio frequency range, and can actually be a hinderance.
Good damping factor is required at low frequencies and 100 or 200 is plenty.
High frequencies do not really require damping at all because the crossover networks in a typical decent quality loudspeaker provide local electrical damping of the mid and hf drivers.
Reduced NFB at higher frequencies can imply increased distortion and lowered damping factor, but this does not have to be sonically damaging, and in practice is a benefit.
I am familiar with an amplifier that does exactly this and it sounds great, so good in fact that a bunch of these are going into a seriously high end international mastering studio.
The figures are as follows.........
Harmonic Distortion into 8 Ohms
1W 100 Hz 0.01%
1W 1 KHz 0.01%
1W 10 KHz 0.05%
10W 100 Hz 0.01%
10W 1 KHz 0.01%
10W 10 KHz 0.05%
100W 100 Hz 0.015%
100W 1 KHz 0.02%
100W 10 KHz 0.20%
Intermodulation Distortion
50 Hz and 400 Hz 0.01%
50 Hz and 4 KHz 0.03%
Damping figures are not given, but I understand that they are fairly high (300 maybe) at low frequencies, and reducing at higher frequencies.
These figures don't look spectacular, but I think not worse than say Nelsons Aleph stuff, and better on paper than a heap of other stuff.
What the figures don't describe is the clearness, relaxedness and fluidness in the mid and higher frequencies, and the higher frequency distortion is not audibly detracting/distracting in any way.
This technique also implies better stability and less reactive load dependence, and these qualities indeed allow sonics detail and blackness eons better than all SS NFB amplifiers I have previously heard.
High NFB at high frequencies implies very high frequency open loop bandwidth if the amplifier is to remain stable into reactive loads.
High NFB at high frequencies can also cause SS sonic harshness and forcefulness.
High NFB at high frequencies is not electrically required, so enjoy the dynamic sonic benefits of lower NFB in the mids and highs I say.
Eric.
I can generally agree BUT if DF gets too low (i.e. output impedance too high) then one ends up with a hevily load-dependant frequency response.
I agree that very high NFB factors, used as a cure-all, are less than optimal. But my opinion is that the NFB factor should be kept constant up to above the midrange at the cost of high NFB at lower frequencies. This would result in an amp that doesn't show increasing THD with rising input frequency (the latter being the case for many SS amps).
Regards
Charles
If this damping factor issue can be done, why not. If it will be limited by the speakers, it's something else, but doesn't have to stop us pursuing it. I'm sure there will be difference, inspite of the speaker limiting.
I'm very intersted in this. Any schematic for hint how this can be done?
I'm very intersted in this. Any schematic for hint how this can be done?
If it is OK with this forum, what is the exact brand and type of amp is beeing talked about
Does Not have To Be Hot Running Class A
Originally posted by phase_accurate
I can generally agree BUT if DF gets too low (i.e. output impedance too high) then one ends up with a heavily load-dependant frequency response.
Agreed, but in my experience relatively minor FR variations are much less sonically damaging than NFB/load induced ringing and higher order harmonics production etc.
I agree that very high NFB factors, used as a cure-all, are less than optimal. But my opinion is that the NFB factor should be kept constant up to above the midrange at the cost of high NFB at lower frequencies. This would result in an amp that doesn't show increasing THD with rising input frequency (the latter being the case for many SS amps).
Agreed, keep the NFB sufficiently high through bass and mids to keep damping appropriate and distortion low, and allow drooping NFB at above mids even though distortion is increased.
Lower order THD (produced/allowed by low NFB ratio) is sonically much more tolerable than even small amounts of higher order harmonics.
Some harmonics production above say 4 kHz is not particularly offensive mainly because higher order products are out of audio band and rendered inadible provided that lower order IM products are not generated also.
Lowered NFB can also imply good IM and TIM distortion performance provided that the open loop bandwidth is relatively high.
To my ear IMD and TIMD are FAR more damaging (and fatigueing) than lower orders of THD.
In practice, good damping/distortion performance at low/mid frequencies, and 'free-er' sounding high mids/highs adds up to very nicely acceptable sonic performance with typical speakers and very high performance large studio monitors also.
With ESL, the capacitive loading can cause a slight drooping response towards the highs, but there are no transient nasties produced as can be the case with high NFB amplifiers.
A droop of 1 or 2 dB at 20kHz may not be perfectly correct, but
sonically works out very nicely.
Eric.
Originally posted by phase_accurate
I can generally agree BUT if DF gets too low (i.e. output impedance too high) then one ends up with a heavily load-dependant frequency response.
Agreed, but in my experience relatively minor FR variations are much less sonically damaging than NFB/load induced ringing and higher order harmonics production etc.
I agree that very high NFB factors, used as a cure-all, are less than optimal. But my opinion is that the NFB factor should be kept constant up to above the midrange at the cost of high NFB at lower frequencies. This would result in an amp that doesn't show increasing THD with rising input frequency (the latter being the case for many SS amps).
Agreed, keep the NFB sufficiently high through bass and mids to keep damping appropriate and distortion low, and allow drooping NFB at above mids even though distortion is increased.
Lower order THD (produced/allowed by low NFB ratio) is sonically much more tolerable than even small amounts of higher order harmonics.
Some harmonics production above say 4 kHz is not particularly offensive mainly because higher order products are out of audio band and rendered inadible provided that lower order IM products are not generated also.
Lowered NFB can also imply good IM and TIM distortion performance provided that the open loop bandwidth is relatively high.
To my ear IMD and TIMD are FAR more damaging (and fatigueing) than lower orders of THD.
In practice, good damping/distortion performance at low/mid frequencies, and 'free-er' sounding high mids/highs adds up to very nicely acceptable sonic performance with typical speakers and very high performance large studio monitors also.
With ESL, the capacitive loading can cause a slight drooping response towards the highs, but there are no transient nasties produced as can be the case with high NFB amplifiers.
A droop of 1 or 2 dB at 20kHz may not be perfectly correct, but
sonically works out very nicely.
Eric.
You can start here-
http://users.ece.gatech.edu/~mleach/lowtim/ .
Pay close attention to power, layout and grounding.
Eric.
http://users.ece.gatech.edu/~mleach/lowtim/ .
Pay close attention to power, layout and grounding.
Eric.
Hello All,
I have a question about simulating the damping factor in spice.
Is the shortest route to do this put a sine wave current source in the output instead of the load, and place a 1k resitor on the input of the amplifier (no drive signal)?
Then choose 1A for the sine wave, and the measured voltage on the output is the output resistance?
Best regards:
Nexus.
I have a question about simulating the damping factor in spice.
Is the shortest route to do this put a sine wave current source in the output instead of the load, and place a 1k resitor on the input of the amplifier (no drive signal)?
Then choose 1A for the sine wave, and the measured voltage on the output is the output resistance?
Best regards:
Nexus.
The issue I see with 'current drive' to speakers is that it's sonic effects may be effectively similar (not necessarily the same) as low damping factor. Plus, it may not give optimal results with speakers that were designed for use with low source impedance (high damping factor) amplifiers.
Another effect I think is of some importance wrt damping factor is that if the amp drives a multiway speaker with passive crossovers, a lower damping factor amp may allow more back emf from each driver to make it to the other (but so would higher impedance speaker cables unless biwired).
Probably damping factors over 200 at the speaker terminals become less important in the real world given practical speaker gauge wires and run lengths which will add series resistance, etc. to this.
I agree with one poster about the detrimental sonic effects of achieving high but nonconstant with frequency damping factors with large amounts of negative voltage feedback. However, my DC coupled OTL design achieves its high damping factor with only moderate negative voltage feedback (26db above 1hz) but uses positive current feedback to achieve damping factors over 500 to 10 khz (when trimmed out). One thing about positive current feedback is that when summed in the same loop as I do with the negative voltage feedback, it actually will reduce, if anything the total inverting feedback signal when it is boosting damping factor into any real world load.
Another effect I think is of some importance wrt damping factor is that if the amp drives a multiway speaker with passive crossovers, a lower damping factor amp may allow more back emf from each driver to make it to the other (but so would higher impedance speaker cables unless biwired).
Probably damping factors over 200 at the speaker terminals become less important in the real world given practical speaker gauge wires and run lengths which will add series resistance, etc. to this.
I agree with one poster about the detrimental sonic effects of achieving high but nonconstant with frequency damping factors with large amounts of negative voltage feedback. However, my DC coupled OTL design achieves its high damping factor with only moderate negative voltage feedback (26db above 1hz) but uses positive current feedback to achieve damping factors over 500 to 10 khz (when trimmed out). One thing about positive current feedback is that when summed in the same loop as I do with the negative voltage feedback, it actually will reduce, if anything the total inverting feedback signal when it is boosting damping factor into any real world load.
Just noticed someone earlier asked how to find out the slew rate, I would like to suggest to use square wave at the input at a quite low frequency of 100 Hz for instance(, less risk to burn the amplifier).
Should be noticed that the input shunt capacitor forming a lowpass filter must first be removed before the test (at least one of the legg since it doesnt have anything to do with the amplifier bandwith and decrease the slewrate at the input before the signal reach the first amplifier stage.
Connect an oscilloscope at the output and notice the speed of the rising or falling flanks.
Slewrate is often given in V/uS, a fair amplifier has at least 20V/uS I think.
I dont want to specify any type of standard how to check the slewrate but it could be good to check for both small signals let say 1 volt output and full swing.
When talking about the feedback and distortion the human ear is much more sensitive to TIMD than THD, even <0,1% THD for all frequencies and power levels is very good.
Should be noticed that the input shunt capacitor forming a lowpass filter must first be removed before the test (at least one of the legg
since it doesnt have anything to do with the amplifier bandwith and decrease the slewrate at the input before the signal reach the first amplifier stage.Connect an oscilloscope at the output and notice the speed of the rising or falling flanks.
Slewrate is often given in V/uS, a fair amplifier has at least 20V/uS I think.
I dont want to specify any type of standard how to check the slewrate but it could be good to check for both small signals let say 1 volt output and full swing.
When talking about the feedback and distortion the human ear is much more sensitive to TIMD than THD, even <0,1% THD for all frequencies and power levels is very good.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.