Roberto, I think you read too much into my message, I tried to make it clear that I dont believe you are out to steal from us, if I did I wouldnt have sent my money because there was already a delay I knew about before this GB.....if you remember I covered this with you in email.
but I sent my money and to a certain extent encouraged others to do the same because I think its a great product and because I trust you, I still trust you to send my PSUs, that is not and never has been in question. I do not believe you are that type of guy, I just think you have not dealt with the communication well at all, telling things that are so easily seen through; probably because you dont want to disappoint people, or maybe think we will run away and take our money with us.
its just that I made it very clear in here and by email, that whatever the reasonable amount of time you needed for them to be ready, thats what I wanted to hear and if there was a problem, I wanted to hear about that too; doing otherwise only serves to make people restless when it doesnt come to pass.
we waited a long time since the beginning of the project to build the best amp we can, waiting some more is no problem.
its simple really, give yourself enough tome to do each batch and try to give a realistic estimate based on the ones you have already built including any supplies orders, as long as the communication is open, none of us have any problem (within reason, it cant go on forever)
this can certainly be recovered, no permanent damage is done. we still love you
just talk to us
Last edited:
for some of this I agree, some is misunderstanding, some has been misinformation, its the second that has caused troubles. It is very important Roberto that you understand I believe you have a good heart, I have seen that from the beginning and my opinion has not changed.
man haha its a love in, searching for phrases that share universal meaning across boundaries can lead to some sappiness
Hi,
All is very clear, certainly in English does not help me. I have to say (as I mentioned above), here I am treated well. however, there are 10 other SMPS ready right now, so 22 total. Sunday I hope to do the test on the last 5pcs then 27pz.
I have taken time to other applications. this is because I wanted to personally carry out these your SMPS. (just do not want surprises)
whether there will be another GB, it is better to program the first time, and then delivery
regards
All is very clear, certainly in English does not help me. I have to say (as I mentioned above), here I am treated well. however, there are 10 other SMPS ready right now, so 22 total. Sunday I hope to do the test on the last 5pcs then 27pz.
I have taken time to other applications. this is because I wanted to personally carry out these your SMPS. (just do not want surprises)
whether there will be another GB, it is better to program the first time, and then delivery
regards
Hi,
I measured between a pause, the linearity error of a class AB amplifier with power transformer 1kVA+2 x100MuF .at + /-75V idle.
Measure is simple (i use 4 ch analyzer) while the input signal is increased linearly up to the clip, are monitored vcc the rail, and Voutput amplifier. Well, error is 3% when referring to linearity. or 30% when applied to the compression curve. (i mean, 750w to 450w are compressed in a log behavior. This is a huge mistake linearity (3%), I wish someone would tell me, in case of huge NFB, as this repair? in this case produces a sub compression.and... has nothing to do PSRR.
the problem is the relationship between the level of the input signal, compared to the decrease of 16V on rail vcc, then output level product, right where it is connected to the NFB. this changes the gain with incorrect proportion in relation to the input signal....Very well.
I measured between a pause, the linearity error of a class AB amplifier with power transformer 1kVA+2 x100MuF .at + /-75V idle.
Measure is simple (i use 4 ch analyzer) while the input signal is increased linearly up to the clip, are monitored vcc the rail, and Voutput amplifier. Well, error is 3% when referring to linearity. or 30% when applied to the compression curve. (i mean, 750w to 450w are compressed in a log behavior. This is a huge mistake linearity (3%), I wish someone would tell me, in case of huge NFB, as this repair? in this case produces a sub compression.and... has nothing to do PSRR.
the problem is the relationship between the level of the input signal, compared to the decrease of 16V on rail vcc, then output level product, right where it is connected to the NFB. this changes the gain with incorrect proportion in relation to the input signal....Very well.
1kVA+2 x100MuF with + /-75V at idle
AP2,
Thanks for doing this test. In order to answer you questions properly more details are needed.
However, let's assume that when you say there is a 3% error, you mean that the supply rails have changed by 3% from +/- 75Vdc to +/- 72.5Vdc as the amplifier's output signal is increased from zero to clipping. If that is correct, then allowing for about 5V drop accross the output MOSFETS (Vds), this would translate into approximately 48Vrms of output. If the output load is 4 ohms this means that the power output is approximately 575W (P=Vsquared over R), or about 288W if the load is 8 ohms). A 3% change in rail voltage seems reasonable for the transformer and capacitors you specified.
By comparison, an regulated power supply could be expected to hold the rail voltage constant despite the change in output signal from zero (idle) to clipping. So, for +/- 75Vdc regulate rails, allowing 5V for Vds, clipping would occur at about 49.5 Vrms, which is about 612W into 4 ohms or about 306W into 8 ohms.
Therefore an amplifier with regulated +/- 75Vdc rails could be expected to produce 37 more watts of power at clipping into a 4 ohm load than an amplifier with unregulated rails which have decreased to +/-72.5Vdc at clipping. 575W is 6% less than 612W.
You mentioned the role that NFB plays in this. A good amplifier's gain within the audio band is determined by the amount of total end to end NFB. The NFB should remain constant regardless of choice of rail voltage. Therefore for any steady output signal below clipping, the amplitude of the output signal should be determined only by the amplitude of the input signal and the gain of the amp.
Regards,
AP2,
Thanks for doing this test. In order to answer you questions properly more details are needed.
However, let's assume that when you say there is a 3% error, you mean that the supply rails have changed by 3% from +/- 75Vdc to +/- 72.5Vdc as the amplifier's output signal is increased from zero to clipping. If that is correct, then allowing for about 5V drop accross the output MOSFETS (Vds), this would translate into approximately 48Vrms of output. If the output load is 4 ohms this means that the power output is approximately 575W (P=Vsquared over R), or about 288W if the load is 8 ohms). A 3% change in rail voltage seems reasonable for the transformer and capacitors you specified.
By comparison, an regulated power supply could be expected to hold the rail voltage constant despite the change in output signal from zero (idle) to clipping. So, for +/- 75Vdc regulate rails, allowing 5V for Vds, clipping would occur at about 49.5 Vrms, which is about 612W into 4 ohms or about 306W into 8 ohms.
Therefore an amplifier with regulated +/- 75Vdc rails could be expected to produce 37 more watts of power at clipping into a 4 ohm load than an amplifier with unregulated rails which have decreased to +/-72.5Vdc at clipping. 575W is 6% less than 612W.
You mentioned the role that NFB plays in this. A good amplifier's gain within the audio band is determined by the amount of total end to end NFB. The NFB should remain constant regardless of choice of rail voltage. Therefore for any steady output signal below clipping, the amplitude of the output signal should be determined only by the amplitude of the input signal and the gain of the amp.
Regards
,AP2,
Thanks for doing this test. In order to answer you questions properly more details are needed.
However, let's assume that when you say there is a 3% error, you mean that the supply rails have changed by 3% from +/- 75Vdc to +/- 72.5Vdc as the amplifier's output signal is increased from zero to clipping. If that is correct, then allowing for about 5V drop accross the output MOSFETS (Vds), this would translate into approximately 48Vrms of output. If the output load is 4 ohms this means that the power output is approximately 575W (P=Vsquared over R), or about 288W if the load is 8 ohms). A 3% change in rail voltage seems reasonable for the transformer and capacitors you specified.
By comparison, an regulated power supply could be expected to hold the rail voltage constant despite the change in output signal from zero (idle) to clipping. So, for +/- 75Vdc regulate rails, allowing 5V for Vds, clipping would occur at about 49.5 Vrms, which is about 612W into 4 ohms or about 306W into 8 ohms.
Therefore an amplifier with regulated +/- 75Vdc rails could be expected to produce 37 more watts of power at clipping into a 4 ohm load than an amplifier with unregulated rails which have decreased to +/-72.5Vdc at clipping. 575W is 6% less than 612W.
You mentioned the role that NFB plays in this. A good amplifier's gain within the audio band is determined by the amount of total end to end NFB. The NFB should remain constant regardless of choice of rail voltage. Therefore for any steady output signal below clipping, the amplitude of the output signal should be determined only by the amplitude of the input signal and the gain of the amp.
Regards
That is 0.26 dB of compression at full power for the unregulated supply compared with the regulated supply; I don't think that will be noticed by normal listeners, keeping in mind that dynamic power compression of moving coil loudspeakers will be magnitudes higher
Last edited:
That is 0.26 dB of compression at full power for the unregulated supply compared with the regulated supply; I don't think that will be noticed by normal listeners
Without doubt!
That is 0.26 dB of compression at full power for the unregulated supply compared with the regulated supply; I don't think that will be noticed by normal listeners, keeping in mind that dynamic power compression of moving coil loudspeakers will be magnitudes higher
Upon second reading of your post I am unsure what you mean by "compression". I was assuming that you meant the unregulated supply cannot support quite as high a signal level at clipping. I totally agree with that.
If my assumption is incorrect, maybe you could elaborate.
Thanks,
sorry for small wrong in digit power.
rest is correct.
Today later i show diagram for better explain,incluse nfb error.
3% is not related to the drop in voltage. (75V drop to 60V) , then clips with 60V (Vcc rail) in this case, for-5V Vds, I get 55Vp.output(4R) @ 1 Khz. I agree, that this problem seems not to be heard. But we must consider that in all these years, we have always listened to, amplifiers, voltage drop, so the concept goes inversely place. "sounds as if the voltage should fall only 1V? . I know this and i'm sure you not imagine effect.
NFB error is: suppose a simple percentage of output signal return at input and then fix a gain. this as ratio, so while input level increment (decrement vcc then ratio change)
this is very marked in class D, where the vP output is the relationship between the percentage of pwm, referred to the rail vcc.
rest is correct.
Today later i show diagram for better explain,incluse nfb error.
3% is not related to the drop in voltage. (75V drop to 60V) , then clips with 60V (Vcc rail) in this case, for-5V Vds, I get 55Vp.output(4R) @ 1 Khz. I agree, that this problem seems not to be heard. But we must consider that in all these years, we have always listened to, amplifiers, voltage drop, so the concept goes inversely place. "sounds as if the voltage should fall only 1V? . I know this and i'm sure you not imagine effect.
NFB error is: suppose a simple percentage of output signal return at input and then fix a gain. this as ratio, so while input level increment (decrement vcc then ratio change)
this is very marked in class D, where the vP output is the relationship between the percentage of pwm, referred to the rail vcc.
Last edited:
Is this a static measurement?
When so, is it possible that under transient load conditions the sag can actually be higher?
I am quoting typical figures for the Voltage measured across the PSU when the amplifier is idling (quiescent state) compared to delivering full power to the specified load.
That cannot be a static comparison measurement.
As for transients. One can pass a transient signal to the load such that the same peak instantaneous power is delivered to the same specified load and in this situation the measured Vdrop will be less, sometimes much less, than for continuous full power.
Although I don't have the equipment to check, I think that using transient signals to deliver more peak instantaneous power into the same specified load will also result in showing a lower Vdrop compared to the continuous full power Vdrop.
This is the basis for measuring and specifying those fictitious power delivery figures, so loved by the "cheap amplifier" manufacturers and retailers. Do they call it IHF power?
That cannot be a static comparison measurement.
As for transients. One can pass a transient signal to the load such that the same peak instantaneous power is delivered to the same specified load and in this situation the measured Vdrop will be less, sometimes much less, than for continuous full power.
Although I don't have the equipment to check, I think that using transient signals to deliver more peak instantaneous power into the same specified load will also result in showing a lower Vdrop compared to the continuous full power Vdrop.
This is the basis for measuring and specifying those fictitious power delivery figures, so loved by the "cheap amplifier" manufacturers and retailers. Do they call it IHF power?
Last edited: