That may be whats making the OP stage to stick. I also wonder if the current available for the OP stage is sufficient for those low impedances you tested into. How does it look into 8 ohms, I don't remember seeing any 8 ohm clip point pics? Remember that the KSA-50 was not famous for working well into low impedences... that need for the Apogees is what brought the KSA-100 into being.
Mark
Mark
MJ15003/4 are not that great complementaries, bandwidth and Cib differ much. I am not familiar with the 21193/4
Power of the KSA50 was around 200 watts continuous in 2 Ohms.
That is around 2.5 times the figure in 8 Ohms, which is a decent value.
And, the original KSA did not show asymmetrical clipping, clipping stepped in gracefully.
If you know the sound of clipping, non-symmetrical clipping is worse.
Maybe Mr Green can measure the various voltages on both sides of the symmetrical line.
Power of the KSA50 was around 200 watts continuous in 2 Ohms.
That is around 2.5 times the figure in 8 Ohms, which is a decent value.
And, the original KSA did not show asymmetrical clipping, clipping stepped in gracefully.
If you know the sound of clipping, non-symmetrical clipping is worse.
Maybe Mr Green can measure the various voltages on both sides of the symmetrical line.
I'd check the drop across the emitter resistors and see if they are within say 10%. I would also check the voltages down both sides of the board as well and see how they check out. It may be possible that you stuffed in a wrong value resistor...... The voltage across both Zeners ahould also be within a reasonable limit too. Your rail drop is insignificant. I'm using just about the same thing with two 680VA toroids but am using a 20 amp variac to set them to 40 volt rails.
I ran mine to clipping into 8 ohms and its not showing that sticking problem and it clips evenly. Can't post photos till next weekend but I will at that time.
Mark
I ran mine to clipping into 8 ohms and its not showing that sticking problem and it clips evenly. Can't post photos till next weekend but I will at that time.
Mark
predicted performance - 4 pair 21193/4
Hi Lgreen,
the SOAR relative to the output capability of your chosen configuration is a lot better than I suspected.
It will drive any speaker from 8ohm down to 1ohm at phase angles upto 60degrees with a case temperature (Tc) at 75degC and still stay within the derated SOAR.
It suddenly struggles as you drop below 1ohm. At 0r6 load the permissible phase angle is only 10degrees but into a resistive load it is predicted to drive over 500 watts at a peak output voltage of 24v7, in excess of 41A. This is continuous sine wave not peak currents.
BTW the 2ohm power is predicted at 3.48 times 8ohm power which is a stupendous result if it happens in practice.
Once you have the little gremlins sorted I think your choices will result in unbeatable bass with the smooth mid and top that ClassA is supposed to give. Keep us informed.
To all
the only thing that worries me slightly is the very low sensitivity of this Krell Klone and similarly the Leach clone. I have a feeling they would both perform better if sensitivity for full power were increased to something in the range of 300mV to 400mV.
Any comments?
Hi Lgreen,
the SOAR relative to the output capability of your chosen configuration is a lot better than I suspected.
It will drive any speaker from 8ohm down to 1ohm at phase angles upto 60degrees with a case temperature (Tc) at 75degC and still stay within the derated SOAR.
It suddenly struggles as you drop below 1ohm. At 0r6 load the permissible phase angle is only 10degrees but into a resistive load it is predicted to drive over 500 watts at a peak output voltage of 24v7, in excess of 41A. This is continuous sine wave not peak currents.
BTW the 2ohm power is predicted at 3.48 times 8ohm power which is a stupendous result if it happens in practice.
Once you have the little gremlins sorted I think your choices will result in unbeatable bass with the smooth mid and top that ClassA is supposed to give. Keep us informed.
To all
the only thing that worries me slightly is the very low sensitivity of this Krell Klone and similarly the Leach clone. I have a feeling they would both perform better if sensitivity for full power were increased to something in the range of 300mV to 400mV.
Any comments?
One thng is for certain and thats that the power supply rails need to be bypassed for best audible performance. Have built many amps over the years but this amp is extremely sensitive to this for some reason. Adding a simple .1 film across each rail at the main electrolytics does make a huge improvement in clarity and increases soundstage immensley. With the Pass amps I've built there is little audible difference in bypassing the rails....
Mark
Mark
sensitivity...
As modern, cosmopolitan, PC guys sensitivity is a big issue...
Luckily in this case it really is just a matter of changing the value of one or two resistors.
The gain of the amp is set by r130 divided by r129...
R129 also controls the time constant of the low end roll-off, and to some extent will control the lowest possible noise floor the amp can achieve. 1k would work here and give a 50% increase in gain, while raising r130 to 30k at the same time would essentially double the sensitivity.
If bypassing caps is your thing, the pair of back to back electros (c102/103) are pretty obvious candidates. They directly modify the gain of the amp, if thay have any 'oddities', it will be pretty obvious. IMHO this would be the one(s) to use your best quality caps on...
HTH
Stuart
As modern, cosmopolitan, PC guys sensitivity is a big issue...
Luckily in this case it really is just a matter of changing the value of one or two resistors.
The gain of the amp is set by r130 divided by r129...
R129 also controls the time constant of the low end roll-off, and to some extent will control the lowest possible noise floor the amp can achieve. 1k would work here and give a 50% increase in gain, while raising r130 to 30k at the same time would essentially double the sensitivity.
If bypassing caps is your thing, the pair of back to back electros (c102/103) are pretty obvious candidates. They directly modify the gain of the amp, if thay have any 'oddities', it will be pretty obvious. IMHO this would be the one(s) to use your best quality caps on...
HTH
Stuart
ok......
Stuart! Good to have you back for a few posts. No bypassing so far but its easy to install so I'll do it and get back to everyone. What values do you recommend with 64K uF per rail filtering?j Also I used "plain" and unexciting caps for C102/103, wondering if a change is in order? I have a lot of "sprague orange drop" caps here for usage if appropriate, or I can take recommendations for better caps. Gain might be changed too.
While I'm traveling and unable to really do anything its good to get that info from Andrew. I hoped my config would be bullet proof and its good to hear that it **should be**. Yes, I have got plenty of bass, ran the amp for 2.5 hours a day ago to check temp and bias into the vandersteen 2Ce and no surprises thankfully...tons of bass, this amp loves the low Freq. range. I don't expect the load to drop <1 ohm at the mentioned phase angle. Maybe 4 ohms worst case. I hesitate to measure into very low <2 ohm impedance due to lack of prot. circutiry. But will do if Andrew/Jacco/Stewart thinks its ok. Clipping is instantly recognizable but prior to clipping the sound is crisp and deep, very deep. Note- commercial amps clip at about "80" to the 2Ce's on my preamp, so increasing the gain on this would actually make the amp a little more consistent with what I already have.
Here is a question for everyone- how much testing would you do before running these Krell's into a set of very VERY expensive speakers that you loved and didn't want to hurt? Saw something about 20-30 hours before, seemed reasonable. But my speakers supposedly are a perfect match to this high current delivery amp, and have 92 dB/w sensivitity. So far <35 mV DC offset so I'm not currently worried about blowing speakers up, but I made a lot of mistakes before so don't want to start again
Stuart! Good to have you back for a few posts. No bypassing so far but its easy to install so I'll do it and get back to everyone. What values do you recommend with 64K uF per rail filtering?j Also I used "plain" and unexciting caps for C102/103, wondering if a change is in order? I have a lot of "sprague orange drop" caps here for usage if appropriate, or I can take recommendations for better caps. Gain might be changed too.
While I'm traveling and unable to really do anything its good to get that info from Andrew. I hoped my config would be bullet proof and its good to hear that it **should be**. Yes, I have got plenty of bass, ran the amp for 2.5 hours a day ago to check temp and bias into the vandersteen 2Ce and no surprises thankfully...tons of bass, this amp loves the low Freq. range. I don't expect the load to drop <1 ohm at the mentioned phase angle. Maybe 4 ohms worst case. I hesitate to measure into very low <2 ohm impedance due to lack of prot. circutiry. But will do if Andrew/Jacco/Stewart thinks its ok. Clipping is instantly recognizable but prior to clipping the sound is crisp and deep, very deep. Note- commercial amps clip at about "80" to the 2Ce's on my preamp, so increasing the gain on this would actually make the amp a little more consistent with what I already have.
Here is a question for everyone- how much testing would you do before running these Krell's into a set of very VERY expensive speakers that you loved and didn't want to hurt? Saw something about 20-30 hours before, seemed reasonable. But my speakers supposedly are a perfect match to this high current delivery amp, and have 92 dB/w sensivitity. So far <35 mV DC offset so I'm not currently worried about blowing speakers up, but I made a lot of mistakes before so don't want to start again