I have noticed today an interesting effect when I use a CRC power supply in the X Aleph (15 volt version with 4 IRFP044) by add 0.22 ohms resisters ( 1 volt across resisters) between the two 30,000 UF Filter capacitors on each rail(30000+ 30000 per rail).
There appears to be a noticable reduction on high frequency bandwidth then using the CRC filtering with a more subude sound stage.
Can any one explain why this is happening or had a similar experience?
Ian
There appears to be a noticable reduction on high frequency bandwidth then using the CRC filtering with a more subude sound stage.
Can any one explain why this is happening or had a similar experience?
Ian
Hello Ian,
Your question is an interesting one and it hasn't received much response so I'll throw in my two cents.
I think that the added low pass section on your power supply is becoming a significant part of your output circuit at higher frequencies. I apologize for stating the obvious but here are my thoughts.
The Aleph X appears to have it's dominant pole at a fairly low frequency. I recall seeing posts where people (yourself included) found that the circuit rolls off at 20 - 50KHz. The output stage of this amplifier also seems to have a higher output impedance than the typical follower stage.
O.K. - now consider the popular view of a power amplifier circuit as one that simply modulates a large DC current source. If you add a low pass RC filter section to that current source then at higher frequencies this may well interact with the amp. Especially if you have nonideal impedances. For instance if the output impedance of the output stage is higher than usual, or if the power transformer has a higher impedance than usual on the secondary, or if the PS filter caps have a high internal impedance. Is the rolloff an additional 6db?
I've always felt uneasy about having these CRCRC type filters on an Aleph-X power supply. Nelson Pass makes it work by using a huge transformer with many times the required current output. He also uses 75 volt caps for a 32 volt supply. This helps reduce the PS impedance as well. If you use a smaller transformer or less than ideal caps then I think you'll see what you're seeing.
I'd try substituting a bigger transformer - which is probably not what you want to hear. I'd also try fiddling with the R's and C's in the power supply to see what happens and if you can control the interaction.
As long as you're on the line there's a question I'd like to ask you. In the Aleph-X schematic you posted way back when, you had used capacitively coupled inputs with 100K to ground on the input pins and then 2.2uf in series with the input resistors. You had also posted questions to the forum and NP about this circuit. What did you eventually end up using?
Hoping this helps.
Graeme
Your question is an interesting one and it hasn't received much response so I'll throw in my two cents.
I think that the added low pass section on your power supply is becoming a significant part of your output circuit at higher frequencies. I apologize for stating the obvious but here are my thoughts.
The Aleph X appears to have it's dominant pole at a fairly low frequency. I recall seeing posts where people (yourself included) found that the circuit rolls off at 20 - 50KHz. The output stage of this amplifier also seems to have a higher output impedance than the typical follower stage.
O.K. - now consider the popular view of a power amplifier circuit as one that simply modulates a large DC current source. If you add a low pass RC filter section to that current source then at higher frequencies this may well interact with the amp. Especially if you have nonideal impedances. For instance if the output impedance of the output stage is higher than usual, or if the power transformer has a higher impedance than usual on the secondary, or if the PS filter caps have a high internal impedance. Is the rolloff an additional 6db?
I've always felt uneasy about having these CRCRC type filters on an Aleph-X power supply. Nelson Pass makes it work by using a huge transformer with many times the required current output. He also uses 75 volt caps for a 32 volt supply. This helps reduce the PS impedance as well. If you use a smaller transformer or less than ideal caps then I think you'll see what you're seeing.
I'd try substituting a bigger transformer - which is probably not what you want to hear. I'd also try fiddling with the R's and C's in the power supply to see what happens and if you can control the interaction.
As long as you're on the line there's a question I'd like to ask you. In the Aleph-X schematic you posted way back when, you had used capacitively coupled inputs with 100K to ground on the input pins and then 2.2uf in series with the input resistors. You had also posted questions to the forum and NP about this circuit. What did you eventually end up using?
Hoping this helps.
Graeme
Thanks Graeme for your thoughtful response.
At the moment my X Aleph is directly connected to the output of my Bride of Zen Blanced Line Stage with the nominal 10uf coupling capacitors inside the feedback loop(with Hendrik X modification).
Your theory is an interesting one.
Previously I was running 2 pairs of IRFP240 with each with 0.44 ohms source resisters per each side of the X Aleph with a straight 60,000+60,000 filtered supply....No problems.
I then got hold of some IRFP044, changed out the IFRP240's and installed the IRFP044 using 1 pair per each side with 0.22 ohms source resistance.
At this point figured I would improve the supply by spliting the 60,000 into 30,000 +30,000 CRC using 0.22 ohms.
This is when I noticed the rolloff of the highs.
Initially I thought it was the extra gate capacitance of the IRFP044 but I have since bypassed the 0.22 ohms resisters with a shortingl ink and the bandwidth is restored.
Your theory could be as there isa significant voltage drop across the 0.22 ohms of about 1.1 volt an at about 5.2 amps so maybe I am asking too much of the filter capacitors.
I might back of the bias with is a bit too high or use say 0.110 ohms for the CRC and see what happens. Bypassing the local driver PCB may also help if in fact there are hgher than optimum impedanced in the power supply.
Thanks again for your response
Ian
At the moment my X Aleph is directly connected to the output of my Bride of Zen Blanced Line Stage with the nominal 10uf coupling capacitors inside the feedback loop(with Hendrik X modification).
Your theory is an interesting one.
Previously I was running 2 pairs of IRFP240 with each with 0.44 ohms source resisters per each side of the X Aleph with a straight 60,000+60,000 filtered supply....No problems.
I then got hold of some IRFP044, changed out the IFRP240's and installed the IRFP044 using 1 pair per each side with 0.22 ohms source resistance.
At this point figured I would improve the supply by spliting the 60,000 into 30,000 +30,000 CRC using 0.22 ohms.
This is when I noticed the rolloff of the highs.
Initially I thought it was the extra gate capacitance of the IRFP044 but I have since bypassed the 0.22 ohms resisters with a shortingl ink and the bandwidth is restored.
Your theory could be as there isa significant voltage drop across the 0.22 ohms of about 1.1 volt an at about 5.2 amps so maybe I am asking too much of the filter capacitors.
I might back of the bias with is a bit too high or use say 0.110 ohms for the CRC and see what happens. Bypassing the local driver PCB may also help if in fact there are hgher than optimum impedanced in the power supply.
Thanks again for your response
Ian
Interesting phenomenom......
Why did you change to a CRC filter in the power supply in the first place? Was it because of hum being audible?
Maybe increasing just the capacity (no CRC) could improve this situation?
What transformer did you use? (Torodial/EI? Regulation (Ri of secondairy winding) would be interesting to know too.
Did you notice any other change in sound between IRFP044 and IRFP240?
Why did you change to a CRC filter in the power supply in the first place? Was it because of hum being audible?
Maybe increasing just the capacity (no CRC) could improve this situation?
What transformer did you use? (Torodial/EI? Regulation (Ri of secondairy winding) would be interesting to know too.
Did you notice any other change in sound between IRFP044 and IRFP240?
I tried reverting back to the straight 60,000 uf with the 044's and it seemed okay.
But whats even more odd is that my other bigger X Aleph on 25+25 volt rails has CLC wit 2 mh 0.070 ohm chokes with the same capacitors 30,000+30,000.....no signs of the highs fading there. But it uses IRFP240's so who knows....another diy mystery.
I will wait a few days and do some more tests when my winter buggs gone........Nelson I miss your Northern California climate...not a hint of a sniiffle when I was over there!
If I could swing it I'd be over there again now.
Ian
Picture Foresthill CA
But whats even more odd is that my other bigger X Aleph on 25+25 volt rails has CLC wit 2 mh 0.070 ohm chokes with the same capacitors 30,000+30,000.....no signs of the highs fading there. But it uses IRFP240's so who knows....another diy mystery.
I will wait a few days and do some more tests when my winter buggs gone........Nelson I miss your Northern California climate...not a hint of a sniiffle when I was over there!
If I could swing it I'd be over there again now.
Ian
Picture Foresthill CA
Attachments
Well I had a good play around with it this morning.
Using the IRFP044's I decided to add an additional 15,000 to the amp end of the CRC*. I think that helped with the stereo imaging a bit. Shorting the 0.22R resister did not change much with the extra filter capacitors but I still felt the presentation was overly rich compared to what I have been accustomed to.
I then changed over one channel to the IRFP240's with the same power supply conditions.(Grey's orginal schematic)
Nelson comments about gain devices effecting the sound were borne out here. In comparisions and I am still assessing this, the IRFP240 appears to have a somewhat thinner presentation and more billiant high frequency response. The IRFP044 is richer sounding with more body in the midrange but otherwise less linear .....The jury is still out but I tend to think that while using only one pair per side of each, neither is better than the other and the IRF240 perhaps offers more lattitude to use in various numbers for the right tonal balance. I doubt if I would use the IFRP044 for a tweeter amp in a bi amp setup and more likely the midrange.
Mean while back at the ranch I think the IRFP044's were masking my subjective impressions overall. I plan to investigate CLC filtering to see what happens there as I will be using the amps for a horn loaded mid and tweeter in my 4345 monitors (JBL 2308 /2307 / 2425 & 2405).
I just tried the extra capacitors again, they certainly help and looking at Nelson's photographs of the AX 200 there appears to be an extra pair of 25,000 75v capacitors and the amp end of the supply.
Thankyou all for your comments.
Ian
Using the IRFP044's I decided to add an additional 15,000 to the amp end of the CRC*. I think that helped with the stereo imaging a bit. Shorting the 0.22R resister did not change much with the extra filter capacitors but I still felt the presentation was overly rich compared to what I have been accustomed to.
I then changed over one channel to the IRFP240's with the same power supply conditions.(Grey's orginal schematic)
Nelson comments about gain devices effecting the sound were borne out here. In comparisions and I am still assessing this, the IRFP240 appears to have a somewhat thinner presentation and more billiant high frequency response. The IRFP044 is richer sounding with more body in the midrange but otherwise less linear .....The jury is still out but I tend to think that while using only one pair per side of each, neither is better than the other and the IRF240 perhaps offers more lattitude to use in various numbers for the right tonal balance. I doubt if I would use the IFRP044 for a tweeter amp in a bi amp setup and more likely the midrange.
Mean while back at the ranch I think the IRFP044's were masking my subjective impressions overall. I plan to investigate CLC filtering to see what happens there as I will be using the amps for a horn loaded mid and tweeter in my 4345 monitors (JBL 2308 /2307 / 2425 & 2405).
I just tried the extra capacitors again, they certainly help and looking at Nelson's photographs of the AX 200 there appears to be an extra pair of 25,000 75v capacitors and the amp end of the supply.
Thankyou all for your comments.
Ian
Transistor Type
Ian,
May I suggest another (not costly at all) experiment for you :
Try to replace the IRFP's with a lateral FET (e.g. 2SK1529), and change the gate resistors (starting value for 4 FETs per channel 1k). You can tune the sonic balance by just changing the gate resistors values up and down.
Patrick
Ian,
May I suggest another (not costly at all) experiment for you :
Try to replace the IRFP's with a lateral FET (e.g. 2SK1529), and change the gate resistors (starting value for 4 FETs per channel 1k). You can tune the sonic balance by just changing the gate resistors values up and down.
Patrick
I also noticed a similar thing when I added 6mh torodial chokes into my X supply. So much so tnat I shelved the X's for now and added the chokes to my pair of Aleph 2's in which a dramatic improvement in the noise floor was achieved and with out the disappreance of any hf response. It actually also made the A-2's sound quite a bit better overall just at the loss of a couple of volts on the rails. I was using 240's in my X's to begin with. Never have used the 044's.
I thought it odd that the hf disappeared literally overnight from my system and now I may have to drag the X's back out and go through the whole process again and listen with and without, and this time take measurements and look at a square wave or two. The strange thing was that the hf rolloff was so noticable that I thought the tweeters on my Dyns had failed for some reason. Puttong the 2's back in the system confirmed they were ok.
Mark
I thought it odd that the hf disappeared literally overnight from my system and now I may have to drag the X's back out and go through the whole process again and listen with and without, and this time take measurements and look at a square wave or two. The strange thing was that the hf rolloff was so noticable that I thought the tweeters on my Dyns had failed for some reason. Puttong the 2's back in the system confirmed they were ok.
Mark
It would not surprise me to find that the difference between the '044 and the '240 would be audible. I went with the '044 in the original version because I intended to bias the absolute pee-widdle-turkey out of the thing and wanted to have plenty of elbow room as far as current. By all means, toss in '240s (or '140s) if it makes you happy.
Changing to lateral MOSFETs isn't so easy as that. The voltage offset is going to be radically different. You'll have to rethink the entire thing to make it work properly. Not that it can't be done; just don't assume that it's a drop-in replacement job.
Grey
Changing to lateral MOSFETs isn't so easy as that. The voltage offset is going to be radically different. You'll have to rethink the entire thing to make it work properly. Not that it can't be done; just don't assume that it's a drop-in replacement job.
Grey
HI
I suggest, that it is the voltage drop whith the CRC ore CLC. It is not much ,but in your Aleph X amps you are very low in voltage anyway. ( lower than the sugested voltage for these fets) That would explain why you dont see the effect in the Aleph 2 s you tested, same for the bigger AX.
Bernhard
I suggest, that it is the voltage drop whith the CRC ore CLC. It is not much ,but in your Aleph X amps you are very low in voltage anyway. ( lower than the sugested voltage for these fets) That would explain why you dont see the effect in the Aleph 2 s you tested, same for the bigger AX.
Bernhard
Mark,
It would be great if you could run some measurements.
I have been thinking about this too, and my thinker is broken.
It maybe the characteristics of the pi filter...I digress that according to text, CLC is better for high current than CRC.With the X Aleph the current is basically doubled so any issues with the supply output impedance may be a mitigating issue
The fact that I was loosing 2.5 volts of total supply voltage would also tend to worsen the issue of gate capacitance, particularly in the case of the 044's. On that hypothesis I plan to double the supply voltage and see if I get the same effect with the 044's.
Ian
It would be great if you could run some measurements.
I have been thinking about this too, and my thinker is broken.
It maybe the characteristics of the pi filter...I digress that according to text, CLC is better for high current than CRC.With the X Aleph the current is basically doubled so any issues with the supply output impedance may be a mitigating issue
The fact that I was loosing 2.5 volts of total supply voltage would also tend to worsen the issue of gate capacitance, particularly in the case of the 044's. On that hypothesis I plan to double the supply voltage and see if I get the same effect with the 044's.
Ian
XA160 PSU with CLCLC ?
In a German review of the XA160, a couple of pictures are shown of the interior of the amplifier, one of which indicates (as in the commentary) that the power supply is not CRCRC, but with inductors and capacitors.
Unless they got it wrong and the inductors are for somethings else. They do look a bit too small though for the current. My guess would be around 3mH, 0.3 ohm, so called I-core.
Thought Ian might be interested to know. Nothing new or unknown in the text otherwise.
http://www.stereoplay.de/sixcms/media.php/188/stp0704PassXA160.pdf
Patrick
In a German review of the XA160, a couple of pictures are shown of the interior of the amplifier, one of which indicates (as in the commentary) that the power supply is not CRCRC, but with inductors and capacitors.
Unless they got it wrong and the inductors are for somethings else. They do look a bit too small though for the current. My guess would be around 3mH, 0.3 ohm, so called I-core.
Thought Ian might be interested to know. Nothing new or unknown in the text otherwise.
http://www.stereoplay.de/sixcms/media.php/188/stp0704PassXA160.pdf
Patrick
CLC
Nelson,
I remember when you first showed pictures of the XA in the forum, they were CRCRC using NTC's as R's. And your remark then was that you preferred R as they did not generate magnetic field.
Would you be kind enough to drop us a hint as to what drove you to ue CLC in the XA160, and how CLC compares to CRC (other than the obvious -- namely less voltage losses with L's) ?
PS I guess I was correct in saying that those were cored coils, and not particularly low in resistance. Is there not an issue with heat dissipation ?
Thanks,
Patrick
Nelson,
I remember when you first showed pictures of the XA in the forum, they were CRCRC using NTC's as R's. And your remark then was that you preferred R as they did not generate magnetic field.
Would you be kind enough to drop us a hint as to what drove you to ue CLC in the XA160, and how CLC compares to CRC (other than the obvious -- namely less voltage losses with L's) ?
PS I guess I was correct in saying that those were cored coils, and not particularly low in resistance. Is there not an issue with heat dissipation ?
Thanks,
Patrick
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.