Tom,
Very interesting. A couple of questions.
Can you describe your speakers? Since every system is different, I would like to get a feel for the speakers, even if one can never really know without listening. I do recall that you feel they are very neutral.
What brand, material, type of capacitors did you use for C10 and C34 in this experiment. In the BOM, the are both Cornell Mica surface mount capacitors.
I don't think I would do it right away, many projects in line ahead of it, but I might give this a try someday.
Jac
Very interesting. A couple of questions.
Can you describe your speakers? Since every system is different, I would like to get a feel for the speakers, even if one can never really know without listening. I do recall that you feel they are very neutral.
What brand, material, type of capacitors did you use for C10 and C34 in this experiment. In the BOM, the are both Cornell Mica surface mount capacitors.
I don't think I would do it right away, many projects in line ahead of it, but I might give this a try someday.
Jac
Very interesting and useful Tom - thanks for your efforts. Did any of those combinations dramatically effect the general stage - depth or width? Since you didn't mention it I'm assuming that element was relatively consistent.
Jac, good questions. I used BOM prescribed parts of different values: CDE silver mica SMT for C34 and TH for C10.
Each speaker features a Scanspeak HDS tweeter, which is perhaps the lowest distortion tweeter in its price class under $100, and two Peerless HDS 4" midrange drivers. Mids/tweeter passive xover designed by Madisound, executed with high quality parts including Mundorf SIO caps in tweeter circuit. I would say very accurate highs with decent mids, but I know there are better.
Bob, my comments about depth refer to soundstage in general. I know your speakers are good at projecting that, and I'm very attentive of it, as well. It was not consistent from one comp configuration to the next. I find that, in general, good depth of image correlates with good width of image, so I didn't mention the two separately. Of course, with two different monoblock amps the stereo soundstage (width) will be distorted or destroyed, but I can tell from mono recording through one speaker whether or not the image has depth. Right now, with one amp 33/18 and the other 33/15, I am already hearing pretty good depth and width outside of and between the speakers. I believe it will only get better when I finally change the other amp to 33/15, and I can hardly wait to be done with this stage of testing and just listen to music.
A couple more general observations. It seems that C10 variation influences midrange presence while C34 has more to do with brightness and detail, and the two combined have a pronounced effect on coherence, or overall smoothness of sound and musicality. Perhaps that has to do with distortion or phase. I think that, as the value of these two components gets closer to each other, they sound worse. I'm sure the two types of compensation work in conjunction, not individually. Interesting that my final preference differs from the Rev C by plus 50% for both component values, but that is not what I was aiming for.
Peace,
Tom E
Each speaker features a Scanspeak HDS tweeter, which is perhaps the lowest distortion tweeter in its price class under $100, and two Peerless HDS 4" midrange drivers. Mids/tweeter passive xover designed by Madisound, executed with high quality parts including Mundorf SIO caps in tweeter circuit. I would say very accurate highs with decent mids, but I know there are better.
Bob, my comments about depth refer to soundstage in general. I know your speakers are good at projecting that, and I'm very attentive of it, as well. It was not consistent from one comp configuration to the next. I find that, in general, good depth of image correlates with good width of image, so I didn't mention the two separately. Of course, with two different monoblock amps the stereo soundstage (width) will be distorted or destroyed, but I can tell from mono recording through one speaker whether or not the image has depth. Right now, with one amp 33/18 and the other 33/15, I am already hearing pretty good depth and width outside of and between the speakers. I believe it will only get better when I finally change the other amp to 33/15, and I can hardly wait to be done with this stage of testing and just listen to music.
A couple more general observations. It seems that C10 variation influences midrange presence while C34 has more to do with brightness and detail, and the two combined have a pronounced effect on coherence, or overall smoothness of sound and musicality. Perhaps that has to do with distortion or phase. I think that, as the value of these two components gets closer to each other, they sound worse. I'm sure the two types of compensation work in conjunction, not individually. Interesting that my final preference differs from the Rev C by plus 50% for both component values, but that is not what I was aiming for.
Peace,
Tom E
Tom, great info and thanks for your painstaking tests. I'll give it a try - I already have 27/15 on one build which I'll modify accordingly in the next mod cycle - I have 33/15 in C0G only, which should be fine for now.
C4 update: I tried auditioning FKP3 vs. MKP1837, but they were on different boards with different compensations (FE & C), so it was Apples vs. Oranges - nothing conclusive. They're both very good, but I have a slight subjective preference for MKP1837 in a pure Rev C (22/10) - it has a certain sparkling musicality and definition without being too bright.
Next, I'm altering a TPA V1.2 board with FE-type 27/22 compensation and MKP1837 at C4, so I should be able to do a more realistic comparison against a V1.4 board with identical 27/22 compensation and FKP3 at C4.
Siva, what resistor do you have at R10? That has a lot to do with how the compensation is perceived. A part that tends to smooth out highs will not yield as much variation of comp sound. A transparent part will be much more revealing of comp changes. Pick your poison. I prefer the open sound of a neutral resistor and the proper comp rather than the smoothing effects of a carbon film resistor and brighter comp.
Dario, I did not try Rev C comp. Have you? I ran out of patience, parts, and time. I'm really tired of modifying amps for now. It seemed that, as C10 and C34 got smaller, sound became duller. I figured going all the way down to Rev C comp would not improve things. Perhaps someone else can try that and let us know. I thought my range of test values and combinations was pretty comprehensive.
One more comment about tests: a little more listening to the 33/18 combination last night revealed extra sparkle, some hi-fi type high frequency emphasis that some may find enjoyable. Personally, I don't care for that zippy sound, but the 33/18 arrangement may be favorable to some ears/systems. Anything larger than 18pF seemed excessively bright in my system.
By the way, don't go to Mouser for 15pF SMT mica caps -- there will be no stock on hand until end of April! I might try buying smaller values and stacking them in parallel, as has been discussed here but never tried. In fact, that might be a handy way to test ascending values. Start with, say, 10pF on the board, then stack additional caps on top of that in increments of 5pF. You could test lots of values with never having to desolder a single cap! Wish I had thought of it sooner. Desoldering these little SMT caps destroys them.
Peace,
Tom E
Dario, I did not try Rev C comp. Have you? I ran out of patience, parts, and time. I'm really tired of modifying amps for now. It seemed that, as C10 and C34 got smaller, sound became duller. I figured going all the way down to Rev C comp would not improve things. Perhaps someone else can try that and let us know. I thought my range of test values and combinations was pretty comprehensive.
One more comment about tests: a little more listening to the 33/18 combination last night revealed extra sparkle, some hi-fi type high frequency emphasis that some may find enjoyable. Personally, I don't care for that zippy sound, but the 33/18 arrangement may be favorable to some ears/systems. Anything larger than 18pF seemed excessively bright in my system.
By the way, don't go to Mouser for 15pF SMT mica caps -- there will be no stock on hand until end of April! I might try buying smaller values and stacking them in parallel, as has been discussed here but never tried. In fact, that might be a handy way to test ascending values. Start with, say, 10pF on the board, then stack additional caps on top of that in increments of 5pF. You could test lots of values with never having to desolder a single cap! Wish I had thought of it sooner. Desoldering these little SMT caps destroys them.
Peace,
Tom E
Proper comp?
A questionable assertion... as I've already wrote the initial alternate compensation was really inviting, sweet, warmer... but it had a way higher distortion and the amp was performing really bad, as Mauro pointed out...
You're simpling masking things you don't like using compensation instead of other parts... a way worse solution since you're impacting the amp performance, on the contrary a sweeter R10 don't.
Sure I did, several times, back and forth, during beta phase with beta boards.
FE compensation sounded consistently better, tighter, sweeter, more natural and with wider/deeper soundstage...
I can understand it but (insisting to) avoiding to try the Rev C compensation you can't confirm or rule out the compensation as the culprit, you've simply equalized the sound to your need giving no real useful info on a possible problem with compensation.
I really appreciate your earing abilities but I can't agree with your method and can't understand why you did refuse to answer to my questions on actual BOM you used or to try the only real checkpoint, the Rev C comp.
Last edited:
T2.5A in a 220Vac country is equivalent to 550W.
That should let a 625VA transformer run full blast using any music programme into Power Amps.
T1.6A is 350W and even that is a bit big for a stereo chipamp that has a maximum total power output of around 100W.
That should let a 625VA transformer run full blast using any music programme into Power Amps.
T1.6A is 350W and even that is a bit big for a stereo chipamp that has a maximum total power output of around 100W.
Hi Andrew,
the 2.5A rating comes from Evolution's mounting guide, I've simply reported it.
I'm using T2.5 fuses with 300VA transformers (245V mains)
If I use a T2 fuse the inrush current at switch on causes them to occasionally blow.
If I use a T2 fuse the inrush current at switch on causes them to occasionally blow.
CL-60 will solve inrush. Andrew is right, of course. No sense in letting a short suck as much juice as most transformers used for this amp will draw.
Dario, I don't want to argue about my findings. If you and other builders enjoy listening to alt comp through a syrupy resistor that sweetens everything, that's fine with me. I used a more transparent part to build one of my amps, and I heard unpleasant sound as a result. Initially, that came as a great surprise. After careful consideration, I began to suspect the alt comp was changing the sound. I cannot prove that, but it seems logical.
I'm sorry I forgot to answer your question about my BOM. I followed your premium BOM in nearly every respect, with premium parts in the signal path. I did not use Mundorf M-Lytic caps for PS, and I used Vishay MF resistors instead of Caddocks at R104/204. I seriously doubt those differences will make a drastic change to the sound, so I don't think my BOM is an issue. The only significant variation was R10, and that makes a huge difference. The Riken sounds fine until you try a more transparent part and realize how much detail you're missing.
I thought my testing method was pretty thorough. I don't know why you object to it. I have the feeling that you would not object to it as strongly if my findings agreed with yours. The comp values I prefer are closer to Mauro's Rev C values, so I don't understand your comments about Mauro's objections and how they might apply to my choices. I stopped short of going to the Rev C values because I didn't sense any additional improvement as I approached them. If someone else wants to do that experiment, I will respect their findings. I might try it myself at a later date.
For now, I'm satisfied that I have used trial and error within reasonable guidelines to find a pleasing sound rather than simply accept what someone else, even you, likes. I wish others were as willing to experiment rather than just accept your preferences without questioning them. That is how you and I reached this point: by experimenting for ourselves. After all, the name of this site is DIY Audio, which means Do It Yourself, not Dario Instructs You.
I respect and admire all the work you've done and the help you've given to so many builders to make a better amplifier. I only want to encourage others to add to your effort, not detract from it. Please do not see my conclusions as a discredit to all your hard work, but rather as an invitation to other builders to try things for themselves.
I am not saying you're wrong and I'm right. I'm saying there are alternatives to your alternatives.
Peace,
Tom E
I'm sorry I forgot to answer your question about my BOM. I followed your premium BOM in nearly every respect, with premium parts in the signal path. I did not use Mundorf M-Lytic caps for PS, and I used Vishay MF resistors instead of Caddocks at R104/204. I seriously doubt those differences will make a drastic change to the sound, so I don't think my BOM is an issue. The only significant variation was R10, and that makes a huge difference. The Riken sounds fine until you try a more transparent part and realize how much detail you're missing.
I thought my testing method was pretty thorough. I don't know why you object to it. I have the feeling that you would not object to it as strongly if my findings agreed with yours. The comp values I prefer are closer to Mauro's Rev C values, so I don't understand your comments about Mauro's objections and how they might apply to my choices. I stopped short of going to the Rev C values because I didn't sense any additional improvement as I approached them. If someone else wants to do that experiment, I will respect their findings. I might try it myself at a later date.
For now, I'm satisfied that I have used trial and error within reasonable guidelines to find a pleasing sound rather than simply accept what someone else, even you, likes. I wish others were as willing to experiment rather than just accept your preferences without questioning them. That is how you and I reached this point: by experimenting for ourselves. After all, the name of this site is DIY Audio, which means Do It Yourself, not Dario Instructs You.
I respect and admire all the work you've done and the help you've given to so many builders to make a better amplifier. I only want to encourage others to add to your effort, not detract from it. Please do not see my conclusions as a discredit to all your hard work, but rather as an invitation to other builders to try things for themselves.
I am not saying you're wrong and I'm right. I'm saying there are alternatives to your alternatives.
Peace,
Tom E
Neither do I, I'm sure you heard that hardness and that your description of sound under different comp values is accurate.
This is not the point.
No problem, thanks for answering now. 😉
These can have a role but not a big one
This can have a way bigger role, for instance PRPs here sound quite good but harsh, which Vishay did you use?
You're over-emphasizing differences among Rikens and Caddocks, IMHO, but nevertheless Caddocks are more detailed.
Don't forget, BTW that Caddocs are THICK-FILM resistors, not too much different from carbon films...
Because it's flawed, IMHO.
You heard a hardness/harshness problem
You suspect that the problem lies in the new compensation since you didn't hear it on MyRef Rev. C
The next logical step it to try Rev C compensation on FE boards, if the problem disappear it's likely the problem is the new compensation, if not you should look elsewhere.
What you did, instead of first checking if your hypothesis had some merit reverting the compensation, was to play with different values to 'tune' the sound.
I've did it too and what it seemed to sound best was actually the worst for performance.
You simply excluded from the start that the problem you were hearing could be originated elsewhere.
Absolutely not, if a problem is present I want to fix it.
Since I and beta testers didn't heard it the first person that can confirm it it's you but you refused to do the most simple, and useful, test.
Compensation doesn't work in that way.
Sorry but I don't think you can talk about trial and error in your case... you didn't check (trial) a lot of things, including C13.
So do I, I just want to be able to repeat their results and that people possibly try what I suggest to further investigate.
I've tried a lot of new parts suggested by partecipants on this thread, some were worth it, some not but I've tried (and spent a lot on money on) them.
Nice one! 😀
Sure but I can't consider your comp a valid alternative without some more tests by your side and at least some sims by Siva.
In both the V1.2 Twisted Pear board pairs, I have Takman REY25 MFR at R10 (a relatively inexpensive option that captures some of the neutrality of the boutique MFRs/foils) - one pair has a classic 22/10 Rev C compensation and sounds very musical in the mids and highs (though it tends to be slightly bass-shy/muffled in the lows). The other has a 27/22 FE-type compensation which is awaiting audition (I may also try the 33/15 on this board later).
I also have a V1.4 board pair with a Panasonic ERD-x CFR at R10 and 27/22 FE-type compensation. It is neutral-sounding, maybe a bit smoothed-out in the highs - however, there is some musicality in the first V1.2 pair that is subjectively compelling, that trumps this V1.4 board set for now.
The last board pair I have in hand is a V1.3 with 27/22 compensation and various other mods that is excessively bright and is currently last in my list of listening preferences, but is just a placeholder awaiting the 33/15 compensation and any other improvements that can be gleaned from the other three.
The takeaway(s) so far: the combination of PRP/Takman REY25 for R7/R10 is an inexpensive, musical and natural-sounding option for Rev C. No reason it can't work with FE-type compensations also. MKP1837 for C4 is a winner for Rev C, and should be fine for FE-type compensation also (it is very transparent, clean and neutral), where I've so far only tried Wima FKP3 (another transparent and neutral choice).
Blowing of a close rated fuse is to be expected if one does not ensure that start up current does not overly heat the fuse element.
Triple the fuse current rating and your fuse does not blow, even when the equipment is severely abused.
Fit a soft start and close rate the fuse. For 300VA, a T1A or T1.2A should be perfectly adequate.
