Hello
I have an Audio Alchemy DITB (dac-in-the-box) with a PS1 power supply. I have done the Kalman Robinson mods which were
1. replace 5 V supply regulators to 1A
2. replace 12 V supply regulators to 1A
3. increase the raw supply caps
4. use caps on bypass supply
The above did improve the DITB considerably but I am curious if the stock OpAmp OP275 can be substituted to something better.....
Question is which one?? OPA627/637, AD varieties, or the Burson discrete OpAmp.
Inputs from fellow members would be most appreciated.
Cheers!
I have an Audio Alchemy DITB (dac-in-the-box) with a PS1 power supply. I have done the Kalman Robinson mods which were
1. replace 5 V supply regulators to 1A
2. replace 12 V supply regulators to 1A
3. increase the raw supply caps
4. use caps on bypass supply
The above did improve the DITB considerably but I am curious if the stock OpAmp OP275 can be substituted to something better.....
Question is which one?? OPA627/637, AD varieties, or the Burson discrete OpAmp.
Inputs from fellow members would be most appreciated.
Cheers!
I have gotten the best sound by using the B-B OPA2134PA. Major improvement also by adding as much electrolytic storage on the supply lines in all areas, esp. right close to the output chip, as will physically fit in the tiny little guy. I've been able to make the lowly DITB totally smoke a Denon DVD3910, Benchmark DAC1, and even equal my fabulous sounding, very, very upgraded Pioneer CLD95(which has very well implemented AD1862 dac chips).
The primary differences between the '627 and the '637 is unity gain bandwidth:
vailable Channels S S
CMRR(Min)(dB) 106 106
>> GBW(Typ)(MHz) 16 80 <<
IIB(Max)(pA) 1 1
Iq per channel(Max)(mA) 7.5 7.5
Number of Channels 1 1
Offset Drift(Typ)(uV/C) 0.4 0.4
Open Loop Gain(Min)(dB) 112 112
... etc. ... from: http://focus.ti.com/docs/prod/folders/print/opa627.html
Personally, I would not bother changing out the Analog Devices' OP275 op-amp unless you just want to see if you can improve this particular A/A DITB. (You might gain some on the op-amp unity gain bandwidth question, but the DITB has bandwidth limiting components in other parts of the circuit. Possible but small improvements in slew rate(s) ... and there might be an overall improvement of the THD+N numbers = 0.00x% v. 0.000x% [the exact quantification without lab quality equipment / analisys = unknown and possibly specific to the your A/A DITB].
The addition of larger voltage regulators was a very good idea = less heat and more PS power "headroom", etc.
The use of larger or better caps on the PS rails was also very smart = better noise floor close to the chips, etc. You might examine the possibility of putting a reasonably sized plastic cap right across the power pins on the output op-amps (pins 4 & 8). Close coupling of a "snubbing cap" with a fast response time often makes a significant difference to any op-amps' performance, sometimes more so than all the other PS and PS rail work other than adding the larger PS caps. (I'm a believer in the generous use of plastic snubbing caps in parallel with electrolitic caps on PS rails = as written by Bob Pease in Electronic Design Magazine.)
vailable Channels S S
CMRR(Min)(dB) 106 106
>> GBW(Typ)(MHz) 16 80 <<
IIB(Max)(pA) 1 1
Iq per channel(Max)(mA) 7.5 7.5
Number of Channels 1 1
Offset Drift(Typ)(uV/C) 0.4 0.4
Open Loop Gain(Min)(dB) 112 112
... etc. ... from: http://focus.ti.com/docs/prod/folders/print/opa627.html
Personally, I would not bother changing out the Analog Devices' OP275 op-amp unless you just want to see if you can improve this particular A/A DITB. (You might gain some on the op-amp unity gain bandwidth question, but the DITB has bandwidth limiting components in other parts of the circuit. Possible but small improvements in slew rate(s) ... and there might be an overall improvement of the THD+N numbers = 0.00x% v. 0.000x% [the exact quantification without lab quality equipment / analisys = unknown and possibly specific to the your A/A DITB].
The addition of larger voltage regulators was a very good idea = less heat and more PS power "headroom", etc.
The use of larger or better caps on the PS rails was also very smart = better noise floor close to the chips, etc. You might examine the possibility of putting a reasonably sized plastic cap right across the power pins on the output op-amps (pins 4 & 8). Close coupling of a "snubbing cap" with a fast response time often makes a significant difference to any op-amps' performance, sometimes more so than all the other PS and PS rail work other than adding the larger PS caps. (I'm a believer in the generous use of plastic snubbing caps in parallel with electrolitic caps on PS rails = as written by Bob Pease in Electronic Design Magazine.)
stephensank's recommendation for the TI / B-B OPA2134 is a good one, although any improvements over the A/A DITB Analog Devices' original will be hard to quantify.
Also: " ... Major improvement also by adding as much electrolytic storage on the supply lines in all areas, esp. right close to the output chip ..." =
" ... I've been able to make the lowly DITB totally smoke a Denon DVD3910, Benchmark DAC1 ..."
One would wonder if this were possible, but sometimes we get an appreciation of the "unaccountable changes caused by an unknown variable" = got lucky?
Also: " ... Major improvement also by adding as much electrolytic storage on the supply lines in all areas, esp. right close to the output chip ..." =
" ... I've been able to make the lowly DITB totally smoke a Denon DVD3910, Benchmark DAC1 ..."
One would wonder if this were possible, but sometimes we get an appreciation of the "unaccountable changes caused by an unknown variable" = got lucky?
I'm getting the impression that you are one of the "if it looks good on paper/insruments, it must sound good" crowd. This is what got the Japanese in trouble into the 80's, resulting in things like the Yamaha M80 power amp. As close to perfect measurements as ever seen on earth, perfect square waves at any sonic frequency, but it is famously one of the most abrasive & awful sounding amps since the Dynaco 120.
I don't mean to insult you at all, if you are of that sort of crowd. It's just that in my 35 years in high end audio service, I have come to the conclusion that if something sounds better despite looking worse on instruments or etc., it's just because we haven't figured out what or how to measure whatever the parameter is that made it sound better.
The OPA2134 is not much better or worse than the OP275 on paper, but I can tell you from using both in many, many different applications that the OPA2134 sounds richer, more natural & more musical than the OP275. The OP275 just sounds a bit "dry" by comparison.
As for the cap issue- the stock DITB does have reasonable size petp film caps at the supply rail pins of every chip on board, so that's not much of a concern. It lacks any significant energy storage near thte output chip, though. Now here is where you might think there's "placebo effect" going on- Adding 220uf at each supply pin of the output chip makes a very nice difference overall. Changing that 220uf to 470uf, which darn well should not affect a low current opamp chip, gives the DITB an absolutely clear advantage in bass extension & punch versus the 220uf. And this is with identical grade Nichicon lytics. I have done blind tests with friends on this & they spot the 470uf equipped DITB every time.
The Denon DVD3910 comparison is quite easily explainable by even just considering power supply issues in the output stage, where Denon did nowhere near the supply work that I did to the DITB. This alone could explain the difference. This is also why most dvd players sound anemic & thin compared to a well done old cd player or 16/44 dac.
I don't mean to insult you at all, if you are of that sort of crowd. It's just that in my 35 years in high end audio service, I have come to the conclusion that if something sounds better despite looking worse on instruments or etc., it's just because we haven't figured out what or how to measure whatever the parameter is that made it sound better.
The OPA2134 is not much better or worse than the OP275 on paper, but I can tell you from using both in many, many different applications that the OPA2134 sounds richer, more natural & more musical than the OP275. The OP275 just sounds a bit "dry" by comparison.
As for the cap issue- the stock DITB does have reasonable size petp film caps at the supply rail pins of every chip on board, so that's not much of a concern. It lacks any significant energy storage near thte output chip, though. Now here is where you might think there's "placebo effect" going on- Adding 220uf at each supply pin of the output chip makes a very nice difference overall. Changing that 220uf to 470uf, which darn well should not affect a low current opamp chip, gives the DITB an absolutely clear advantage in bass extension & punch versus the 220uf. And this is with identical grade Nichicon lytics. I have done blind tests with friends on this & they spot the 470uf equipped DITB every time.
The Denon DVD3910 comparison is quite easily explainable by even just considering power supply issues in the output stage, where Denon did nowhere near the supply work that I did to the DITB. This alone could explain the difference. This is also why most dvd players sound anemic & thin compared to a well done old cd player or 16/44 dac.
stephensank & FastEddy: many thanks for your posts and comments/suggestions. I greatly appreciate it.
I must say that stephensank's recommendation of adding 470uF lytic at each Supply pin of the output seems a superb idea. And indeed, I have always felt the AA-DITB to be not so stellar on the low bass area that my B&W speakers can easily & comfortably reach.
I would appreciate if you can tell me exactly how to go about this mod. I presume I will require 2 x 470uF electrolytic caps (what voltage) and also should solder them where? Sorry for bothering you guys but the idea of strengthening the bass sounds fantastic to me. Hope the 470uF wouldnt hurt the OP275 ........
I have 2 x 470uF 25V panasonic electrolytic caps with me and where exactly should I solder these??
Cheers!
I must say that stephensank's recommendation of adding 470uF lytic at each Supply pin of the output seems a superb idea. And indeed, I have always felt the AA-DITB to be not so stellar on the low bass area that my B&W speakers can easily & comfortably reach.
I would appreciate if you can tell me exactly how to go about this mod. I presume I will require 2 x 470uF electrolytic caps (what voltage) and also should solder them where? Sorry for bothering you guys but the idea of strengthening the bass sounds fantastic to me. Hope the 470uF wouldnt hurt the OP275 ........
I have 2 x 470uF 25V panasonic electrolytic caps with me and where exactly should I solder these??
Cheers!
I would go with the first V+ to G and V- to G.arupg said:
You may find a difference (good or bad) by adding your second option of V+ to V-, but since a plastic film cap should already be here then I think this may be too much.
According to the Marantz thread guys/gals,Oscon (digital) BG(analogue) caps should be considered for the electrolytic local decoupling and if there is room, very local regulation of both digital chips and analogue chips.
AndrewT
Thanks for your inputs. As Stephensank recommends, whats your take on using a size like 470uF on the V+ to G and V- to G for the OP275? Do you in anyway feel this is too large a size to spark off things like unstability or oscillations etc.
I am more worried about the side-effects of such DIY work as I have expensive hifi components & cant afford to either mess-up or fry them or the dac or the Opamp itself since I live in India and these sort of things are rather a bother to fix due to complete lack of parts etc.
It has to be a SURE SHOT diy or else all hell will break loose...
Thanks & regards
Thanks for your inputs. As Stephensank recommends, whats your take on using a size like 470uF on the V+ to G and V- to G for the OP275? Do you in anyway feel this is too large a size to spark off things like unstability or oscillations etc.
I am more worried about the side-effects of such DIY work as I have expensive hifi components & cant afford to either mess-up or fry them or the dac or the Opamp itself since I live in India and these sort of things are rather a bother to fix due to complete lack of parts etc.
It has to be a SURE SHOT diy or else all hell will break loose...
Thanks & regards
Stephensank's & AndrewT's advise is well founded and will certainly improve the response of the output op-amp ... Using larger electrolitics very close to the power pins of the op-amp may of course obviate the need of any paralleled plastic "snubbing" additions ... but, as long as the soldering iron is hot and the gadget is open on the bench ... 
" ... I am more worried about the side-effects of such DIY work ..." As long as no ground loops are accidently imposed and there are no shorts to the power supply rails (like using underrated caps or solder blobs crossing traces, etc.) ... then confidence should remain high ...
" ... I am more worried about the side-effects of such DIY work ..." As long as no ground loops are accidently imposed and there are no shorts to the power supply rails (like using underrated caps or solder blobs crossing traces, etc.) ... then confidence should remain high ...
Ok, thanks for the inputs. Appreciate FastEddy's comments and boost of confidence!
Am yet to get a response from stephensank on the 470uF 25V lytics. Would appreciate if you can confirm successful implementation of 470uF for the V+ to G and V- to G of the OP275. Do note from that there's already something like 0.1uF PP cap installed at stock between V+ to G and V- to G. So does that mean that these 0.1uF PP caps come off, in goes the 470uF and then try... or leave the 0.1uF and add the 470uF on top on these.....
Also, If need be, put in PP caps between V+ and V- What values??
Finally, will be careful as per FastEddy not to create ground-loops or shorts.....
Hope this works well!
Am yet to get a response from stephensank on the 470uF 25V lytics. Would appreciate if you can confirm successful implementation of 470uF for the V+ to G and V- to G of the OP275. Do note from that there's already something like 0.1uF PP cap installed at stock between V+ to G and V- to G. So does that mean that these 0.1uF PP caps come off, in goes the 470uF and then try... or leave the 0.1uF and add the 470uF on top on these.....
Also, If need be, put in PP caps between V+ and V- What values??
Finally, will be careful as per FastEddy not to create ground-loops or shorts.....
Hope this works well!
" ... Do note from that there's already something like 0.1uF PP cap installed at stock between V+ to G and V- to G. So does that mean that these 0.1uF PP caps come off, in goes the 470uF and then try (?) ..."
The purpose of the 0.1 uF cap (100 nF? ... plastic or other type?) is where the improvements are to take place, but I would recommend that this cap NOT be removed ... and that the added 470 uF (or other) cap be added to this in parallel ... (It might be that this cap is doing or will do be doing what I have suggested earlier = acting to "snub" the added, larger cap, if effect.)
Again, be careful about allowing under rated cap voltages being used ... Ex: if your rail to rail voltage is +/- 18 VDC, that's a 36 VDC differential between the op-amp power pins (4 & 8) ... so I would try if available to use 63 VDC or 100 VDC rated caps for this purpose here, even if that means dropping the desired capacitance back to a lesser amount in the interest of physical space savings ... say, fall back to 220 uF rather than 450 uF and up the rating to 50 VDC, max.
(Applying voltages up close to the absolute limits of electrolitic capacitor is never advisable = this just generates more heat and narrows the safety factors. Although you appear to be directed toward quality audio caps for this purpose, all cap manufacturers state these maximum voltages for very good reasons ... proper engineering considerations suggest a 2 to 1 safety factor, at least. I would use 3 to 1 and more on +/- 18 VDC rails ... using 63 VDC or 100 volt rated caps ... you never know when lightning is going to strike ... ;>(
AndrewT: " ... Relocate the 100nF ON THE POWER PINS. Then you have pads to secure your additional electrolytics. ..."
This is a good idea, but keeping the 0.1 uf (100 nF?) cap is certainly advised in either case.
The purpose of the 0.1 uF cap (100 nF? ... plastic or other type?) is where the improvements are to take place, but I would recommend that this cap NOT be removed ... and that the added 470 uF (or other) cap be added to this in parallel ... (It might be that this cap is doing or will do be doing what I have suggested earlier = acting to "snub" the added, larger cap, if effect.)
Again, be careful about allowing under rated cap voltages being used ... Ex: if your rail to rail voltage is +/- 18 VDC, that's a 36 VDC differential between the op-amp power pins (4 & 8) ... so I would try if available to use 63 VDC or 100 VDC rated caps for this purpose here, even if that means dropping the desired capacitance back to a lesser amount in the interest of physical space savings ... say, fall back to 220 uF rather than 450 uF and up the rating to 50 VDC, max.
(Applying voltages up close to the absolute limits of electrolitic capacitor is never advisable = this just generates more heat and narrows the safety factors. Although you appear to be directed toward quality audio caps for this purpose, all cap manufacturers state these maximum voltages for very good reasons ... proper engineering considerations suggest a 2 to 1 safety factor, at least. I would use 3 to 1 and more on +/- 18 VDC rails ... using 63 VDC or 100 volt rated caps ... you never know when lightning is going to strike ... ;>(
AndrewT: " ... Relocate the 100nF ON THE POWER PINS. Then you have pads to secure your additional electrolytics. ..."
This is a good idea, but keeping the 0.1 uf (100 nF?) cap is certainly advised in either case.
FastEddy: Thank u so much for your kind concern. All points mentioned are noted. As per Stephensank's Nichicon Muse suggestions, I'm curious which grade he used to obtained the results as claimed "gives the DITB an absolutely clear advantage in bass extension & punch".
Presently, I've only general purpose Panasonic lytics of 470uF 25V & none of the exotic Elna Silmic, Cerafine, Blackgates or Nichicon Muse KZ or FG or FX.
My PSU is +/- 15 VDC, so across Pin 4 & 8 of OP275, its 30 VDC. I assume 63V cap should do here. Across V+ & G or V- & G of OP275, do u reckon 25V would suffice? Or better with 35V or 63V? The tiny casing of the DITB & the extremely tight fit of the PCB is a major problem accomodating the caps.
If I decide to go with the Panasonic general purpose caps, would this grade of cap hinder performance? My main aim that motivated me to doing this mod is to get 'superior bass performance and slam' out of this dac.
best regards
Presently, I've only general purpose Panasonic lytics of 470uF 25V & none of the exotic Elna Silmic, Cerafine, Blackgates or Nichicon Muse KZ or FG or FX.
My PSU is +/- 15 VDC, so across Pin 4 & 8 of OP275, its 30 VDC. I assume 63V cap should do here. Across V+ & G or V- & G of OP275, do u reckon 25V would suffice? Or better with 35V or 63V? The tiny casing of the DITB & the extremely tight fit of the PCB is a major problem accomodating the caps.
If I decide to go with the Panasonic general purpose caps, would this grade of cap hinder performance? My main aim that motivated me to doing this mod is to get 'superior bass performance and slam' out of this dac.
best regards
" ... If I decide to go with the Panasonic general purpose caps, would this grade of cap hinder performance? ..."
Probably not, but Stephensank's remarks about using higher quality should be adheared to ... especially if you don't plan on adding any plastic caps. These higher end caps do sound better, although sometimes the remarks about quality can get rather subjective ...
Probably not, but Stephensank's remarks about using higher quality should be adheared to ... especially if you don't plan on adding any plastic caps. These higher end caps do sound better, although sometimes the remarks about quality can get rather subjective ...
What I have always done is just as one poster suggested, to relocate the 0.1u yellow film caps under the board & put the 470uf caps in their places, as I also do near the dac & other chips with as large value lytic caps as will fit the tiny spaces. In the case of the dac audio ps pins & the opamp ps pins, the 470uf 16v Nichicon VR caps I used had to be tilted about 45 degrees to fit inside the tiny box. I hated how it looked, since I like my upgrade work to look very orderly & neat, but it was worth the difference in sound.
After 30+ years of replacing failed and/or crappy electrolytic caps, I can tell you that Panasonic, Sanyo/Oscon, Elna(incl. silmic & cerafine), Rubycon(incl. blackgate), Illinois Capacitor, Samwha, Mallory/Jamicon, Sprague, etc., have terrible long term reliability compared to Nichicon & Nippon/United ChemiCon. And I have yet to hear better performance than Nichicon, with ChemiCon often being a quite close second. Rubycon & Elna have, by far, the worst reliability of caps made in Japan, so it makes me nauseous when "upgraders" sell people on BlackGate & Cerafine. Panasonic & Sanyo are almost as bad.
If, on this next DITB I am about to upgrade, I decided to transfer it to a larger enclosure, I will add larger & better film caps accross all of the power line electrolytics. Those little polyester 100nf caps AA used are barely adequate, in my eyes, for HF bypassing, but they are about the best that will fit. Although, actually, now that I think back a bit, the last DITB I was able to add a pair of unusually skinny, but quite long, polypropylene film/foil 1.5uf/100v caps from each of the opamp's ps pins to ground, by "flying" them above the chip on their leads & grounding them far from the chip, but close to the main supply ground. I was not completely comfortable with the long lead length, nor the "remote" ground, but it worked perfectly. I recall that I strapped a decent size polystyrene parallel with each of these, too. I think they were 0.015uf/33v Mial. On this & several previous DITB, I tossed the stock AA power unit & instead used a BurrBrown 501(think that was the number) "brick" modular +/-15V regulated power unit. So, the 12 & 5v regulators in the DITB were the second regulators.
No matter how big the electrolytics you fit right up close to the pins of an opamp, or within a discrete gain stage or etc., I consider it absolutely essential for good sound to parallel them with as good & large film caps as will fit in the available space. Even the best lytics have increasing equivalent series resistance not much above the so called audible range.
After 30+ years of replacing failed and/or crappy electrolytic caps, I can tell you that Panasonic, Sanyo/Oscon, Elna(incl. silmic & cerafine), Rubycon(incl. blackgate), Illinois Capacitor, Samwha, Mallory/Jamicon, Sprague, etc., have terrible long term reliability compared to Nichicon & Nippon/United ChemiCon. And I have yet to hear better performance than Nichicon, with ChemiCon often being a quite close second. Rubycon & Elna have, by far, the worst reliability of caps made in Japan, so it makes me nauseous when "upgraders" sell people on BlackGate & Cerafine. Panasonic & Sanyo are almost as bad.
If, on this next DITB I am about to upgrade, I decided to transfer it to a larger enclosure, I will add larger & better film caps accross all of the power line electrolytics. Those little polyester 100nf caps AA used are barely adequate, in my eyes, for HF bypassing, but they are about the best that will fit. Although, actually, now that I think back a bit, the last DITB I was able to add a pair of unusually skinny, but quite long, polypropylene film/foil 1.5uf/100v caps from each of the opamp's ps pins to ground, by "flying" them above the chip on their leads & grounding them far from the chip, but close to the main supply ground. I was not completely comfortable with the long lead length, nor the "remote" ground, but it worked perfectly. I recall that I strapped a decent size polystyrene parallel with each of these, too. I think they were 0.015uf/33v Mial. On this & several previous DITB, I tossed the stock AA power unit & instead used a BurrBrown 501(think that was the number) "brick" modular +/-15V regulated power unit. So, the 12 & 5v regulators in the DITB were the second regulators.
No matter how big the electrolytics you fit right up close to the pins of an opamp, or within a discrete gain stage or etc., I consider it absolutely essential for good sound to parallel them with as good & large film caps as will fit in the available space. Even the best lytics have increasing equivalent series resistance not much above the so called audible range.
" ... I can tell you that Panasonic, Sanyo/Oscon, Elna(incl. silmic & cerafine), Rubycon(incl. blackgate), Illinois Capacitor, Samwha, Mallory/Jamicon, Sprague, etc., have terrible long term reliability compared to Nichicon & Nippon/United ChemiCon. ..."
... and these are (some of) the manufacturers thatoften slip a little of their long term voltage rating = zzzzztttt... ... If you use 'em, get the ones that have the higher voltage rating whereever possible ... 50 VDC should be 63VDC, 63 VDC should be 100 VDC ... etc.
... what he said.
... and these are (some of) the manufacturers thatoften slip a little of their long term voltage rating = zzzzztttt...
... If you use 'em, get the ones that have the higher voltage rating whereever possible ... 50 VDC should be 63VDC, 63 VDC should be 100 VDC ... etc.... what he said.
FastEddy said:" ... I can tell you that Panasonic, Sanyo/Oscon, Elna(incl. silmic & cerafine), Rubycon(incl. blackgate), Illinois Capacitor, Samwha, Mallory/Jamicon, Sprague, etc., have terrible long term reliability compared to Nichicon & Nippon/United ChemiCon. ..."
... and these are (some of) the manufacturers thatoften slip a little of their long term voltage rating = zzzzztttt...
... what he said.
It's good design to double the rating. If there is 15 volts on the leads, then get a 35V rated cap.
I have never had a BG or Nichicon fail in the field.
Steve N.
Empirical Audio
Manufacturer
I have seen BlackGates fail. They are better reliability than Rubycon's lesser caps, but not great. I've replaced thousands of Rubycons. But in 30+ years, I can still count on one hand(and have a thumb remaining) the number of Nichicon caps I've had to replace due to age or materials failure. And there's not actually any I can remember replacing due to lightning/surge damage, but I'm just guessing I'm not remembering those. With Nichicon, although I habitually use higher than needed voltage if they're handy, my experience shows that they can tolerate easily ten percent over their rated voltage pretty much infinitely, so I never worry if a 16v Nichicon is used on a 15v line or etc.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.