OK, Nic, apologies if I took you up wrong - interested to hear how you find the PH regs - BUT this gets pretty expensive when you begin to add up the cost of these PH regs, doesn't it?
I'm looking at a cheaper route to achieve low noise, high quality, high current 3.3V supply - LiFePO4 batteries without regulators
I'm looking at a cheaper route to achieve low noise, high quality, high current 3.3V supply - LiFePO4 batteries without regulators
Well, it should have cropped up during testing & been obvious - it's one of the first things to check on any circuit - any components running hot. In the case of op-amps, it usually signifies oscillation which is a flaw.
Just staying within max ratings would not be a desirable design goal, I believe.
Dude, what are you talking about? The opamp is acting as a buffer, and runs warm. This is 100% normal. It does not run any where near it's max ratings. It is not oscillating. Before you make such comments, you should probably have some clue as to what you are talking about.
Listen, don't get patronising with me like your partner, Russ - I wasn't saying that your op-amp was oscillating, but that this is a normal test in any new circuit & should have been obvious & investigated before the product was launched. You had two boards releases to get this right - the 24bit board it was warm, the 32bit board it was HOT & not warm as you say
Why is it HOT ? Do you know? Have you investigated? If it's normal, why did Russ post a fix?
Come on get real and deal with things in the real world not the TP world - stop living in denial - the PS is not the best incarnation & the evidence is overwhelming!
Yes - it does make it a quite costly as 5 PH regs adds about 200$ to the price (in addition to the JSR's I already had). Worth it? I don't know - I guess its the old story about improvements costing more the higher the quality of the starting point.
Nic
Hmmm, I know what you mean about the last 10% or is it 1% improvement costs a lot more than the previous 90% or 99%. It's up to each individual to decide how much improvement they are getting from these PS mods - 1%, 10%, 20%, more?
Look, I don't want this to be blown out of proportion any more than you do (I think). Yes, it was fully tested. With different topologies. Russ posted a tweak, which is not the same as a fix. It is a different topology. We knew full well it would run at the temperature it runs at. It is absolutely fine.
I am not trying to be patronizing, but you seem convinced that this is some sort of design flaw, when it is not. It was a design decision, made after lots and lots of testing. It was not overlooked, not a mistake, not an error.
There have been no failures of the opamp buffers in the field. None. The approach yields very low noise and very low Z, which is what we were after. Are there other approaches? Yes. Could we have made it more complex and more expensive? Yes. It is what it is. Can we please move on.
There is an old saying that if you reduce the temperature of an IC by 20 degrees C, you will double the life of the IC. I don't know where the original research for that saying came from.
I wouldn't call the design flawed just because the opamps run hot. The opamps were dissipating a quarter watt or so. Now the heat is in the resistors. That is a worthwhile improvment even if it doesn't affect the sound much.
The question is, do you want to replace the opamps in 2020 or in 2030?
Will you still be listening to this DAC in 2020? I have run through 5 DACs in the last decade. That is not counting the ESS Sabre demo board and the 3 DACs I have built up myself of my own Sabre design.
I put the tweak on one of my Sabre DACs. It may have improved the bass a little bit. It's hard to tell. It may allow the opamp better control over the bass. It didn't seem to affect the higher frequencies or the soundstage. I didn't bother with trying to do an A-B test.
I wouldn't call the design flawed just because the opamps run hot. The opamps were dissipating a quarter watt or so. Now the heat is in the resistors. That is a worthwhile improvment even if it doesn't affect the sound much.
The question is, do you want to replace the opamps in 2020 or in 2030?
Will you still be listening to this DAC in 2020? I have run through 5 DACs in the last decade. That is not counting the ESS Sabre demo board and the 3 DACs I have built up myself of my own Sabre design.
I put the tweak on one of my Sabre DACs. It may have improved the bass a little bit. It's hard to tell. It may allow the opamp better control over the bass. It didn't seem to affect the higher frequencies or the soundstage. I didn't bother with trying to do an A-B test.
Why post a "tweak" that has absolutely no improvement potential - the sound won't change; the lifetime of the op-amp won't change - what's the "tweak" for? You can't have your cake & eat it too. Either it is a tweak to improve something or it isn't!
Very low Z on the output can oscillate, dude - I'm sure you checked this, dude & will tell us it doesn't oscillate.
You didn't really design this board taking all possible options into account and test them as you say, did you! - you simply followed the ESS eval board layout, so come clean & stop the pretence. It's amazing that from all the choices you could have made, you ended up with a final design that resembles the ESS eval board so closely, isn't it
Last edited:
I've tried ALW Jung types, Teddy regs, Burson regs, Audiocom regs, various other diy shunt and series (still not the Salas one)
For me I don't feel the need to faff around with anything else.
These are dead stable no matter what caps on the output unlike certain other discrete types and the guy who designed them isn't full of ** which is good enough for me😉
Leo, could you put a percentage figure estimate of improvement since implementing the PH regs on both Buffalos?
Um, I think Russ mentioned that it lowers the opamp temps. That would be an improvement. That does not mean the original design is deeply flawed.
A better question might be, why try so hard to find tweaks for a board you do not own?
I have not insulted you, but you seem to be fine with insulting me and leveling false allegations.
Perhaps you should design your own and wow us with your brilliance.
[EDIT] To every one else, sorry for the noise.
Last edited:
I don't really know regarding percentage, all I do know is that I don't enjoy the music as much if I revert back to alternatives
With other regulation I found the sound to be a little tiring after a while, I now find it just flows which is exactly the thing I was after🙂 this becomes more noticeable after longer listening
Both dacs as standard are very good, just not quite there for me, I've a lot of other dacs to compare and do tend to be picky, beauty with diy is we can take things further if we feel the need
Bit of a Freudian slip there, Brian 😀
Maybe, just maybe, you would see this thread as a way to investigate/learn how to improve the product instead of continually jumping in to defend it?
Last edited:
I just hope this thread don't turn bad and end up like the others, also not put people off posting tweaks to try, for me this one posted by Russ is worth doing, it only costs a few pence to try, normally I'm quick to try mods but thought I'd leave this one for others to try first mainly because this Buffalo32 is not stock and may or may not effect the results of the tweak😕
Maybe somebody else with a stock Buffalo32s could try it and post results?
OK - so back on subject - tweaking buffalo.
I will try to power Avcc with two local 3.3V PH shunt regs after removing the two buffer op-amps. For powering these two regs I have two options:
1) The 5V JSR which is already feeding my VD section (which is made up of two 3.3V PH shunts and a 1.2V PH series reg).
2) The +15V VA rails powered by JSR's (+ a series resistor to reduce the voltage a bit). There is local access to these rails.
Is there any particular reason to think that one way will be better or safer than the other?
Thanks,
Nic
I will try to power Avcc with two local 3.3V PH shunt regs after removing the two buffer op-amps. For powering these two regs I have two options:
1) The 5V JSR which is already feeding my VD section (which is made up of two 3.3V PH shunts and a 1.2V PH series reg).
2) The +15V VA rails powered by JSR's (+ a series resistor to reduce the voltage a bit). There is local access to these rails.
Is there any particular reason to think that one way will be better or safer than the other?
Thanks,
Nic
Hi NicMac, the PH shunt will need some dropout voltage, and I'd think the 5V rail will be too low. On the other hand running it from 15V will result in a lot of power dispated in the reg - it'll get pretty hot. I'd suggest running it from 15V and using a power resistor to drop it down to around 7V. To do this you'll need to find out how much current the PH reg draws - I'm sure Paul will tell you.
Then use:
Resistor = 8 / Current
Doing this will keep your reg cool. A 3 Watt resistor will be fine.
Then use:
Resistor = 8 / Current
Doing this will keep your reg cool. A 3 Watt resistor will be fine.
Thanks Spartacus, I already have two PH 3.3V shunts and a 1.2V PH series reg running of the 5V rail nicely. I think the PH miniregs have 1.5V (shunt) and 3V (series) dropout. I could however raise the voltage of the supply to 6-7V for some more margin.
I will ask Paul about the current draw of the 3v3 shunt regs (I seem to recall 150 mA @ 5V).
As I'm planning to try the Placid shunts for the bipolar VA supply in place of the JSR's I was just thinking that I might not want to put this additional load on the VA+ rails. Of course I could just adjust these to accommodate the extra current draw - but still.
I will ask Paul about the current draw of the 3v3 shunt regs (I seem to recall 150 mA @ 5V).
As I'm planning to try the Placid shunts for the bipolar VA supply in place of the JSR's I was just thinking that I might not want to put this additional load on the VA+ rails. Of course I could just adjust these to accommodate the extra current draw - but still.
Hi All,
As I said when I presented my little mod. It purpose was to show a very simple way to turn the linear regulator (or buffer if you will) into a shunt regulator. Please don't read any more into it than that. 🙂 Some of you expressed a preference for shunt regulation so I wanted to show how the Buffalo could be made to easily have one. 🙂
The AVCC opamps run well within spec and should not have any problems at all.
One further note, if you want an even more "stiff" (higher output Z) current source, then grab some JFETs that can handle about .5W and measure the IDss. When you find a couple in the 35ma range you should be able to use them as a CSS. Or you could get a couple 35ma CLDs. Then use those in place of the resistors. I would still recommend the zener for safety.
I posted to give you guys something fun to try if you wanted to do it. I don't much care to attach my subjective review to a given cct change.
Take the post in the friendly spirit of DIY it was intended. Sometimes I like to give people something fun to try. What each individual will get out of the experience is not something I could control even if I wanted to. Posting the mod was very certainly *NOT* any sort of suggestion or remediation of a design flaw. The design as been thoroughly vetted.
Cheers!
Russ
As I said when I presented my little mod. It purpose was to show a very simple way to turn the linear regulator (or buffer if you will) into a shunt regulator. Please don't read any more into it than that. 🙂 Some of you expressed a preference for shunt regulation so I wanted to show how the Buffalo could be made to easily have one. 🙂
The AVCC opamps run well within spec and should not have any problems at all.
One further note, if you want an even more "stiff" (higher output Z) current source, then grab some JFETs that can handle about .5W and measure the IDss. When you find a couple in the 35ma range you should be able to use them as a CSS. Or you could get a couple 35ma CLDs. Then use those in place of the resistors. I would still recommend the zener for safety.
I posted to give you guys something fun to try if you wanted to do it. I don't much care to attach my subjective review to a given cct change.
Take the post in the friendly spirit of DIY it was intended. Sometimes I like to give people something fun to try. What each individual will get out of the experience is not something I could control even if I wanted to. Posting the mod was very certainly *NOT* any sort of suggestion or remediation of a design flaw. The design as been thoroughly vetted.
Cheers!
Russ
- Status
- Not open for further replies.