Jan-Peter said:
A good improvement would be to use a external power supply to feed the op amp bufferstage in the UcD180/400 module.
To tell you the truth I don't know why the UcD700 do sound better. We use in all amps exactly the same modulator PCB. So this can't be the difference. Probably the more power full power supply?? In the UcD700 is used 4x10.000uF/100V capacitors. One rail voltage has charge of 81Joules at 90VDC (0,5*2*10.000uF*90^2=81Joules). I woudn't short circuit this by accident with a screw driver......
Or all the parallelled switched high voltage BCC/Vishay capacitors on the UcD700 PCB???
For the moment I take it as it is and we are happy with the performance
Jan-Peter
Looking at what is different about them I can see a few suspects that may be responsible for the improved sound too. Mentioning the auxiliary supply for the front end buffers is a good hint as to one reason why the 700 has a better soundstage/focus and dynamics.
I've seen the changes snubbing and local decoupling caps can make so you must have found a good combo there with so many in parallel too.
I see a different and maybe smarter routing of the output coil from the half bridge switching node, this now seems routed in a more balanced way being positioned along the axis of both halves of the symetrically laid out half bridge. I think that might help limit EMI?
You also mention added shottkey's /snubbing around the mosfet. Did you go so far as to turn the body diodes off by fully bypassing them? I can see that influencing the sound a great deal.
This isn't still a 2 layer board is it? I"m curious to know how many pre production test PCB revisions there's been?
Thanks,
Chris
classd4sure said:
oh, oops. sorry, i was honestly thinking about the UCD400 thread. its all mushed together since i read all of them in the course of a week.
ill re-read, but i dont remember seeing it... my mistake. oh well, at least im subscribed to the thread now
edit:
yep, its in there, i just glanced over it, sorry.
Yes, this is needed to have a low EMI.
Indeed we use a Schottky diode in series with the Fet and a Schottky anti-parallel. This improves the performance and reduce the EMI, could this be the reason for the sonic improvement……I don’t know….
Indeed we have now switched to a 4-layer PCB. The pre production runs were 4 times, most of them were to avoid errors before we go to production. The design was after the first proto already quite good (=Bruno’s excellent design work, I am only the Hardware Engineer…).
Jan-Peer
It's the sonic performance...you know....what you can't measure...
When we have done the first production batch we will show the average value.
One of the proto's showed a THD performance of 0.005% at 500W 4 ohm
We don't like to show diectly on the web one measurement of the best proto. We like to show performance what we can guarantuee....
Ha, ha....we will see.....
Jan-Peter
Yes you are right, you can have one topology with higher THD that sounds better than another topology with lower THD.
BUT using the same topology (like you have already said is the case with UcD700) , a higher THD is always a sign of problems, and will never lead to a better sound. You know it, and i know it. But of course a challenge for your marketing dept.
Jan-Peter to be honest you have aired this kind of imaginary results numerous times in the past.
Yet your production modules show far weaker performance in real life. Anyone with a THD instrument and a production UcD module can confirm this for him self.
Talk is cheap, product performance speak!
It’s the same technology, but because we have now the opportunity of a 4-layer PCB we could use the internal capacitive coupling between certain layers to improve the drive of certain parts in the circuit. Besides this there are more secrets parts in the hardware part of the design. Of course you can understand that I can’t give the details…..
R-u-b-b-i-s-h….we have too much Audio Precisions measurement gears to prove all our measurements. (two Systems 2, and two ATS-2!!!)
I like to do the discussion on the technical aspects, and not "mine is better" discussion.
Start your own thread about your FANTASTICE THE LOWEST THD IN THE WORLD CLASS-D amplifier, as to discuss how bad our new products is. Without ever seeing it our even listen to it!
Thank you!
Jan-Peter
I am sorry, Jan Peter, no need to get upset 
I am not talking about : FANTASTICE THE LOWEST THD IN THE WORLD CLASS-D amplifier, i didn't mention it at all, recently.
Also i am not saying your new product is bad.
All i'm pointing out is that everyone in here talk about how much improved your new amplifier is compared to your old design.
When in reality your own datasheet says the opposite...
I am not talking about : FANTASTICE THE LOWEST THD IN THE WORLD CLASS-D amplifier, i didn't mention it at all, recently.
Also i am not saying your new product is bad.
All i'm pointing out is that everyone in here talk about how much improved your new amplifier is compared to your old design.
When in reality your own datasheet says the opposite...
Hi,
Thanks for the reply Jan-Peter and I would imagine everything contributes in some small way, alot of mods lean towards limiting or lessoning the effects of EMI, I've personally found EMI highly detrimental and I can imagine that limiting it in every way possible would be worth some improvement, especially at higher power levels where it becomes more pronounced. Hence your parallel decoupling caps, coil placement, snubbing, multi layer pcb and SMT, symmetrical layout, fully balanced circuit.. from start to finish. All things to look at before even glancing at specs, it's looking smart not just shinny.
4 Revisions doesn't seem like much at all, I'm sure future revisions will push it further yet.
Have you experimented with many mosfets before settling on the philips ones? Were also chosen with EMI in mind? Or for total cost? I seems like the output stage isn't taking full advantage of the latest technology and yet maybe that's exactly what is required in a robust solution I don't know, thanks for sharing what you do anyway.
Lars, I realize why you're saying what you are, and I don't think it will work, those kinds of posts really don't help your case, and just looks desperate.
You know, I know enough about the module to almost envision the circuit, it all makes perfect sense, that is to say it's not built around a compromise. What I know of your amp is very little, but feel safe to say it's overly complex for what it needs to be.
How about a total part count for both modules, just to say?
Could you then explain why all those extra parts make yours better?
In my personal view is you've always had the hard case to sell in that the modules didn't use post filter feedback, you insult my intelligence with your marketing techniques and measurements, while a straightforward honest listening test would be of interest to me, the games really aren't.
In a recent post you've actually said to me that I've been "convinced" by someone that post filter feedback is a requirement. Actually I'm convinced by study, research, experimentation and obvservation. While I can't obviously get around to that with every possible aspect in class d, with enough research one can draw conclusions on certain solutions that simply make sense.
Post filter feedback is one of them, the intrinsic phase delay brought on by the coil which you claim to be a problem in the sound hasn't managed to convince me at all that I don't enjoy what I hear or that it is missing in any aspect.
Why would a higher thd signify a "problem" when it's at a power level where higher thd should be expected?
As far as marketing techniques go, I know you have the lowest THD of any known amp in the world, what happens when you turn it on?
I've also asked you in the past what you've done that's new when you were indirectly hinting you had the amp of tomorrow. So you've claimed to tackle old problems in new ways, is there a patent reference where I can learn about it? Do you care to discolose anything even vaguely as to what makes your solution better... can you give me anything at all that might sway me to your way of belief other than simply your word and wild claims which is all you've provided us with thus far?
Not to be a total bore, but who's left at LCaudio to revise the modules as problems crop up on users of these modules anyway, basically, what support will it have? Past module releases from LCaudio haven't been bullet proof.
If you think powered modules should have lower THD and anything less means there's a problem, that means there's an underlying problem that will remain uncorrrected on all your previous modules which use the same topology, but you'll sell them a kit to fix it, would you mind telling us the idle current in the kit as opposed to the standard module? You've said in the past the extra cost of said module/kit is explained by the hands on tuning of critical components, are you telling me this wasn't being done before, and you're now selling this (that being proper biasing and setting of operating points) as a feature? Lars.... Since you quite who's doing that now anyway?
I really don't expect any straightforward answers so don't worry about dissappointing me any further, we just can't have this kind of behavior here and keep it worthwhile at the same time.
I do it because I care Lars.
Regards,
Chris
Thanks for the reply Jan-Peter and I would imagine everything contributes in some small way, alot of mods lean towards limiting or lessoning the effects of EMI, I've personally found EMI highly detrimental and I can imagine that limiting it in every way possible would be worth some improvement, especially at higher power levels where it becomes more pronounced. Hence your parallel decoupling caps, coil placement, snubbing, multi layer pcb and SMT, symmetrical layout, fully balanced circuit.. from start to finish. All things to look at before even glancing at specs, it's looking smart not just shinny.
4 Revisions doesn't seem like much at all, I'm sure future revisions will push it further yet.
Have you experimented with many mosfets before settling on the philips ones? Were also chosen with EMI in mind? Or for total cost? I seems like the output stage isn't taking full advantage of the latest technology and yet maybe that's exactly what is required in a robust solution I don't know, thanks for sharing what you do anyway.
Lars, I realize why you're saying what you are, and I don't think it will work, those kinds of posts really don't help your case, and just looks desperate.
You know, I know enough about the module to almost envision the circuit, it all makes perfect sense, that is to say it's not built around a compromise. What I know of your amp is very little, but feel safe to say it's overly complex for what it needs to be.
How about a total part count for both modules, just to say?
Could you then explain why all those extra parts make yours better?
In my personal view is you've always had the hard case to sell in that the modules didn't use post filter feedback, you insult my intelligence with your marketing techniques and measurements, while a straightforward honest listening test would be of interest to me, the games really aren't.
In a recent post you've actually said to me that I've been "convinced" by someone that post filter feedback is a requirement. Actually I'm convinced by study, research, experimentation and obvservation. While I can't obviously get around to that with every possible aspect in class d, with enough research one can draw conclusions on certain solutions that simply make sense.
Post filter feedback is one of them, the intrinsic phase delay brought on by the coil which you claim to be a problem in the sound hasn't managed to convince me at all that I don't enjoy what I hear or that it is missing in any aspect.
Why would a higher thd signify a "problem" when it's at a power level where higher thd should be expected?
As far as marketing techniques go, I know you have the lowest THD of any known amp in the world, what happens when you turn it on?
I've also asked you in the past what you've done that's new when you were indirectly hinting you had the amp of tomorrow. So you've claimed to tackle old problems in new ways, is there a patent reference where I can learn about it? Do you care to discolose anything even vaguely as to what makes your solution better... can you give me anything at all that might sway me to your way of belief other than simply your word and wild claims which is all you've provided us with thus far?
Not to be a total bore, but who's left at LCaudio to revise the modules as problems crop up on users of these modules anyway, basically, what support will it have? Past module releases from LCaudio haven't been bullet proof.
If you think powered modules should have lower THD and anything less means there's a problem, that means there's an underlying problem that will remain uncorrrected on all your previous modules which use the same topology, but you'll sell them a kit to fix it, would you mind telling us the idle current in the kit as opposed to the standard module? You've said in the past the extra cost of said module/kit is explained by the hands on tuning of critical components, are you telling me this wasn't being done before, and you're now selling this (that being proper biasing and setting of operating points) as a feature? Lars....
Since you quite who's doing that now anyway?I really don't expect any straightforward answers so don't worry about dissappointing me any further, we just can't have this kind of behavior here and keep it worthwhile at the same time.
I do it because I care Lars.
Regards,
Chris
Thanks Chris! That's exactly what i have been saying.
This is true, however the instructions are now better, and this has helped the break-in situation a lot. If the modules are treated right, they are bullet proof!
But on the other hand, their warranty is bullet proof! If there is a problem with any module, it is replaced without question. Even modded ones, or where the problem is clearly caused by misuse.
As you know i no longer work in LC, and i have absolutely no commercial interest in the audio industry any longer. I am here as a hobbyist, just like the rest of you guys. Personally as a consumer i wouldn't trust a producer who says:
If you want to get a good sound from our module, we encourage you to mod it.
But if you do so, we no longer give you any warranty.
I simply fail to understand how you as a customer is willing to accept this treatment, from one of L C Audio's distinguished competitors. Of course this is just my personal view of how it looks, and everyone is entitled to their opinion. I haven't been at their board meeting to whitness when this approach was decided.
It is at half power on both modules, as claimed in the datasheet. So i don't think higher THD should be expected.
Yes if you are using the same topology, and you get more THD this means you have less control of switching times and dead time. Yes you are right, This to me says audible problems.
Does that also include the safety features, and protection circuits?
Or just the parts in the signal path?
All the best from
Lars
Hi Lars,
You've very selectively quoted and replied, I can appreciate that.
In fairness my view on the modded modules, if Hypex's views of staying neutral to keep the commercial buyers happy doesn't ever interfere with the quality of my module, then I'm fine with that.
I'd be much more upset having to pay alot more money for a signiature series which was hand biased by the master himself before being wrapped in royal satin and shipped off. Or in alot of commercial like features which would be of no use to me.
No doubt there'd be something on it I'd like to do differently, I'd rather be pulling off 2 cent caps than 20$ ones.
Your service track record has been good and I give you respect for that. With you gone just prior to the release of their latest product, a good supply of replacements might not make up for a simple revision, not where people blowing their speakers are concerned, or will LCaudio replace those as well.
Really they should have addressed such issues before they/you?? made wild claims and went to war.
Don't worry at some point I'm sure both amps will end up in the hands of reviewers, I won't care.
I remember first reading the documentation for the Zaps and it was filled with mods for this and that, honestly it made it look like one bandaid after another.. and you say you've fixed the issues with better documentation? A good product is one which is robust against foolishness all the wrong ways it can be done. Nothing is perfect within reason of course, it would be foolish to claim such things, don't you think? Hypex has one warning on their data sheet about polarity reversal, how do the Zaps fair in that arena?
As far as comparison of part counts, modulator sections vs protection et tout, why not have a contest and do both categories, can even throw in a pro and con area just for fun.
Not to correct your quote but Hypex never once claimed that modding the modules was in fact required to get good sound, nor to encourage anyone to do it. They simply said it could be done, and that in release form they're far from being "all out". They'd have to be fools to support user moded modules, perhaps in circumstances it would be reasonable of them to replace a module if the problem didn't stem from the modifications and I'm sure in such a case they would do so. How many free replacements do you feel they should supply before the average 10 year old learns ohms law and good grounding schemes?
"Good sound" will always remain largely subjective. Why would I want to pay extra for what someone else thinks is best?
I'm not much of a lego block kind of hobbiest so I don't mind burning my finger tips either and I don't require step by step photo's to get it done either. I think I stand a pretty good chance at modding it and having it work, looking forward to the parts arriving, if this is mad I've seen worse. I'm mad my soldering iron is a piece of **** that was fixed with a revision that's sold as a new version, same iron only now it's user rebuildable with a parts list for it, since it keeps breaking. That make me mad. If only commercial guys could get the rebuildable version.... that would make me mad too. Spending 10$ on the parts to fix it myself and taking the 2 minutes to get it done, instead of 120$ on a new iron each time it breaks, yeah I won't mind that.
Let's put the specs away and talk circuits or something interesting what say?
You've very selectively quoted and replied, I can appreciate that.
In fairness my view on the modded modules, if Hypex's views of staying neutral to keep the commercial buyers happy doesn't ever interfere with the quality of my module, then I'm fine with that.
I'd be much more upset having to pay alot more money for a signiature series which was hand biased by the master himself before being wrapped in royal satin and shipped off. Or in alot of commercial like features which would be of no use to me.
No doubt there'd be something on it I'd like to do differently, I'd rather be pulling off 2 cent caps than 20$ ones.
Your service track record has been good and I give you respect for that. With you gone just prior to the release of their latest product, a good supply of replacements might not make up for a simple revision, not where people blowing their speakers are concerned, or will LCaudio replace those as well.
Really they should have addressed such issues before they/you?? made wild claims and went to war.
Don't worry at some point I'm sure both amps will end up in the hands of reviewers, I won't care.
I remember first reading the documentation for the Zaps and it was filled with mods for this and that, honestly it made it look like one bandaid after another.. and you say you've fixed the issues with better documentation? A good product is one which is robust against foolishness all the wrong ways it can be done. Nothing is perfect within reason of course, it would be foolish to claim such things, don't you think? Hypex has one warning on their data sheet about polarity reversal, how do the Zaps fair in that arena?
As far as comparison of part counts, modulator sections vs protection et tout, why not have a contest and do both categories, can even throw in a pro and con area just for fun.
Not to correct your quote but Hypex never once claimed that modding the modules was in fact required to get good sound, nor to encourage anyone to do it. They simply said it could be done, and that in release form they're far from being "all out". They'd have to be fools to support user moded modules, perhaps in circumstances it would be reasonable of them to replace a module if the problem didn't stem from the modifications and I'm sure in such a case they would do so. How many free replacements do you feel they should supply before the average 10 year old learns ohms law and good grounding schemes?
"Good sound" will always remain largely subjective. Why would I want to pay extra for what someone else thinks is best?
I'm not much of a lego block kind of hobbiest so I don't mind burning my finger tips either and I don't require step by step photo's to get it done either. I think I stand a pretty good chance at modding it and having it work, looking forward to the parts arriving, if this is mad I've seen worse. I'm mad my soldering iron is a piece of **** that was fixed with a revision that's sold as a new version, same iron only now it's user rebuildable with a parts list for it, since it keeps breaking. That make me mad. If only commercial guys could get the rebuildable version.... that would make me mad too. Spending 10$ on the parts to fix it myself and taking the 2 minutes to get it done, instead of 120$ on a new iron each time it breaks, yeah I won't mind that.
Let's put the specs away and talk circuits or something interesting what say?
No, Bruno did choose for the Philips 1,5 year ago, and we did not yet tested other Fets. Although Fairchild has now a very nice new one, FPDP61N200. Suppose we would use this Fet we can increase the power in lower impedance, however we have to redesign the output choke for that. I have already ordered some samples of 300V Fets from Ixys, the design is working so nice that I am planning to see how the performance is by 2 x 125VDC.
Actually we have learned so much new things that this will open a path for true High Power UcD amplifiers. No plans AT ALL for a product, but it’s nice to play with even higher power …
What do you mean with: “isn't taking full advantage of the latest technology”?
For the both auxiliary supplies I would go for a current 250mA pro voltage. The heatsink of 200W will be already an overkill, we still test the UcD700 in free air on a plate of aluminium of 200x200m, 3mm thickness.
Jan-Peter
Jan-Peter said:
No, Bruno did choose for the Philips 1,5 year ago, and we did not yet tested other Fets. Although Fairchild has now a very nice new one, FPDP61N200. Suppose we would use this Fet we can increase the power in lower impedance, however we have to redesign the output choke for that. I have already ordered some samples of 300V Fets from Ixys, the design is working so nice that I am planning to see how the performance is by 2 x 125VDC.
Actually we have learned so much new things that this will open a path for true High Power UcD amplifiers. No plans AT ALL for a product, but it’s nice to play with even higher power …
What do you mean with: “isn't taking full advantage of the latest technology”?
For the both auxiliary supplies I would go for a current 250mA pro voltage. The heatsink of 200W will be already an overkill, we still test the UcD700 in free air on a plate of aluminium of 200x200m, 3mm thickness.
Jan-Peter
Five years... it was a vague way of saying maybe with a newer and better Fet you can get away with less of them, and possibly simplify other areas like the schottky's and maybe other snubbing? Smaller, cheaper and more efficient, just taking advantage of newer and better technology that way.
Don't worry I realize R&D doens't come in a free catalogue.
I looked at IXYS stuff before and found they'd be hard to drive well requiring large gate currents, how are your samples in that area?
The redesigned choke worked out well for the 400, it might be a nice option here too, your later revisions may be smaller and more powerful but low cost and robust to start with was a fine choice by me.
I'm glad you're getting it out there, and it's nice to know that you as the manufacturer of the module doesn't think it's perfect and can never be improved on, or waste my money on features that aren't of use to me or of sound design, or charge me extra for setting the bias properly and calling it a high end version. It's exemplary keep that up
Also, I think I'd prefer more robust unit so I can worry about having to baby it less and in return can easily put up with simplified single page datasheet /user instructions.
I hope you meant secret circuits and not "secret parts"... that lends to magic and voodoo and ..... I know you meant circuits.
Regards,
Chris
phase_accurate said:
That thought crossed my mind. Wanted to try one of those drivers since I first heard of DSD and saw their 5MHz class D eval., very cool read.
I see your point about the PCB layout being considered a secret part, and it obviously makes it or breaks it so there's no arguing with that logic. I just wanted any myths about glowing beads made out of organic unobtanium that absorbe all nerby radiation or any such thing out of the way before they began, and before the commercial guys ran away with it as their theme song.
Regards,
Chris
Actually the both is true...
(but indeed you are right, it should be circuits)We don't like voodoooooo in our design, we try to keep it scientifically
The gate charge are more of less the same as the current used Fet. So this will not be the problem anymore....
Promised: when I have something interesting to show about UcD+Ixys I will post it.
Jan-Peter
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.