Lars,
Very good to change the pricing system. Will order from you after the 30th of June, being the date for the change. I have the UCD's 180 and 400's but I would like to listen to your Zap 2.3SE also.
I always like fairness in sales, if possible, and you listened to your custommers which is the way it should be.
Very good to change the pricing system. Will order from you after the 30th of June, being the date for the change. I have the UCD's 180 and 400's but I would like to listen to your Zap 2.3SE also.
I always like fairness in sales, if possible, and you listened to your custommers which is the way it should be.
fiak said:Hi everybody.
Going back where everything started.
Yes, it deserves to pay a little more for the ZapPulse.
A simple look at the modules is enough.
Anastasio.
So what looks so much better on the ZAPs??? The varistor that is needed to keep things safe (not needed on UcD)? The high-end caps? I think the comparator is probably pretty cheap was that an LM319?? They seem to cost less than a dollar. On the other hand the AD8620 on the UcD costs lots more, even the standard OPA2134 is more expensive.
So please explain what looks so much better on the ZAP so I can understand your statement.
Best regards
Gertjan
Hi Gertjan.
What I mean is that the Zap module looks to be more inspired where the UCD one looks computer disigned.
I also mean that there are dedicated fans to both copanies, maybe we can arrange for a football meeting.
Personally I am grateful to both HYPEX and LCAudio because amongst others they offer a product to the
DIYer not particularly involved in amp disign to enjoy HIGH END quality music, if the rest of the equipment support that.
Anyway we must answer Gregs question about compatibility and whether there are any other similar products.
My best, Anastasio.
What I mean is that the Zap module looks to be more inspired where the UCD one looks computer disigned.
I also mean that there are dedicated fans to both copanies, maybe we can arrange for a football meeting.
Personally I am grateful to both HYPEX and LCAudio because amongst others they offer a product to the
DIYer not particularly involved in amp disign to enjoy HIGH END quality music, if the rest of the equipment support that.
Anyway we must answer Gregs question about compatibility and whether there are any other similar products.
My best, Anastasio.
Lars Clausen said:Gertjan: There are no LM319's on the ZAPpulse
The comparator is a balanced discrete circuit, not an IC.
If I remember correctly, in the zappulse 2.2SE there was an opamp or comparator that receives the input signal. I also think that comparator opamp got feedback via a 1M resistor from the output stage before the LC filter. I think I found it back, that was an LM6172, corrrect? I can't check it since I don't have the modules anymore. With the 2.2SE, I had too much noise for my active speakers, it seems you have solved that now with a low noise version. I also did not like the switch-off behavior that you have probably solved as well in the 2.3SE by automatically stopping the oscillation when powering down, correct? So the issues that I had are probably non-existing in the newer version.
Anyway, at hat time, the zappulse did not work out for my specific application so I used the UcD180 for my mids and tweeters. At that time, those UcDs were the most suitable modules for me.
I do like it that you seem to use good quality caps, that is a nice touch. However, why don't you take feedback after the LC filter (patent issues?), putting the LC filter in the feedback loop has many advantages such as a more constant output impedance that stays low up to much higher frequencies and no issues with the Q-factor of the LC filter that depends on the impedance of the load. I know your bandwidth is wider, but this comes at the expense of more HF output as the oscillation frequency is less suppressed. Is such a wide bandwidth needed? Most tweeters don't do that much above 20kHz and CD players give no info above 22kHz, although some players try to generate some info beyond 22kHz such as the TEAC players. SACD goes beyond 22kHz but also drops of slowly above 22kHz.
Anyway, I have listened to your modules on my passive B&W CDM9NT speakers. Your modules sounded good, better than my conventional amp better than my Tripath based Marantz. I could not clearly say which was better UcD180 or ZAPpulse, I did not listen long enough since my priorities were with the active system.
Thanks and best regards
Gertjan
Gertjan: About why i take feedback before the LC filter, and not after.
My base philosophy (who everybody can see on my homepage) and it applies to both conventional and Class D amplifiers is:
Keep any feedback loop delay as low as possible.
This way you can achieve an open, and alive playing style, which i like. You will also get a much more airy and liquid top compared to any amplifier with a 'slow' feedback loop. Simply because your feedback loop will compare what comes out of your amplifier with what will come out in 500 - 2000 nS. So your feedback loop will compare the signals on two different spots of the time line of the audio signal. The longer delay, the more the top sounds like a plastic cup, not like a metal high hat for example. And amazingly enough, also the dynamic of the bass gets flatter with longer feedback delays.
I would rather listen to that one that compares with 500 nS delay, than the one with 2000 nS!
Adding the LC filter in the feedback loop vastly adds to feedback delay time, and it shows in the sound. I do admit that the post LC feedback scheme does give a more dampened midrange, but this doesn't really appeal to me as much as the airy top.
Another thing is load stability. It is correct that a post LC feedback will compensate for load impedance and capacitance.
Until a certain level. If you add more capacitance to the load, it will affect your switching frequency, until it gets so low, the choke starts to saturate, and the amplifier blows. A pre LC feedback like my own product can not 'see' the load, and you can add any amount of capacitance to the load as you prefer, it will remain stable.
The feedback loop can 'see' the load.
Any amplifier analog or digital that relies on feedback that can 'see' the load has a flat and rubbery bass reproduction. Whereas a system where the feedback can not see the load has a upfront and dynamic bass reproduction. This is what i prefer, but of course other people like the flat and rubber-like bass more. It's a matter of taste.
Just my experience from 20 years of designing audio amplifiers for a living.
If you want to play with this theory, try the simple circuit below. In pos 1 the feedback loop can see the load, and reacts accordingly. In pos 2 it can not see the load (so much) and you will get a more dynamic bass. As long as you are within clipping level, the damping factor of the system will be the same. But the bass reproduction will change a lot.
My base philosophy (who everybody can see on my homepage) and it applies to both conventional and Class D amplifiers is:
Keep any feedback loop delay as low as possible.
This way you can achieve an open, and alive playing style, which i like. You will also get a much more airy and liquid top compared to any amplifier with a 'slow' feedback loop. Simply because your feedback loop will compare what comes out of your amplifier with what will come out in 500 - 2000 nS. So your feedback loop will compare the signals on two different spots of the time line of the audio signal. The longer delay, the more the top sounds like a plastic cup, not like a metal high hat for example. And amazingly enough, also the dynamic of the bass gets flatter with longer feedback delays.
I would rather listen to that one that compares with 500 nS delay, than the one with 2000 nS!
Adding the LC filter in the feedback loop vastly adds to feedback delay time, and it shows in the sound. I do admit that the post LC feedback scheme does give a more dampened midrange, but this doesn't really appeal to me as much as the airy top.
Another thing is load stability. It is correct that a post LC feedback will compensate for load impedance and capacitance.
Until a certain level. If you add more capacitance to the load, it will affect your switching frequency, until it gets so low, the choke starts to saturate, and the amplifier blows. A pre LC feedback like my own product can not 'see' the load, and you can add any amount of capacitance to the load as you prefer, it will remain stable.
The feedback loop can 'see' the load.
Any amplifier analog or digital that relies on feedback that can 'see' the load has a flat and rubbery bass reproduction. Whereas a system where the feedback can not see the load has a upfront and dynamic bass reproduction. This is what i prefer, but of course other people like the flat and rubber-like bass more. It's a matter of taste.
Just my experience from 20 years of designing audio amplifiers for a living.
If you want to play with this theory, try the simple circuit below. In pos 1 the feedback loop can see the load, and reacts accordingly. In pos 2 it can not see the load (so much) and you will get a more dynamic bass. As long as you are within clipping level, the damping factor of the system will be the same. But the bass reproduction will change a lot.
Attachments
Lars Clausen said:
Hi Lars,
ABout your base philiospy of keeping feedback loops short. OK, the explanation sounds convincing. However, my conventional Accuphase e407 which is considered a high-end amp by many, and sure the price was high-end, has a very wide freq. response and therefore fast feedback but any class D amp that I have tried sounds better than the Accuphase. How to explain this???
Class D has something magic that other amps don`t have and I don`t think it can be explained by the speed of the feedback loop or frequency response extension.
In fact, using the UcD180, the highs sound incredibly real (on my active system) so I don`t have the feeling that super fast feedback and extremely extended freq. response plays the most important role there since my Accuphase has for sure a more extended frq. response than the UcD but the UcD sounds better. Even the Marantz Tripath based amps sounds better to me than the Accuphase (but more noise on the Tripath) and your ZAPpulse sounded better than the Tripath.
I think the problem that you mention on capacitive loads causing the oscillation frequency does not hold for the UcD since the oscialltion frequency is not determined by the LC filter but by propagation delay in the active components and the phase compensation in the feedback loop. Maybe you can get a hand on the AES paper of Bruno where he explains that as far as I know.
About the rubbery bass!!!. I now use the Tripath based marantz for my woofers. Tripath has no feedback after the LC filter. UcD has feedback after the filter. I plan to use UcD400 in bridge for my woofers anytime soon. Of course it is a completely unfair comparison, but I expect that the UcD will give a tighter/better controlled bass.
To be honest, I can hardly believe that feedback before or after the filter makes such a big difference in bass. We really should then come up with a measurement method that can identify it. Some kind of dynamic measurement, for example sine bursts or something???
I think your example with the switch is a bit strange, in pos 1 there is a 8 ohm resistor in series with the speaker and you fully rely on feedback to reduce it. In pos 2, you have shorted the resistor. Of course pos 2 will be better. Am I missing something here?
Thanks and best regards
Gertjan
by the way, If you have any idea why any of the class D amps I have heard so far sound better than conventional amps, please let me know. Really interested in knowing that.
IMO this statement is too restricted and therefore leaning towards the incorrect side.
I'd rather say that the sound of an amp is better if it is not SOLELY depending on NFB (and that accounts for even more parameters than just its output resistance).
The intrinsic output resistance of switching amps is very low. At low frequencies this accounts for both pre- and post- filter NFB. IMO this is one reason for the effortless bass of class-d amps.
Pre-filter NFB doesn't take filter nonlinearities into account and is giving unpredictable FR at the top end.
Post filter NFB doesn't necessarily make the NFB - loop slower if done correctly.
As soon as you have a class-d amp with post-filter NFB takeoff on offer I will buy a pair of modules from you !
Regards
Charles
Charles: Sure you are absolutely right, my statement is only accurate in context of this discussion, and i admit it might be slightly over-simplified.
In general your version is of course adresses a much broader area of problems
If i can make a post-NFB amplifier that still sounds good in the top and bass i will, but by now i have not been very impressed. I use a CD from Niels Lan Doky to test the tops. It has some nice rolling cimbales. On the post NFB they sound like the treble membrane is a plastic cup. As does any other clear high-hat sound from any recording. But on a pre-NFB it sound exactly how i want it, you can hear the brass and with a seamingly infinite air in the top.
But of course i am trying out every option to constantly improve my amplifiers.
Like my boss said: Remove as much THD as you can, but please don't destroy the good sound ..
In general your version is of course adresses a much broader area of problems
If i can make a post-NFB amplifier that still sounds good in the top and bass i will, but by now i have not been very impressed. I use a CD from Niels Lan Doky to test the tops. It has some nice rolling cimbales. On the post NFB they sound like the treble membrane is a plastic cup. As does any other clear high-hat sound from any recording. But on a pre-NFB it sound exactly how i want it, you can hear the brass and with a seamingly infinite air in the top.
But of course i am trying out every option to constantly improve my amplifiers.
Like my boss said: Remove as much THD as you can, but please don't destroy the good sound ..
Charles: Actually i am my own boss, but my company is owned by an investment company, that also owns other well known audio brands, so that's why the story about 'my boss'. Actually he is the chairman in our board, and has nothing to do with the daily business in LCAudio. But anyway the point was what he said
About the 'half pre and half post' solution yes i have experimented with that, but i got a quite unstable amplifier out of it. It seems it didn't really like the phase shift in the output filter.
Do you have any good ideas?
About the 'half pre and half post' solution yes i have experimented with that, but i got a quite unstable amplifier out of it. It seems it didn't really like the phase shift in the output filter.
Do you have any good ideas?
Charles: It appears to be quite simple to move the feedback loop on existing ZAPpulse modules, to post LC filter mode. You only have to change a few components. So with a little tampering you can actually try it out yourself, and see which sound fits most to your liking. Pre filter NFB or post filter NFB.
However we will NOT change our module to post filter mode by default, simply because they (IMO) sound better with pre filter mode. (Like they are now).
BTW have you been working on the hybrid loop? And if so, please let me know where you are posting. Very interesting!
However we will NOT change our module to post filter mode by default, simply because they (IMO) sound better with pre filter mode. (Like they are now).
BTW have you been working on the hybrid loop? And if so, please let me know where you are posting. Very interesting!
Which is better????
I can honestly say that neither is. They are diferent, period.
As for comparing them side by side..........same PSU, etc...........
Can be misleading.
I know that making very small changes to the HF roll-off.........say something in the area of 0.1 to 0.25 at 20 kHz.........can make a radical difference in the perceived sound. Could account for some of what Lars is describing in another thread.
I know, just don't ask to me 'splain how I know. OK?
Jocko
I can honestly say that neither is. They are diferent, period.
As for comparing them side by side..........same PSU, etc...........
Can be misleading.
I know that making very small changes to the HF roll-off.........say something in the area of 0.1 to 0.25 at 20 kHz.........can make a radical difference in the perceived sound. Could account for some of what Lars is describing in another thread.
I know, just don't ask to me 'splain how I know. OK?
Jocko
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.