| opp |
Hi – tried to make this tread Link come a live in the loudspeaker forum, little success though. I am new to this forum, hope you will bear with me, if this “reposting” is improper.:cannotbe:
As DIY’er I am always trying to optimize my investment in Hi-fi. (guess that’s a common goal /necessity for all on this forum). Amplifier design is a question about compromises, so my reasoning is: if you have an active xover, it must be possible to optimize the amplifier choice for each frequency range.
I have a renovated Quad 405, UCD 180 and a 10 year old NAD 30 watt, for a 3-way active system. It is the latter that has to be replaced.
My hypothesis is that the treble need special attention, on the other hand the power demand is low. It must be possible to find a most favorable amplifier for this purpose – that is really the topic.
I imagine a small high quality class-a (10-20 watt), is that a good choice? If yes – can someone recommend a design with a commercial available PCB?
Or should I go for another design? |
|
|
| darkfenriz |
| quote: | Originally posted by opp
I imagine a small high quality class-a (10-20 watt), is that a good choice? |
Probably good choice, but it is frequently overlooked that despite low RMS or average power, trebles sometimes demand high pulse current.
So it is good to have some power/output swing margin for tweeter amp. |
|
|
| amplifierguru |
Hi opp,
You are correct that amplifiers can be chosen on their merits for the respective frequency band - big heavy duty AB for bass, medium finessed AB for mids, and smaller (but enough for transients) low distortion for top.
I would apportion say 200W 4ohm BASS, 150W 8ohm MID, and 50W treble.
Your BASS amp should be clean clipping and have competent PS and heatsinking for continuous duty.
The MID amp can be more finessed for lower duty but clean with good PSRR and low THD.
The treble amp could be a good quality AB amp with no significant complex non-linearities and low THD, or a Class A.
Hope that helps,
greg |
|
|
| bogicp |
hi opp,
I'm also new "in town".
One thing that is in my mind when I read your message, is that you must choose tweeter driver first. Not all tweeters can be driven with 10-20 watts with succesful sound pressure level in respect to other drivers... if you cannot define tweeter driver you must expect more than 20W power for it.
Best regards
-boggy |
|
|
| opp |
Yes - the tweeter sensitivity is an important detail, as bogicp points out. The tweeters will be Vifa XT25 or maybe the new Seas Crescendo, both with a sensitivity around 91-92 dB, the woofers is approx. 88 dB. This fact (together with the lower power demand I the treble) points towards a small amp. But the argument that treble demands high peak current/voltages points in the opposite direction.
Maybe a compromise is a design with nearly the same voltages/current level as the woofer amps, but equipped with transformer and cooling according to the actual needed average watt-effect (the expensive parts). The demand for peak current could be handled by large reservoir caps – would this work? Well - probably not with a class-a design - or?:confused: |
|
|
| bogicp |
hi opp,
I probably speculate, but you cannot easily determine that you need only current... Class A has problem with overheat if you use larger supply voltage... but if you only need current I think that you can obtain good transient response with good capacitors (you must have extra supply caps, polypropylene 1u near to output transistors). Also, you must do PCB design very carefuly. Try to keep this tweet amp small as possible, with short and thick lines, this is only way to obtain good transient response and not degrade good components with bad PCB. But you probably know, these days, chips is better way to obtain good transients at high frequencies... they don't have that big PCB or wiring problems, all critical things are in chip :)
Hope this helps.
Best regards
-boggy |
|
|
| opp |
I definitely get some very fine inputs in this forum.:happy1:
Do you think a chip-amp can bee superior to a class-a? Is Hypex UCD180 a chip amp in this regard – it certainly is small and sounding excellent? |
|
|
| bogicp |
I don't know, you must read specs, possibly you may work on some simulations, and build prototypes... then hear it all and decide...
no one in this forum (and world) has same way of thinking, and this is only way to do some better things... look what other works, but think with your head :)
First thing you can do is some googling for "signal integrity", learn this carefuly, and try to implement some of principles in high speed design in your tweeter amp, and you will hear difference.
Best regards,
-boggy |
|
|
| opp |
Have been away googling - thought-provoking for me.
Tanks - boggy
:wave: |
|
|
| bogicp |
No problem, :) one thing is in my mind also... what do you think about building not-that-big-power but very FAST tweeter amp, with bandwidth wider than 100kHz, and mount it on tweeter magnet?
Think about phase in broad range amplifiers, can we hear differences?
Best regards,
-boggy |
|
|
| opp |
| What’s really worrying me at the moment is the active filter. I have build a 24dB Linkwiz, when first measuring it, I thought I did something wrong. A pulse at the input was not at all summing up to a similar pulse at the output (summing bas, mid and treble). PC simulation gave the same result, so my circuit is OK. Still this active system delivers the best sound I ever had.:confused: When the tweeter amp. is in place, the next project is a digital xover, I hope it can deliver good pulse response. |
|
|
| bogicp |
Look at http://www.ocf.berkeley.edu/~ashon/...e/phaseaud2.htm
Later added:
http://www.audioholics.com/techtips...seDistortio.php
there is some words about phase distortion.
As I see, you use LR4, fourth order Linkwitz-Riley filters, in your crossover... and they do phase distortions, you cannot obtain regular impulse response with fourth order analog filters (or digital IIR implementaion.. this is same problem...)
Some people think (hear?) that this phase deviation is audible, some not...
With digital FIR you can obtain expected impulse response but drawback is pulse "ringing"...
When you do impulse response equalisation you insert delay in signal, because all phase/delay differences will be equalised to worst in range,.... when you look at low frequencies, delay is about 10ms...
and if you use your speakers in some real-time work (playing, mixing, mastering....) you can feel latency. etc...
try some googling about "loudspeaker phase distortion" and you probably find that your crossover work good... but who knows? :)
best regards
-boggy |
|
|
|
