CFA vs VFA and marketing bs

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I saw Marantz is pushing CFA as a superior design now with a larger feedback loop lower distortion etc. I don't remeber if it was TNT or that Italian webpage but they compared a CFA to a VFA but added a few things like current mirrors and CCS with data supporting the opposite. Using the same components will the benefits of a complex VFA circuit be worth the extra work and will it be able to push heavier loads?
 
I could not resist .... :D

"will it be able to push heavier loads" (BS #1).

That , depends on the current stage design (EF2/3 BJT / LFET-VFET) , how many pairs and device SOA.

Superior design ( semi BS) :) ....
Some VFA's can exceed 200V/us (fast slew).

I have just noticed subtle differences between a fast VFA and a CFA. This seems to
be speaker dependent. The CFA is more sensitive to impedance changes in
the loudspeaker /X-over it is driving. This can be good if the CFA-speaker
combo is permanent. Not as good if you are testing many speakers on one
particular CFA.

Some CFA's are so fast as to sound brighter and more detailed in the HF region.
The differences are SO subtle as to just require a slight tweak to the DSP source.

Conversely , a multiple pair output stage with a >800VA power supply will
sound Superior (faster , more detailed) regardless of VFA/CFA.

A CFA usually has less components (so what). After many CFA's and VFA's
- the lines have definitely blurred .

I'd say Marantz just wants to sell amps. I'd also say the forum's "Badger" (VFA) would
sound comparable to the best marantz.

OS
 
Administrator
Joined 2004
Paid Member
Hi Pete,
I also can't resist a comment.

Hi Jared,
The quality of an amplifier depends greatly on PCB layout and build quality. You can have identical amplifiers going by the schematic. They could easily sound totally different even if they used the exact same parts too. So much depends on things not in the specs or even service manual that the other argument of CFA vs VFA is completely irrelevant.

Having said that, I do have to say that Marantz of old and Marantz of recent history has made some of the nicer amps being sold. Also, look at Ostripper's designs as I believe they are capable of at least equal performance to most commercially available amplifiers. The Badger is supposed to sound excellent. I haven't heard one myself, but I have heard the same thing about it from several other members.

-Chris
 
Well the main amps that are a primary concern are actively crossed lcrs. So i need 9 more amps. 3 500 watt and 2 150 watts (4ohms). The speakers are very sensitive and supposedly for highly sensitive drivers the amp has to be vast. The nice amp i have now is a lfet design - was a kit so it's hard to know the circuit. PCB wise I am probably only putting the psus on them - point to point on the amp unless I'm using an omp amp. I'm better with microcontroller and processors and want to stay away from dsp but still control parameters. Would you recommend a tone control circuit for the high frequency driver?
 
Administrator
Joined 2004
Paid Member
Hi Jared,
Before you get all excited here, you should consider one important thing. S/N ratio. You are planning to have a number of very high powered amplifiers connected directly to the drivers - so no loss in any crossovers. To add to this problem that you don't know you have yet, you are using very efficient drivers.

Look at common signal to noise ratios and figure out what that will be in terms of noise power, or voltage. That's one problem with mating high powered amplifiers with high efficiency drivers - using crossovers! You are running direct, no losses in a crossover to help out here.

Next issue, turning all that on and off along with the source components. You don't want to get that wrong, and you don't want to run it all the time. As well as the nightmare of a power fail when the system is running.

Lots to think about. I have experience with customers who bought Carver PM and M 1.5's along with Klipsch La Scala speakers. They all complained about the hiss level. That was only in stereo mode driving horn speakers with the factory crossover in place. Your system will be much worse for noise. Just give it a thought before diving in with both feet. To be honest with you, amplifiers that deliver 100 watts would be more than enough.

Another thought. If you get the entire system cranked, how much current does everything draw? :yikes:

-Chris
 
Well I have a 500 watt amp (4 ohm) that's awesome. It's has a 800va toroid with additional windings to power the micro controllers and processors I set up to handle things like soft start crossing etc temp shut off DC sensing - etc. the 500watt is for 2 10s at 8 ohms. They can actually handle 800watts at 99 db efficient so figured 50o was more then enough and allows me to run them serial or parallel. Also figured the lower the power levels on a high powered amp would minimize the the nonsense with signal to noise. The 5.25s are 75 watt peak at 8 ohms so again that's how I came up with those numbers. The tweeter is 75 watts. So figured same amp for the 5.25s will be fine. I know the drawn will be heavy and each one of these speakers will have a separate breaker in another sub panel with isolated separate ground from the rest of the house. I have a few other things I added to it. I can voice a speaker and sub - good with automation type things also and plan on adding a few other cool things on there. Yea I'm going over board with power since 256 watts will give me 124 spl but it is a special project for a special application. All the amps will have a master controller - all parameters will be controlled via ip but with no DSP. Hit you on PM and we can chat a little more about it.
 
Administrator
Joined 2004
Paid Member
Hi Jared,
Higher power always makes the level of noise worse for an equivalent S/N rating. Every single time.

Most drivers cannot handle the power they claim to. It's an old story that keeps repeating itself. A 1" voice coil can withstand about 30 ~ 35 watts of constant power - and that's with aluminum formers and high temperature epoxy. If you can find out what the voice coil former is made from, and the diameter of the voice coil, you can figure out what the true power handling capacity will be ... best case. Ever see a real Dale 250 watt, 8 ohm dummy load? Those must be mounted on a heat sink to be able to dissipate that amount of power. If those speakers could do it for real, we would be using those as dummy loads (no cone and contact voice coil gap). We aren't.

Is that 800 watt amplifier a stereo unit by chance? If so, that's 1,000 watts out of an 800 VA transformer. Not accounting for losses of about 50% (making the total power 2,000 watts), how can you extract more power from an amplifier than it takes in from the AC power? If it is a class D amp, your losses go way down, but you still have them (losses).

Now for the really ugly truth. With a speaker, when you get it hot the IR losses go up. A lot. You can lose 2 dB, maybe more. That has got to put a cramp in your style!

To create high SPL levels, you need to run lots of smaller, efficient speakers. A wall of horns would be the ticket, like the ones used in rock concerts of old. I haven't been to a concert since the early 1980's, so I can't say much about now. You're trying to create concert levels, you're going to have to go about it in the same way. It's no mistake that concerts are set up the way they are. Hope that helps. Keep the VC temperature down!

-Chris
 
When building amps, generally there is importance in being able to drive a dynamic impedance without placing the non-linear loads onto the VAS. CFA may be more sensitive to this reflected non-linear load Z, IOW, the speaker. I find that complementary push pull CFA works great, with some specifics. :spin: Perhaps maybe common base or common gate VAS. J-fets are nice.:hug: Certainly a relatively low Z (reference to IPS currents) global loop. Maybe even throw in some common mode loops and some CCS's and such; if done analytically, to relatively equalize the Gm of the two halves, positive and negative components, maybe reduce some asymmetrical distortion components at the expense of some gain. Plenty of gain in SS components.;) We all know that the Gm curve of a N-ch Jfet is not going to be equal to a P-ch Jfet even of relatively similar overall Gm. How to reduce the differences.....:scratch2:

Driving an OPS that exhibits a constant, very light load to the VAS regardless of the ever changing load impedance is an advantage, especially with CB or CG arrangement. Something like a "gain cell" follower, a power buffer circuit, local error correction and all that stuff contained within and such and yada yada, ect ect, so on and so forth.....:scratch1::scratch2::yes:
 
Disabled Account
Joined 2004
I found this interesting as well as a touch bizarre.

Schmid “Why ‘Current Mode’ Does Not Guarantee Good
Performance” (invited after an ISCAS talk in May 2002)

... deconstruction of the term “current-mode” by
showing that it is empty of inherent meaning and only has meaning
in the context of science as a social process.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.