That is the case with the ND. Servo changes the diamond currents , gives
me 6db more OLG with 3-4 mhz ULGF.
I'm running 27pf caps at the VAS (3.4mhz ulgf) , rock stable. I tried 22pf
with just a hint of overshoot on the square wave.
VFA's can barely be pushed past 1mhz without adding a dash of lead compensation to adjust phase at HF.
An example is the Symasui - 39pf miller equals 1.4mhz. But , you need
2.7p - 3.3p (at NFB - lead) to keep margin and overshoot acceptable.
OS
Well, it makes sens. And, you know, I'm everything but an expert in simulators. Do you think it makes a real difference in comparison and can lead to opposite results IRL ? (I don't care here about the absolute values, only relative differences).
Anyway, it is an old story, for me ;-)
Last edited:
I use this since decades: Miller for max but flat bandwidth at HF with no bump in the response curve. Then low pass filter in input to kill *any* overshoot on square waves. Best for both stability and TIM.
After this, I often reduce the corner frequency of the input filter further, for best results in listening.
(if I try max MHz *in the loop*, I don't want any amp to pass signals at very high frequencies that, anyway, I cannot hear. See what I mean ?)
For CFA-ND amplifiers.... instead of the 12V Zeners setting the the front-end input circut + servo voltages.... if we spend the money for a transformer based regulated +/- 15V power supply..... it could also power options:
--- high quality power up and speaker protection circuits
--- optional D-to-A converter.....optional digital Xover
--- optional analog Xover opamps
--- Integrated preamp with volume (and tone) controls
------Remote control
For VFA-Symetri.... there is still a small JFET cult.
--- high quality power up and speaker protection circuits
--- optional D-to-A converter.....optional digital Xover
--- optional analog Xover opamps
--- Integrated preamp with volume (and tone) controls
------Remote control
For VFA-Symetri.... there is still a small JFET cult.
Attachments
Oh, thank-you very much. May-I add your image (with a credit, of course) ?
In fact, I just added this because some inquisitor of the LTP church pretended that VFAs are noiser. Witch is stupid: same gain in IPS, same currents one transistor VS two.
Go one step further and use LDO regulators for the front end and servo.
Have you thought of the timing issue ? Power off especially ... what if
the 12V supplies collapse before the VAS/OPS rails do ?
I have considered adding the excellent DC and over-current protect
from Vzaichenko's "21st century" to a complete ND-OPS offering.
Just analog .... the micro might scare off some builders.
The Symetri , same deal. I think it would be good to have the PCB
powered by +/- 60V plus just a single 12V for the protection (SS relay ).
I will start on the power stage part of the PCB. Would 76 X 248mm be
good (Badger size) ?
The Badger has a lot of extra space in it's layout , no problem fitting
quite a few additional refinements.
PS- Yes JFET's for the Sym - Symasui simulation is just as impressive
with FET's ... 12V Cascode makes for a perfect swap.My Sym was
designed for either BJT/FET.
Edit - Line ??? I just saw it - you combined the spook/symetri , and added jfets.
I actually did this while away for the spooky offering.
As I previously stated , the symasui is 1/2 the spooky. Better odd order cancellation
and PSSR ,less parts. Clips alright ... the spooky with a symetri VAS would do better.
The spooky/symetri don't cancel odds as good as ND/symasui (cancel's ALL).
OS
Last edited:
You could have a look to this too:
www.esperado.fr - Protection d'amplificateur ultime
Don't be impressed by the apparent complexity, there is a lot of added features in it: Soft start and silent stop, temperature warnings and fans managements etc.
(May-be it was a mistake to satisfy all the requests in this forum?)
Just play with the menu "clear all/See all" (and the inputs in between) to see the various parts.
First, clear all, and add/remove each function's parts you want to see together.
If you want only the protection, a single quad OPA (TL074) + one extra transistor do the protection job + soft start in my personal amps. A microscopic board and a mechanical relay.
The advantage of this protection is the speed and universality. Whatever the problem, Clipping, DC, Oscillation, short circuit, DC from your preamplifier, it fires in µs.
And you will never be able to hear any commutation noise when power on, power off, even without the soft start circuit, because, if any, the protection itself will fire before your speaker have time to move.
No other system, based on integration of the output signal is able of this speed, of course. Neither able to protect your tweeters against any amp oscillations at MHz, even little.
An other advantage is it is universal, don't need any modification of the amp boards, and can be, even, run in an external box, without even opening the amp's one.
I had used-it in a BIG PA system during years -you know the worse situations you can imagine- and It had always saved both my amps and the speakers, despite the big margins i took to reduce the risk of unwanted lose of service.
The idea is original, i would be proud if it inspire-you.
Last edited:
But esperado - simple (below - 4 trannies) !!
Just enough to save a speaker from DC or overcurrent.
-Powered by the rails.
-one 5-12V aux supply (relays)
-NO IC's.
Just my posted DC sensor with a 3 input diode OR gate -
delay/DC/current. This triggers the SS relay.
Soft start is a separate PCB.
Diodes , mpsa92/42's , caps and the FET's.
Nothing to scare first time builders off .
PS - been using this circuit for 8 months , speakers never moved
more than a mm before the relay activated.
OS
Just enough to save a speaker from DC or overcurrent.
-Powered by the rails.
-one 5-12V aux supply (relays)
-NO IC's.
Just my posted DC sensor with a 3 input diode OR gate -
delay/DC/current. This triggers the SS relay.
Soft start is a separate PCB.
Diodes , mpsa92/42's , caps and the FET's.
Nothing to scare first time builders off .
PS - been using this circuit for 8 months , speakers never moved
more than a mm before the relay activated.
OS
Attachments
There's a few designs on the go placing this at the speaker binding posts. Either end eliminates extra wire connections.
Attachments
But we still have the rails and the overcurrent opto.
And a timed monostable multivibrator triggered by a rail turning on.
So this 6-7 transistor circuit needs to connect to emitter resistors , the rails,
and a 12V aux. supply.
Might be better to have it all integrated.
OS
If you are doing current sensing as well it does make sense to do it all on the amp.
I've got a system half designed to operate on a data bus with a microcontroller. A shielded 4 wire does a loop around the amp and takes care of it all.
The micro would make some fearful.
Navy Electricity and Electronics Training Series (NEETS), Module 9-3 - RF Cafe
That's the turn on delay (above link) ... just two trannies and a cap.
I'll tie in a dual color led (red=protect/green=good) no fancy flashing.
All the "21'st century" amp protection can be simple analog. All
the micro does is integrate the soft start and tie it all together.
If we wanted to tie this in with anything external , just a simple output
to report any protection event.
My first business is to layout the IPS/OPS on one board while leaving
2-3 sq. inches for these "enhancements".
OS
Yes, go ahead.
The vast majority of noise is related to the input stage topology and little else. So it's easy to make super-high performance designs and then discover they're noisier than most.
Source resistance from red to blue is:
Red - 100R
Orange - 1k
Yellow - 2.5k
Green - 5k
Blue - 12.5k
Last edited:
Done. Tell-me if everything is ok. Or send-me a private message if you want some changes.
(thanks a lot, again)
I've still been living dangerously with no DC protection. I need to implement something before I upgrade my woofers. 4 quality 15" drivers adds up to some money. Jwilhelm was kind enough to share his board with me but the # of parts was daunting concidering I run up to six mono amps for my system. Something small and simple like the the circuits being discussed are what I want.
E
E
That Vzaichenko DC detector saved me 4 times so far.
A couple times forgetting to tighten PD+/ND- , or touching the servo
input .... or not having a ground on the IPS.
I only have the full "21'st century" protect on my main stereo amp.
My sub still needs at least a DC/delay setup.
Designing this will kill many birds with one stone.
OS