There were many ideas started back then... microphones, accelerometers sensors, capacitive detection. ETC. Many just to get flat freq response. But my simple approach which started with the ACE-BASS got me thinking and it does work and works well enough for low freq distortion reduction that I am wondering what would be even better in reducing THD of the speaker. Thats why I am here. When I saw and bought the book by Esa, it got me interested all over again.... hopefully a lot has improved in the almost 30 years since I last had a thought about it. So, Rock On.
Show me what you did for Yamaha. OK?
THx-RNMarsh
Last edited:
It was the AST/YST speaker that Yamaha sold in the UK based on our drive units. The Service Manuals should be available.
Little different really from your circuit. Because we'd done a lot of work on electronic EQ and Negative Output R, they were surprised that we could adapt their circuit to match our drive units and their box with ease .. in fact faster than they could.
No. They are completely opposite approaches. But some of their advantages, eg THD reduction, are the same.
In my #4331 post, I sneer at Current Drive cos it doesn't give ALL the advantages of ACE bass etc.
Negative Output Z (or even just R) can be designed. The present Current Drive systems are fudged. 😱
I suppose I should concede that Marshy's Negative Output R/AST/YST etc is sorta motional feedback though not to the extent of ACE bass. 🙂
Current Drive has no motional feedback component.
_____________
Plis dun mek fan orf mi jus cos i kunt reed en rite 😡planet10 said:
Last edited:
Will current drive reduce distortion by a like amount? If not - what is the advantage of current drive?
THx-Marshy 🙂
THx-Marshy 🙂
I think the best way to get ' motional feedback '
is to drive 1 coil of a DVC driver and use the output
of the 2nd coil as feedback ...........
Hard for me to equate that with a current-source amp,
but maybe I just lack the math ?
is to drive 1 coil of a DVC driver and use the output
of the 2nd coil as feedback ...........
Hard for me to equate that with a current-source amp,
but maybe I just lack the math ?
This depends on EXACT details of the drive unit. IME, the amount of THD reduction is roughly the same. Don't forget measuring distortion in speakers has some uncertainty too. 😡
Current Drive only has 1 advantage over Negative Output R (like your method). It's resistant to voice coil heating. For the other 'elegant' methods like yours and ACE bass, you need to take other measures to deal with voice coil heating.
Current Drive's biggest CON is that you can't design a system for good performance .. you just tweak it by trial & error. The 2nd biggest CON might be its nearly useless for Vented boxes.
Voice coil heating creates compression not necessarily distortion ( adding of harmonics).
My experience with the current drive version of Erik Stahl's patent claims are excellent so why look further?
Electrostatic speakers if directly connected to a high voltage power amp ( w/o transformer) are essentially current driven. The driving force is ideally the same over the total area and is proportional to "charge displacement current" .
Thus most of this discussion misses a point that is the "one point drive" of a plane or piston is a theoretical ideal that is not achievable for electrodynamic speakers there are no absolutely rigid membranes.
German physicist Trendelenburg who wrote a book an acoustical properties of loudspeakers already in 1928 ( out of print yet an absolute must read ) invented and had built an electrodynamic speaker with a 2 by 2 meter steel alloy membrane with some 100 if not 1000 voice coils evenly spaced and of course- it cannot be otherwise - current driven. Don't know how he had solved the suspension problem but he did as these types of speakers were indeed built and used in sports arenas .
Thus the theoretically exact model to discuss current vs voltage is an array of "infinitely many" "infinitesimally small" electrodynamic speakers and the answer is ultimately clear: current drive.
My experience with the current drive version of Erik Stahl's patent claims are excellent so why look further?
Electrostatic speakers if directly connected to a high voltage power amp ( w/o transformer) are essentially current driven. The driving force is ideally the same over the total area and is proportional to "charge displacement current" .
Thus most of this discussion misses a point that is the "one point drive" of a plane or piston is a theoretical ideal that is not achievable for electrodynamic speakers there are no absolutely rigid membranes.
German physicist Trendelenburg who wrote a book an acoustical properties of loudspeakers already in 1928 ( out of print yet an absolute must read ) invented and had built an electrodynamic speaker with a 2 by 2 meter steel alloy membrane with some 100 if not 1000 voice coils evenly spaced and of course- it cannot be otherwise - current driven. Don't know how he had solved the suspension problem but he did as these types of speakers were indeed built and used in sports arenas .
Thus the theoretically exact model to discuss current vs voltage is an array of "infinitely many" "infinitesimally small" electrodynamic speakers and the answer is ultimately clear: current drive.
Bring us all up to the schematic or the patent being referred to, please.
[ I used RTR electrostatic tweeters direct drive off the plates of Dynaco tube amps... worked very well]
THx-RNMarsh
[ I used RTR electrostatic tweeters direct drive off the plates of Dynaco tube amps... worked very well]
THx-RNMarsh
current drive is entirely possible also with electrodynamic drivers, the main problem area is the the base region, where vented or even sealed enclosures will pose a problem as it makes resonance peaks in the impedance, and thus posing a problem for the current amplifier. in the other areas speaker must be designed with a smooth impedance curve and possibly with the use of current diversion crossovers.
Kgrlee could you provide a little insight how to make a V(out) amplifier having a negative output impedance. I see this as a tool to give the amplifier a bigger hammer to deal with the back EMf from the speakers.
Kgrlee could you provide a little insight how to make a V(out) amplifier having a negative output impedance. I see this as a tool to give the amplifier a bigger hammer to deal with the back EMf from the speakers.
Can you provide more details of the 'current drive version of Stahl's patent'?
Details please.
It so happens that for Rice & Kellog's invention, you DON'T want a rigid membrane. You want it to break up above about kr = 2.
This little understood fact is the reason why their invention is still pre-eminent for everything from cheapo stuff to the very highest quality Golden Pinnae systems.
I've done a lot of work on the 'ideal' pattern of breakup but in fact, the cheapest & nastiest transistor radio speaker already approximates (??!) this behaviour. 🙂 OK. I"ll admit some breakup in nicer fashion than others.
Marshy's little circuit is it. It's that simple.MiiB said:
Rod Elliot has the maths and also important caveats. The main one is that if the Negative Output R = DC speaker resistance, the system is unstable.
http://sound.westhost.com/project56.htm
Du.uuh! Just realised Marshy's circuit ISN'T Negative Output R but Positive. Explains why he gets more bass. So YES, Mr. Marsh, your circuit is sorta Current Drive cos it has significant output R.
Both Positive & Negative Output R give less acoustic distortion. Rod doesn't measure this being an unwashed amp man obsessed with amp distortion 😱
If you use a lot of Negative Output R (bearing in mind Rod's caveats about stability bla bla) you lose bass cos the speaker is overdamped. If you now use EQ to bring the bass back up, you have Yamaha AST/YST.
Last edited:
Details found here
US Patent 4,118,600
Application No 780,454
Filed Mar 23, 1977
Loudspeaker lower bass response using negative resistance and impedance loading
Inventor Karl Erik Stahl
the only reason why breakup is desirable is that you decrease the area radiating with frequency, thus maintaining dispersion and SPL. This is however associated with other some major flaws and drawbacks that is practically impossible to neglect. The major one being that absorption and breakups is dynamically nonlinear and you end up with driver that "sound" very different with different SPL's. This has of course nothing to do with current drive, but is related as Current drive would demand a different driver design.
In current (present) driver designs we rely on damping (mechanical, "semi" electric and acoustic(pressure)) to control the response and power handling, with current drive other measures of damping must be taken, as the current drive will override those mechanisms
Thanks for the Elliot Link
In current (present) driver designs we rely on damping (mechanical, "semi" electric and acoustic(pressure)) to control the response and power handling, with current drive other measures of damping must be taken, as the current drive will override those mechanisms
Thanks for the Elliot Link
Last edited:
This isn't a 'current drive version of Stahl'. As I pointed out, Negative Output Z (of which Stahl's ACE bass is a sophisticated example) is the complete opposite approach to Current Drive.
It works very well but it IS affected by Voice Coil heating. YMMV
This evil breakup is the reason why Rice & Kellog's invention is the BEST solution from cheapo to Golden Pinnae.
You can't neglect it but it is not a flaw. The earliest treatment I've seen of this is Jordan in Wireless World early 70s or late 60s.
Can you suggest another type of driver/operating principle that actually has usable flat response over nearly 3 decades? Let alone good directivity and other stuff that's important for good sound? And can be designed as opposed to fudged?
You can design Rice & Kellog drivers that sound the same up to the very highest SPLs. In fact no other principle has been proven to do this.
Ok. Then Stahl wrote a bunch of nonsense. I give up. Stahl didn't understand his invention.
Du.uuh! In case you haven't noticed, I'm a big fan of Stahl. I've used his stuff and it works very well.
But it IS affected by voice coil heating. This will change the response of a Stahl sub but so what? It will still have more & better boom boom than lesser stuff.
But I'd rather it didn't change its response 🙂 You don't have to be bothered.
Besides, I brought it up cos Marshy was asking what advantages does Current Drive have over the Negative/Positive Output Z stuff. Freedom from voice coil heating effects is probably the biggest advantage of Current Drive.
David Birt's Loudspeaker Power Amplifiers with Load-Adaptive Source Impedance has a good discussion and some solutions that I've applied to ACE bass.
Last edited:
kgrler why is three decades desirable..?? apart from dispersion which I respect then electrical damping is way more linear than any mechanical energy absorption (which is what breakups is all about). Soft membranes are dull and veiled sounding compared to stiff membranes, which then (in most cases) suffers from material resonances. All here is a choice of which set of compromises you choose to work with. My take is stiff membranes very stiff...so stiff that 1. order resonances are pushed 3 decades over the passband. and also mechanically damped so the 1.resonance is not a peak, but more like a 3 dB bump.
Very little directly, because of the slow time constants of the voice coil heating and cooling.
IMHO, some of the major distortion mechanisms in moving coil speakers are:
1) The movement of the voice coil back and forth in the magnet gap modulates the BL factor, which causes distortion. This is mostly odd order assuming the voice coil is properly centered in a symmetrical magnetic field, which it probably isn't. This not only causes distortion of the bass, but also intermodulation between bass and higher frequencies.
2) The current through the voice coil modulates the strength of the magnetic field in the magnetic circuit, causing mostly even order distortion.
3) The movement of the voice coil back and forth in the magnet gap modulates the voice coil inductance as well. This mostly causes intermodulation of the high frequencies by the low frequencies, since most of the movement is caused by low frequency input while the high frequencies are most affected by the changing inductance.
4) Non-linear suspension stiffness causes distortion of low frequencies.
It's the easiest, but probably not the best. Note that distortion mechanisms 1 and 2 above will affect the feedback coil as badly as they affect the driving coil.
If you refer to the attached figure coming from post #96, the circuit provides positive output resistance : if the the resistive value of the lowers, the amp output voltage lowers, meaning that the amp has some internal resistance.
A negative output impedance is characterized by an output voltage increasing when the resistive value of the load decreases.
A little change in the schematic, using positive feedback, can easily provide it :
The concept of damping factor for amplifiers is somewhat in difficulty with negative resistance output, as the DF should become more than infinity if strictly applied.
Using negative output resistance, one must be very careful that the load impedance must always be higher than the absolute value of the negative output impedance of the amp.
As far as I understand it, the Erik Stahl concept uses two feedbacks :
- the first one provides a negative output impedance equal in absolute value to the Re and Le values of the driver.
Re and Le then appear to be eliminated and there is now access to drive the motional impedance.
If the driver has a small thermal constant and a high thermal resistance, a compensation for temperature variations of the voice coil is needed. (Ian Hegglun has proposed a smart solution in Electronics World, May 1996). With reasonable listening levels and large drivers, it seems it can be dispensed with. However a great difficulty remains, it is the compensation of Le value which is not constant at all with the position of the voice coil in the gap and with frequency.
- the second feedback circuit adds apparent components to the motional impedance. As they are electronically defined, there are more linear than those of the real driver. The whole circuit provides a means of lowering distortion.
.
Last edited: