Now I understand what you are presenting, thank you. But I still remain sceptical of the audibility of the break up mode in the 3" dome case. The 3rd harmonic is generated by the modulation of the motor function and radiates because of that. Any break-up where parts of the dome move out of phase with each other will actually attenuate the radiation where the two effects are coincident and therefore reduce the audibility of the distortion.
Quite possibly, but my experience is that a solution via direct radiation avoids any horn colouration - which may in fact be minimised but presents a whole new set of design difficulties and potential problems. Maybe another advantage to add to the 3" dome based solution is that it is well-matched polar-wise for a crossover to a conventional tweeter.
Last edited:
The extent of many of the various advantages of current drive is dependent on the particular situation. But in all cases the two distortions discussed here, namely thermal compression and impedance modulation due to eddy currents in the pole pieces, are eliminated completely in all cases from the final transfer function.
I am not sure of your question regarding the "low frequency limit", however. Current drive requires equalisation of the fundamental system resonance where drive units are normally optimised for voltage drive. Given this equalisation, however, any low frequency cut-off can then in principle be designed.
Last edited:
yes it was very late and my question was not well stated, but you partially gave me an answer anyway.
Let's say i want a driver to cover from 250/300hz to 1/2khz. The driver is a typical pro unit with high efficiency (98db), 67g mms,3 demodulating ring,... in closed box (with Linkwitz Transform as it's not specified to be in closed box) with dsp/processing hardware unit in multiamp scenario for domestic listening.
For a voltage source amplifier this is a frequency range where damping factor is still relevant about the character of the system.
So it makes me wonder if in case of current source amplifier driving this driver there would be mismatch in behavior with the low end way which will be voltage amplifier driven? And if gain are possible and in which area (given i grasp the point about thermal and eddy current)? Or is it pointless?
For compression driver when i read through the articles and manual of Nelson Pass Firstwatt F1 i see a lot of similarities between the way he present full rangers and how i see a compression driver so my question about this particular kind of drivers.
About noise Pos, the spec of F1 are outstanding and even if very low power in domestic situation and given efficiency of CD this should be great.
If you want to try an amp designed for this purpose rather than playing with resistors, apart from Firstwatt's F1 and F2, Joe Rasmusen have a design based on an lm chipamp which look not too complicated in the amplifier section.
Last edited:
From the measurement I have seen the ~4.5kHz breakup in the ATC dome seems to be well dampened.
That is not the cas with the JBL kevlar driver, and even less so with metal cones. OT here though, which counts as one additional benefit of that dome compared to cone alternatives.
There is no such thing as horn coloration anymore when you correctly use a properly designed waveguide. At least that has been my experience with that particular 6" PT waveguide, among others
High Qms drivers will require very precise (driver-matched and high Q) EQ if used near fs, and any instability (aging, break-in, excursion, ???) can result in an acute artifact in the final response.
Another point to consider are external factors: even if you EQ the driver and end up with a precisely correct and stable transfer function, the driver is not electrically damped at fs, so any pressure caused by external/indirect sources (box modes, room modes if the speakers are near a pressure node, like an in-wall speaker, ...) could trigger the undamped resonance.
And this would be difficult to measure.
Of course with both issues, the higher the Qms the more acute the consequences.
This could be avoided by gradually turning back to voltage drive when approaching fs. No idea about how to implement this though: me just want to buy an off-the-shelf unit
Yes I have been lurking there, but amplifier construction is not my thing.
The solution I am using now (well, it is still in its box...) is to use a 100V amplifier with a series resistor.
I found that using that 100V unit with enough resistance to reach the minimal 50 ohms specs was a good match with the 16 ohm driver I intend to use, and results in pretty much the same available output power for the driver as Joe Rasmusen's solution.
The 70V version would probably be a better match for a 4 ohm driver.
In the end output impendance is still lower than Joe Rasmusen's amplifier, but it should be enough to get most of the aforementioned advantages.
Last edited:
Not always in my experience but it is the other side of crossover so not well excited either.
Hence the use of Birt's self-balancing bridge or at least a modification of it that is capable of accurate motional feedback - of the voice coil at least.
The ATC dome and its derivatives use a double suspension that uses an enclosed volume and vent. Whilst I refuse to take responsibility for those who choose to experiment with their own drivers, the fundamental resonance and Qms can be changed to make the dome more favourable for current drive from the start.
Not with motional feedback.
There are published means to do this but this approach also eliminates some of the other current drive advantages such as the displacement related ones just where they are most beneficial. A modified self-balancing bridge is the best system I have ever encountered.
It's quite straight forward : At DC and LF (and at HF, needed for stability) make the feedback dominated by output voltage, while at the frequencies of interest (> fs, up to a few kHz) the current portion should dominate. Needs a few R's and C's in the feedback network, but it's not overly complicated. Output impedance is low and resistive up to fs, then raises like an inductor, then it tops out and starts falling again like a capacitor until it is resistive at HF. If needed, the LF impedance can even be made negative (while still resistive) for more motional feedback within the driver (velocity controlled), this sometimes is lowering suspension-related distortion in underhung compression drivers when operated near fs (but make sure you don't overdrive it).
Never heard of this, and Google won't spit out something useful. Care to share a link or something?
Personally I've had only moderate success with using 'pure' current drive around fs with the high Qms cone drivers I've tried, serious jump resonance ill-effects raised its ugly head every single time (there is a whitepaper on this by some Tymphany guys).
Also, my design strategy usually is to make the driver respond aperiodic to any external disturbance (including its own distortion) so I tend to leave some electrical damping left to reach total Q's of no more than 0.8 or so around fs.
OK, I'll buy it, just tell me where to click
I have read that you designed current drive amplifiers for some entry level Adams monitors. Maybe that particular part can be get from Adam's service center?
Last edited:
I don't have the references to hand (I will add them later when I find them) but David Birt has two easy-to-read AES Journal papers on the subject. There were also two patents granted I seem to remember - although these may well have expired by now.
To make things stable, please see my earlier posts on drive unit modifications. I also ended up using segmented pole pieces to prevent eddy currents and therefore leave a classical series resistance and inductance that was to a large extent compensated in my modified system. Without this (or ATC's polymer-based non-conducting pole pieces) you are stuck with a semi-inductive component that proves very difficult to compensate and therefore produce an adequately stable overall system.
Also worth keeping in mind is that the raw ATC units are very linear starting points and I am somewhat sceptical the same approach would be as successful in attempting to compensate drive units that are not so well engineered in the first place. Some form of compensation for suspension non-linearity may well be advisable as also shown by Birt (possibly in the same papers if my memory is good enough) and also shown in an open loop scheme by Klippel.
Hope that helps for now...
Last edited:
Please excuse my crashing into this thread without any reading at all, but the recent incremental advance designed by YG Acoustics may potentially come into heavy play here. They use a light alloy sub-frame underneath the dome which provides additional support to the softest parts. In a mid dome where a large center area approaches flat, and therefore has minimal strength, doing something like this might provide even greater benefit than it does at tweeter dimensions.. Just a guess.
@pos, I think the power amplifier might have changed now to standard voltage drive, to be more cost-effective...
Might be easier to buy one of those standard chip amp PCBs and modify the feedback network... I'll be glad to help out in the design process (it quite depends on the actual driver being used).
Might be easier to buy one of those standard chip amp PCBs and modify the feedback network... I'll be glad to help out in the design process (it quite depends on the actual driver being used).
This one is the only I could find : "Loudspeaker Power Amplifiers with Load-Adaptive Source Impedance"?
No access to AES papers at this time... hard to tell from the abstract what really going on, relating to current drive.
Thanks a lot for your offer, I appreciate it, but the gustiest thing I have ever been able to do when it comes to electronic DIY was assembling a set of hypex amplifiers.
I am a lost cause and I chose to give up on these things for my own mental sanity (that and woodworking....)
Adding motional feedback to the equation look like a daunting task for the regular DIY guy.
At least to me...
I know some guys in France done it but they are 'specialist' and really advanced on many subject audio related.
Here is an exemple:
Zippy - Enceinte intégralement asservie - 29923222 - sur le forum Discussions Générales - 1472 - du site Homecinema-fr.com
JCB is a nice acoustician too which had some very nice control room done in Paris...
In his blog their is some entry about this but it's in french:
http://www.delapucealoreille.com/
At least to me...
I know some guys in France done it but they are 'specialist' and really advanced on many subject audio related.
Here is an exemple:
Zippy - Enceinte intégralement asservie - 29923222 - sur le forum Discussions Générales - 1472 - du site Homecinema-fr.com
JCB is a nice acoustician too which had some very nice control room done in Paris...
In his blog their is some entry about this but it's in french:
http://www.delapucealoreille.com/
It is more complex but easier when starting with drivers that do not rely on motional feedback for linearisation. Its mention in this thread, however, was not intended to motivate DIYers but in support of the other questions in this thread - especially why it is not more widespread in professional loudspeakers. But if drivers are deemed linear enough to start with why would anyone bother when they can make the sale anyway?
Does the 3"ATC able of 120 dB spl peaks (not continuous) without a large bafle and without too much distorsion please ? or maybe on a more restricted Frequencies range ?
A good alternativ to a high-mid/low tweeter horn + CD ?? Or does the limitation is about its relativ low efficienty (trade off should be to drive it alone in an activ setup) ?
A good alternativ to a high-mid/low tweeter horn + CD ?? Or does the limitation is about its relativ low efficienty (trade off should be to drive it alone in an activ setup) ?
I believe it does in its recommended operating band in 2pi space as typical of soffit-mounted monitors - so it just depends what you mean by a "large baffle". As expressed in this thread, I believe it is still the best solution in a multi-way loudspeaker of this sort and its nominal bandwidth still presents the best compromise (although others here have expressed otherwise). It is not a low-efficiency driver, however.
I also think that any attempt to build a high quality monitor capable of 120dB+ peak SPLs with low distortion should always be active to avoid unnecessary compromises: Any attempt to do otherwise would likely be compromised in such a manner that the use of expensive drive units becomes questionable.