warp08
Headphoneus Supremus
- Joined
- Feb 1, 2010
- Posts
- 2,211
- Likes
- 44
At Can Jam/RMAF 2011 last weekend, I had the pleasure to sit down with the designer, Jerry Harvey himself to discuss the recent design changes that generated quite a bit of controversy over what has been considered by many as a project in jeopardy. As one of the early adopters of this technology, I myself wanted to learn more about what has been changed and for what reason and how did it affect the performance of the final product.
Let me state up front that the post represent a best effort on my part since I am not an EE. I will send this URL back to Jerry to review and correct any inaccuracies, which are almost inevitable, given the complexity of this product, especially to a layperson like myself. The scanned in hand-drawn diagrams have been drawn up by Jerry himself and posted here with permission. Aspects or as a whole of this design has been submitted for one or more patents, since the shipping version of the JH-3A amp--scheduled to be distributed to customers who have preordered the product--features a new approach, Jerry calls the "Reverse Active Crossover Design." I have to give credit to him, because not many designers would invest this kind of time during a busy international show--or any other time for that matter--to provide this level of detail to a customer. In return I have promised to try to capture what I could to try to dispel some of the incomplete or secondhand information circulating on this forum by capturing it here as a high-level reference.
This post is a continuation of my previous one from this morning in the JH-3A thread, so If you have accidentally bumped into this one first, I would recommend you jump to the other one now and read it. It is short and to the point, as well as being complimentary to this one, which focuses on the technical design aspects only.
Many of us on Head-Fi tend to disregard JHA's focus of innovation being split between the Pro market--artists, recording engineers, etc.--as well as of late, the high-end consumer portable audio market, namely us. JHA has been very successful in this endeavor, but their core competency primarily had been their signature custom IEM products. The JH-3A system has been their first foray into amplification and it has been quite a learning experience for the company and a less than optimal experience for those who preordered it, primarily because of multiple delays. One of the main drivers of these delays of late has been the desire to incorporate feedback from a handful of customers like myself as well as other JHA customers who have been actively lobbying for a way to be able to use their JH-3A-enabled custom IEMs with other headphone amps as well, instead of having to dedicate them exclusively to the 3A.
Since I already own a pair of conventional JH13 Pro and JH16 Pro custom earpieces in addition to the JH16 Pro shells that have been delivered as part of my JH-3A system, this is not a concern for me. This old annotated photo below illustrates the commonly known major components of the conventional JH16 Pro design (and, as it turns out, incorrectly identifies the chipset on the right as "3-way crossover circuitry.")
However, up until last Friday, I have believed that the new JH-3A/16 shells contained no passive crossovers, because the JH-3A "Active Crossover" system built into the old design and controlled by the DSP DAC chipset's programming. The combination of these components had been jointly responsible for managing the signal feed into the 8-BA driver JH16 custom shells connecting them via an 8-wire crossover cable, terminated by an 8-pin Lemo connector. This was only half the truth, as it turns out. Little known fact is that JHA uses custom-designed BA drivers for Low and Mid range frequency bands that contain built-in crossover technology. Research shows, that achieving accurate time delay and phase coherence is most difficult at the higher frequencies, so the High (tweeter) drivers are optimized two ways: a passive crossover circuitry as well as physical placement. They are a lot closer to the eardrums than the Mid and Low driver pairs/quads depending on the model. JHA BA driver suppliers are prohibited by contract to ship these custom drivers to anybody else, so they are only produced in low volumes and more expensive than generic BA drivers used for similar purpose. That's part of the "JHA edge" as well as time accurate tuning for phase correctness, i.e., so one particular frequency is never 180 degrees apart in phase, therefore canceling each other causing a "dip" in the frequency curvature, perceived by the listener as a lack of mid or bass gain levels within the affected octave(s).
As it is shown in the diagrams below, crossover systems designed to operate multi-driver speaker systems--in these case miniature Balanced Armature (BA) drivers operate as part of a "network." That means 3-way IEMs--which operate in 3 distinctive frequency bands (Low-Mid-High) must include dedicated circuitry to perform the necessary filtering between the different bands. As you can see, a theoretical (i.e., illustration purposes only) breakdown of the
The two diagrams--the top one representing the "Active Crossover (AC)" and the bottom one the novel "Reverse Active Crossover (RAC)" configuration reveal that the basic design did not change. Critical aspects of sound delivery of custom multi-driver ear speakers in a coherent way include optimizing the following attributes for EACH frequency band (Low-Mid-High) regardless of digital or analog input:
- EQ
- Gain
- Time
- Phase
Both in the AC and the revised RAC design, attributes 1-3 have been controlled by the DSP chipset and adjusted independently by its programming, which includes
At the high level, the difference is that the crossover filtering is done exclusively in the digital domain (DSP). In reality this is more of a theoretical representation since as I described earlier, most of this processing is done in the DSP anyway and the only thing that had prevented JHA of accommodating existing shells into the design via an adapter was the fact that there was no High passive crossover circuit inserted in the shell design. With RAC, this problem is solved, but at what cost? Essentially, the concern over processing power between the AC (full-DSP) and RAC design is invalid, because the precise DSP-based control over EQ, Gain and Time attributes of each frequency bands ensures that in this configuration the actual workload is minimal and the existing conventional shells can be incorporated into the RAC design without major surgery. Retuning, as well as "plastic surgery" to modify the shells and their wiring to a 4-pin socket config, however, will definitely be needed. There will be an option to purchase a Lemo-to-TRS adapter so the modded shells could now be made to work with other amps, as opposed to being dedicated to the JH-3A only. This is definitely lowers the price barrier a bit for prospective JH-3A owners while retaining the versatility of their existing investment at only a minor degradation of sound quality due to using any adapter in the signal path. Depending on how expertly this is accomplished, the end result should be all but imperceptible for most listeners.
JH 13 Pros unfortunately still won't be accommodated because the revised gain structure and microcode of the amp is simply too powerful to accommodate with "only" a dual-stack Low driver configuration that would not result in distortion. So, the JH16 Pro design will be the only shell being carried forward, at least for the time being. That actually makes a lot of sense since the bass is very easily controlled via a knob to match the preference of the listener depending on what the track calls for.
Finally, I would like to share one reference URL I have found useful in understanding at least at the surface-level the science behind this. The article is a bit dated, however, I have found it to be much better written than most on Wikipedia, at least for someone who dabbles in IT Enterprise Architecture for a living and to paraphrase Jude (owner/administrator of this forum) "could not solder his way out of a wet paper bag."
At any rate, I hope to have helped a bit dispelling the fog around this latest twist in the JH-3A saga. Time will tell how many people will embrace this promising technology. All I can say is that I have not met anyone who had listened to the demo unit at Can Jam over the weekend who did not like what she or he had heard. That includes Tyll@InnerFidelity, who decidedly wore a quite different facial expression than during his signature Ultrasone Edition 10 video review...
