Reliability Issues

Dear SLOG,

This is the first time that Sound Lab has contributed to the SLOG site. In following the dialogue on SLOG we are aware that there has been some concern over some issues lately that we feel are important enough for us to address. We have been reticent to become involved directly as we respect the right of Sound Lab owners to freely state their opinions without us being accused of a conflict of interests in what we might contribute. However, we have been invited to speak up at this time in order to shed light on the prevailing issues.

One issue is that it has been speculated that we have started using a "porous mylar film". In fact, we are using the same roll of mylar that we purchased six years ago. The film is so thin that a 50 pound roll lasts for years. I've never heard of "porous mylar" and I'd be interested in learning what it is. We use Type C mylar, which is one of the most refined grades. It is mainly used in the manufacture of capacitors.

Another issue is that some speakers have been reported as having lost efficiency or are in the process of doing so. The speculation has been that Sound Lab has been losing quality control and its products are not what they used to be. The truth is that there were indeed two problems that resulted from the introduction of two new technical features which were intended to improve our products. Furthermore, the truth is that these two problems have been isolated and corrected and, as a result, our products are indeed better than ever. One problem had to do with a component that slowly failed in the new high-energy bias supply introduced last year and the other had to do with the limited introduction of a new insulating material that failed, which was introduced about a year and a half ago. I would like to address both of these items in some detail in order that both the problems and their solutions are clearly understood. I should mention that the scope of the problems was much smaller than one would believe after reading the comments on SLOG. The numbers of speakers involved were a fraction of the total products manufactured. Hopefully, my comments will put speculation to rest.

Concerning the bias supply problem, initially we talked with a few customers and from their comments it appeared to us that the bias controls were failing. A few bias controls had indeed failed and we felt that, since we assumed that the bias supply was reliable, sending new controls to replace the original ones should solve the problem. It was anticipated that this would also save customers the inconvenience of having to send backplates to us for repair. Our best guess was that possibly line transients could have pushed the controls past their limit. The circuit was revised to take stress off of the control, and new controls were sent to the customers involved, with instructions on how to install the new circuit. The results were mixed. In two cases the new circuit solved the problem but there were a few cases that didn't make sense unless the latest bias power supply circuitry was at fault.

Having the dust settle from our involvement at the Consumer Electronics Show, I was able to spend time to look deeper into the problem. I found the source of the seemingly unsolvable bias problem in a DC to AC step-up converter that we purchase from another company. This circuit, which is the heart of the bias supply, employs a tiny step-up transformer. Upon inspection of a defective unit, I found that the secondary coil of this transformer was open. Upon taking the transformer apart I found that the attachment of one of the secondary coil leads was burned away from its terminal and that a carbon track had been created, which made a high-resistance path to the terminal. This fit the circumstantial evidence perfectly. Customers had reported that the bias potential slowly reduced until the control was advanced as far as it could go and that the speaker was still losing efficiency. I took a second defective transformer apart and found exactly the same thing, the same lead and the same terminal. The only difference was that the carbon path was not as extensive as that of the first unit, which strongly implied that it could be a progressive situation just as the customers had stated. Close examination showed that the lead wire to the open terminal was intact before the carbon trace. This indicated that the transformer coil, itself, was not failing as there was no charring on the coil. It appeared to me that it was an attachment problem of the lead to the terminal. The most probable explanation is that in some units the lead isn't reliably soldered to the terminal and that the high voltage at this point in the circuit easily jumps the small gap and starts the burning process. As proof of this theory I found that a small percentage of the transformers didn't show continuity on the secondary wind using an ohm meter between the terminals. These units would surely fail. The percentage of units showing this open circuit was fortunately small, so I don't expect that there will be many failures.

The vendor of the converter has been contacted and is looking into the problem and assures us that it will be quickly solved. In the meantime, we are doing an accelerated test on some of the converters that we know are OK to test the theory that if continuity is indeed present between terminals on the transformer secondary, that they will work properly indefinitely. This is supported in that the large majority of the circuits in the field are doing fine.

In retrospect, we could have found this problem sooner if we had demanded that the backplates be returned to us for inspection when we first heard of the problem. This will be our policy from now on. If any customer is experiencing slow degradation in the sensitivity of their speakers, please send your backplates to us for inspection. If your speakers emit a crackling sound when the bias control is advanced, then your backplates should be OK. Before shipping your backplates, please contact us first for directions on how to properly package them for shipment in order to avoid shipping damage.

Probably due to the problem described above, the notion that Sound Lab quality control has been permitted to degrade has been entertained on SLOG. I can understand that this is one possible conclusion, but it is not the correct one. If the first backplates that exhibited the problem were returned to the factory, the problem of degrading sensitivity would have been solved almost a year ago. The most logical conclusion would be that we didn't insist that the backplates be returned to us, and that our intent to try to have customers fix the problem in the field was not wise. I predict that when the defective units have been replaced by units that are proven to be good, there will be no more incidents of this problem.

Concerning the second problem, about two years ago we decided to try a new stator insulation material on a few speakers. We felt that this new material, a type of nylon, was superior to our original material and was worthy of a limited field test. We carefully tested this material to our satisfaction, and it appeared to have all of the proper electrical characteristics along with a greater resistance to abrasion compared to the original material that we had been using for over 20 years. Several pairs of speakers were built using the newer material, and after a period of time (weeks to months) we started receiving reports that the material was failing. To me, this was astonishing as we had no indication of this in our lab tests. We immediately started running some accelerated endurance tests on the material (higher than normal bias and signal potentials) and we confirmed that over an extended period of time the nylon started giving off a distinct presence of ozone which, due to its powerful oxidation effect, led to the ultimate failure of the material. Apparently, the porosity of the material was such that it was able to initially handle the high potentials involved, but that it also permitted air to permeate it which led to the destructive ozone production.

Permit me now to back up in time. Since the inception of Sound Lab in 1978, we have always looked for better materials with which to insulate the stator electrodes of the panels. About three years ago, which was well before the use of nylon, we tested a promising material that we had obtained from a specialty extruder. At that time our concern was focused on the effects that humidity had on the stator insulation. Since this particular material was noted for its extremely low moisture absorption, a pair of speakers was built and shipped to a customer who lived in an area that had a challenging environment. The results were very positive, but due to the significantly higher cost of the material, and the fact that our standard material was actually quite good in humid climates, we didn't pursue this material further. However, we did make note of the extremely low hygroscopic characteristic of the material along with some other of its attributes.

Back to the present. When we became aware of the nylon problem we did two things. First, we started a program of rebuilding every unit that used the nylon insulation. Secondly, we decided to look again at the material we had sent to Singapore three years earlier, since it has survived the harsh salt-air, humid and hot environment as well as the long, very cold (-70 F) air-shipping phase. The company that made the material was contacted, and after some refinements (alloying) of the material we ordered a rush sample to try. We received the material and immediately gave it an accelerated endurance test. No problems were found. The decision was made to go with it instead of the original material. The new material not only had superior moisture resistance, but it had superior electrical characteristics and was very inert to most chemicals, including strong acids. Furthermore, it is very tough and is very resistant to abrasion, which was a significant improvement over the original material. We used this material in the rebuilding of the speakers that had used nylon. We felt very confident that it was the proper choice.

With solid statistics to back me up, I can say that we made the correct decision. With over a year of experience (actually three years if the Singapore units are considered) the results have been most gratifying. What's astonishing is that a pair of M-1s (or A-1s/U-1s) use more than half a mile of insulation, and we now have many miles of the new insulation in the field with great results.

In order to eliminate any speculation that the new material has resulted in a loss of sensitivity, we compared a U-1 panel, using the original material, with an identical panel using the new material. Armed with a calibrated real-time analyzer and a calibrated sound-level meter, we compared the sensitivities of the two speakers, balancing acoustical factors and carefully equalizing microphone distances. The panel with the new material consistently showed a slightly higher sensitivity of 1.2dB. This was demonstrated to several people in our lab with the same results. The slightly higher sensitivity of the newer material is within experimental tolerances, so we can at least conclude that the sensitivity of the new material is, indeed, consistent with that of the original material.

During this last year there have been other improvements that do not show from the outside. We have now started using water-jet machining to cut out the components of the panel's infrastructure. The benefits are closer tolerances, no machining chatter marks, and no heat produced during the machining process which can distort the material and create carbonized leakage paths. Also, in key high-stress joints in the framework, we now use an industrial cyano-acrylate adhesive that is considerably stronger than the standard plastic adhesive that was used for years. Furthermore, the horizontal ribs that support the stator grids have water-jet machined precision notches to support and critically align the stator wires. These improvements have improved both the strength and precision of the panels.

We, at Sound Lab, are committed to continually look for ways to improve our products. I suppose that there is always an inherent risk factor in this process. One is faced with the choice of not attempting to improve something that works acceptably or to take a calculated risk in trying new procedures, materials and designs in an attempt to offer superior products. With amplifiers, cables, music sources, etc., continuously improving, we really have no choice but to continue research and development if we wish to keep up with other technological advances in our field. We attempt to never release a change until we are confident that it represents a genuine improvement and that we have fully done our homework. However, if a change results in a problem, we are also committed to making things right with our customers.

It is hoped that this note will ease the fears and suspicions that have arisen recently. You are invited to contact us directly on any questions or issues that you may have in order to get first-hand clarification. We are always open to suggestions that might improve our products and our customer relations, and we apologize to those who have had to endure the frustration caused by the problems I've addressed. We are deeply committed to your musical enjoyment. Thank you for your continued loyalty to our products and efforts.

Sincerely,

Roger West,
President
1-28-2003