Black Lion Audio introduces T4BLA optocell for optical comps

Already referenced in WA 2a mod thread!

$90-100, but you have to order from one of their dealers not direct from bla.

I congratulate you on the release of this product. While compatible with my units I can't recommend it as I have not tried it yet. I look forward to doing so

However, describing Bill Jone's work (the master and most knowledgeable man in this planet on optical cells) as sorely lacking is almost offensive and I am not saying this because we use his work. I remit myself to facts.There is no better T4B in the world than the one made from Bill Jones (Kenettek)

I should know as

1) It sounds 100% (yes, 100%) identical to my vintage UREI's 2A's optical cells
2) I have auditioned every single T4B from every manufacturer in the planet and choose Bill's by a mile.

Funny enough,  we are recording sound samples this week of the usual suspects as well at ours all at the same settings and will be using a comp3a as the test bench due to the simplicity of the circuit and lack of tubes.   Seemed like the most neutral option to use as a test rig.

Gosh, I really hate that I have to get involved with this discussion. I usually try to keep a low profile and let my work speak for itself, but some of the claims made here are so outlandish that I feel I have to set the record straight.

I have a T4BLA in house and put it through our standard quality control tests, the same tests every Kenetek T4B goes through multiple times before it goes out the door. The T4BLA failed every quality test. Here is a brief explanation of the most important tests and how the unit failed.

The two most important tests are pulse test and the dynamic range test. The pulse test shows attack and release times. A T4B is hit with a 1 kHz pulse for 0.5 seconds and then the response is measured for 11.75 seconds. The response is measured for 0.25 seconds before the light pulse hits in order to show the condition of the T4B in its initial state. The X-axis is time in seconds and the Y-axis is resistance in ohms, shown on a log scale. As you can see, the two photocells in the T4BLA have starkly different pulse responses. They don't match at all. Not even close. What this means is, the meter is not going to show what's actually going on with the gain reduction in the unit. The audio will actually be more dynamic and responsive than the meter indicates. Also, the meter will be indicating gain reduction when none is actually taking place. In a Kenetek T4B, these curves will be so close they almost look like one line.

The next graph, the dynamic range graph. This test measures the attenuating/compressing action of the T4B across the entire spectrum from no compression to full compression. The X-axis is number of volts AC RMS applied to the electroluminescent panel. The Y-axis is resistance on a log scale. Once again, the unit I have in house, which is brand new, out-of-the-box, show drastic differences in the audio and meter cell behaviors. Not only do they show dramatically different light-to-dark curves, they also show different thresholds. In the case of this particular T4BLA, it means that the meter is going to show compression when no compression is actually going on. It also means that the amount of gain reduction indicated will not be what's actually happening with the unit. So, this T4BLA will cause the meter to indicate more gain reduction than is actually happening.

Finally, the meter cell in this T4BLA fails the "time to dark" test, which tests how long the photocells recover to their fully dark state after the signal has been removed. The standard for Kenetek T4B's is 180 seconds (3 minutes). In this T4BLA, the meter cell takes around 750 seconds (12.5 minutes) to reach a fully dark state. This means that the meter will be indicating significant gain reduction when no gain reduction is actually taking place. It also measures cell memory, which affects the transient response. All photocells have a memory, but cells with excess memory have poor transient response.

I felt the need to expose this information to the pro audio community so that buyers can make an informed decision, and to provide a rebuttal to the claims that the T4BLA is "an improved version compared to what is currently available".

Thank you,

Bill Jones
Owner

The Stam SA2A I got to use for a while was truly great. I can’t imagine wanting to change it at all.

Been biting my tongue since the initial post. Will attempt to continue to do so.

First of all, we are not here to argue whether or not Kenetek makes a good module or not. We doubt anyone would. We actually salute the effort he puts in and wish more companies were willing to do testing of that quality. We both have pretty stringent test and quality standards so we feel that this effort and energy would be better spent arguing against cheaply made and badly tested modules that get by on gimmicks or suspect marketing.

From our lead design designer Jesus Ortiz:

I don’t know how accurate Kenetek's claims can be. He’s claiming that we don’t pass his tests. That is, the QC process for his modules. The essentially arbitrary values in the tests made for his specific modules. I feel that it wouldn’t be a stretch to say that HIS module is bad because it doesn’t pass OUR test values. It just depends on what side of the table you’re sitting on. While I understand the Kenetek unit has been around longer, why would we assume that his unit is all the correct values and the ultimate benchmark as to what a correct T4B is? Again, it depends on which side of the table you’re on getting the benefit of the doubt. (1)

In his first graph labeled T4BLA Pulse response. I believe it’s supposed to show us the attack somehow but the time scale is in seconds. It’s just not possible to verify an attack that’s originally spec’d at better than 1ms with a scale in the seconds. I included the scope graph of a pulse at 1KHz of both units shifted to be seen side by side. The white one on the right is the T4BLA and the one on the left is the Kenetek. I don’t know that the variations are as drastic as we’re being led to believe. In addition, it seems like both graphs in his pulse response graph are actually pretty similar as long as one doesn’t necessarily pick and choose what they see. The response of the module is the response of the compressor. Yes, there is a vertical variation to the two devices but we should be talking about behavior. Behavior has a lot to do with the shape of the resistance curve instead of just the resistance at specific points. We need to see it as a whole picture taking into account the graph shape and slopes. In a later graph he refers to the affected transient response but in truth, both units seem to be very similar in this case.

For the second graph, it would probably be good to discuss the bigger picture and think of the T4B elements in the actual 2A circuit. Gain reduction using the optoresistors happens because it makes a voltage divider with resistor R6. As the EL pad lights up, it reduces the resistance of the optoresistor which means more signal is lost at R6. R7 is usually shorted across when using compression so for simplicity we’ll assume that we’re using compression. That and the resistance of the optoresistor is actually in parallel with pot R1 which is 100kΩ. That said, if we look at Mr. Jones’ graph of dynamic range response and begin with the 10 value on the x-axis, we can see that one of the modules has a resistance of about 100MΩ and the other is closer to 10MΩ. On first glance yeah, it’s a factor of 10 so it looks like a lot. Now apply that to the 2A. The 100MΩ in parallel with the 100kΩ pot equals 99.9KΩ. The 10MΩ value in parallel with the 100kΩ pot equals 99.010kΩ. Using that with the 68kΩ R6 means that at this voltage point, the unit with 100MΩ has a reduction of about 4.51dB. The other unit with a resistance of 10MΩ would have a reduction of about 4.54dB. That seemingly huge difference equates to a .03dB difference in the gain reduction level. For perspective, that’s around 0.3% variation. In truth this voltage level is not very telling in the scope of the 2A’s performance. It’s essentially no compression. (2)

If you plug the modules into an 2A the compression curve of the unit looks like the image below. The yellow line is the T4BLA and the Magenta is the Kenetek unit. This graph is the common output vs input linearity that most use to show a compressor’s behavior. Both T4Bs were plugged into one of our B172A’s 2A side. For consistency only one B172A was used to avoid possible variations. No settings were changed when switching the T4B elements and each unit was given 5 minutes of about 10dB gain reduction, then in this case 200 seconds of no signal in order to clear the light pad’s charge. I tried this test in various different variations and kept on noticing one interesting thing. If you look at the area where the knee is, there’s what seems to be a bump on the Kenetek curve. I tried several different way but that bump persisted. It seems to be an overshoot of the reduction element. In some variations of this test it was actually worse and the knee of the unit was actually much sharper. The curve doesn’t seem to recover from it so I feel like because of the overshoot pushing harder, it wouldn’t be a stretch to say that the Kenetek unit is actually the one that’s giving incorrect performance. The graph below is of the two T4Bs run with no clearing time allowed. Essentially it went from the 10dB (the AP was sending about 15dBu into the unit) of gain reduction to the -20dBu input signal that begins the automated graph. The green is the T4BLA and the Cyan is the Kenetek T4B. One would assume that an issue with the “Time to Dark” of the unit would show with a drastic variation in the initial signal level on the lower left corner of the graph since the T4Bs are going very quickly from heavier compression to about 20dBu below the threshold point. As can be seen the two units are almost exactly lined up. Also, testing the predicted meter functionality in the way that Mr. Jones points out isn’t the most accurate way as if I remember correctly, the meter element has an integration time closer to 1.4ms. Put differently, it rolls off above around 700Hz to maintain smooth response. Above that point it gets averaged out. (3)

On the opposite side of the spectrum, this graph give the two T4Bs 800 seconds to clear or return to “Dark.” The Red is the T4BLA and the Green is the Kenetek T4B. The bump just past the threshold point is still there but actually more pronounced with a harder knee on the threshold point. In fact, in all the graph I took with all the variations, the T4BLA maintains a pretty consistent smooth knee. (4)

While I’m very proud of our ability to objectively design and analyze products, we’ve never been a company that relies overly heavily on graphs and datasheets for our design processes. In truth when we were designing the T4BLA and tuning the various elements in the module, we were only doing listening tests. We were tuning to the specific audible behavior of the unit. Not really the linearity of it’s resistance on a graph. That’s actually how most of our products and mods go. It wasn’t until we were close to where we liked it’s tonality and behavior that we began doing actual electrical analysis to ensure consistency and catch and odd behaviors or issues. Certain things like attack and release times should never be left solely to subjective testing. Maybe if we had Mr. Jones’ test criteria before we began the design process we could’ve purposely designed it to pass his tests but we figured listening to it was a more practical way.

Mr Jones makes a good T4B module and so do we. Like most components some folks may prefer one over the other when they hear them. This isn’t a black and white thing. There’s plenty of grey area to explore. When you consider the terrible quality of some of the other T4B elements we tried, we’re in excellent company. Considering how we both are with testing, I feel that the customer can’t lose. That’s how it should be. They shouldn’t have to worry about the technical grading and numerical description on a datasheet. They should be able to know that the unit is a quality piece and worry only about the sound of it.

Still biting my tongue, which is getting a bit sore.

I will say that one unit has sides that match, one doesn't.

Hey Paul give Anthony at ADL a shout, the man has a reputation for being very kind when it comes to service pricing.

Thanks guys, I can’t send it to the U.S. at present but I’ll see if I can get a pin out. BLA are having a laugh with their claims. They should do their homework.

Returning emails would be a start and being truthful would help.