T4B research: LDR behaviour: -Interesting!

As part of the Bloo compressor kit project, and by way of ensuring that the product really is every bit authentic to the original, we've been doing some more in-depth T4B research.

Some discoveries have been most enligtening. We ran some tests along the lines mentioned in the UAudio webzine article on T4B aging and response patterns. The Bloo T4 behaves in exactly the same way. On our bench today, we had (amongst other experimental variants) three original UA T4B's, and three Bloo recreations. One thing which struck us was the difference in distortion on the various T4s. -Not compression-related distortion, -distortion with no GR applied. -Removing the T4 on our test "mule" gave a baseline distortion figure of 0.08% THD at 1kHz, +4dBm in/out level, 600Ω in/out impedance.

Fascinating thing: The distortion creeps downward after you pass tone through the optical unit for a while! Looking at the schematic, it became apparent that the unit doesn't even have to be powered on to achieve this, -it seems that just having a signal present across the shunt LDR improves the distortion over time!

The distortion improvement can be dramatically accellerated by "blasting" the LDR with signal. (Note: Signal, *NOT* light!!! -Blasting it for a long period with high-Gain Reduction settings on a working compressor will burn up theEL panel, not to mention stress the driver tube!) In about 2 minutes, we reduced the distortion on one T4b from 0.4% to an indicated 0.08%... (essentially unreadable above the noisefloor and residual distortion of the test mule unit) by blasting +30dBm into it for 2 minutes.

However, the distortion seems to creep back in after the T4 has been unused for a while, and right now I'm wondering if longer, slower 'Burn-in' produces a longer-lasting result...

Anyhow, suffice to say that on one of the first T4B modules which were made, the distortion appears to be permanently too low to measure, and our current line of thinking is that it might be because of how long it sat on the bench with tone being fed through it, running sweeps, checking thresholds, etc...

Anyhow... here's a practical benefit for anyone with an LA-2a or clone: If you're going to be using it on anything delicate, you might like to try running 1kHz through it overnight at a decent level.... put the 'compress/limit' switch into 'compress', and leave the unit powered off, with plenty of signal fed to the input. -By morning, I reckon you should have a well burned-in T4B module, and I'm pretty certain that I can guarantee you will have unmeasurable T4B distortion...

I'll post more as I discover it, but this is really interesting to me. -Anyone have any idea why this is happening, on a physical level? -What is it about cadmium LDRs that is happening here?

This is also nothing to do with the behaviour mentioned in the UA webzine article by the way... this is very different. The UA article only discusses optical reactions and behavours... I'm not even totally sure that they know about this behaviour...

Keith

..I should add that I suspect any kind of sustained, high-level signal might 'burn-off' the distortion, as it were. If you don't have a 1kHz oscillator, a looped CD going into it at a decent level overnight might have the same effect... we'll see.

I wish I could get hold of a minty-fresh, never-powered-up-since-it-left-the-factory, new UA T4 and measure it in its virginal state... I'd be curious to know if it too distorts until it's been 'broken-in'... I rather suspect that it would!

Anyhow, I think we're going to search for a 'worst-case' T4B... if we can find one with 1% or more distortion, then we'll feed it some Black Sabbath overnight and see how it feels in the morning!!!

Keith

Strange but interesting!
Obviously a side effect not planned on by Lawrence.
I wonder what a frequency sweep across one of those cells would reveal.
I have some really weird rejects from the box of 100 cells that I tested.
One cell starts out slower than the others when hit with light, then passes the ref cell up!
Another one passes the ref cell up and then slows down and the ref cell passes it by.
The rest are either too fast or too slow.
I do a listening test as a final verification. Some cells will cause the compresor to sound harsh during compression.

You ought to try just exposing a T4 to daylight, then putting the lid back on to see how the distortion is.

Cells take about three days to settle to their normal off resistance.

I plotted some measurements of a series of T4s (including the one which you made for me!!! -thanks again and again!!!) into Excel, and graphed the results.

The light was not an issue, strangely enough! Yours behaved exactly as the UA versions did! One of the UA had matched cell readings better than anything I could have imagined... (An early one I suspect!) the other had poorly-matched cells indeed. There are two others here which have been used a lot, and I should pull them and measure them as well if I get chance.

Oh yes, I built a jig using an octal relay base, fed with audio from a crown 800 Watt amp... (should be just enough, I reckon! :wink: ) and measured both cells simultaneously using a pair of fluke meters... makes things very easy indeed.

Keith

im putting the one that has .4% dist on over night.... well see how it affects the LDR's long term

keith check your pm's

> The distortion creeps downward after you pass tone through the optical unit for a while!

Yes. I never measured, but I could hear it. And it isn't permanently low: my opto-boxes liked to have program running for a half hour before the show. If I used them every day, this was not critical, but if they sat idle for a month it would take 15-30 minutes to sweeten them up again.

CdS and kin are very funky mixtures, as much mechanical as chemical. The nearest thing is the carbon mike, except opto-chemical instead of pressure. Carbon mikes also improve with a little use, settle with disuse. In some situations they sinter and you have to beat them back to looseness.

Distortion will generally be lower at lower level, any age/history. The high-level optos on the PA system are much harsher than the one on the recording rig, and in retrospect I should have run them even lower. But even with 2V peaks, they are not real obvious when going to tape. It was CD that unmasked the distortion. It is almost pure 3rd, right?

[quote author="PRR"]It is almost pure 3rd, right?[/quote]
'Zactly.

So pure in fact, that (Scenaria might remember this nugget, but I didn't make a big deal of it) when I turned up the monitor control on a 'bad' T4B, it sounded like I was listening to a pure 3kHz sine, instead of the usual 3 + some phase-linked 5 + some 7 + some... you know the sound.

Since the Neutrik had recently been exhibiting some tendency to sudenly switch frequencies without my telling it to (later traced to dry air & static... if I rubbed a shoe on the floor and 'sparked' the front panel it caused the Neutrik to swap level, waveform, frequency or some random setting) and it was so clear a waveform that my instinct was "It's jumped frequency" rather than "that's the distortion component..."

Your point about carbon mics did bring up a thought I had, -if they are shaken or abused, it might be linked, I'm still looking into it. -With one T4, it stayed good after a 'burn' for several minutes, when left in the unit, passing silence. If however, we removed it and re-plugged it -even if only for a minute- some or most of the distortion was back when it was re-plugged... That one puzzled me for a minute, so I'm going to try 'tapping' a few to see if there's anything to that.

Of course if it were physical, then LA2as that sat in racks in studios might not suffer as much as those in mobile racks...

More info as I dive into it... I might not be able to get to it today if I end up staying off work.

Keith

[quote author="SSLtech"]Anyone have any idea why this is happening, on a physical level? -What is it about cadmium LDRs that is happening here?
[/quote]
I'm not an expert on this by any means - but I do recall from semiconductor device physics class that Cadmium is not thermally stable - and if it gets warm, then lattice distortion takes place and the crystal bonds alter, and this appears to be reversible. (this is a gross over-simplification, btw). Strikes me that if you apply a voltage across a piece of it, you will certainly excite the lattice - there's not much doubt about that. But if this is a thermal effect, and it's caused by conduction, then one might perhaps wonder whether you could achieve the same effect faster by increasing the conduction? I.e, illuminate it and apply your conditioning voltage (which will increase the current, and therefore rate of temp rise, through it).

Might also be worth checking (if you are able to) what the ambient temperature of the cells is when you test. I wonder if there's any correlation between the distortion measurements and this?

But it is a fascinating observation about the distortion, nevertheless!

Aah.. We're aiming for a indirectly-heated LDR topology here..?

Finally something new in opto compressors..

:razz:


Jakob E.

[quote author="gyraf"]Aah.. We're aiming for a indirectly-heated LDR topology here..?

Finally something new in opto compressors..

:razz:
[/quote]
Why not? (if it works) :wink:

shh

were already working on this concept :razz:

> re-plugged it -even if only for a minute- some or most of the distortion was back

Clean the socket.

Or don't use octals. They really are not good for this. They are fine with significant current flowing (plate and heater power), and do OK with zero current (grid). But a little tarnish distorts signal currents, and these currents are not high enough to burn-through the tarnish.

Yes, I know the octal TB box is part of the mystique. But I think it would be better with a proper low-level contact, or even soldered.

[quote author="PRR"]Clean the socket.[/quote]
Good thought. Will do, if only to eliminate from enquiries. This is currently a convenient way to swap and compare, beyond the T4B module mystiqe...

Contact tension is good, and contacts appear clean, however, I'll give everything a scrub, just to be sure.

Keith

Reminds me of the old "crystal oven" to keep oscillator frequencies stable in the days before phase-locked loops...

What's interesting is that some of the older, 'well-worn' units don't seem to ever exhibit this effect when we test them. The same socket in all instances, so I'm ignoring that as a factor, I'll clean the plugs however, to remove them from suspicion.

I have the option of borrowing an older and apparently very little-used T4B from another lab member, and I've just found out that another friend of mine has a box of new-old-stock original T4Bs that they're hoarding as spares for their collection of vintage LA-2a's... I'm going to write and tell them about this effect and ask him if he can either run a distortion test, or send me a couple on the strict understanding that I test them, learn what I can and then send them back unmolested.

Keith

I have had the one that was at .4% under constant signal since last night around 9pm

I figure on letting it burn in for some time to see if theres any permanent affect...

alrighty..

it didnt make a darn bit of difference..

I just dropped the T4 on the AP... it first measured about .3% dist and within 30 seconds has setteled around .11%

I powered the compressor down.. pulled the T4 out and the dist dropped to .08% (negligable at that point as the compressors opperating dist sits at .08%)

so I powered down a 2nd time.. pulled the T4 out again and reseatted it... with the same dist numbers of .08%

I should mention that theres +4dbu@600ohms 1Khz going into the box..

im really impressed with the low dist values at this point..

now I just put in a ua reissue T4.. it was sitting on the shelf overnight...same values... .08% THD

just tried another bloo T4 with the same results.. .08% dist...

I think the conclusion as PRR suggested already.. when using this compressor I would power it up and send some signal into it for a minute or two.. if the box was innactive for weeks or months then it might need signal present for a longer period of time to drive the dist into its settling value..

thermal heat doesnt seem to affect it too much... I applied a blow dryer to the T4 and could get a shift in values...from .08% to about .18% but setteled back down to its home base in about 10 seconds once the heat was removed..

I am thinking this...

it might be wise to provide a space of ventaltion on the top to allow breathing room .. not that you would necessarily hear that amount of distortion but im sure squeezed in a rack the heat would build up and shift the THD values on the LDR's

I used one of those Miller type NOS mil spec ceramic sockes that you have to use a crowbar on to pry out the T4. Since I was doing a lot of plugging and playing.

Well, if these T4 cells cut down on distortion, what about using a cell across an audio input on a mic pre or something to cut down oistortion?
:?

no whats really being seen is they actually inrease dist until they settle down... once they settle down they appear have no affect on the boxes dist at all.. which obviously is a good thing..

the T4 pulled from the compressor and it measures between .07 and .08% THD into 600 ohms at 1Khz

so essentially there is no change once the T4 settles down