Rubber or silicone gaskets for tube socket bases to reduce vibration?

[dchang0]

The slightly-microphonic EF86 tube has been tamed!

I ended up using just two of the tiny o-rings and not four (one on top and one underneath each of the two mounting tabs) because it was too difficult to get the tiny o-rings to fit with the big o-ring in the way.

The aluminum shield base is not screwed down tightly. It can be tilted about 10 degrees. The ceramic socket itself does not tilt much, since it is sitting on the flat o-ring underneath it. The o-rings are not compressed much--I tightened the screws down just enough to make all the mating surfaces touch. The nylon insert locking nuts make this possible.

With this setup, I can continue to use my brand-new $30 Tung Sol gold pin EF806S until its guts eventually vibrate loose and it becomes audibly microphonic even with the shock mount.

Oh, one thing I found out is that it is definitely better to buy the next size down of tiny o-rings (McMaster Carr SKU 1182N005 (3/32in I.D., 7/32in O.D.)). These 1182N006 (1/8in I.D., 1/4in O.D.) o-rings can shift around under the screw head as it is tightened. I had to push the o-ring into place to center it under the screw.

The three layers of heat shrink do help with the microphonics; I will leave them there for now in spite of them trapping heat. Eventually they will be replaced with the $5/pair large silicone o-rings sold as tube dampers or the $29 Herbie's Audio Lab Ultrasonic Guitar tube dampers, both of which will expose more glass to the air for cooling.

Another alternative I saw somewhere recommended by an amp builder is to cut the heat shrink tubing into 1/4" rings and put those at the top and bottom of the tube--same idea as with the large silicone o-rings but it would allow the tube shield to be installed.

(I do not actually know if the Herbie's Ultrasonic Guitar tube dampers work or are worth the expense. Many people vouch for them, but how much do they reduce microphonics by? I can quantify what the heat shrink trick did based on how much more I could turn the volume up.)



THANK YOU TO EVERYONE FOR YOUR HELP!

[trobbins]

Not many take the effort to find a mechanical solution - but it sounds like you enjoyed the challenge.

The EF86 has internal microphonics above 1kHz. Adding some extra weight to the valve itself may continue to lower the microphonic output - although once the microphonic output is not noticed then it's difficult to gauge how much it still colours the sound, as it will always be there to some extent.

I've used a closed cell neoprene rubber strip, adhesive backed, 3mm thick with good effect. A hole punch and a scalpel blade to match the noval base, and nyloc nuts to provide a nominal amount of compression.

Mullard identified the 'statistics' of identifying a good microphonic valve amongst a batch, which is consistent with what people find nowadays:
http://dalmura.com.au/projects/Micro...n%20valves.pdf

[dchang0]

Thanks for the detailed article--very interesting read.

Do you happen to know whether the Herbie's UltraSonics really work? I imagine they do, but for six times the cost of plain old silicone o-rings, they probably aren't six times better...

[trobbins]

If you can get some weight on the valve, and also use it to dampen the glass tube along its length, then that can be worthwhile I reckon.

No I'm not in to that kind of commercial product. But not everyone wants to obliterate or cover an aesthetically pleasing label that they can refer back to a nuanced review of how a particular manufacturers EF86 can make a big difference.

At some stage I was going to play around with adding a vibration source (ie. a mobile phone with vibrate on, or a speaker with the cone loaded with extra mass and a white noise signal) to an amp chassis and monitor the no-signal output with a spectrum analyser, and do some before/after tests and tube-rolling to look for comparison changes.

Ciao, Tim

[dchang0]

It would be interesting to see what you measure, if you do decide to add a vibration source.

Maybe debunk some of the myths that surround certain brands of EF86, or maybe support them...

[trobbins]

Statistics are important to appreciate - which is why the Mullard report provides a good level of confidence. Any other reporting where the sample size of each option is low, and the history of each sample is not identifiable, makes for speculative reporting.

The vibration source is a quandary - I do have some new, but spare, compact long throw drivers (Peerless 830986) that I could attach to a chassis via a magnet, and add some temporary mass to the voice coil cover that would still allow somewhat linear excursions across a wide frequency range. I guess the intent is not so much to make perfectly linear vibration generator, but to make a generator that is not too 'peaky' and provides enough vibration across a useful frequency range (not just for EF86, but also large output valves appear to have noticeable microphonics at low frequencies - such as the 807).

[dchang0]

Statistics are important to appreciate - which is why the Mullard report provides a good level of confidence. Any other reporting where the sample size of each option is low, and the history of each sample is not identifiable, makes for speculative reporting.

The vibration source is a quandary - I do have some new, but spare, compact long throw drivers (Peerless 830986) that I could attach to a chassis via a magnet, and add some temporary mass to the voice coil cover that would still allow somewhat linear excursions across a wide frequency range. I guess the intent is not so much to make perfectly linear vibration generator, but to make a generator that is not too 'peaky' and provides enough vibration across a useful frequency range (not just for EF86, but also large output valves appear to have noticeable microphonics at low frequencies - such as the 807).

Good point. We would not likely be able to get many samples of any particular tubes nowadays, especially the highly-sought-after and expensive NOS ones... Oh well, one can dream...

It would be useful to some enterprising tube seller to come up with a reliable test for measuring microphonics so that they could truly (financially) guarantee a certain maximum level of microphonics for any particular tube they sell. Right now, I am going on the promises/claims of various sellers that their tubes are low or no microphonics and then finding out with disappointment that I have a throwaway tube...

What would be the limitations of simply using the speaker in an already-existing combo amp? For instance, take an AC15, attach the speaker to some other amp to produce your white noise, then measure the EF86 in the AC15? Could this serve as a viable vibration generator that works in a useful (to guitarists) frequency range with minimal effort? Or would there be some limitations that render the findings skewed or useless?

[TimmyP1955]

For those who didn't look at the video linked to earlier, I grabbed a screen shot: