Pentode noise - what to do...

[Mark Buckingham]

I have a homebrew amp that started as a Rt66 type of thing, but into el34's.

The amp is in a tweed super chassis & cabinet. 2x10" weber signature speakers.

After having microphonic/rattling issues with two different ef86's, I tried a 5879. This had less rattling, but didn't sound as good... The mid's and high's that made the ef86 special were gone. Additionally, after playing for a while, the rattling came back.

So... I disconnected the cathode bypass cap on the pentode, and put 1/2 of a 12au7 in front of it. There is a cap and a gain control between the triode and the pentode. This sounds good, but doesn't cure the microphonics on the pentode. I was thinking if the pentode wasn't the first tube, and if it's gain was reduced, then the noise floor would be high enough to not matter. Instead, even with the gain from the 12au7 up (or even with a 12ax7), the 5879 rattles like heck.

If you can ignore the rattle, it sounds pretty good though... On the other hand, I can't ignore the rattle.

I've tried O-rings on the pentode. I haven't tried shock mounting it yet, but it's only 6 inches or so from a speaker, so I'm not sure if that would help. (And even if it would help... Would I need to get a new pentode? Once these things start rattling, will they rattle no matter what you do?)

I guess I have a few options at this point:

1) drop the pentode, put a cascoded 12au7 in there instead. I don't know if the tone stack would have to change. (It's still the Rt66 type.) I also don't know that a 12au7 wouldn't be microphonic, but I wouldn't expect it to be.

2) weber now sells a shock mount adapter for 9 pin tubes. I could try that.

3) start over (again).

Any ideas?

Thanks.

[stokes]

Are you sure the rattle is from the EF86?If so try a different tube.I dont have a lot of faith in these tube dampers.They may do some good to help prevent a tube from getting to the point that the tube gets shocked to the point that some components inside get loose,but if the components in the tube are already rattling, damping the outside wont likely help.

[Mark Buckingham]

Are you sure the rattle is from the EF86?If so try a different tube.I dont have a lot of faith in these tube dampers.They may do some good to help prevent a tube from getting to the point that the tube gets shocked to the point that some components inside get loose,but if the components in the tube are already rattling, damping the outside wont likely help.

The rattle is definitely from the pentode. I tried a NOS tesla, and a new svetlana ef86... I then moved on to the 5879, and it was noisy out of the box... Not _as_ noisy, but still too noisy. When I tap on the tube, I hear the rattle...

So... three different tubes, all with their own unique rattle. :-)

I guess if I thought shock mounting the tubes would work, I'd do it, but I think putting a pentode 6 inches from a speaker might be asking too much.

[Rob Mercure]

Hey Mark,

The 5879 was specifically designed to be the "front end" - input - for an amplifier and is an extremely rigid microphonic free design Tektronics used them extensively in their tube scopes. But I'm quite sure that it wasn't designed to be part of a combo amp so you may just be subjecting the input tube to too much speaker vibration. Also, since your 5879 sample seems to be only one tube, you could have a bad example of that design and don't forget any contribution due to tube socket tension/corrosion. Hmmm, haven't been following the latest productions - has someone recreated the 5879 or is this NOS? If not NOS I'd suspect a bad recopy (you can not "reissue" what you didn't "issue" to start with). And it never hurts to shock mount tube sockets but as far as the "o-rings" go you first need some special lubricant to make these work - just sent me $33.95 and I'll catch a copperhead in the back yard and I guarantee you the extra virgin pressing.

Rob

[Mark Buckingham]

Rob-

The tube I used was NOS... I'm pretty sure they're out of production.

I guess it couldn't hurt to try another one. I probably sure try the shock mount thing at the same time...

Can you look at this thing from weber's site, and let me know what you think?

https://taweber.powweb.com/store/chassis/sm9.jpg

I think it might work... I just want to get to a "done" point with this amp. :-)

[Rob Mercure]

Mark,

Well, the simple graphic looks OK but I really can't tell much. This was discussed recently and simply mounting the tube socket tabs between rubber washers works and may provide a quick way for you to decide whether to invest in a more expensive solution - you should be able to remount the socket using the washers without unsoldering the connections. Oh, and don't neglect the possibility of a microphonic component other than the tubes fooling you.

But since you mentioned tapping on the 5879 you simply may have a bad tube. While these were a "premium" tube in their day that doesn't preclude a bad production run or the unit being kicked around a few times over the years.

A slight aside (not aimed at you Mark): It really annoys me when some "expert" trashes or lauds a particular tube make/type based on a single or couple of samples. When you're dealing with tubes that haven't been make in a few years you have no idea whether a particular sample has been well treated. I've got a couple of NOS 8417s in the box that I "inherited" when a friend died in 1987 which I know had set on a shelf for 5-6 years. But in the 20 years I'd had them I've moved several times and despite my best efforts I wouldn't swear that they had not taken a knock in transit that would affect them. I'll find out when I finally use 'em but if they're not "perfect" I'm not going to blame GE thirty years later for a bad batch!

Rob

[Mark Buckingham]

Thanks Rob. I'll see if I can jam a couple of washers in there. Maybe hunt down another tube also...

I wasn't blaming the 5879 for the noise by the way... I figured it was my own fault for the way I mounted it.

[Rob Mercure]

Hey Mark.

Didn't mean to imply that you were - but was, poorly I'll admit, trying to tell you not to be discouraged when a sample size of one doesn't work. I myself have butted my head against the wall far too many times trying to get a defective part to work and blaming myself for the failure. I like "double redundancy" - that is, unless I've got another, almost identical, circuit to try a tube in I like to try at least two other tubes to ensure that the tube's defective. But I'm old enough to have pulled TV tubes off the parts shelf and had them be defective only a few months from the factory and/or have tubes, such as an 12AT7, that worked well at audio but were unstable at radio frequencies. Which means if you do have a funky 5879 hang onto it - it may make a dandy "third" stage in something.

Rob

[Mark Buckingham]

Some new results

Mark,

Well, the simple graphic looks OK but I really can't tell much. This was discussed recently and simply mounting the tube socket tabs between rubber washers works and may provide a quick way for you to decide whether to invest in a more expensive solution - you should be able to remount the socket using the washers without unsoldering the connections. Oh, and don't neglect the possibility of a microphonic component other than the tubes fooling you.

Rob

Just an update... On your suggestion, I looked for other sources of noise, and found out that the speakers were rattling against the baffle at certain frequencies. Pulled the speakers, and put in some weatherstripping, and put them back in. I'm getting a lot less noise from the cabinet. (Thanks for the idea.)

I wanted to figure out if the pentode was really noisy or not though, so I wired a standard 12ax7 stage up, pulled the pentode, and jumpered the 12ax7 to where the pentode connected to the tone stack...

Basically, the noise went away... I lost quite a bit of gain too, of course. Strange thing: I like the sound of the triode better. It actually works really well with the Rt66 tone stack. The only real problem with this is that there's only enough ooomph from the preamp to just start to make the power section distort.

So... I'm going to add another 12ax7 stage after the tone stack, between the volume control and the phase inverter. (Or maybe between the tone stack and the volume control.) -- I think that should be enough to bring things up enough.

Right now, I'm running my old tube driver in front of the amp, and it sounds great, so I think I'm almost done with this thing. :-)

[Marvino]

Hi Mark, I don't know whether you're still reading the thread, anyway here's some tricks I've learned with pentode preamps and getting good quality audio.
-from the thread, yes JJ makes the EF806S pentode preamp with the same quality specs as the early ef86s' but with better electrode structure IMHO.
-Make sure you have a quality power supply, S/B Pi choke filtered and loaded with bleeder resistor. Use a RFI filter at the AC input.
-Pentodes with the suppressor grid not internally connected to the cathode, should be connected to ground or better yet to filtered -bias (yes) if your amp is fixed bias.
-don't try to get a ton of gain from your pentode. Set the screen voltage divider with a large plate to screen resistor, and like 150k to gnd. Also decouple the screen to ground with good quality 22uf and a foil .1uf caps.
-decouple the plate resistor from the supply by adding a 10k resistor in series with the B+. Decouple the junction of the 10k and the plate resistor with another 22uf and a .1uf to ground.
-Use coax from the guitar jack to the input network.
-pentode preamp grids are VERY sensitive. Cathode biased input grids are VERY SENSITIVE to tiny changes in self bias. Excellent common star grounding around the circuit is critical.
-If you look at some very old pentode preamp input networks, you'll find the preamp tube DC isolated from the guitar input with a high quality foil .1 to .47 uf cap. This allows the tube's cathode resistor to hold a steady bias voltage and maintain headroom for your guit pickup signal. Today's guitar pickups are very hot and can easily swamp a pentode way beyond parameters designed for 40 to 50 year old pickup signal levels. You realize your guitar pickup, level control, cord, etc. is part of the resistor coupled network to the preamp grid. Consider using an Auricap for input DC blocking and signal coupling to the input grid.
-Make sure your filament circuit is floating and balanced (no higher than 6.3VAC rms) and should be DC biased at about 70 clean volts to the filament CT. Just make a resistor divider at the end of the PS string (330K, 100K 1w's), de-couple the junction with a 22uf and a .1uf. connect the +bias to the fil winding CT or if none, use two 100 ohm resistors accross the fil winding, connect +70v to the junction of the 100 ohm res'.
- Use a higher cathode resistor than you see such as in an old AC30 circuit. This give you less gain, more headroom and more sustain. Yes it's fine to use a good quality cathode by-pass capacitor for a nice round quality, I prefer NPO ~ 33 to 40 uf. I like a lot of lows; especially for alternate tunings.
-Loading of the pentode ouput makes for susceptibility to microphonics. Loosely couple the pentode plate of the EF806S or your choice with a high quality cap in the range of .0047uf with a 470k to gnd to a paralled 12au7 stage set for low gain. Here you will get some signal current, enhance that round sound, and be able to drive a tone stack without loading the pentode.
Yes you can run a pentode quietly. Don't expect to run this in a little combo cabinet with the pentode hanging next to the voice coil!
Best wishes and good luck! Marvino

[Mark Buckingham]

Thanks Marvino, those are good tips.

I've actually un-pentoded the amp... I just wanted to get some closure on this thing & I didn't want to order more tubes or spend any more time goofing around with the thing.

I actually ended up with this:

input->(1/2 12ax7)->resistive divider->gain control->paralleled 12at7->Rt66 tone stack->volume control->phase inverter.

It works pretty well, not quite as nice as a single ef86, but still pretty nice.

Sometimes, you just have to be "done" with a difficult project. :-)