Help for near-final EF86 TMB Mod Classic?

[whboyd]

JMPGuitars_18W_EF86_TMB_as built near final.pdf

Looking for observations/suggestions on my new build. Haven't yet done the final cleanup/solder touchup, etc. so be kind but any suggestions welcomed.

I built this on a Mojotone turret board which I modified to suit Josh's schematic & EF86. B+ was high, so again followed the suggestion & used two 12V zeners between the B+ & OPT CT. This brought voltages down (see chart).
My concerns/questions are several:

• Even though Josh said the best location for a choke was the dumpster I went ahead & left it in cause it was already installed - 7H 75mA 500ohm and added a 1K resistor in series after it to correct voltages between the first two sections of the cap so it affects the EL84 screens & everything afterwards but not the plates.
  Concerned a bit about the voltages on EF86 & 12AX7s - thoughts? Amp sounds great on both channels but do I need more voltage on the EF86 or either of the 12AX7s?
  Tried cathode bias resistors 125-130-150-180 & 200 and settled on the 125 as the bias calculator shows dissipation of about 87% - only getting 8.5V on cathodes. Your thoughts on this?

I used a Hammond PT and Pacific Magnetics OPT (odd they only put an 8 ohm sec. winding - no 4 or 16).

The amp is extremely quiet unplugged and even with a guitar plugged in, as long as I move it away from flourescents, etc. So I think the routing and shielding has paid off (and maybe even the choke? ) Vox did also use chokes in the orig. AC15s & AC30s but they choked the whole supply so they were huge & heavy- 'nuff said about that!

Thanks in advance for your time & expertise - when i've got it totally finished i'll update this thread with photos & maybe a sound clip or two.

[JMPGuitars]

I'd like to see if Bieworm can jump in here and give you some feedback.

I will only mention two things in the meantime:
1. You have a lot of stray solder bits and stuff that desperately needs to be cleaned up.
2. Nicked/melted wires need to be repaired or preferably replaced.

Thanks,
Josh

[whboyd]

Thanks Josh, with lots of iterations and replacing parts it got pretty messy - any suggestions of best way to clean it out? Would i maybe use Flux remover on a rag? Also the wire supplied by Mojo seems to have a very low melting point. Towards the end I switched to teflon wire which doesn't melt but most was already done. If you have time to send a pic with particular burns/nicks you see it would be helpful. At 72 the eyes aren't quite what they used to be!

Look forward to Bieworm's comments.

[Bieworm]

Will do after the weekend. I only have my phone here with me. And I forgot my glasses. Bedt to go over the photos with my laptop.
Stay tuned!

[whboyd]

At Josh's suggestion I did clean the board & chassis, and replaced a few wires (will def. replace the output jack I melted while doing some of this LOL!) So I replaced the photos with new ones post-cleaning.

Amazing how you can work on an amp for weeks looking at it for hours and then once you take a photo & post it, you see THOUSANDS of things you hadn't noticed! Anyway looking forward to feedback from Bieworm & others so don't be shy!
Thanks to all.

[Bieworm]

OK, after examining the pics I can see that the classics are all there..
- parallel wire runs from the board to the tube sockets. We don't like that kind of lead dress, because even though your amp might appear noise free there is still a big chance of unwanted oscillations. These might be in the higher frequencies and not audible for the human ear, but they still might cause issues. Be aware of that.
- the soldering technique in general needs a lot more practice. Solder joints should be even and concave. Try to avoid stacked solder joints (IE at the speaker input jack)
- soldering components inside the hole on top of the turrets is also not preferred. There is zero mechanical connection and you wouldn't be the first having a component falling out of it. Be aware that turrets might be dirty inside and might give bad connections or solder not sticking to the turret at all.
- burnt wires point out that there has not been carefully moved with the soldering iron. Watch out for burnt components here!
- I would recommend to seal off the ends of the cut isolation of the wires with heat shrink. Now there are ripped ends all over and it's pretty hard to see if there possibly are unprotected parts of the wires that could be hazardous for shorts. IE the heater wires can be very dangerous for they carry a lot of current.

Why are you concerned with the voltages on the EF86 and 12AX7? Could you give us these voltages?

[whboyd]

Thanks so much for the pictures!– they are very helpful and I have a lot of work to do to clean this up! I will probably replace all of the leads going from the board to the tube sockets, And many of the soldering joints. I had to use Q-tips to clean the board with flux cleaner, so yes, that is a little bit of cotton that you pointed out.

My concerns about the voltages at the small signal tubes is on the voltage chart which I uploaded. For some reason, it’s a PDF and when I uploaded it, it didn’t appear as an image in line with the others, but is visible as a link at the top should I convert it to an image file and repost it? The EF 86 has low voltage, as does the PI, and the TMB tube seems somewhat high. Please look at those and let me know if I should try to adjust those voltages with the resistors to the appropriate sections.

Also, I’m a little perplexed about the cathode voltage on the EL 84. I did not get a lot of voltage variation as I used the different resistors – 125, 130, 150, 180, and 200 (never did see more than about 11 V).my question is, with the 125 installed I’m showing about 8.7 V on cathodes, and the biased calculator indicates that to be about 87% of maximum dissipation. do you think I’m good with that or what would you suggest? Also, I’m using an older set of Mullards so should I try some different and maybe more modern tubes? It looks like they are not closely matched since there are different plate voltages using a shared cathode resistor. Anyway, your thoughts on that would be helpful. And thanks again for the pictures – they are extremely helpful even if somewhat embarrassing. But I do want to get this right so we’ll put a lot more work into correcting as many of these areas as possible. i’m a 72-year-old T1 diabetic and I’ve had a lot of laser work on my eyes so even with magnifying glasses and a headlamp I don’t see some of these things as well as I used to, but I will take my time and redo them and I’m encouraged that this amp sounds so good as it is. But I will get it right – This may be my last amps after a lifetime of enjoying this hobby l :Thanks again Bieworm-I really appreciate it and look forward to your reply

[whboyd]

A couple more thoughts…
Great idea to use heat shrink on the ends of the wires to avoid the melted back insulation. I’m thinking about possibly replacing all the tube sockets so wonder if you have a suggested brand/type? These sockets came with the Mojotone small parts kit, and I found them to have smaller holes and be much more prone to melting than the Belton or military surplus two sockets I’ve used in the past. Are you guys recommending ceramic or something else these days?

Also, do you have a picture of an example of the type of lead dress that’s preferred, to avoid the parallel wires you mentioned? I have checked the output of this amp with an oscilloscope , and see no evidence of any high frequency oscillations but I’d like to get it as optimal as possible if I’m going to redo a lot of this work. Thanks!

[Crabman]

I can’t comment on your EF86 voltages, but you are right that the voltage you’re seeing at your El84 cathodes doesn’t make sense considering that your screen voltage is up around 320 V. If you look at an EL84 spec sheet you’ll see that while they don’t show grid curves for screen voltage above 300 V, in your case you’re at 320 V so you should be seeing somewhere around 12V on the cathodes for close to 100% dissipation at idle. The fact that you observe 8.5 V with dissipation calculated well below maximum tells me something is off. Under normal circumstances with 340V on the anodes and 320 V on the screens you would probably cook your tubes with the cathodes biased at 8.5 V.

However, your dissipation calculations appear to be correct based on the info you supplied. Not sure what would cause this since I don’t consider myself a leading expert, but it’s possible that something is off with your EL84s. Check that you measured your voltages properly. You could consider changing out your EL84s if you have a spare pair sitting around. If you decide to do that, remove the 125 ohm cathode resistor first and replace it with something closer to 180 ohms (or more) to avoid biasing new tube too hot.

There could be a better explanation that I haven’t considered?

Dave.

[whboyd]

Thanks. Ill put in a 180 ohm & switch to different tubes then post new voltages.

[whboyd]

Opinions on how i configured the zeners in my power supply? Its slightly different from Josh’s “option 1” but the two 12v zeners dropped the B+ about 22.5volts. I had an xtra 32uf cap section so placed them after that Didnt seem to reduce their effectiveness.

[JMPGuitars]

Questions and Observations

Which version of my EF86 TMB Modern Classic did you build?
There are two. V1 is based on the Vox style combo of the original EF86 Lite. V2 is my take on the EF86.

Your Zeners are after the first filter stage, even though it's not the same as our typical first filter stage. Even if they weren't, the only difference would be their efficiency.

Assuming that you measured the voltages correctly, and entered the values correctly in the calculator, it's probably okay. Dissipation target is 85%. The 12V target is an average voltage using higher cathode resistance. The presence of the choke could be affecting this. End result is what matters. If you want to confirm whether or not it's the choke causing the difference, then you can bypass the choke and temporarily replace the cathode resistor with 180Ω to be safe. Or not, and just look at the tubes to make sure they're not red-plating and move on with your life.

The variance in your EL84 plate voltages isn't that significant. Zaphod_Phil would argue that less than perfectly matched power tubes sound better.

Most important is still to clean up the amp. Blow out any fuzzballs with a strong air blower.

Get yourself a digital microscope with a nice big screen (10" minimum), and you can keep having fun. Or if you have the budget, get a fancier scope like a Mantis stereo microscope.

Belton micalex are my preferred sockets.

If you're going to remove wires to repair them, replacing them with better wires (better insulation) would be preferred.

Per Bieworm's soldering comments, review the videos linked in my signature.

Thanks,
Josh

[whboyd]

Getting there -
After the suggestion of swapping EL84s I noticed one was very loose and ultimately I believe had a broken socket pin on a heater which was intermittent. Replaced both EL84 sockets with Mil-surplus ELCO sockets (dated 1953 )which are really great and huge improvement on the Mojotone sockets - I will be replacing ALL of them for sure!

Anyway this may explain some or all of the biasing mystery. I now have it back together with a 195ohm cathode resistor and will add the voltage chart below - Looks like it's pretty dead on except V1 is a little high, V2 slightly low (negligable?), and the PI a little low. I'm wondering if it's worthwhile to keep tweaking resistor values to get closer to Josh's target or leave as is? I built the V2 version except adapted with 6CA4 rectifier (and the dropping zeners you recommended). Since I started with the Mojotone board & small parts kit, I had to modify the layout to fit and replace lots of parts (their jacks/switches/sockets/pots etc). I also replaced most of the signal caps & many resistors so there's not much left of the Mojo in my build.

As soon as I'm satisfied I have the circuit operating properly I will clean it all up again and address all the issues you guys have kindly pointed out. Great suggestion about the digital microscope - i think I have one never used! Looks like i'm getting close - then on to the cabinet & speaker!

[Crabman]

Glad to see your voltages are back in the right ballpark. Makes more sense now .. Keep in mind that now things are where they need to be, you can play around with your bias. It's ok to bias your power tubes at maximum dissipation when using cathode bias (according to the Valve Wizard). You will notice some tonal variations as you transition between 80% and 100% dissipation. My preference is bias at close to max dissipation. The difficulty is finding an available resistor value that will put you where you want to be. The rules are different when using fixed bias .. should not bias above 70%.

Another point to note is that Robinette's calculator assumes that only 5.5% of current measured at the cathode is from the screens. You will find that more current than this flows through the screens on the EL84's at idle. You can calculate it from your posted voltages. What this means is that you are probably biased just under 80% and not 85%. The 180 ohm cathode resistor will help increase your dissipation and will probably make the amp sound a bit better ... smoother distortion (this is just my opinion though).

As far as some voltages being a bit off .. probably not a big deal. This is discussed in Merlin's books. There are plenty of Marshall amp modders out there that fixate on specific voltages for the various gain stages and claim that the amp will not sound right unless within specific tolerances. I'm not completely convinced of this.

[whboyd]

Thanks for your views crab man! I will experiment with the 180 home (and I think I have some 150s and 175 as well but was reluctant to do so because of the guidance that the target would be about 85% according to Robinettes Calculator. I had also suspected that the variances in voltages of the smaller signal tubes was probably not a big deal since components age and drift a good bit anyway overtime and sometimes people even find that they think they sound better when they’ve drifted away from their original value!

I may put the 180 ohm resistor back in, finish my cleanup of the chassis, sockets and board and revisit my sloppy soldering and then try to get this back into a box with a nice speaker (I have a pristine Vox celestion silver Bell from the 1960s waiting for it) and see how it sounds.

I am curious about the suggestion from Bieworm and Josh that I need to work on parallel signal wires - as I would have thought that the closest path would be best, but if I need to redo some of the board to socket leads, it would be helpful to see a good example of how to run those in an aparallel fashion.
As always – thanks for your observations and suggestions – I just love this stuff! Started with Heath kits and hafler &Dynakits and modifying my old hallicrafters short wave receivers in the 1960s, and have built 5E3 deluxes 5F6A baseman amps, numerous vox clones, and even a 18 W Marshall back in the 1990s to which I added reverb and Tremelo. I have a garage full of “projects” and always thought I would get to them when I retired which I’m very close to now so Hope I can keep it going! At some point, I may reach out to the group to see if anybody wants a lifetime supply of resistors, capacitors, tube, sockets, vacuum tubes, Transformers,… Lol!

Meanwhile, I’m gonna get this one finished up nicely and off the bench so I can play it.
Thanks again for your support!

[JMPGuitars]

Wow, those tube sockets were junk, huh??

Short wires are good, but the wrong wires being parallel to each other can cause big problems, including noise (keep those plate wires away from the grid wires).

85% is what we generally prefer. You can bias however it sounds best to you within reason, but understand that maxing out the dissipation will reduce tube life, and can also add to noise issues or make crossover distortion or oscillations worse.

Your voltages are close enough now that I wouldn't be concerned with the preamp tubes unless it doesn't sound good. If you don't like the way the dirt from the EF86 sounds (too muddy?) then you would want to raise the voltage up a little. That will also slightly increase the headroom. If you like how it sounds now, don't mess with it...unless you're bored?

[Bieworm]

For the layout you can look at some examples on this website. There are lots of examples of good practice, but here's one of my builds so you can get an impression how I do the lead dress.

viewtopic.php?p=258211#p258211

[Crabman]

I need to walk back the comment about class AB cathode bias being O.K. at 100% dissipation. It's really only acceptable when using cathode bias in single ended designs. As Josh pointed out, we should aim for 85% to be on the safe side. Sorry about any confusion there.

When I built my Trinity 18 Watt kit it came with a 150 ohm cathode bias resistor and I measured 12V at the cathode and figured all was o.k. Years later, I bothered to actually calculate it only to find that it was at 110% dissipation! It didn't red plate and sounded very good but I'm sure I've shaved some lifetime off my EL84s. I have since put a 180 ohm in its place. Dissipation now sits around 90% ... probably higher than most since my plates are at about 355 V.