Tubemeister 18 // … the meltdown fix

auto-bias circuitry ?! … hmmm, is that reliably possible …?!

I’m always curious to see what kind of approach these rocket engineers will take to complicate the basic tube amp design … as if the Traynor auto-bias fiasco wasn’t enough to scare everybody in the industry out comes Hughes-Kettner with a line of amplifiers that sports a micro-controller based biasing scheme … since an ATMEGA88 AVR chip contains maybe 80,000 transistors that’s quite a jump in circuit complexity, beyond the dozen or so tube elements we tend see in typical tube amps …

01-tm18-guts

04-tm18-controller

the amp had fried a power tube and let it run long enough to cook the tube socket real good as well … inductors L1 and L2 apparently had suffered from heat also, but everywhere else things looked ok visually …

02-tm18-l1l2

03-tm18-friedsocket

the amp belongs to a friend, and I told him before I started that there would be no guarantee the amp would work reliably because of the cooking, let alone just getting it going again … he digs the tone of this amp so much he asked me to try anyway …

I first determined that one of the power FET’s acting as gates on the power tube cathodes was screwed (stuck open, but my guess is that it had probably shorted beforehand while the corresponding tube cooked // who knows, it’s weird …) … choice: either I replace the busted FET, or not … but not knowing if the “smart” bias circuit hadn’t done its job correctly I decide to forgo trying the circuit out with a new FET and bypass the whole works (and accompanying issues) and install a good old-fashioned manual bias circuit instead …

since we don’t know how the AVR chip is programmed, or whether the right kind of FET’s were even chosen for the job, chances of bias failing again was a distinct possibility … it failed once, could fail again … I always shoot for maximum reliability and so figured this was the right approach to take

don’t let the complexity of the mother board fool you, this is doable :)~

first thing I did was remove the power FET’s // I thought I’d replace them at first but like I said changed my mind after thinking about it more … if you follow the edited schematic (shown below) you’ll see what resistors needed to be lifted or removed in order to completely disengage the auto bias board from the output stage … that is, R76 and R77 (10k control lines from AVR board) are completely removed, as are R44 and R45 (220k grid bias) …

the Cathodes of both power tubes are first hard wired to TP11 (GND) // this needs to be done with strong wires and generous amounts of solder // the Test Point pads are perfect for landing fair size globs on the board … in principle, shorting out Q14 and Q15 respectively (removed) and R89/R91 (1r5 bias measuring resistors no longer used) … taking them out of the picture completely so to speak // more importantly, giving the cathodes the best connection to GND possible … nothing beats a solid piece of wire for reliability

05-tm18-tp11

06-tm18-cathodes

I then installed 200k grid bias resistors from the tube socket pads // close enough to the stock 220k value

07-tm18-gridbiasers

then, just a matter of installing a Fender style 10k panel mount bias pot in series with a 22k/1watt resistor fed by TP10 (-59v) … the bias pot is then grounded at the chassis by soldering the bottom lug to the pot cover // just like in a Fender amplifier

08-tm18-tp10biaspot

09-tm18-back

that was basically it for the bias circuit … a circuit comprised of only 4 parts, a pot and three resistors // somebody please calculate the statistical reliability factor for me …

at last I added a 25k rev-log pot in series with the light-strip’s current limiting resistor so that the annoying blue glow could be dimmed down a fair bit …

matched/balanced power and PI tubes were supplied by www.eurotubes.com (my main tube supplier) … the bias was performed using tranny shunting … the amp worked perfect, the owner said it sounded better than before (everything is subjective) … glad it worked out

SCHEMATICS:

tubemeister18-modded-schem-mods

tubemeister18-schneider-back-board-schematic-l1-l2

tubemeister18-schneider-bias-schem

if you feel like retracing the steps I took in your very own Tubemelter 18 please make sure to double check everything ten times before proceeding with circuit alterations // reversing a wrongly removed part might prove difficult here … these “loose” notes are meant as a VERY ROUGH GUIDE, please proceed with caution

all pictures and schematics posted at highest resolution …



5 comments on “Tubemeister 18 // … the meltdown fix
  1. Lance Morgan says:

    This was a helpful guide. Glad you got it all back up and running. I’m planning on doing your lighting modification using the 25k rev-log pot because the glow gets on my nerves as well. Any chance I could get a before/after close up photo of that part of the project? The schematic helps but just thought it’d be helpful to see a little more detail on the actual resistor and pot in question. All the best – LM

  2. jcm says:

    HI LANCE … TO BE CLEAR, LEAVE THE ORIGINAL 680r RESISTOR IN THE CIRCUIT AND “ADD” A 25K POT IN SERIES IN THE CIRCUIT … ALL I’M DOING IS RUNNING A WIRE TO THE BACK SIDE OF THE CHASSIS WHERE THERE IS ROOM FOR THE POTENTIOMETER

    TUBEMEISTER 18 - LED DIMMER CIRCUIT

  3. Lance Morgan says:

    That worked like a charm! Thanks again. It’s nice to have some control over the lighting and make it more subtle.

  4. EAN says:

    Very nicely done! Do you know what part I would use to replace L1 and L2 on the powersoak? I see they’re 1mH inductors 5%, but I’m not sure of any other specs and can’t find any info online.

  5. jcm says:

    HI EAN, … THE ONLY OTHER SPEC YOU WOULD NEED HERE IS THE POWER DISSIPATION SPECS … IF YOU LOOK IN MY PICTURES, THE L1 AND L2 INDUCTORS LOOKED COOKED IN THE AMP I WORKED ON – SIGN OF “HIGH ENOUGH” CURRENT … SINCE THE AMP PUTS OUT AROUND 18 CLEAN WATTS AND AROUND 25 FULL-DIRT WATTS (FROM A PAIR OF EL84’S) i WOULD GUESS LOOKING AT THE POWER-SOAK CIRCUIT THAT YOU DON’T NEED MORE THAN A 5WATT RATING FOR THOSE … WHAT’S IN THE AMP NOW LOOKS LIKE 1WATT INDUCTORS, BUT I COULD BE WRONG … MAYBE 2WATT INDUCTORS IS OK PROVIDED AMP DOESN’T GO INTO MELTDOWN OR FREAKOUT … I ALSO WONDER IF THEY WEREN’T PUT IN THERE SIMPLY TO ACT LIKE FUSES – I KNOW IT MIGHT SOUND SILLY, BUT THAT’S ALWAYS A POSSIBILITY … I SAY THIS BECAUSE I DON’T SEE 2mH TOTAL BEING ENOUGH INDUCTANCE TO MAKE A NOTICEABLE EFFECT ON FREQUENCY RESPONSE IN THAT CIRCUIT – THO I’D HAVE TO DO MORE ANALYSIS TO CONFIRM … IF IT TURNED OUT TO BE THE CASE, YOU COULD REPLACE THEM BY A SHORT (VERIFY IF SOUND IS SAME OR NOT – CONFIRMING INDUCTORS AREN’T THERE TO SHAPE SOUND), OR BY A FUSE …

Leave a Reply

Your email address will not be published. Required fields are marked *

*

This site uses Akismet to reduce spam. Learn how your comment data is processed.

360+ Pro Fuzz/Preamp

360+ProButton

360+ DIY Kit

360+diyKitButton

Mu3 DIY Kit

360+ProButton

IFMTA (book)

FWboard-Button

IFMTA (PDF)

FWboard-Button

Champ (PDF)

uChampButton