Announcement

Collapse
No announcement yet.

6L6GC, wired one-pin off CW kills bias probe

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • 6L6GC, wired one-pin off CW kills bias probe

    Having found a set of Groove Tube 6L6GC’s set aside in a box from years ago and needing a replacement set for the present Fender Twin Reverb on the bench, I proceeded to check them out.

    I normally have my Compu-Bias Octal Probes in V8 & V9 positions, checking just one pair of tubes at a time for matching purposes, going thru the ever-changing collection of 'pulls' accumulated. This time adding a matched pair of Sovtek’s with their other pair from the quad set being way north and south. The first pair (of these Groove Tubes) installed were matched, yielding 24mA. The second pair installed (hot-switching in Standby, with no HT present), after warming up, produced a small audible pop, with the bias meter now indicating hundreds of mA as I immediately switched back to S/B The current reading remained coming down very slowly, something I’d never seen happen before.

    I pulled that tube out, and reached for a different one, leaving it’s partner in V8 position. After those warmed up, switched out of S/B and again had this hundreds of mA reading on V9. Uh oh. Powered down, removed the tubes, then the bias probes, and checked the cathode resistor in the suspect probe. OPEN. Ah, rats!!! Went looking for my spare probe (I recall it being questionable), and had replaced it. Nowhere to be found.

    Went on line to contact Randy Lineberger at Compu-Bias to see about getting a new pair of Octal probes. Phone number not in service. Started to just order a pair, then stopped. Checked their distributor, only to find that firm was no longer handling Compu-Bias products. Uh oh. I use this instrument way too much to call it quits over one probe failure….still not yet knowing why it failed.

    Deciding to finally try to cut this damaged probe open and see if I can restore it, I got my narrow X-Acto saw blade installed in the red handle, mounted the probe into my vise, and began carefully cutting along the glue line between the socket and tube base, then stopped.

    Did a little more digging on line to see if there was any report of Compu-Bias going out of business. Couldn’t tell for sure, but did note a lot of people were unable to raise them….and those reports were a couple years or more old. Then, I thought to check my last emails with Randy, and found he had given me his cell phone number. Called it, and got his son’s wife on the line. Told her who I was, and wanted to know if they were still in business, having seen the web site is still up. YES…..they ARE still in business. Michael, Randy’s son is doing the work in the basement of their house, and I spoke with him to see if he had any probes on hand to ship. Yup.

    So, just in case there are other Compu-Bias probe user’s out there, (303) 829-3412 will get you thru to Michael Lineberger who’s ‘minding the store’, so to speak.

    Went and placed the order for another pair of probes, just to be safe, needing one ASAP as it is.

    I resumed cutting open the damaged probe successfully. Then, un-soldering the 8 header pins, I carefully pulled the socket away from the tube base, not all the way, needing to see just how they constructed it. Careful positioning of the Dale CMF series RN60D 10 ohm 1% cathode resistor, a CMF series RN60D 1.00M 1% in series with an RN55 8.25k 1% resistor, set between the Plate and the Cathode, with a wire from the tap (plate voltage scaling network, and a couple leads across the 10 ohm cathode resistor, all connecting to a mini-DIN 8-pin male connector to feed the dedicated hand-held bias meter.

    I opened the probe up all the way, after first un-soldering the socket-side of the 10 ohm resistor. Then, finished de-soldering the other end of it, where it was joined with the 8.25k resistor, both tucked into Pin 8 header side of the probe.

    All I had in precision 10 ohm resistors were the 3W silicon resistors for servicing Ampeg SVT CL’s. Twice the size, but with careful placement, and using Teflon sleeving on the pin-end of the resistor, I was able to park it into place, re-assembled the probe, soldered all the leads back into place, then seeped glue into the open joint I made cutting it open and clamped it into position to cure overnight.

    Next, looked at the tube that caused all this. Odd. There were two leads poking out of the missing pins of the base (Pins 1 & 6). Looking at one of the good tubes and this one, I saw what happened. At the manufacturer, this tube was installed into the header one pin off….Clockwise. That put the heater onto………Pins 3 & 8!! Wanna kill a bias probe….that’s where ya do it….put the plate voltage thru the 10 ohm resistor to Gnd!!

    Then felt really stupid……blinders on…..DID YA THINK TO LOOK AND SEE IF THE HEATER WAS GLOWING???? NOPE. That cost me a few hours of work, lost labor time and $65 for another set of probes.

    At least I found how to get back in touch with Compu-Bias by phone. They do, IMHO, make a really nice instrument….been using one since 2009, with only one probe failure (lost heater voltage). And now I see how to pull them apart…very carefully.

    I just tried the Meter with the repaired probe. Reading went thru the roof again. Uh oh. Disconnected & removed the repaired probe, then tried the other probe in position 1. Same problem. The mis-manufactured tube KILLED the Meter, briefly applying high voltage to the low-voltage sense circuits for Probe 1. I did find my repair job was successful, at least, and have two working probes again. I tried leaving the Probe 1’s DIN connector installed (it turns on the meter), but didn’t connect it to the amp, and checked to see if Probe 2 worked. Works fine. Still, has to now be sent back to Michael to repair.

    So, THAT was an expensive wake-up lesson.

    I at least have a back-up…..many moons ago, I built an adapter to use the two Bias Probes on my HP 3467A 4-Ch Logging Multimeter (to work above 600VDC with KT-88's). I set that up to scale the Plate Voltage as a 1000:1 divider, while reading the current thru the 10 ohm resistor. So it will scan Volts-Current-Volts-Current and repeat, while I send the now damaged Meter back to get repaired.

    Now...what about the tube? Ah, yes. Having seen what they did, I though perhaps I'd just unsolder the pins and see if I could remove the header, rotate it and re-attach it, not knowing if a high-temperature adhesive is used. A little tug on the tube base, and......bloody header came right off!! Wasn't even properly soldered! Removed the solder from the pins, re-installed the tube bottle CORRECTLY, after first cleaning each of the tube wires from the glass bottle. Soldered them back in, cleaned up the overflow, inspected it carefully, then plugged it into the one working probe/channel on the meter. Matches the other 3 tubes.....so, I did at least get a matched quad set of 6L6GC's out of all this effort. The expensive way!!

    Thank You Groove Tubes for your mis-wired tube!!!
    Logic is an organized way of going wrong with confidence

  • #2
    Amazing, kudos for your efforts!
    How did GT manage to test / match it?
    My band:- http://www.youtube.com/user/RedwingBand

    Comment


    • #3
      What adhesive did you use to reassemble the bottle?

      Comment


      • #4
        I presume GT DOESN'T do full testing of their 'matched sets'. Otherwise, HOW would this ever make it past TEST?? Spot check, I guess. I presume these were installed in an amp that immediately blew the AC Mains fuse, and went to the Broken shelf. I further presume that at that time, I had a fresh quad set of tubes popped them in, verified all ok with bias probe meter, and returned the amp to inventory, put the 4 tubes in a foam-lined box for later testing. Later, like years later.

        Having found that these tubes didn't use any adhesive in the assembly, just relatively good fit, relying (once again) on solder connections for mechanical support, I didn't stop to put any RTV or other adhesive into the base.
        Logic is an organized way of going wrong with confidence

        Comment


        • #5
          Maybe GT did their process with an 'as received' button base, then replaced it with a full base afterwards?
          That it matches the other 3 indicates that it probably have worked properly at some stage.
          My band:- http://www.youtube.com/user/RedwingBand

          Comment


          • #6
            Yes.....it occurred to me as well, since this last 'dirty-deed' tube matched the other three tubes, that the tube matching process is done BEFORE the header base is put on. That's the ONLY WAY that tube could have been tested. Otherwise their set-up would have been damaged, while NO heater glow would have been present!
            Logic is an organized way of going wrong with confidence

            Comment


            • #7
              I honestly can't believe that they would take the time to rebase any tubes. How old were these, before or after the Fender buyout?

              Much more likely that they batch tested and then assigned a number to an entire lot. This is probably why so often the match is not so close as we would expect.

              Comment


              • #8
                Unless they don't want to waste bases on all those tubes they have to throw out .
                Originally posted by Enzo
                I have a sign in my shop that says, "Never think up reasons not to check something."


                Comment


                • #9
                  Musta been before Fender bought them? No Fender logo on them. Actually, they were marked 6L6G-S (the amp with them is back in the road case, waiting for a P/S inductor to arrive from east coast, hum bal pots to also arrive from east coast. They wouldn't re-base the tubes.....they'd fixture them for testing without any bases. They looked physically just like a set of JJ 6L6GC's I found I had set aside for later scrutiny.

                  Semiconductor industry does the same thing...all testing is done while still on the large wafer with precision probing fixtures...gotta see what they actually grew with the dies/maskings. Only so many dies, but produce a wide range of part numbers, based on yield results.
                  Logic is an organized way of going wrong with confidence

                  Comment

                  Working...
                  X