Unusually strong KT66 pair

Discussion in 'Glowing Bottle Tube Amp Forum' started by HH1978, May 8, 2019.

  1. HH1978

    HH1978 TDPRI Member

    Posts:
    13
    Joined:
    Feb 6, 2016
    Location:
    Brussels, Belgium
    Hello,

    I've been offered locally a pair of Marconi KT66 at a decent price, but the test results are unusually strong :

    Ia: 117 ma ; Gm : 7,1 ma/v for tube 1 and Ia : 125ma ; Gm : 7,2 ma/v for tube 2. Average NOS reads 65ma and 6,3 ma/v

    Should I go for it, or will it be impossible to bias correctly whithin the range of a JTM45 bias pot? Or is there another reason why I should avoid it?

    Many thanks,

    Hugues
     
  2. dan40

    dan40 Tele-Afflicted Silver Supporter

    Posts:
    1,140
    Joined:
    Aug 19, 2015
    Location:
    Richmond Va
    First you need to know if the sellers tester is in calibration or not. An unusually high reading may indicate his tester is not reading properly unless he knows his tester well. I am also not sure that those test figures from his tester will guarantee a matched pair. Ask the seller if he has tested them in an amp for idle current and how far apart are they. I would also want assurance that they are not noisy or microphonic before shelling out a bunch of cash.
     
    Wally likes this.
  3. HH1978

    HH1978 TDPRI Member

    Posts:
    13
    Joined:
    Feb 6, 2016
    Location:
    Brussels, Belgium
    Thank you!

    The seller is experimented and reputable, so I believe his tester is calibrated. He also has a couple of other KT66, in singles, that read all in the average area. That said, he's a HIFI guy, not really interested in guitar amps, so I doubt he tested the tube in a JTM45, or would be able to do so. I'm not experimented at all regarding HIFI amps, so would that be acceptable test conditions?
     
  4. dan40

    dan40 Tele-Afflicted Silver Supporter

    Posts:
    1,140
    Joined:
    Aug 19, 2015
    Location:
    Richmond Va
    As long as he has tried them in an amplifier (hifi or guitar) and found them to be quiet and somewhat close in idle current, I would feel comfortable. Hifi guys are even more picky about close current matching, so I would think that what he is offering would be a matched pair.
     
  5. deus56

    deus56 NEW MEMBER!

    Age:
    47
    Posts:
    2
    Joined:
    May 5, 2019
    Location:
    Texas
    Here are a pair of NOS[​IMG] Marconi's from my collection.
     
    aerhed, bftfender and dan40 like this.
  6. Silverface

    Silverface Poster Extraordinaire Ad Free Member

    Age:
    67
    Posts:
    9,063
    Joined:
    Mar 2, 2003
    Location:
    Lawndale CA
    No. his results are strong using WHAT tube tester? Those look like results from a Korean or Vietnam War vintage Military TV7/U or TV/10 tube tester - not a modern high voltage unit. Those same testers are the ones most commonly used by Hi-Fi tube sellers on eBay. They are not suitable for guitar amp power tube testing.

    Typical tube testers only put a maximum of 185volts on the plates - far less than typical guitar amp conditions but not out of line for older Hi Fi amps. Hi fi guys may be picky, but test results at guitar amp-like 450VDC - and matching at that plate voltage - is generally VERY different than at 185 VDC.

    Tube testers must also be calibrated every 5 years or so or the results are very questionable. I've owned a couple dozen 1960's Hickok, Bk, Eico and various 1950's military testers over the years, and the ONLY thing they are good for when testing guitar amp power tubes is weeding out bad/shorted tubes. They can't be used to compare quality levels unless they are specialty high voltage testers AND calibrated regularly.

    I weed out bad tubes with a tester, but all matching and quality comparisons are done in actual guitar amps using a Weber Bias Rite (with the plate voltage reading option ) or similar.

    I would not trust those test results.
     
    bftfender likes this.
  7. HH1978

    HH1978 TDPRI Member

    Posts:
    13
    Joined:
    Feb 6, 2016
    Location:
    Brussels, Belgium
    Thanks,

    The seller uses an Amplitrex. But in absence of a proper real life test, I'll try to source another pair.

    By the way, I just check the bias with my current tubes (Shuguang) and had strange readings. Same plate voltage on both tubes (460v with a Mullard GZ34 and 440v with a chinese GZ34), but 10ma difference on plate current (46ma and 36ma, which means one tube is running above the safety limit, though not to the point of red plating). When I exchanged tubes positions, I got the same readings in the same positions, which means the amp is the issue. I checked the resistors values and found out that one of the screen resistors has drifted from 470 ohms to 225 ohms. Could that explain the different plate current readings?
     
  8. Silverface

    Silverface Poster Extraordinaire Ad Free Member

    Age:
    67
    Posts:
    9,063
    Joined:
    Mar 2, 2003
    Location:
    Lawndale CA
    That *IS* one of the modern high-voltage testers. Very unusual to find a seller using one. Is the seller a commercial dealer or a small local shop? Just curious if we might have a good source for properly tested tubes, but would not want to overload a small shop.
     
  9. dan40

    dan40 Tele-Afflicted Silver Supporter

    Posts:
    1,140
    Joined:
    Aug 19, 2015
    Location:
    Richmond Va
    That could cause the difference. Do you measure the same negative DC bias voltage on pin 5 of both sockets? You could remove the screen resistor long enough to recheck the idle current to see if it changes. Just move your screen voltage supply to pin 4 after removing the resistors and recheck the current draw. Be sure to replace the resistor with a proper pair after testing.
     
  10. bftfender

    bftfender Friend of Leo's

    Posts:
    3,659
    Joined:
    Dec 21, 2017
    Location:
    York PA
    what jtm45 ya have ?
     
  11. HH1978

    HH1978 TDPRI Member

    Posts:
    13
    Joined:
    Feb 6, 2016
    Location:
    Brussels, Belgium
    It's a block logo clone I bought from a guy in France. Not sure who made it, the seller told me it was a french sound engineer. With the exception of a master volume at the back, it seems true to original schematic, and the guy used old Phillips mustard caps and Allen Bradley resistors. Transformers and choke are from Magnetic Components, and pots are Alpha.

    However, I replaced the screen resistors with new 2W 470 Ohms. One of the original ones had melted and crumbled to pieces when I took it out, and the other had drifted to 778 Ohms. I checked the resistance across pin4 and 6, 468Ohms on each. I also checked the negative voltage on pin 5, sounded fine at -46V, and tracked it down to the bias diode rectifier. All fine it seemed. When I reinstalled the power tubes, plate voltage read 440V on both tube, but plate current on one tube went through the roof at more than 80ma, and the other went down at around 20ma, the first tube red plated and the rectifier (the chinese one of course, for testing, not the Mullard) arced and I shut the amp down immediately.

    At this point I'll probably bring the amp to my tech as I don't have a variac or a light bulb limiter and don't want to blow other tubes or components. But out of curiosity, any idea about what went wrong?
     
  12. dan40

    dan40 Tele-Afflicted Silver Supporter

    Posts:
    1,140
    Joined:
    Aug 19, 2015
    Location:
    Richmond Va
    If you are reading the same bias voltage on pin5 of both sockets, most likely cause was a bad tube. It's probably what toasted the original screen resistor. If your bias voltage on both sockets does not match, you have a problem in the bias circuit. You can take this reading without power tubes installed if you like.
     
  13. HH1978

    HH1978 TDPRI Member

    Posts:
    13
    Joined:
    Feb 6, 2016
    Location:
    Brussels, Belgium
    Thanks Dan,

    I measured the bias voltage on pin5, -46V on both sockets, so there shouldn't be a problem in the bias circuit. Plate voltage was also the same on both tubes. Is there a safe way to test screen voltage?
    Also, is it safe for the amp to test it with only one power tube? One of my KT66 is definetely bad or dead. I ordered a new pair from TAD, but I would have concerns putting in fresh tubes before I'm sure the amp won't toast them.
    NOS tubes is not an option anymore for the moment, for the same reason.
     
  14. dan40

    dan40 Tele-Afflicted Silver Supporter

    Posts:
    1,140
    Joined:
    Aug 19, 2015
    Location:
    Richmond Va
    You can power the amp up with one tube but voltages will be slightly higher because of the lower current draw of the single power tube. I wouldn't play it like this though. If you would like to confirm screen voltages, just measure pin 4 of each socket. If all of the voltages look good, I would set the bias voltage for the highest negative voltage the pot will allow, and then install your new set of power tubes and bias them up where you want.
     
  15. HH1978

    HH1978 TDPRI Member

    Posts:
    13
    Joined:
    Feb 6, 2016
    Location:
    Brussels, Belgium
    Thank you! I really appreciate the help :)

    I tried tonight, using just one power tube. Now I have one socket giving a solid 460V on the plate, but the other beeing instable, and only in the 350-370V range. I suspect I have a short or a bad contact somewhere, but have not been able to find where yet.
    Grid stoppers are not exactly the same actual value, 5,7K on the first socket, 6,8k on the second, but I don't think that could explain the difference.
    I also checked the coupling caps from the PI, both read the same value on the multimeter (105nF, so within tolerance). )

    Also, looking carefully at the screen supply, I noticed my amp doesn't have the 1K 2W resistor on the schematic. I doubt it is really an issue as the amp has been working fine for years though.

    [​IMG]
     
  16. dan40

    dan40 Tele-Afflicted Silver Supporter

    Posts:
    1,140
    Joined:
    Aug 19, 2015
    Location:
    Richmond Va
    Measuring plate voltage with an unbalanced load on the OT like that will likely cause weird readings. The grid stoppers are only there to help stop any oscillation in the circuit and do not cause any voltage drop. Being a few ohms different on those won't affect anything. I would just check all socket voltages with no power tubes installed. If everything reads good (it will be a bit higher without power tubes) and you know the OT is working properly, you should be safe to install a new set of power tubes and bias them up. Just start with a high negative bias voltage (cold bias) until the amp warms up a bit and then set the bias to correct values.
     
  17. dan40

    dan40 Tele-Afflicted Silver Supporter

    Posts:
    1,140
    Joined:
    Aug 19, 2015
    Location:
    Richmond Va
    While it is powered up, check the power tube side of the PI couplers for any DC voltage. One side will have B+ voltage, but the power tube side should be reading zero or no more than a couple mv of DC. If you do see higher DC voltages on the tube side of one of them, it means it is leaking DC voltage through it and upsetting the bias of the power tube.
     
  18. HH1978

    HH1978 TDPRI Member

    Posts:
    13
    Joined:
    Feb 6, 2016
    Location:
    Brussels, Belgium
    Thanks!

    I'm not sure to understand the last post. The couplers are the 0.1uf connected to the 82k and 220k resistors for one, 100k and 200k for the others, is that correct? If so I've got 268V on the PI side and negative bias voltage on the power side for the first one, and 225V and negative bias voltage for the second.

    I tried with a fresh pair of tubes and the plate voltage went back to normal. The minimum bias is a bit too hot, which is strange as I have a 25k pot in series with the 56k resistor on the schematic, but I found that the resistor between the power supply and the bias diode rectifier has drifted to 240K when it should be 180k. Should I replace that resistor or change the bias pot for a 50k (or both)?

    I tried the amp with the new tubes and it's working, except it tends to squeal when I jumper the channels, which it didn't do before.
     
  19. HH1978

    HH1978 TDPRI Member

    Posts:
    13
    Joined:
    Feb 6, 2016
    Location:
    Brussels, Belgium
    I replaced the resistor before the diode bias rectifier with a proper 180k one, and now the bias pot has a correct range. I biased at 65% plate dissipation and tried to play through the amp. Sounds as great as it always did, no more squeal.
    I'll keep the bias probes on for a while, to see if it stays stable, and if so, I'll put the Mullard GZ34 back in place.

    Thank you so much for the help!
     
    aerhed and dan40 like this.
  20. dan40

    dan40 Tele-Afflicted Silver Supporter

    Posts:
    1,140
    Joined:
    Aug 19, 2015
    Location:
    Richmond Va
    Glad to hear you got it playing again!
     
IMPORTANT: Treat everyone here with respect, no matter how difficult!
No sex, drug, political, religion or hate discussion permitted here.


  1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.