Tube Depot Matching Accuracy

0 Members and 1 Guest are viewing this topic. Read 3189 times.

Roger A. Modjeski

Tube Depot Matching Accuracy
« on: 1 Jun 2013, 03:50 pm »
Now and then I get to examine matching from other companies. In this case it was four Sovtek EL-84M tubes that came with an RM-10 that was in for upgrading to MKII status. The tubes were in their original boxes with a Tube Depot sticker on the box with the number 22 (the matching number). There are no electrical units so we don't know if 22 is milliamps or volts or shekels. Whatever they are they were 20% off from the highest to the lowest tube. This is not what I call good matching.

corndog71

  • Full Member
  • Posts: 919
  • Some people call me Rob.
Re: Tube Depot Matching Accuracy
« Reply #1 on: 1 Jun 2013, 04:48 pm »
Having bought many tubes from them, I'm not surprised.  Here's a pic of the last ones I bought - also a matched quad of EL84s.  Looking at the sticker I would guess the number is supposed to represent the bias point. 



They were an improvement over the Valve Art tubes I got from Antique Electronics but not dramatically.  I've been seriously considering buying a fresh set from you to compare.  But I'm saving up for another amp kit that I want to build and am trying to stick to that goal.

I appreciate your comments and investigations, Roger.

Roger A. Modjeski

Re: Tube Depot Matching Accuracy
« Reply #2 on: 1 Jun 2013, 04:51 pm »
Having bought many tubes from them, I'm not surprised.  Here's a pic of the last ones I bought - also a matched quad of EL84s.  Looking at the sticker I would guess the number is supposed to represent the bias point. 



They were an improvement over the Valve Art tubes I got from Antique Electronics but not dramatically.  I've been seriously considering buying a fresh set from you to compare.  But I'm saving up for another amp kit that I want to build and am trying to stick to that goal.

I appreciate your comments and investigations, Roger.

What amp did the go into? There are often ways to confirm the matching in the amplifier.

corndog71

  • Full Member
  • Posts: 919
  • Some people call me Rob.
Re: Tube Depot Matching Accuracy
« Reply #3 on: 1 Jun 2013, 05:36 pm »
I put them into my home-built version of the ST35.  I've incorporated Shannon Park's Driver circuit (http://www.diytube.com/st35/ST35_REVD_2009.pdf) which calls for a 12AX7 for voltage amp and 12AU7 for phase splitter although I'm currently using a pair of NOS 12AU7 tubes I got from Andy at Vintage Tube Services. 

I'm sure you'll be amused by the giant blue caps in my power supply.  They are Clarity Cap TC series 700V film caps. (http://www.claritycap.co.uk/products/tc.php) I wanted to see if the lower ESR would make any audible difference over the panasonic electrolytics I had been using.  Pretty sure it did.  The caps are 45uF, 110uF, 110uF.

I also put in the Enhanced Fixed Bias (EFB) circuit by Dave Gillespie.   As per Dave's instructions I set the bias at 27mA per tube.

What do I need to do to test the matching?  I already have test points for the plates available.  I can always add more.  I'm actually considering rebuilding it in the more traditional way just so I can reach the tube socket pins easier.  When I built this I just wanted to do something different.


Roger A. Modjeski

Re: Tube Depot Matching Accuracy
« Reply #4 on: 1 Jun 2013, 05:57 pm »
Do you have individual cathode or some other resistor per tube to measure across?  In any amplifier you can use the output transformer primary resistance as a test point from center tap to plate. First measure the resistance with the amp cold. If you are lucky the CT to plate of each tube is equal then you can just measure the voltage per tube and the real-time difference plate to plate. If they are not equal one can either calculate the current using ohms law or add a small resistor to the low one to make them equal. This resistor will not change the performance of the amp in any way. I've don't it often. In my amps the winding resistances are equal and I always measure and re-match the tubes in any amp sent to me to minimize the difference to better than 10%. Past 20% the transformer starts to saturate and the tubes should be replaced.

In the Dynaco 35 watt amps it is essential to replace the single cathode resistor with some better scheme as you have done. Have you considered the servo bias ideas that adjust each tube individually and automatically?

corndog71

  • Full Member
  • Posts: 919
  • Some people call me Rob.
Re: Tube Depot Matching Accuracy
« Reply #5 on: 3 Jun 2013, 09:25 pm »
Ok, just to be clear on this, I should measure the resistance of the output transformer (http://www.dynakitparts.com/store/schematics/Z-565.pdf) from the plate which looks to be connected to either the blue or blue/white wire to the CT which looks to be the red wire. 

Hmmm... I tried it and can't get a reading.  Maybe I need to disconnect the red leads from the bias circuit.
« Last Edit: 4 Jun 2013, 01:37 am by corndog71 »

Roger A. Modjeski

Re: Tube Depot Matching Accuracy
« Reply #6 on: 4 Jun 2013, 04:30 am »
Ok, just to be clear on this, I should measure the resistance of the output transformer (http://www.dynakitparts.com/store/schematics/Z-565.pdf) from the plate which looks to be connected to either the blue or blue/white wire to the CT which looks to be the red wire. 

Hmmm... I tried it and can't get a reading.  Maybe I need to disconnect the red leads from the bias circuit.

From the link you sent me  http://www.diytube.com/st35/ST35_REVD_2009.pdf  there is a 10 ohm resistor to ground at the bottom of the cathode resistor string. (R43,44,45,46). He gives you bias instructions. You don't need the output transformer method when you have known cathode resistors. It is good to check their value before turning on the amp as they sometimes drift in value. The value is critical to the accuracy of the measurement. Now if you want to know if the tubes you have are matched you could set all the bias pots to the same spot  and see how the currents match up.

For the transformer method, nothing is disconnected or changed in the circuit.

corndog71

  • Full Member
  • Posts: 919
  • Some people call me Rob.
Re: Tube Depot Matching Accuracy
« Reply #7 on: 4 Jun 2013, 05:34 pm »
From the link you sent me  http://www.diytube.com/st35/ST35_REVD_2009.pdf  there is a 10 ohm resistor to ground at the bottom of the cathode resistor string. (R43,44,45,46). He gives you bias instructions. You don't need the output transformer method when you have known cathode resistors. It is good to check their value before turning on the amp as they sometimes drift in value. The value is critical to the accuracy of the measurement. Now if you want to know if the tubes you have are matched you could set all the bias pots to the same spot  and see how the currents match up.

For the transformer method, nothing is disconnected or changed in the circuit.

I'm using the driver part of that circuit.  The bias circuit I'm using is this...


Roger A. Modjeski

Re: Tube Depot Matching Accuracy
« Reply #8 on: 4 Jun 2013, 06:10 pm »
I'm using the driver part of that circuit.  The bias circuit I'm using is this...




Disconnect the cathode ties between R1,2 and R3.4. Now you have 4 separate 10 ohm resistors to measure across. Why he put them in parallel makes no sense at all. You can leave them untied. Your neg probe can go to the "bias tp common" the plus probe to each cathode. Press hard or use clips to make a good connection.  You are looking for something around 30 mV on the 200 mV (not mA scale). If you want the exact dissipation then measure the cathode to plate voltage and multiply by the current.

corndog71

  • Full Member
  • Posts: 919
  • Some people call me Rob.
Re: Tube Depot Matching Accuracy
« Reply #9 on: 4 Jun 2013, 07:08 pm »
Disconnect the cathode ties between R1,2 and R3.4. Now you have 4 separate 10 ohm resistors to measure across. Why he put them in parallel makes no sense at all. You can leave them untied. Your neg probe can go to the "bias tp common" the plus probe to each cathode. Press hard or use clips to make a good connection.  You are looking for something around 30 mV on the 200 mV (not mA scale). If you want the exact dissipation then measure the cathode to plate voltage and multiply by the current.

The cathode resistance is supposed to be 5 ohms 1W but it could be inferred that all he had on hand were hand-picked 10 ohm 1/2W resistors and so that's what he used.  R6,7 are both 180K but in series to reach the needed value/wattage because that's what he had on hand.
« Last Edit: 4 Jun 2013, 08:10 pm by corndog71 »