Rough TX-300

Discussion in 'Fisher' started by calman46, Apr 29, 2018.

  1. EngineerNate

    EngineerNate AK Subscriber Subscriber

    Messages:
    2,740
    The Bournes 3329 are great. Totally sealed, low PPM, and nice smooth adjustment that holds well once it's set.
     

     

    Please register to disable this ad.

  2. calman46

    calman46 AK Subscriber Subscriber

    Messages:
    1,731
    Location:
    MA
    Ok time to order some parts . 3300uf 63v or 100v to replace the 3000uf 40v originals filter caps of bridge .Stay with 1000uf filter cap off the other rectifier but increase the voltage to 63v or 100v ? The new bias trimpots and Fred's mod resistors are all of them 1/2 watt ? The original trimpots are 2W on the TX-300. I was thinking about adding and inrush current limiter or a Triac to save the power switch ,what do you think? any suggestions ?Thanks
     
  3. larryderouin

    larryderouin Turn it UP, POP? PLLUUEEEZZZZZEE Subscriber

    Messages:
    21,760
    Location:
    Glen Burnie Md.
    3300uf 50v is sufficient. You're only pushing 34.5V thru there. C14(1000uf) can go up to 50V also as the B- is -38V. No real need for 63 or 100v. I believe all of Fred's resistors mod calls for 1/2w, but not sure on the trimpots. check with him. Fisher may have put in 2W trimmers just because they were only thing available that fit.
     
  4. fred soop

    fred soop Super Member

    Messages:
    1,965
    1/2 watt is fine. Those trimmers are dissipating around 50 mW. No idea why Fisher used those monsters. A good sealed trimpot in the original could have saved them a lot of blown amplifiers due to dirty open frame pots. The 3.3 Ω and 6.8 Ω can be 1/4 watt or 1/2 watt. The other bias resistors, 180 Ω, 120 Ω, and 270 Ω (your values may vary) should be 2 watts. Todays 2 watt resistors are smaller than the old 1/2 watt, so replacement of all of this will save space. You may need to remove the resistors to get to the 100 µF capacitors anyway, and they are cheap enough. Just replace them.
     
  5. calman46

    calman46 AK Subscriber Subscriber

    Messages:
    1,731
    Location:
    MA
    img076.jpg taking a look at the actual TX-300 bias circuit and comparing it to the schematic the TX-300 I have is different (#13097D). First the trimpots only have two tabs ,one being the wiper the other tab has been removed. The schematic shows connections to both sides of the trimpot with the wiper also connected to both sides. So I'm not sure if that's is just how Fisher drew it (rookie here). The schematic also has a 22 ohm resistor from the wiper to the other tab on the pot. There are actually two 22 ohm resistors in parallel there. img076.jpg
     
    Last edited: May 11, 2018
  6. fred soop

    fred soop Super Member

    Messages:
    1,965
    There is a service note suggesting changing the 22 Ω resistors to 8.2 Ω. (2) 22s in parallel would be 11 Ω, close to 8.2 Ω. The idea was to provide some limit to the resistance if the pot goes open.

    The pot wiper is only connected to one side. Going the other way from the wiper leads to the 22 Ω resistor.
     

     

    Please register to disable this ad.

  7. calman46

    calman46 AK Subscriber Subscriber

    Messages:
    1,731
    Location:
    MA
    I thought it might be some kind of safety mod. Can you tell by the TX-300 schematic the actual trimpot connections ? It looks like all 3 tabs too me. I think I'm going to rebuild this bias circuit with new parts . Any idea if this mod was effective ? ,thanks
     
  8. larryderouin

    larryderouin Turn it UP, POP? PLLUUEEEZZZZZEE Subscriber

    Messages:
    21,760
    Location:
    Glen Burnie Md.
    Calman; Can you take a picture of the bias pot (both sides if possible or just the connections please) I'm curious how this pot is mounted.
     
  9. calman46

    calman46 AK Subscriber Subscriber

    Messages:
    1,731
    Location:
    MA
    The terminal strip with all the trimpot connections was riveted to chassis with no circuit connection. I removed the bottom of one trimpot and it looks good inside ,nice clean . But these have to go, right?
    20180512_123401.jpg 20180512_123528.jpg 20180512_123600.jpg
     
  10. larryderouin

    larryderouin Turn it UP, POP? PLLUUEEEZZZZZEE Subscriber

    Messages:
    21,760
    Location:
    Glen Burnie Md.
    They should. They were notoriously unreliable. They'd get dead spots or go open and then the fireowrks started.
     
  11. fred soop

    fred soop Super Member

    Messages:
    1,965
    Here is the link for the modified circuit. The resistor values for the 120 Ω and 180 Ω are not exactly the same as the 100 Ω and 220 Ω in your unit but the difference is not critical.

    Schematic

    p3
    [​IMG]

    Overview - 4 bias adjustment networks are visible near the back of the unit

    1407
    [​IMG]

    Closeup of the bias adjustment network

    1424
    [​IMG]
     

     

    Please register to disable this ad.

  12. calman46

    calman46 AK Subscriber Subscriber

    Messages:
    1,731
    Location:
    MA
    I was going to use your mod Fred until I found that my TX-300 has its own. If I use your mod I omit the two 22 ohm resistors on the pot add the 3.3ohm & 6.8ohm and used the values for the TX-300 for the 100, 220,270,330 ohm resistors and change to 2watts ? or just change the wattage on the 100 &220? I can do the work I just don't have the theory knowledge .
     
  13. fred soop

    fred soop Super Member

    Messages:
    1,965
    The 100, 220, 270, and 330 resistors are already 2 watt. I replaced mine because they would have to be removed to get to the electrolytic capacitors underneath and it was cost and labor efficient to replace rather than attempt to salvage.

    You would only need to remove the 22 Ω and the pots (and the emitter resistors) and replace with the revised network. BUT, please note that my emitter resistors are 0.33 Ω rather than the 1 Ω of the original. That will change the bias adjustment. I worked out the 3.3 Ω and 6.8 Ω resistor values by testing. Those values allow the bias to be adjusted between half and double the proper value with the network going between 2.22 Ω and 4.5 Ω with the halfway point being 3.36 Ω. I replaced the emitter resistors mainly to have low tolerance units there. Low tolerance is not necessary for operation but is critical to accurate measurements. Since the original resistors were being replaced anyway, the new value was chosen. Keeping the 1 Ω emitter resistors will change the values in the bias adjustment network.
     
  14. calman46

    calman46 AK Subscriber Subscriber

    Messages:
    1,731
    Location:
    MA
    I get it . I misinterpreted post #24 about the 2watt . I have some 0.47 ohm 5w, 5% WW. I could try for emitter resistors . I need to get the power supply up first . What do you think about what Fisher did with the 8.2 ohm or 22 ohm x 2 ? Is it enough to save the output transistors ?
     
  15. fred soop

    fred soop Super Member

    Messages:
    1,965
    I calculate the bias chain of resistors at approximately 63 mA (assuming ±38 V power supply). That would be 660 mV at the base of the transistor connected to the transformer, assuming the pot went open. (You may note that the math is incorrect, but the transistor itself would draw approximately 3 mA base current, leaving 60 mA for the 11 Ω resistance) Subtract 200 mV for Vbe leaves 460 mV across the emitter resistor. With a 1 Ω emitter resistor, the idle current would be 460 mA or almost 1/2 amp. Power dissipation at idle would be approximately 10 watts per transistor. Changing the emitter resistors to 0.47 Ω would increase the idle current to almost 1 amp and dissipation to 20 watts! The transistors might survive short term but would overheat quickly. Not good.

    Using 0.47 Ω emitter resistors would still require a different bias voltage than the 0.33 Ω that I used. I could do some calculations but there are variables that I can't calculate other than by direct measurement. For example, I don't have the exact value for Vbe or Ibe of the transistors and that will have a large effect on the bias. Also, if your power supply is, in fact, lower than indicated in the service manual (±34.5 V, my 600-T is actually closer to ±39 V) this will affect the voltage at the bias network.

    If you want to work with the original bias pots, here is the process:

    1 - Power off.

    2 - Mark the position of each pot as accurately as you possibly can.

    3 - Mark which pot goes in which position.

    4 - Remove the pots.

    5 - Clean the pots with whatever method you prefer.

    6 - Connect an analog ohmmeter between the wiper and end terminals.

    7 - Adjust the pot over its range. The resistance MUST change smoothly. If there is ANY hiccup, the pot is not clean. Go back to step 5.

    8 - Reset to the marked values and reinstall.

    9 - As a final check, measure resistance from the wiper one more time. All the circuits I've seen show 10 ohm pots so the final measurement should be less than 10 ohms in proportion to the wiper position, but there could have been other values used.

    10 - The actual adjustment procedure involves tuning for minimum IM distortion using the analyzer that all of us have sitting on the bench. If you stay with the original adjustment, check the DC offset at the output. If it is unacceptably high, it is corrected by changing the bias of the upper and lower transistor pairs with respect to each other. This should be a very small adjustment and not enough to greatly affect the bias. Split the adjustment between the 2 pots.

    11 - Proper bias current at idle is 30 mA. Depending on the actual circuit, it may be possible to measure the voltage across the emitter resistor. However, some variations will have this resistor also handling current for the bias network. A direct measurement may not be trivial.
     
  16. calman46

    calman46 AK Subscriber Subscriber

    Messages:
    1,731
    Location:
    MA
    I was thinking about using the original pots when I learned that Fisher had did a mod with the 8.2 ohm (or 2x 22 ohm ) on the pot . But I'm not sure how safe the output transistor are with this mod. On the emitter resistors ,why is a 1 ohm or a 0.33 ohm OK but not a 0.47 ?
     

     

    Please register to disable this ad.

  17. fred soop

    fred soop Super Member

    Messages:
    1,965
    Any of the values for emitter resistors would be ok, but the bias has to be set properly for whatever resistance is used. 0.47 Ω would be fine, but not with the resistor values used in my bias adjustment network or the original network. My explanation in post #35 pertains to worst case with failure of the bias pot. The transistors would overheat quickly with any value of emitter resistor, including 1 Ω.

    The circuit used in the TX-300, 600-T, Heathkit AA-21, and a few others, is from the RCA Transistor Manual from around 1963. The circuit was short lived and by 1966-1967, most power amplifiers were using the Lin topology, which is still used today. It has had much study, modifications, and derivatives over the years and much more is known about that type of circuit than the one we are dealing with here. One of the things that has been determined is that minimum crossover distortion is reached when the voltage across the emitter resistors is approximately 24 mV. This holds true regardless of the resistor value or the resulting current. So, the value selected is a tradeoff. A lot of very high powered amplifiers use 0.1 Ω because at peak output current, this low value will minimize the resistive losses and allow for slightly higher power. But, the tradeoff is very high idle dissipation. These large amplifiers typically also have huge heat sinks and parallel output transistors so the high idle dissipation is not a serious problem. 0.33 Ω seems to be a good compromise, but 0.47 Ω is also reasonable. The 0.33 Ω calculates to 72 mA of idle current, which would be approximately 3 watts of idle dissipation per transistor with ±40 V rails. With an 8 Ω load, the peak current would be 5 A (assuming no losses and no power supply sag). That would be 1.65 V that would be across the resistor rather than the load, so the larger the emitter resistor, the greater the power losses.

    Now, back to our circuit. Since it was short lived, there is not 50 years of research and development as with the Lin circuit or its variations. I don't know the optimum bias, though Fisher set it at 30 mA. I also don't have distortion meters to do a measurement and calculation. So, all of this is educated guesswork.
     
  18. calman46

    calman46 AK Subscriber Subscriber

    Messages:
    1,731
    Location:
    MA
    Ok I will be using your mod as originally planed . Thank you for taking time and explaining it . I have a 1962 RCA transistor manual .I'll see if it has this circuit in it .
     
  19. fred soop

    fred soop Super Member

    Messages:
    1,965
  20. calman46

    calman46 AK Subscriber Subscriber

    Messages:
    1,731
    Location:
    MA
    I have some(20) 400v 6A diodes and a 400v 25A bridge that I want to used . They are over kill but I think this is a good project to use them on. I won't be doing any soldering yet so if you have any suggestions let me know, thanks
     

Share This Page