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.

    Dismiss Notice

EHHA Rev A - Interest Thread

First
 
Back
46 47 48 49 50 51 52 53 54 55
57 58 59 60 61 62 63 64 65 66
Next
 
Last
  1. runeight
    You can always dial up the bias a little, maybe 125mA or whatever you can get within the range of the bias pot. Just watch the temps on O/P devices.
     
    For me, I don't mind running them hot, especially since they are mosfets. But I know that many of you prefer otherwise.
     
    125mA will bump the power dissipation from ~3W to ~3.75W. I don't know how much it will drop the input to regulator voltage.
     
  2. ujamerstand

    From memory, my measurements at TP3 was around 9.6V. The AC line in my area was 122.8VAC at the time of measurement. That's probably why it's a touch higher than your readings.
    Quote:

    I've got the bias set at 122mA right now. (From 270mV / 2.2ohm, is this correct?) The pre-reg voltage stabilizes at ~38.6V under the higher current draw. The heatsinks for the output devices gets hot to the touch, I'll have to get a temperature sensor to see if it is less than half of the mosfet's junction temperature. I'll bump the heatsinks to 2.5 inch version otherwise. [​IMG]
     
    Quote:
     
    So what do you think? Is the SumR iron recommended? Underspec it a little?
     
     
  3. luvdunhill
    what about mounting a pair of chassis mount bridges and then build up a CRC that feeds into C1 (bypass D1-D4)? I think a little more capacitance might be nice to have and this would allow you to deal with your excess voltage. Grab a chassis mount resistor for the "R" as well. Even with doing that, you should consider the LM317HVT version in case something were to go wrong.
     
  4. luvdunhill


    Quote:

     
    It doesn't work that way. If the heatsinks were at half the max junction temperature, the device would be dead. For a TO-220 device, you want to be able to keep your finger on the heat sink for at least 3 seconds before having to pull away. This will be about Tj = 80C.
     
  5. ujamerstand
    I am using LM317HVT. Mouser did not carry lm337hv though, digikey has it, but not in TO220 packaging. Honestly though, the heatsink feels warm to the touch for the regulators, but not hot. If I understand this correctly, the main problem here is with the power dissipation from the high voltage drop, right? Or is there something else I'm missing here? Edit: reading the data-sheet on LM337, it does say it has a maximum input voltage at -40V. -_-
    Quote:
     
    Huh. I see. As a rule of thumb, what would be a safe temperature to operate the components in regards to The junction temperature? On amb's beta22 boards and heatsink page he recommends that the junction temperature should never go above 100degrees Celsius for a mosfet with maximum junction temperature at 175 degree Celsius, that's where I got the half the junction temperature thought from.
     
     
  6. ujamerstand
    Also, a little off topic here, but how do you calculate the voltage drop over particular values over the CRC regulator? Say if I use the same values for the capacitence in the bijou, 470uF, maybe a 5 ohm resistor, the Vloss would be the current draw of output stage* 5 ohm = ~0.6V?
     
  7. runeight
    So based on two builds with the SUMR transformer we have high, but not death, input to regulator voltages. What is the primary spec on the transformer? Is it 115V or 120V?
     
  8. luvdunhill
    The maximum input voltage is usually 37V for most LM3X7 types, unless you shop around. As you suspected, the issue is not voltage drop, maximum input voltage which is just one of the design parameters for the device itself.
     
    100C is too hot for me, and I suspect this number was thrown out as the devices used in the beta22 and sigma22 are especially rugged devices, and the designer was familiar with how hard to push them. I'd recommend 80C for a TO-220.
     
    Your suspicion is correct for the Pi filter. Perhaps check out PSUD2 for modeling this sort of stuff:
     
    http://www.duncanamps.com/psud2/ 
     
    Also, once you have the resistor dropping the correct amount of voltage, make sure you also then calculate a suitable power rating. I'd recommend double the wattage from the calculation that you perform using quiescent current draw (assuming this is a class A amplifier)
     
  9. jdkJake


    Quote:


    120VAC
     
  10. jdkJake


    Quote:

    I am running the LM317T Fairchild part from Digikey. It is rated at up to 125C.
     
    Last time I measured:
     
     
    temps (top center of each heat sink).

     

    MOSFETS at 220mV:   38.25V at the regulators

                                       31.35VAC at the secondaries

                                   ~44C to 45C

     

    I might bump the MOSFET a bit, but, I am mostly gonna keep an eye on it for now. Down here in Florida, the local power company (Florida Flicker and Light aka Progress Energy) is notorious for playing with the power levels. I suspect they are high right now as we come out of the summer months and peak HVAC demand. I would not be a bit surprised to see 121VAC or less by months end. Just no way to predict it.

     

     

     
     
  11. ujamerstand
    Thanks for the link, though it should be http://www.duncanamps.com/psud2/index.html 
     
    As for the particular device I used (Fairchild LM337T), the datasheet says the maximum adjustable output voltage is -37V, but it does not mention the maximum input voltage. What IS mentioned is that the maximum voltage differential between Vi and Vo is 40V. Should I really be worried over this?
     
  12. jdkJake


    Quote:

    Case in point, tonight, right now, I am measuring 123.3VAC at the IEC inlet.
     
    MOSFETS are at 220/230mV respectively and input to the regulators is at 38.3V.
     
    Might be totally different in a few hours once peak demand settles down (usually after 11pm EST).
     
     
  13. jdkJake


    Quote:

    I see what you mean, the spec is more concentrated with Input/Output differential:
     
       http://www.fairchildsemi.com/ds/LM/LM317.pdf  
     
    BTW, ujamerstand, are you using the tube LEDs? I am dissipating over 1W per board just to keep those bada-- LEDS putting out 4000mcd a piece.  [​IMG]
     
  14. ujamerstand
    Nope. No pretty leds for me. Realized that I don't like that with my minimax build. More of a natural tube glow kind of guy. [​IMG]
     
    AC line readings tonight is 124.1VAC. 
     
  15. jdkJake
    Fair enough. They do put off a nice subtle glow when unlit from underneath.
     
    Man, your power is almost as whacked out as ours. Might be easier to just regulate the AC coming in than try to burn off excess.
     
First
 
Back
46 47 48 49 50 51 52 53 54 55
57 58 59 60 61 62 63 64 65 66
Next
 
Last

Share This Page