Home |
Search |
Today's Posts |
#1
|
|||
|
|||
is there a paper on how to calculate amount of feedback?
doug wrote: I finally succumbed and added a feedback network to my PSE 6l6 amp just put a 22K resistor between the OPT and the K of first triode in the driver. Took out the WAAAY too prominent mids that I just couldnt stand but I still have to decide how much feedback to use. not sure if I like the result yet or not. much flatter but seems to have lost a lot of the high end that I liked so we will have to see after some trial and error would love to hear some inout from the group though Doug A straight connection of 22k between OPV anode and IPV cathode may upset the biasing of the IPV, since DC flows down the 22k, which displaces the working point of the tube. A 10 uF cap in series with the 22k could be tried, but then you have to worry about LF stability. You DO NEED to work out what reduction of sensitivity of the amp occurs after FB is applied, so if you need 1v input instead of 0.25v to get the same output level, then that's a gain reduction of 4 times, or an applied 12 dB of NFB. If the Ro without FB is high, then the accoustic output of the speaker varies up and down in ways ther designer didn't allow for, since it has been made to suit low Ro amps. The prominent mids could be due to the speakers making more than normal mid levels due to the high Ro. Prominenet mids means an absense of bass and treble. So what is the response of the amp with no FB? FB tends to extend and flatten the response, as well as lowering Ro from say 4 ohms down to 0.5 ohms with 12 dB of applied FB. Maybe you only need 6 dB. Patrick Turner. |
#2
|
|||
|
|||
Thanks Patrick,
Time to break out the meter and not be so lazy. I will start measuring this weekend and report back what I find Doug "Patrick Turner" wrote in message ... doug wrote: I finally succumbed and added a feedback network to my PSE 6l6 amp just put a 22K resistor between the OPT and the K of first triode in the driver. Took out the WAAAY too prominent mids that I just couldnt stand but I still have to decide how much feedback to use. not sure if I like the result yet or not. much flatter but seems to have lost a lot of the high end that I liked so we will have to see after some trial and error would love to hear some inout from the group though Doug A straight connection of 22k between OPV anode and IPV cathode may upset the biasing of the IPV, since DC flows down the 22k, which displaces the working point of the tube. A 10 uF cap in series with the 22k could be tried, but then you have to worry about LF stability. You DO NEED to work out what reduction of sensitivity of the amp occurs after FB is applied, so if you need 1v input instead of 0.25v to get the same output level, then that's a gain reduction of 4 times, or an applied 12 dB of NFB. If the Ro without FB is high, then the accoustic output of the speaker varies up and down in ways ther designer didn't allow for, since it has been made to suit low Ro amps. The prominent mids could be due to the speakers making more than normal mid levels due to the high Ro. Prominenet mids means an absense of bass and treble. So what is the response of the amp with no FB? FB tends to extend and flatten the response, as well as lowering Ro from say 4 ohms down to 0.5 ohms with 12 dB of applied FB. Maybe you only need 6 dB. Patrick Turner. |
#3
|
|||
|
|||
doug wrote:
I finally succumbed and added a feedback network to my PSE 6l6 amp just put a 22K resistor between the OPT and the K of first triode in the driver. Took out the WAAAY too prominent mids that I just couldnt stand but I still have to decide how much feedback to use. not sure if I like the result yet or not. much flatter but seems to have lost a lot of the high end that I liked so we will have to see after some trial and error It may be easier to just try different values of resistance for that 22K resistor. Temporarily connect a pot of value of around 100K and vary it to see if you find a "sweet spot". Also be aware that the triode's operating point may shift, and that you may need to adjust upwards its cathode resistor to get the operating point back to where it was. Also a small amount of capacitence to ground across the cathode resistor may bring back some of the highs you lost. Again trial and error is the only real way, as you probably don't know exactly what frequency you're starting to lose the highs. Do this after setting the feedback resistor. |
#4
|
|||
|
|||
"doug" wrote in message . ca... I finally succumbed and added a feedback network to my PSE 6l6 amp just put a 22K resistor between the OPT and the K of first triode in the driver. Took out the WAAAY too prominent mids that I just couldnt stand but I still have to decide how much feedback to use. not sure if I like the result yet or not. much flatter but seems to have lost a lot of the high end that I liked so we will have to see after some trial and error would love to hear some inout from the group though Doug Hi, Doug, You stuck in the 22k and it sounded better or different. Some bits were better. Some bits were worse. Feedback is not penecillin. It doesn't kill all known germs and like most medicines, there are side effects.... It cannot make a crap design sound good. It may make a good design sound better... It is not always the path to personal happiness.... It could turn your amp into a transmitter and NASA will be banging on your door 'cause you just ****ed up the $70 zillion Mars mission and Voyager 9 is now heading for a quiet suburban street in Alabama !! Patrick is dead right in that you can't just hang a 22k from the OPT secondary back to V1 cathode because it's in parallel, down to ground, with the V1 cathode resistor. What effect this has on the working point of V1 needs to be established. V1 may now be in meltdown !! You need, ideally, a scope, an AF generator, a voltmeter, a pile of paper and a pencil. You draw out the schematic, run the amp in it's original state amd note down all the fixed voltages at the valve pins. If you're happy with this no load/idle state, that's fine.... Then run the AF generator and scope together to establish a frequency/gain graph comparing the input to the OPT secondary. Ideally this will be a flat straight line from 0Hz to 100kHz It won't be, but at the moment you are just collecting information No scope ? No generator ? Not a problem. You are talking about how it sounds. Ears are pretty good. If you can't measure it, listen to it. Come back and tell me what you can generate and what you can measure. If you only have a $10 Walmart multimeter, I can point you in the right direction regards jim |
Reply |
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Forum | |||
Doppler Distortion - Fact or Fiction | Pro Audio | |||
Powerful Argument in Favor of Agnosticism and Athetism | Audio Opinions | |||
Paper for printing CD inserts? | Pro Audio |