ST-70 mono makeover

[deicide67]

If I wanted to have another go at trying feedback with the LED bias, should I just add 100 ohms between the LEDs and the ground and use the point between the LEDs and the resistor to inject the FB?

Thanks!

[DF96]

What your measurements appear to show is a highish output impedance so voltage rises at speaker impedance peaks (16Hz (bass resonance?) and 2000Hz (crossover or tweeter resonance?)). Was this with or without feedback? My guess is without - as this lumpy frequency response into a real speaker is exactly what I would expect to see without feedback.

If you are using LED bias for the first stage then inject feedback below the LED, although above probably won't do any harm either.

[deicide67]

Hi,

Yes, this is without feedback. Yep, it is all easy to see looking at the impedance profile where these things occur.

If I inject below the LED, do I need to put a 100 ohm resistor between the LES and ground?

What value would you start with for a feedback resistor?

Thanks everyone for all your help!

[20to20]

Originally Posted by deicide67

Hi,


What value would you start with for a feedback resistor?

Thanks everyone for all your help!

Blair,

Have you determined what the gain of the amp is?

[deicide67]

Hi 20,

At 1K with .101V input, I get 11.42VAC output. So, 113.07 is the total gain factor right?

11.42vac/.101vac= 113.07

[DF96]

Yes. What gain do you want? Or, what feedback factor do you want? What values of feedback and cathode resistors gave you the bass problem?

[20to20]

Originally Posted by deicide67

Hi 20,

At 1K with .101V input, I get 11.42VAC output. So, 113.07 is the total gain factor right?

11.42vac/.101vac= 113.07

Yep, just don't mix your scope against VOM readings. (peak vs. RMS)

I did a quicky calc from your table and came up with 225, ( 4.5/.02 ) so not sure why those numbers are so much higher.

[deicide67]

Here are the results from a few FB resistor values. Looks like 33K is the most linear to me.

I had to inject the FB between the LEDs to get it to work. Is that a no no?

[20to20]

Originally Posted by deicide67

Here are the results from a few FB resistor values. Looks like 33K is the most linear to me.

I had to inject the FB between the LEDs to get it to work. Is that a no no?

The table shows your gain increases with FB. That cannot be allowed as that would be positive FB.

LED's are a gimmick. Get your amp working with conventional means first.

If you use a pot for the FB resistor you can dial down the gain to make your input sensitivity realistic and give you the output power max you want. Then set it with a fixed resistor. But remember as you change the FB resistor which is in parallel with the driver cathode resistor you will shift the biasing. So start with a slightly higher K resistor value and then add in the FB signal through the pot. If you are just stabbing different FB resistors onto the cathode you are shifting the bias point without realizing it.

[bear]

Often, amps are set up so that about 1v input drives them to full output... this is NOT a requirement by any means, just what is approximately the usual situation.

_-_-bear