I suggest you strip this down a bit.
Delete C3, C6, and C7
Set R1=0, R12=10k, R19=10k
and let's see what happens.
😎
Delete C3, C6, and C7
Set R1=0, R12=10k, R19=10k
and let's see what happens.
😎
Not at .5 Hz it doesn't.
Me, I don't use Zobels (like .1uf and 5 ohms at output), and I sleep like
a baby.
A fussy baby...
😎
Me, I don't use Zobels (like .1uf and 5 ohms at output), and I sleep like
a baby.
A fussy baby...
😎
You are right, why should that affect the stability of the feedback loop?
A 1uf, 100K RC filter will have 30 degrees phase shift at 2.6Hz.
But, as I mentioned in the earlier post, my AC Bode plot might incorrect since it is the In to Out response of the amplifier, not that of the feedback loop itself. I think I am doing the phase margin tests incorrectly. But that does not explain why the actual build oscillates under some condition.
I am using Tenma 32V, 3A switching regulated bench supplies for my testing. When the oscillation hits the rails, the supplies limit the current. With the 1 watt signal levels, the amplifier should never hit the rails.
Your comment eliminates a design variable. Fussy is usually good; it generates focused attention and respect. Flip side, how would you enable your stable F6 circuit to oscillate at 0.5Hz?
Wow!
With gain=1, the Bode plot phase margin will be 180 degrees. All of the open-loop gain is going to negative feedback. This will not oscillate.
Obviously a voltage gain stage is needed ahead of the input stage if it to be driven "ordinary" audio sources.
What closed loop gain do you think is appropriate for this amplifier?
OK. My capacitor selection is somewhat limited, but I have some electrolytics scavenged from a PC power supply. These are extreme values: 4700uF! They seem to work well but the THD vs Freq plot is about the same as the right hand plot in post #1178. I have not managed to trigger any oscillation. With an RC of 47 seconds, it takes a long time get reach bias voltages (like 5 minutes). I will obtain and try some other cap values like 100uf and 200uf.
lhquam,
if you disconnect the transformer from one of the outputs does it still oscillate/motorboat?
- just for instance the top one gets no modulation, it will be a BA-1 output stage.
I like a method of bias with just a MP cap!
- imho the real reason for the effect might be that
- you use a bootstrap for the top bias
- the bias of the top is C4, 44μF connected to the output (and I just note that value differs from C15).
- sonically this is an electrolytic in the signal path!!🙄
- an alternate method would be to connect the bias to the output and insert a big bootstrap in the middle of the R20 (1k4+1k4) (and treat it similar with SAME cap in a split up R21).
if you disconnect the transformer from one of the outputs does it still oscillate/motorboat?
- just for instance the top one gets no modulation, it will be a BA-1 output stage.
I like a method of bias with just a MP cap!
Your finding that a CLG = 1 gives a stable amplifier is an eye opener, but is counter to that commonly known. Why is it not oscillating? Please post your refined schematic after implementing the suggestions by Mr. Pass.
Your schematic is pleasantly simple. Does the underlined mean after implementing [or not] the suggestions of Mr. Pass in your original schematic?
Yes, after implementing the suggestions of Mr. Pass
Just trying to abstract all datas.
Bigger caps and no cap across zenners.
Not different from Ihquam schematic but drawn differently.
Just trying to abstract all datas.
Bigger caps and no cap across zenners.
Not different from Ihquam schematic but drawn differently.
Last edited:
lhquam: Recall that kasey197 has used a step up transformer [other than Jensen] to further increase OLG without an additional gain stage in the front end.
Nelson explained that he made a typo in post #1182 and was not suggesting a closed-loop gain of 1. I will proceed under the assumption of a gain of around 6.
Does the underlined mean metalized polypropylene? Are you still working with your schematic in post#1023.