Ampeg B-18 issue/debug

[aborgman]

I had a feeling that your problem might lie in the NFB loop once the OPT seemed to be OK. I was unaware that the speaker wiring had been messed with. You still have a problem but if the 8Ω workaround works for you i'd leave it as is.

I may do that - but as an engineer it just pains me not to know why.

One last question: if you take the speaker out from the 16Ω tap with the NFB on the 8Ω tap what happens?? There is something tickling the back of my mind about Loops connected outside the speaker connections. Damned if I can remember precisely what it was though.

I'll try that on Monday and see what happens. I'll have hunt down my 16R dummy load.

 

[aborgman]

One last question: if you take the speaker out from the 16Ω tap with the NFB on the 8Ω tap what happens?? There is something tickling the back of my mind about Loops connected outside the speaker connections. Damned if I can remember precisely what it was though.

8R speaker connection, 8R feedback -> good

8R speaker connection, 16R feed back -> has issue

16R speaker connection, 8R feedback -> has issue

16R speaker connection, 16R feedback -> has issue

Yep... looks like it must be something between the 8R center tap and the 16R tap.

 

[aborgman]

8R speaker connection, 8R feedback -> good

8R speaker connection, 16R feed back -> has issue

16R speaker connection, 8R feedback -> has issue

16R speaker connection, 16R feedback -> has issue

Yep... looks like it must be something between the 8R center tap and the 16R tap.

Hmmm... turns out it's broken even when using the 8R tap for feedback - it's just not evident at 1KHz. You only see the problem below about 400Hz.

 

[beans-on-toast]

If you have the model 470 or 471 OPT in your B18, the DCR specs for the secondary 8 and 16 ohm taps are 0.4 and 0.7 ohms respectively.

 

[aborgman]

If you have the model 470 or 471 OPT in your B18, the DCR specs for the secondary 8 and 16 ohm taps are 0.4 and 0.7 ohms respectively.

I see:

.443 ohm - 8R tap
.581 ohm - 16R tap

Which seems pretty close... but I really have no idea what this transformer is.


Also of note is how it responds to frequency with various feedback used.

If using the 16R tap for feedback -

noise is evident at every frequency. The lower the frequency, the lower power level necessary to see the noise.

Noise looks like this:

If using the 8R tap for feedback -

800 Hz and above, full power achievable with no noise showing.

Below 800Hz - the lower the frequency, the lower power the noise shows up. Power before noise was ~19W @ 20-80Hz, gradually increasing until hitting full power around 800Hz.

Noise looks the same as in 16R feedback case

If using no feedback -

With no feedback it was possible to achieve full power at any frequency above 40Hz.

Below 40Hz the wave began distorting in the 15-20W range - but the distortion was not the distortion seen in the cases with feedback attached. This distortion looked as follows:

So...

With no feedback I never see it, but I do see distortion at low frequencies.

With 8R driving the feedback - I see the characteristic noise, but only at frequencies below ~800Hz. Below 800Hz the noise is evident at lower power the lower the frequency.

With 16R driving the feedback - I see the characteristic noise at all frequencies, but it is evident at lower power the lower the frequency.

 

[beans-on-toast]

Those are interesting results.

The DC secondary resistances look reasonable.

The thing about output transformers is that they are not ideal. There are inherent parasitic reactances, losses, and non-linearities that dramatically affect the performance of the amp. High flux densities in the core causes distortion in the low end. This distortion increases with lower frequencies. Sound familiar? Inductances and capacitances cause additional distortions. Connect the transformer to a real speaker and frequency dependent distortions are even more pronounced. These are things that give an amp its characteristic tone. Some of what you are seeing is normal. It would be nice to perform the same tests with another transformer and see how they compare.

My instinct tells me that you should try lowering the feedback resistor value to maybe 10K and see what happens.

Following that I would summarize your results and send them to a transformer specialist like Paul Patronete ([email protected]). I've discussed transformer issues with him on a number of occasions and he has been very helpful. Tell him that you are trying to determine if your transformer has a problem and ask his opinion. Also for the cost of shipping, they would probably check out your transformer for you.

 

[John D]

Ferrite beads?

 

[aborgman]

Those are interesting results.

The DC secondary resistances look reasonable.

The thing about output transformers is that they are not ideal. There are inherent parasitic reactances, losses, and non-linearities that dramatically affect the performance of the amp. High flux densities in the core causes distortion in the low end. This distortion increases with lower frequencies. Sound familiar? Inductances and capacitances cause additional distortions. Connect the transformer to a real speaker and frequency dependent distortions are even more pronounced. These are things that give an amp its characteristic tone. Some of what you are seeing is probably normal.

What I'm seeing with feedback detached looks relatively "normal" to me. Low frequency distortions are there - but they are the sort I expect to see from transformer saturation, and are only there at really low frequencies.

The behavior with feedback attached? At this point it's a total mystery to me.

My instinct tells me that you should try lowering the feedback resistor value to maybe 10K and see what happens.

Hmmmm.... now it gets REALLY weird.

In an attempt to see what effect changing FB resistor size has, I connected a pot as the FB resistor.

With the pot I can get acceptable output without noise as long as the FB resistance is above about 7.5K. Below about 7.5K ohm and I start to see the same old noise - but now it is showing up on the UPPER half of the wave.

If I keep decreasing the FB resistor size, I can eventually get the noise to show up on both halves of the wave.

Even stranger - using the exact same leads I used to connect to the pot but connecting to a regular old metal film resistor... and the problem is there.

Pot set to 15K - no issue.
Plain metal film resistor at 15K - problem.

Inductance in the pot acting as a LP filter? If so, I wonder why adding a cap to ground from the FB stage making a LP with a 20KHz -3dB point had no effect...

Following that I would summarize your results and send them to a transformer specialist like Paul Patronete ([email protected]). I've discussed transformer issues with him on a number of occasions and he has been very helpful. Tell him that you are trying to determine if your transformer has a problem and ask his opinion. Also for the cost of shipping, they would probably check out your transformer for you.

This thing is so weird I really don't know what to think... I'll definitely drop him a line.

 

[BassmanPaul]

Also for the cost of shipping, they would probably check out your transformer for you.

This is what I would do. I suspect the OPT in my B15 has a problem too. Paul sounds like a really good guy to talk to. My dealings with him have all been excellent.

Edit: If you do this and it is faulty you can have him rewind the transformer there and then. IIRC he can give you a secondary tapped for multi impedances.

 

[beans-on-toast]

Hmmmm.... now it gets REALLY weird.

In an attempt to see what effect changing FB resistor size has, I connected a pot as the FB resistor.

With the pot I can get acceptable output without noise as long as the FB resistance is above about 7.5K. Below about 7.5K ohm and I start to see the same old noise - but now it is showing up on the UPPER half of the wave.

If I keep decreasing the FB resistor size, I can eventually get the noise to show up on both halves of the wave.

Even stranger - using the exact same leads I used to connect to the pot but connecting to a regular old metal film resistor... and the problem is there.

Pot set to 15K - no issue.
Plain metal film resistor at 15K - problem.

Inductance in the pot acting as a LP filter? If so, I wonder why adding a cap to ground from the FB stage making a LP with a 20KHz -3dB point had no effect...

Are all these tests performed with a non-inductive dummy load? I'm wondering what the distortions look like with a speaker attached.

Put the cap parallel to the feedback resistor. This is also a LPF but without the leg to ground which can cause problems.

At this point, it really looks like this is something that you can tune out with some tweaking. It might help to take a look at this [Invalid or Expired Link Removed] and note the differences in the feedback and PI.

 

[aborgman]

Put the cap parallel to the feedback resistor. This is also a LPF but without the leg to ground which can cause problems.

...and this works.

Cap in parallel with the feedback resistor completely kills the oscillation.

Cap to ground (standard 1st order RC LPF) did nothing.

I went with 220pF, because I had it laying around - and because it's effect within the desired passband is minimal. It only makes output down .5dB @ 15KHz.... and completely obviates the problem.

 

[BassmanPaul]

...and this works.

Cap in parallel with the feedback resistor completely kills the oscillation.

Cap to ground (standard 1st order RC LPF) did nothing.

I went with 220pF, because I had it laying around - and because it's effect within the desired passband is minimal. It only makes output down .5dB @ 15KHz.... and completely obviates the problem.

Yeah!!

 

[beans-on-toast]

I'll second that, great news!

 

[Brian K]

Glad it worked out. Out of curiosity, what kind of volume were you getting with the B-18 and the newer cab? Was it significantly more than the original B-18 cab? Able to keep up with loud drums?

Thanks,
b