That wasn't exactly the problem, however I did notice a difference with the 1k resistors versus the 100 ohm resistors in that when one cable is unplugged it merely reduces the audio level instead of slightly changing how the audio sounds.
The real problem is one tape I was playing. Someone must have somehow had one channel out of phase when recording the tape as other tapes sound fine.
The real problem is one tape I was playing. Someone must have somehow had one channel out of phase when recording the tape as other tapes sound fine.
Last edited:
So doing a stability test involving putting a .1uF cap across the output as the only load I get the following.
With a .1uF cap on the 16 ohm tap I get a frequency of 78.1Khz.
With a .1uF cap on the 8 ohm tap I get no oscillation.
With a .1uF cap on the 4 ohm tap I get a frequency of 163.9kHz.
Shorting the input to ground doesn't stop it.
On the 16 ohm tap disconnecting the input cable reduces the amplitude.
On the 4 ohm tap disconnecting the input cable increases the amplitude and it starts to oscillate and stop at some relatively low frequency.
How do I fix that instability assuming it is not the Edcor output transformer?
Also why does shorting the input to ground not stop the oscillation?
Would installing the stock plate chokes fix the instability?
With a .1uF cap on the 16 ohm tap I get a frequency of 78.1Khz.
With a .1uF cap on the 8 ohm tap I get no oscillation.
With a .1uF cap on the 4 ohm tap I get a frequency of 163.9kHz.
Shorting the input to ground doesn't stop it.
On the 16 ohm tap disconnecting the input cable reduces the amplitude.
On the 4 ohm tap disconnecting the input cable increases the amplitude and it starts to oscillate and stop at some relatively low frequency.
How do I fix that instability assuming it is not the Edcor output transformer?
Also why does shorting the input to ground not stop the oscillation?
Would installing the stock plate chokes fix the instability?
As you are finding, nice looking square waves are of lesser concern than how it measures and performs. As I've read this entire thread, I think you're in a bit of a pattern of "shot gunning" solutions --trying/tweaking things without fully realizing their implications. If this were my amp, I would do a bit of a reset. I'd remove all feedback and compensation and start afresh. Tap the feedback from the 16Ω output for starters. Apply 10 dB feedback. Then tune the amp with that amount of feedback. You do not need or want a variable resistor in the feedback loop.
You will very likely need a cap across the feedback resistor and either a Zobel across the secondary or a step filter on the plate of the voltage amplification stage. You might not be able to achieve flat response out to 20 KHz with these output transformers while still obtaining unconditional stability. LF stability is just as important as HF stability. You will very likely need to compromise.
Also you should not need a compensating cap from the plate of output stage to the cathode of the voltage amp stage. This is the 5 pF cap in your case. Remove it. I am fairly confident you can achieve all the tuning you will need with simply 10 dB of feedback, a phase advance cap, and a step network. As mentioned, you're very likely going to need to make some compromises. With these output transformers, I doubt you will be able to achieve a flat response to 20 KHz AND unconditional LF/HF stability AND nice looking square waves (which are of lesser concern anyway) AND a level of feedback over approximately 10 dB.
You will very likely need a cap across the feedback resistor and either a Zobel across the secondary or a step filter on the plate of the voltage amplification stage. You might not be able to achieve flat response out to 20 KHz with these output transformers while still obtaining unconditional stability. LF stability is just as important as HF stability. You will very likely need to compromise.
Also you should not need a compensating cap from the plate of output stage to the cathode of the voltage amp stage. This is the 5 pF cap in your case. Remove it. I am fairly confident you can achieve all the tuning you will need with simply 10 dB of feedback, a phase advance cap, and a step network. As mentioned, you're very likely going to need to make some compromises. With these output transformers, I doubt you will be able to achieve a flat response to 20 KHz AND unconditional LF/HF stability AND nice looking square waves (which are of lesser concern anyway) AND a level of feedback over approximately 10 dB.
Spent some time with the amp yesterday and got it as good as I am gonna get it with the Edcor transformer.
Starting with a larger value for the 5pF cap was a good idea. I tried 10pF and then 15pF. It seems like 10pF is what the amp wants given the greatest stability was had with that value.
I then played around with the cap value across the feedback resistor and settled on a .0047uF cap across two of the three 1k feedback resistors in series with the feedback pot which got the amp flat to 20kHz.
The amp is now stable into .1uF at 4 and 8 ohms. I still get an oscillation of 73.5kHz at 16 ohms, but trying different things just brought back the oscillation at 8 ohms. So I'm going to chalk that up to the Edcor transformer.
I think the fact that I cannot get equal voltages between the 4 ohm tap and ground and the 16 ohm tap and 4 ohm tap points to the transformer being the problem and whatever is causing that may be why I cannot get the oscillation stopped on the 16 ohm tap into .1uF of capacitance. Makes me wonder if the taps are the ohms they say they are particularly the 16 ohm tap.
The only further improvement I think possible will be finding that Acrosound transformer.
The amp is set for an input sensitivity of 1Vrms.
Final schematic
How the square wave looks is very important as before I even tried making the square wave look good the amp sounded and measured good being it put out 25 watts from 20-20kHz, but wasn't exactly right given the amp was not stable into .1uF of capacitance which means that if a speaker was connected with a long cable or was somehow capacitive it could cause an oscillation that could damage the speaker and/or tubes.
10kHz waveform.
4 ohms 1V/DIV.
8 ohm 2V/DIV.
16 ohm 2V/DIV
The square waves tend to point to the 16 ohm tap as having an issue.
Starting with a larger value for the 5pF cap was a good idea. I tried 10pF and then 15pF. It seems like 10pF is what the amp wants given the greatest stability was had with that value.
I then played around with the cap value across the feedback resistor and settled on a .0047uF cap across two of the three 1k feedback resistors in series with the feedback pot which got the amp flat to 20kHz.
The amp is now stable into .1uF at 4 and 8 ohms. I still get an oscillation of 73.5kHz at 16 ohms, but trying different things just brought back the oscillation at 8 ohms. So I'm going to chalk that up to the Edcor transformer.
I think the fact that I cannot get equal voltages between the 4 ohm tap and ground and the 16 ohm tap and 4 ohm tap points to the transformer being the problem and whatever is causing that may be why I cannot get the oscillation stopped on the 16 ohm tap into .1uF of capacitance. Makes me wonder if the taps are the ohms they say they are particularly the 16 ohm tap.
The only further improvement I think possible will be finding that Acrosound transformer.
The amp is set for an input sensitivity of 1Vrms.
Final schematic
How the square wave looks is very important as before I even tried making the square wave look good the amp sounded and measured good being it put out 25 watts from 20-20kHz, but wasn't exactly right given the amp was not stable into .1uF of capacitance which means that if a speaker was connected with a long cable or was somehow capacitive it could cause an oscillation that could damage the speaker and/or tubes.
10kHz waveform.
4 ohms 1V/DIV.
8 ohm 2V/DIV.
16 ohm 2V/DIV
The square waves tend to point to the 16 ohm tap as having an issue.
No doubt you can make it unconditionally stable at the expense of some HF bandwidth. That may or may not be important to you. The output transformer probably has different HF characteristics depending on which tap you utilize, which could explain the behavior you're describing.
Just don't place too much emphasis on the beauty of the square wave. It can lead to tilting at windmills (chasing imaginary enemies).
Indeed I could, but not being flat to 20kHz would be a problem.
On another thread on another forum ir was determined that the Edcor transformers aren't quite as good as people like to think they are.
In the topic that talked about the Edcor transformer (was a thread about modding a Magnavox 9300 amp) a photo was posted of what an ideal 10kHz square wave should look like for a properly compensated tube amp and mine looks real close to that.
It was mentioned that there's an acrosound transformer that is better than the Edcor I used so once I find one for a reasonable price I'll use it instead.
That said I do know the Edcor has a problem given the 4 ohm tap should be the center of the secondary winding and it doesn't appear like it is. Further testing will show for sure.
So I'd need a way to measure the impedance of the transformer.
Im thinking to measure the impedance I can put a 3.8K resistor across the full primary and use a Dayton Audio DATS V3 speaker tester to measure the impedance of the 4, 8 and 16 ohm taps.
You are way overthinking this. Your focus on unconditional stability is obviously influenced by the gurus on another board. What practical operating condition is going to present a .1uF only load on a speaker terminal? There is always R, L, C with speaker wire and speakers in place.
The person who advised me of that test I'm sure knows what he's doing based on his many posts related to modding Magnavox tube amps to fix their shortcomings.
I once had a small solid state chip amp that was fine with a short speaker cable, but with a little over 12' cable showed instability when a sinewave was fed to the amp above a certain level.
I once had a small solid state chip amp that was fine with a short speaker cable, but with a little over 12' cable showed instability when a sinewave was fed to the amp above a certain level.
Here's the topic I was talking about.
https://audiokarma.org/forums/index.php?threads/more-fun-with-magnavox-the-9300-series.687735/
https://audiokarma.org/forums/index.php?threads/more-fun-with-magnavox-the-9300-series.687735/
I've had more fun with more 9300's than a normal person has a right to have. Guru's need a problem to solve, that no one ever thought of, to be a guru. But is stability a ''problem'' or one of those perfectionist's problems? Existential?... ''problems.''
Last edited:
The 6BQ5's are Westinghouse, GE, Matsushita, and Zenith, no mods except line voltage and high voltage resistor additions to bring the PS output B+ and heater voltages into line with the original 117v input voltages without the preamp loads. Quiet and sweet. No ''shortcomings.''
Now if the amp had come with the original output transformer I would have likely restored it stock (20 watts at some unknown distortion value) and been satisfied with it not even thinking I could have better performance.
What I need to do is fine another of these amps, restore it stock and see how it measures.
What I need to do is fine another of these amps, restore it stock and see how it measures.
It starts out... ''In other words, the thing has still got to resemble and represent the essence of what the original 9300 series was designed to be.''
And ends with a completely differnt thing, unless you consider all amps with a 1 input tube and two PP output tubes, the same.
That's a seperate issue from the stability testing, but it's part of the soup.
At first I think it was moreso seeing what could be done to fix problems with the stock circuit then evolved over time to the best quality within reason.
I think the big problem is the stock Magnavox output transformers. I'm thinking it was the lower resonant frequency of them with it being around 6kHz.
And as I have discovered the Edcor transformers aren't quite as good as I originally thought.
That said through dealing with this amp I have learned how to properly compensate an amplifier.
I think the big problem is the stock Magnavox output transformers. I'm thinking it was the lower resonant frequency of them with it being around 6kHz.
And as I have discovered the Edcor transformers aren't quite as good as I originally thought.
That said through dealing with this amp I have learned how to properly compensate an amplifier.
I didn't as mine came without the output transformer, however others who regularly deal with these Magnavox amps have noticed problems with the output transformers.
I did notice that with a Magnavox output transformer I got for mine which I found out later was actually for the low frequency output of an AMP-150, a sinewave would start to distort at frequencies above about 2kHz.
I did notice that with a Magnavox output transformer I got for mine which I found out later was actually for the low frequency output of an AMP-150, a sinewave would start to distort at frequencies above about 2kHz.
''...I did notice that with a Magnavox output transformer I got for mine which I found out later was actually for the low frequency output of an AMP-150, a sinewave would start to distort at frequencies above about 2kHz.''
Does a 150 drive its bass channel with PPP 6V6?
Does a 150 drive its bass channel with PPP 6V6?
Yes PPP 6V6 for the lows and a single ended 6V6 for the highs.
It's interesting how they did the crossover which uses coils and caps.
If I ever wanted to use this amp in a BI-amp setup I can duplicate the passive crossover.
It's interesting how they did the crossover which uses coils and caps.
If I ever wanted to use this amp in a BI-amp setup I can duplicate the passive crossover.
Last edited: