When I last scoped the output it basically looked like a sine wave but with the peaks flattened off a little, I'm assuming this is what clipping looks like on a scope?
Will scope this unit again once the parts arrive.
Will scope this unit again once the parts arrive.
Yes, flattened peaks is clipping. Should be symmetrical.
Make sure to use correct load resistor.
So you should see flat peaks but if they're sloped high at the front and low at the back then the output transformer is saturated. What is happening is that the wave tries to make the input waveform shape on the output transformer, but there's not enough flux to couple the large wave energy requirement. The output then slopes down as the flux hits that maximum limit.
When the clipped peaks are linear but slanting downward this is no indication of core saturation.
Rather it's a consequence of the limited bass response of the OT and a signal frequency being less than 10 times the lower corner frequency of the OT.
The same effect can be observed with a square wave signal at levels far below clipping.
In other words it's a linear effect caused by phase shift.
Core saturation is non-linear and looks different.
Rather it's a consequence of the limited bass response of the OT and a signal frequency being less than 10 times the lower corner frequency of the OT.
The same effect can be observed with a square wave signal at levels far below clipping.
In other words it's a linear effect caused by phase shift.
Core saturation is non-linear and looks different.
Sorry I was thinking more like this than high or low frequency impact on square waves (ignore the subject matter - it's more the waveform shape):
I spoke about clipped voltage peaks/tops (on both sides with a balanced PP amp).
As core saturation never causes voltage clipping, the clipping itself must come from the tubes.
My explanation above referred to the slanting of the tops.
As core saturation never causes voltage clipping, the clipping itself must come from the tubes.
My explanation above referred to the slanting of the tops.
It just occurred to me that the waveshape in your post looks familiar.
In fact I've seen exactly that kind of asymmetrical saturation (positive going voltage collapsing to zero) when overdriving the reverb driver of a Fender amp. Not sure if it was the reverb transformer or the tank transducer saturating, but I think the latter.
With a resistive load the transformer did not saturate.
Ive just come back to this amp after making all the fixes and largely have what I wanted sorted, however one area i have overlooked is the bias of this amp.
The bias supply voltage is too high and ive had this thing over biased, Ive dropped the 47k resistor in series with the bias trimmer down to 22k and can get it down to -38V, which is still a little too high I feel, I am only getting about 70W out of this amp with no distortion on the low gain.
Looks like I will have to mess around with the voltage divider resistors on this to get the range I require.
It was biased even more than that this whole time, around -45V or so, which was way too high.
Ive been advised that the phase inverter design ive got here is not really that suitable to drive these tubes enough and that I should change it back to the original marshall 2203 phase inverter with an ECC83.
Any thoughts on this? It turns out that this LTP phase inverter im using is just the same design as used in some fender amps.
I also may need to address the level of negative feedback feeding into the phase inverter.
Not easy to understand. Too high do you mean too negative the voltage of the bias, or too high the current?
The two bring to opposite directions.
If you don't have a loop, the preamp has plenti of swing to drive PI and output tubes.
Place a 250k log potentiometer instead of the resistor between the secondary of the output transformer and the PI.
Check you have no voltage on it. Turn it until you like it. Measure the value. Substitute with a resistor with the closest value.
Yes sorry, when I mean highly biased, im meaning higher negative bias voltage.
I was told that the ECC81 im using wont be enough to drive these tubes and I should go with an ECC83 like the marshall used.
See my discussion on the marshall forum here, from post #87 onwards.
https://marshallforum.com/threads/b...-by-far-your-most-valuable-tool.102523/page-5
Funny you say that, because thats exactly what I did and ended up settling on 4.7K, seems to sound OK and you will see my scope screenshots in the thread I linked above. I dont see that the NFB is causing any issues from what I can tell.
If I can get 100 clean watts out of this thing, it would be great however.
[/QUOTE]
I also get a bit confused when biasing this, if i use the scope method, I can remove the crossover distortion with a bias as much as around -49V, this is still way too high bias according to the datasheet for 807.
I ended up going by the method to measure the voltage drop across the output transformer primary and calculate the current by ohms law, seems to be a fairly accurate way to bias this, im measuring 50mA per tube now while idle, does this sound right? Thats what im getting biased at -49V.
Yeah not too focused on watts, Im happy enough with its sound but I believe I should be getting more on the low gain, it feels a little on the quiet side quiet to me, on the high gain with distortion, its got plenty of power and I get 160W.
I also get a bit confused when biasing this, if i use the scope method, I can remove the crossover distortion with a bias as much as around -49V, this is still way too high bias according to the datasheet for 807.
I ended up going by the method to measure the voltage drop across the output transformer primary and calculate the current by ohms law, seems to be a fairly accurate way to bias this, im measuring 50mA per tube now while idle, does this sound right? Thats what im getting biased at -49V.
Your tubes have a max plate dissipation of 25W. Target 20W. 17,5 if you want to be safer. Maximum power won’t change.
With his 4.7k/4.7k NFB divider the power amp gain from PI input to speaker output is significantly lower than 2.
Probably something like 1.5.
So for an output voltage of 30Vrms (113W@8 Ohm) the PI input signal needs to be around 20Vrms.
Not sure if the preamp can deliver that level after the tonestack.
It's simply too much feedback for an instrument amp.
That's why I suggested to chase the tone and not what the scope shows.
That's why I suggested to chase the tone and not what the scope shows.
50mA idle current is already too much for good tube life.
The grid bias voltage from the datasheet refers to a screen voltage of 300V.
With a screen voltage of 400V+ the grid voltage needs to be more negative.
You got good advice at the marshallforum. Use it.
Agree.
But less NFB means less presence boost.
Problem is the low forward gain due to the 807s and the 12AT7.