The SE EL84 "All Out" Build Challenge thread
- Thread starter Redboy
- Start date
That is the 2 roll off scenarios for a .068uf cap and a 0.15uf cap.
Personally I would not be wanting to expect such a very low bass roll off for a flea watt amp. Then we go for a sub if it is deemed necessary.
Ok, although typical, DB figures give the false impression of little loss. But 30% loss is significant.
Those two responses you gave have the X axis in log format, Although that format is typical, it gives a
false impression of the response. Posting in linear format gives a quite different, more obvious "look"
into what is happening. Can you switch to linear?
1. The response in both are decreasing multiple octaves, into the hundreds of hz. Rane corp says such is perceivable.
Now if the deviation were less than 1/3 octave, the ear probably won't hear such.
2. Notice the phase shifting into the hundreds of hz as well. This changes the rise time (attack time), affecting the
sound; at least in a good system.
In general, not very impressive low frequency responses, if one looks closely. And that sub will not completely fix
the problem.
Again, I can understand a budget, cost factor, but the schematic does not give the expectations the OP was seeking.
cheers and all the best
pos
Last edited:
I wouldn't spend my time doing a test like this. I'm more pointing out that the grid leak resistor and cap value can be changed together to maintain the same corner frequency and with a low impedance driver behind this, you could attempt to make some comparisons.
-1dB at 15 Hz vs. -1dB at 8 Hz.? We call this splitting hairs...
My point exactly. I'm observing some pedantic and condescending vibes coming across here . . .
The OP's idea is to go all out with an EL84 SE design, but I do read "within practical limits" between the lines. It's totally possible to go to extremes with this exercise, but I think it would be veering besides the point. I do believe fun and "without shit" still apply.
And I agree with you on this. From the counter POV, I am more concerned with HF rolloff with the 12AX7 which poses a bit more of a workout for the brain.
Having said that, the 12AX7 remains a nice sounding combo with the '84.
There's also the factor of a low damping amp and what happens when you put a reactive load on that amp. The LF response is likely to rise a bit down there anyway due to those factors, so in practicality these differences are going to be incredibly difficult to measure at the speakers, let alone hear.
Indeed. The cathode feedback winding helps a little in improving this.
I see no reason for such a test if you are not that concerned with musical quality.
Yes, I would go with a sub if this is the circuit you wish.
Cheers and good luck.
pos
Last edited:
I am not sure who "we" are.
I am thinking of the OP who wished respect for the EL84, thus an excellent design.
Graphs using a linear X line would have shown the viewers just how poor those size
capacitors/resistor time constant is vs frequency. We also have to deal with
grid current leakage, which essentially lowers the value of the grid resistor.
As a comparison, In a truly high fidelity design, we deal with electronics with tenths of
a db across the audio spectrum.
As an example, - 0,1db frequency response means a gradual downward variation
(generally at the extremes frequencies) in the -60db region. In our systems, one can hear
variations in the -130db range. That is true transparency.
Just to let Redboy and the public know this schematic, with any tube, would not satisfy Redboy's wish.
As far as damping factor, the true damping factor is Not speaker impedance (Z) divided by amp output
impedance Z. That is good for marketing, but not anywhere near accurate. One has to consider the voice coil
DC ohm value.
cheers and all the best.
pos
Last edited:
With a truly excellent design and parts, the difference is startling. And it does not mean selling one's arm.
But it does mean lots of time and effort, even with an education and experience.
Solving a few equations does not beget great sound.
pos
Last edited:
I would have to reveal results of 40+ years lab work, costing over $150,000 in expenses and time. Are you willing to
repay me this expense for my work??
I attempted to help by suggesting an increase the bass response for a more natural sound. By the way, a 0,15ufd
is not enough. That 430k grid resistor is not 430k ohms as the input, stray, and Miller capacitance lowers that value
as the frequency increases. There is also grid current leakage to contend with.
A least I tried to help. I shall bow out now.
pos
You are probably too cool for this forum.
Negatron,
Is there a DIY section or a place to profit from your knowledge and designs in Sheol? This is a build challenge, not a criticism challenge.
I really hope you don’t rub off on @paul_b and the other experienced builders here, otherwise we will owe them for all their free help, knowledge, experience and designs!
Unless I’m running a company and planning to profit from my work/designs/pictures, I’ll continue to post them on the haven free of charge.
Respectfully,
Greenhorn amp builder in training
@paul_b - Observing your initial comments on the transformer, I am much in agreement with you.
There has been much spoken about a big transformer vs a small one. Overall, lower phase shift and wider bandwidth favor the bigger one, but as you say, it's a double edged sword.
A bigger core will need less turns for the same number of henries, but the best solution is always to get the window as full as possible for a give core size. I have also noted that a slightly higher internal plate impedance needs a higher count of henries as does a lower impedance like the 2A3 or 300b.
With my proposal, I am looking at a higher B+ (370V) and lower current - much as per the late Roger Modjeski's designs. My calcs have pointed me to a 6.4k primary load and as a reference, a 43mm stack of M6 EI76 can get a -3db bass corner at 7 Hz, primary inductance of 34H and a winding resistance of around 230 ohm.
With the U/L tap I mentioned and with the cathdoe feedback considered, I estmate the Rp of the EL84 to be in the region of 1k to 1.2k. I still have to calculate that bit properly. Output impedance calculates out to around 1.5 ohm.
There has been much spoken about a big transformer vs a small one. Overall, lower phase shift and wider bandwidth favor the bigger one, but as you say, it's a double edged sword.
A bigger core will need less turns for the same number of henries, but the best solution is always to get the window as full as possible for a give core size. I have also noted that a slightly higher internal plate impedance needs a higher count of henries as does a lower impedance like the 2A3 or 300b.
With my proposal, I am looking at a higher B+ (370V) and lower current - much as per the late Roger Modjeski's designs. My calcs have pointed me to a 6.4k primary load and as a reference, a 43mm stack of M6 EI76 can get a -3db bass corner at 7 Hz, primary inductance of 34H and a winding resistance of around 230 ohm.
With the U/L tap I mentioned and with the cathdoe feedback considered, I estmate the Rp of the EL84 to be in the region of 1k to 1.2k. I still have to calculate that bit properly. Output impedance calculates out to around 1.5 ohm.
