Plh

[Nelson Pass]

Originally posted by Ejam
At what setting do you see a linearly dimishing distortion profile, where 2nd is biggest followed in decenting order, 3rd, 4th, 5th .... Hiraga placed great importance on this profile.

You can find such a setting for a given output level and frequency,
but you will discover that the content will not remain consistent
elsewhere. This is a fact of life in general, and not unique to
this design. As a result, I can't hand you a specific setting.

Unlike Hiraga, I have not found a magical profile, and duplicating
his has not necessarily resulted in a sweet spot.

Very simple amplifiers and full range speakers are excellent test
beds for evaluating such phenomenon, but I have not found
customers and critics consistently preferring 2nd, 3rd, or a
Hiraga mix. The only consistent result is that very few people
like 4th order or higher.

The specific phase of the distortion components potentially has
value, particularly for 2nd harmonic, and you can experiment with
this listening to a single-ended amp by reversing the input and
output phase simultaneously.

It's important to note that the larger problem is intermodulation
distortion. It results from the same nonlinear phenomena, but
is a real problem in complex and dynamic passages, I think more
so than simple harmonic distortion.

[MikeW]

I listened to some different albums tonight. That pot is a real pain in the ***. I can see different settings for different songs and albums. Do you have a digital circuit that reads the songs and sets the pot for you? It's bad enough it takes me a week to decide if I like an amp, now I'll have to listen for a month or two.
Then change the power supply, output MOSFET's and start all over again.

[Kent Smith]

I think I agree with Nelson about the harmonic issue. Recently, I have experimented with the same triodes in SE and PP. I found that I favored push-pull simply because I could get more power for the music before running into objectionable harmonics. Yes, there were differences in the sound.... but I couldn't say that one way was more "accurate" or better than another overall. You may disagree of course, and maybe that just proves that this is more a matter of taste. In any case, it's not an issue of 2nd or 3rd for me.... too much of either is bad!

Perhaps the IMD issues should be further investigated.

Thanks for the observation, NP! I continue to learn....

[jh6you]

I just read Japanese Stereo Sound No.155, 2005.

They selected 21 commercial speaker sets, and let three audio critics to submit their listening-test reports.

At the same time, in an anechoic (reflection free) room blinded from the listening-test place, they did measurements of all the speakers to get the following data:
- Frequency response curves
(at 0, 30, 60, 90 and 180 deg, and vertically at +/-10 deg)
- Impedance curves
- 2nd harmonic curves
- 3rd harmonic curves

Main purpose was to see any similarity between the listening-test and the measurement reports.

There were many words in conclusion. Nevertheless, one word was that we could learn through ears (listening) more than through eyes (measurement).

I think this conclusion will apply for this project too . . .
And, I am fully looking forward to hearing from Lumanauw and others about their listening impressions.

Regards
jH

[angshudas]

Since the PLH has an output impedance of 3 ohms, won't the impedance of the speaker alter the frequency response, making the bass sound bit boomy.

What type of speakers are recommended for the PLH.

Angshu

[Nelson Pass]

Originally posted by angshudas
Since the PLH has an output impedance of 3 ohms, won't the impedance of the speaker alter the frequency response, making the bass sound bit boomy.

Quite right. This amplifier is most suitable for kind of speakers
popular with single-ended tubes and such. Lowther, Fostex,
and a host of high efficiency "wide range" drivers do quite
well with this sort of impedance.

If you want a lower output impedance, you will want to increase
the open loop gain and/or use more feedback, like the original
JLH. You can do this by reducing the Source resistor values,
running the output stage 50/50 push-pull, biasing the input
stage with constant current sources, using a bigger input
device biased at higher current, and so on.

[angshudas]

Originally posted by Nelson Pass


If you want a lower output impedance, you will want to increase
the open loop gain and/or use more feedback, like the original
JLH. You can do this by reducing the Source resistor values,

Dear Mr. Pass,

Thank you for the confirmation.

Can the source resistors of 0.47 ohm, be reduced to 0.2 or 0.1 ohm resistors instead.

I am experimenting with BJT input stage as I cannot get small signal MOSFET locally. For the output I am using IRFP840.

Angshu

[Nelson Pass]

Originally posted by angshudas
Can the source resistors of 0.47 ohm, be reduced to 0.2 or 0.1 ohm resistors instead.

They can be reduced to 0 if you like. In my favorite example,
I was using two outputs in parallel with 1 ohm each. If you take
that down to 0, you should pick up about 10 dB of open loop,
and your DF will go up to something like 10.

[stappvargen]

Would halving the resistances for the phase splitter (Q1), and thereby doubling the quintesent current for Q1, have any positive or negative effects? In my simulations I see no difference, but then again I can't simulate THD.

Also, thought about reducing the resistor values used in the feedback by a factor 10 (R1, R2 and also R7) along with the gate resistors (from 221 to 100 Ohm) in an attemt to leasen the negative effects of gate to source capacitance of the FETs. Is there any other drawback to this other than a lower input impedence?

Any feedback appreciated!
/Niclas

[Nelson Pass]

Depends on what you use for Q1, but in general the Mosfets
will have higher transconductance and lower distortion as the
bias current is increased. At higher bias currents, I would start
thinking about using a current source to bias from the positive
rail, so that you aren't using a bootstrap circuit to drive too
low an impedance.