Quick measurements of Forte 1a, a 50wpc class A amp by Nelson Pass

[pkane]

Nelson updated Forte with more of the audiophile tweaks in later versions. 1a, while operating in pure class A mode, has a significant amount of negative feedback compared to its younger siblings. I had the amp on the bench to try to measure ultrasonic and RF interference from DACs into the amp, and since it was already hooked up, decided to take some quick measurements. All measurements were made at 1w into 8Ω.


PLEASE NOTE: An updated set of measurements with better power isolation, bias adjustment, and a different DAC were performed here (link). This resulted is a much better set of amp measurements. Some of the original measurements were updated in this post.

1kHz and harmonic distortion measurement using REW and Apogee Element24 as ADC @ 192kHz:

I couldn't find a way to get rid of the 120Hz frequency.

Here's no signal going into the amp, but the amp is on (nearly the same as amp being off!)

Frequency response is pretty good, about 4dB down at 40kHz, 3dB down at 10Hz:

Multi-tone test shows increased noise due to IMD towards higher frequencies:

And for good measure here's the 1k square wave:

I didn't do any stress tests (at least not intentionally!) this time, and don't have the load simulator to measure output at greater powers.


I did take some Picoscope measurements (also posted in another thread) to see what it look like above the audible range. This is playing a 10-22k sweep in REW, also at 192kHz:

 

[pkane]

A picture of the insides. Sorry about all the dust, it's been in storage for over 20 years. That capacitor on the right is fine, it's just a reflection that makes it look like it's bulging

 

[SIY]

What does distortion vs power look like?

 

[pkane]

What does distortion vs power look like?

Didn't measure this. As is I had to replace the fuses a couple of times and on the first attempt sent 0dBFS into the amp. The 20w 8Ω resistors were not very happy. They were hot to the touch for the next 20 minutes

I guess I could measure this up to, say, 5w in some reasonable increments.

 

[SIY]

If it's really Class A, the dissipation at full power won't be higher than the dissipation with no signal. Or is it the dummy load that complained?

 

[maxxevv]

Would the measurements have looked better had those wires been properly twisted and shielded ? The circuit board shielded from the toroidal transformer ?

 

[pkane]

If it's really Class A, the dissipation at full power won't be higher than the dissipation with no signal. Or is it the dummy load that complained?

Dummy load — a ceramic resistor.

 

[kaka89]

I think this looks pretty good considering it is an old device?

 

[SIY]

Dummy load — a ceramic resistor.

Would you like a real grown-up dummy load? I feel like I owe you for all the use I've gotten out of your software.

 

[pkane]

Would you like a real grown-up dummy load? I feel like I owe you for all the use I've gotten out of your software.

Kind of you to offer! Probably wouldn't be the best use of the dummy load, as I don't have that many amps to measure Forte just happened to be the smallest one that was easy to carry upstairs.

 

[pkane]

Would the measurements have looked better had those wires been properly twisted and shielded ? The circuit board shielded from the toroidal transformer ?

Maybe a little. Most of the harmonic distortion wouldn't change with twisted wires or shielded transformer. The mains frequency and related noise were there even when the power was turned off, so I have to believe it's not related to the active circuitry inside. Likely forming some kind of a ground loop with the rest of the test equipment (computer, Apogee interface, etc.)

 

[restorer-john]

Would you like a real grown-up dummy load? I feel like I owe you for all the use I've gotten out of your software.

That's so sweet SIY.

You know we are in a room full of geeks when someone is offered a big boy's dummy load instead of a beer.

 

[restorer-john]

And for good measure here's the 1k square wave:

Can you show the generator square wave and the reproduced square wave on the one trace/screen, that way we can see what the amplifier does to it.

 

[NTomokawa]

Holy harmonics, Batman!

Very impressive-looking power block. Reminds me of the Krell KSA-50.

 

[pkane]

Can you show the generator square wave and the reproduced square wave on the one trace/screen, that way we can see what the amplifier does to it.

Let me see if I can figure out how to do that with REW.

 

[pkane]

What does distortion vs power look like?

All I can get with the little Apogee before it gets to the clipping point by the power amp output (W on the horizontal axis), both channels into 8Ω:

 

[pkane]

Can you show the generator square wave and the reproduced square wave on the one trace/screen, that way we can see what the amplifier does to it.

Not sure I got this right, but I think I did

 

[pma]

1kHz sine spectrum looks rather like underbiased class AB than class A. Like typical crossover distortion. Very poor result.

It looks like if you used a soundcard to generate square wave signal? Such test would make no sense. The amp square response goes before the input, in other words input signal is delayed behind output? I do not get it.

 

[pkane]

1kHz sine spectrum looks rather like underbiased class AB than class A. Like typical crossover distortion. Very poor result.

It looks like if you used a soundcard to generate square wave signal? Such test would make no sense.

Look up some other Nelson Pass class A designs

https://www.audiosciencereview.com/...ass-a-power-amplifier-review.9741/post-261454

Or look up some more recent designs as measured by Stereophile. None of them will break the SINAD records (except maybe on the low side). And I like Nelson's amps and have enjoyed them for a very long time

The square wave was generated by REW, converted into analog by a high quality 24/192k DAC in Apogee Element24. What puzzles me is that the square wave coming out of the amp looks better than the original digital version of it from REW. I'll have to try to do this from scratch another time, as the digital waveform appears to be missing a lot of the higher frequency components compared to the one captured at the output of the amp.

The delay you are seeing is simply because I had to record digital and analog in two steps. There's no sound card.

 

[Doodski]

1kHz sine spectrum looks rather like underbiased class AB than class A. Like typical crossover distortion. Very poor result.

It looks like if you used a soundcard to generate square wave signal? Such test would make no sense. The amp square response goes before the input, in other words input signal is delayed behind output? I do not get it.

Maybe it's time for a idle current and bias calibration. Is this your first rodeo @pkane?