Amp Lab – Quilter Labs Bass Block 800

Gain & THD+N Standout

In initial power up testing, the THD+N values are very low on this amp. It is a clean amp. Also, when pushed to the point of clipping, the output it is at a power level that definitely puts it into the 800w range. As with a lot of amps, turning the gain down to 0 produces negative gain numbers. So, for this testing, I am counting the minimum gain to be 2 ticks above the minimum value.

Similar to the Gain knob, when the Master is all the way down, it gives negative results on the gain map. It is only interesting to test at levels that actually amplify the output. It shows a very low THD+N all the way up to the maximum output of the amp, and then a sharp increase, as the output becomes clipped. Similarly, the gain map shows a sharp knee where the output of the amp is reached. Pretty straightforward class-D amp behavior. 

Frequency Response Standout

The Contour sweep is where things get interesting. When the Contour control is turned all the way down, it creates a 10 dB cut centered around 475Hz. When it is turned up to maximum, things get a little weird. There is a serious dip at 5kHz that essentially cuts out that frequency all together. The rest of the response looks like a low-pass filter that does not affect frequencies below 300Hz. More on this in the next section; let’s just move on for now to the rest of the results.

Next up in the frequency response testing the “optimally flat” test. For the 80Hz to 8kHz range that we are interested in, with all the controls set to the midpoint, the response is already +/- 0.35 dB, which is extremely good. I was able to the results down to +/- 0.297 dB by going down 1 tick on the Contour knob, but this is really getting silly. Needless to say, this is a very flat response and better than most.

Input Impedance

The Passive input measured 424 kOhm impedance with a 1kHz sine wave. The Active input measured 29 kOhm impedance with a 1kHz sine wave.

[Pat Quilter note: “I am not sure about the method used to measure the impedance externally, but the net impedance across the audio band should be 1meg (two 2meg resistors in parallel) on both active and passive. If impedance is measured by inserting a known resistance in series with the signal and looking at the voltage drop, it is important to do this well below the limiter threshold, but well above the class-D switching residue at 250kHz (or use the output signal from the preamp).

 On the Bass Block 800, the Active input is padded 10dB using a 22K/10K resistive divider, so the Active input’s impedance is 32K, while the Passive should be 1meg.”]

[Editorial response: Our standard methodology for measuring input impedance utilizes a 1kHz sine wave input. After double-checking our input impedance measurements, we observed results consistent with our initial measurements. However, we did note that if we ran the impedance test using a 100Hz input, measured impedance was approximately 980Hz.”]

Waveforms

Looking at graph 8, both traces look very clean and indicative of an output with very low THD+N. When zoomed in on these signals though, a strange interference pattern is visible. Under normal operating conditions, it is very small, compared to the output signal, and therefore does not interfere with the output. Some careful measurements show that this is a 38.25kHz sine wave with an amplitude of about 21.7 mVrms. My first thought was a problem with the test bench, but I hooked up another amp (different brand) with this exact setup and the signal was not there. Also, the 38.25kHz signal is not there when the amp is powered off, it is only there when the amp is on, and it happens on both of the Quilter Bass Block amps that we have in for review. Interestingly, this signal is even present on the output of the amp when there is no input signal at all being put into the amp.

[Pat Quilter note: “Like virtually all class-D amps, there is a visible carrier frequency on the speaker waveform; in our case, at 250kHz. l wonder if this is being ‘folded’ down to 38kHz on the scope trace, which might be possible when using a digital audio test rig.”]

 [Editorial response: “We suspected that this might be a possibility, but we have had many similar class-D amps on our Audio Precision test bench, and it appears there is something unique happening with this particular amp.”]

I am pretty sure that this interference signal is responsible for the strange frequency response graph when the Contour is set to maximum value. When the output is very small, due to the low-pass filter effect of the Contour knob, the interference signal becomes significant. The scope trace shows that the interference signal is more pronounced than the 5kHz signal that the amp is putting out, and the AP analyzer only sees the 38.25kHz portion, and not the 5kHz portion, causing the outputs to essentially look like it is 0.

[Pat Quilter note: The response of the Contour control near ‘10’ does have the frequency dip, as noted, which is an artifact of using each end of the pot to control a rather different frequency spectrum. The additional attenuation at 4-5kHz is basically imperceptible.”]

 [Editorial response: “We wanted to mention the 38.25Hz signal to help explain the particular curves seen on the figures created during our bench testing. As for the 5kHz attenuation not being perceptible, as noted, we could clearly hear impact, as discussed below.”]

As a further verification of the reality of this problem, I hooked the amp up to a bass cab. No connections were made to the AP unit, except as a signal generator for the input into the amp. Setting the Contour to max and listening to the amp, there is an audible output at 4kHz. It goes down somewhat in volume at 4.5kHz, but there is no audible speaker output at all at 5kHz. By the time you get up to 6kHz, the audible output is back at a level similar to the 4kHz level.

Power – 1kHz Burst and Continuous, 50Hz Burst and Continuous

Last up is power testing. Before getting to power results, it is important to note that this is the first amp I have had in a long time that did not have any issues with shutdown during the 1kHz power testing. Most amps will overheat at least enough that they temporarily shut off output. The really bad ones just blow up during testing. There were no overheating problems at all with this amp at 1kHz. I think it is partially due to how fast the fan kicks into high speed when the amp is driven hard. It seems that the fan speeds up as soon as the output is driven hard, instead of waiting for heat to build up, first. I am very impressed with the heat handling capability of this amp. It goes to show how well thought-out the heat management system is.

As always, the hard part of power testing is figuring out what we should call “max power.” We will continue to use the clipping standard as the limit for our testing. We are continuing the use of the 50Hz test signals, as well as the 1kHz test signals, because I think that I gives a more realistic number for use on a bass amp.

For the 1kHz signals, clipping occurs about 4%THD+N, so that is what we a calling the upper limit of this amp. A strange thing occurred with this amp in that at frequencies below 100Hz, the output had a much higher THD+N at high output levels. The amp distorted the input signal significantly before clipping occurred. If the 4% value was used, the test kicked out at half the rated power. Going back and rechecking it confirmed that the THD+N value that corresponds to the start of clipping on a 50Hz signal is 11.7%. So, for this testing, we will use 4% for the 1kHz testing and 11.7% for the 50Hz testing. This gives fair max power ratings, but keep in mind that the amp is technically more distorted at the lower frequencies at high volume levels.

Conclusion

I have mixed feelings on this amp. I really like the construction and overall quality of the build. For the most part, it is a clean amp that has the advertised power. I just have a hard time with the 5kHz hole in the output when the Contour knob is turned up and the distortion at low frequencies when the power is turned up.

Fig. 1 Gain – Graph of gain with master set at midpoint

Gain = minimum
Gain = midpoint
Gain = maximum

Fig. 1 RMS – Graph of output RMS voltage level with master set at midpoint

Gain = minimum
Gain = midpoint
Gain = maximum

Fig. 1 THD+N – Graph of output THD+N ratio for with master set at midpoint

Gain = minimum
Gain = midpoint
Gain = maximum

Fig. 2 Gain – Graph of gain with master set 2 indents above minimum

Gain = minimum
Gain = midpoint
Gain = maximum

Fig. 2 RMS – Graph of output RMS voltage level with master set 2 indents above minimum

Gain = minimum
Gain = midpoint
Gain = maximum

Fig. 2 THD+N – Graph of output THD+N ratio for with master set 2 indents above minimum

Gain = minimum
Gain = midpoint
Gain = maximum

Fig. 3 Gain – Graph of gain with master set at maximum

Gain = minimum
Gain = midpoint
Gain = maximum

Fig. 3 RMS – Graph of output RMS voltage level with master set at maximum

Gain = minimum
Gain = midpoint
Gain = maximum

Fig. 3 THD+N – Graph of output THD+N ratio for with master set at maximum

Gain = minimum
Gain = midpoint
Gain = maximum

Fig. 4 Gain – Graph of gain with all knobs set at midpoint

Active input
Passive input

Fig. 4 RMS – Graph of output RMS voltage level with all knobs set at midpoint

Active input
Passive input

Fig. 4 THD+N – Graph of output THD+N ratio for with all knobs set at midpoint

Active input
Passive input

Fig. 5 Gain – Graph of gain

deep = minimum
deep = midpoint
deep = maximum

Fig. 5 RMS – Graph of RMS output voltage level

deep = minimum
deep = midpoint
deep = maximum

Fig. 6 Gain– Graph of gain

contour = minimum
contour = midpoint
contour = maximum

Fig. 6 RMS – Graph of RMS output voltage level

contour = minimum
contour = midpoint
contour = maximum

Fig. 7 Gain– Graph of gain

Optimally flat = minimum
All knobs at noon = midpoint

Fig. 7 RMS – Graph of RMS output voltage level 

Optimally flat = minimum
All knobs at noon = midpoint

Fig. 8 Scope – Scope trace of output with 1kHz sine wave input

Fig. 9 Scope – Scope trace of output with 5kHz sine wave input and contour set to maximum

Fig. 10 Scope – Scope trace of output with 80Hz sine wave input

Fig. 11 Scope – Scope trace of output zoomed in to show detail

1-5 (unacceptable to impeccable)

On-bench