Detailed Measurements and Technical Review of the Crest Audio Vs900 Power Amp

I have to admit, I’ve got the bug once again. After building and listening to my little Sonorous 10S Subwoofer over the last couple of weeks, I’ve suddenly got this need for more, more bass, more low-end bump, more rattling, more shaking, more chest-thumping boom. So I’ve started hatching a plan to build a bigger, better, more HT-centric subwoofer, where the size of the enclosure has no limits, no bounds. But before I get into that part of this project (saving that for another post), we’re going to check out the new/old amplifier I just picked up which will power this monster subwoofer. So per my usual browsing on the web, I tend to check out my local Craigslist (aka CL) at least a couple of times a week to see what kind of audio gear people trying to unload. I connected with a guy who was selling a bunch of pro-style amps from Carver, Peavey and Crest Audio that he had collected over the years. None of them really seemed all that interesting for a subwoofer build, though his pricing was reasonable.

But after digging a little deeper, I ended up looking up the datasheets on each of his amps and was surprised to find the Crest Audio Vs900 amp he was selling had some decent specs: class A/B architecture, 2-ohm stable per channel, bridgeable, 1100W max rated power into 4-ohms bridged (at 1 kHz, 1% THD), claimed frequency response down to 10 Hz (none of this 30 Hz high-pass filter nonsense some pro amps have) plus just a nice clean front panel with nothing but a couple of status LEDs. So with a subwoofer from Parts Express already in mind, I thought this amp would make a great pairing (can you guess what sub I’ve got in mind?). He said he’d let it go for $100 and I said sure, let’s do it. If this thing can do its rated power, we’re talking about 9 cents per watt here. So let’s see if this 20-year old, 46-pound boat anchor can actually deliver.

So I fired up REW and pulled out my 400-watt load box, which is just a bank of four 100 watt, 4-ohm load resistors wired in a parallel/series configuration. I picked them up for Amazon for about $10. I’ve got a 45:1 resistive divider network that effectively will take 40 Vrms down to 0.9 Vrms which is just about right for driving the input to my Presonus Firestudio Mobile with some headroom (which can take up to +10 dBu or 2.4 Vrms). Obviously this setup cannot measure the rated power of this amp in bridged mode or into to 2 ohms, so all of my measurements will be limited to one channel driven into 4 ohms. The measurements I was able to collect are as follows:

Maximum power (RMS) into 4 ohms (one channel driven) at 1 kHz
Maximum power (RMS) into 4 ohms (one channel driven) at 40 Hz
THD (total harmonic distortion) from 1W up to rated power
Frequency Response from 5 Hz to 20 kHz (overall flatness from 20-20 kHz)
Channel Imbalance

First things first, anything you buy off CL may not work at all, let alone meet any kind of performance parameters it once had 20 years ago. The guy said he tested it and it worked fine, so I was really hoping it didn’t go straight into protect mode the second I turned it on. At power-up, lights came up, fan spun up and the protect circuit kicked out after a couple of seconds as normal. This was a good start! I played some music through it to a small speaker, just to see that it made sound and that the volume/gain knobs worked okay. Then it was off to the test bench to really check it out. I tested CH B first and ran a quick frequency response plot at about 1 watt. Holy cow this thing was flat! I’d tested some smaller Class D amps recently and was expecting the typical low-cut/high-cut sort of thing, but was impressed to find the response down only -0.3 dB at 20 Hz and -0.4 dB at 20 kHz. I couldn’t even see the -3 dB point since my sweep started at 5 Hz and at 5 Hz the response was down only -2.3 dB. CH A was a touch better at -2.0 dB at 5 Hz. CH B also had a little wobble in the response that was not present in CH A. I suspect it’s due to some possible aging of a 100uF electrolytic cap in the pre-amp stage that makes up the high-pass circuit for this amp. I could replace it, but the effect is small enough that I probably am not going to worry about it. I ran some additional frequency response plots at higher power levels up to 50 watts. The plots shown here are at 50 watts.

With the frequency response plots you can also grab distortion over frequency using REW. Though it’s not as accurate as using the RTA at a single frequency, if you increase the measurement length from the standard to either 512k or 1M, you get some better fidelity on the distortion side of things, but still not as good as RTA. But this does provide a general idea of the overall distortion over frequency and can identify problems or other areas the amp may have at frequencies that you may not notice otherwise when just using RTA. In this case Ch A is slightly better below 20 Hz than Ch B but nothing too crazy. Again, points to possible aging of the inline electrolytic cap (this would be C202 in the schematic) in this channel. I did not run full sweeps up to full power since the protect circuit did kick in when running the sweep starts at 5 Hz. I could have upped the start frequency, but went ahead with RTA measurements instead.

So I ran tests from 1W up to 450W at 40 Hz and 1 kHz on both channels (separately) at about 50W increments and grabbed screenshots from REW RTA with the noise floor set at 130 dBc. This is about the limit of my setup with 128k FFT Length and 4 averages. This kept each measurement to about 6 seconds which helped prevent build-up of heat in the amp or my load resistors during testing. At least once I got above 100W or so. Tests were run about 1 minute apart, give or take the amount of time it took me to grab the screenshot and adjust the gain for the next measurement. I’ll probably only upload a few of them, since basically between 5W and 350W there isn’t much going on. Things only start to get interesting around the 400W mark. At 1 kHz into 4 ohms, one channel driven, I measured exactly 450 watts RMS at 0.097% THD on CH B. This was excellent! And an order of magnitude lower distortion than the spec! Of coursse, this was one channel driven not both as the spec stated, but still, this was impressive. I pushed the amp another 10W to 0.2% THD and the red clip light came on indicating it was time to turn it down. So I’d say max power for CH B was easily 450 watts with the clip light coming on at just around 460 watts.

Next up was CH A. At 1 kHz into 4 ohms, one channel driven, I measured the same thing, exactly 450 watts RMS at 0.099% THD. Only this channel I was able to drive to 468W before distortion hit 0.28% but the clip light never came on. So I left it at that, I did not push it harder at 1 kHz. So all in all, the datasheet peformance was met, to my limited ability at measuring only one channel at a time, the amplifier performed admirably. Another thing to know is some pretty high mains hum present, especially at 120 Hz (2×60 Hz here in the US). This wasn’t audible to any extent in my tests with a subwoofer, but probably due to aging caps or just the fact that the first amplifier stage runs on ±108 Vdc rails. So there’s not a lot of headroom to convert that AC waveform to pure DC, other than with a fatty transformer, and bridge rectifier and the biggest caps as you can afford to fit in your design. I feel like Class D amps with SMPS power supplies tend to do a better job with mains rejection than Class A/B can. But this would be good enough for most applications I believe.

So 1 kHz test tones are all fun and games, but this thing is going to do subwoofer duty, so let’s re-run the measurements at 40 Hz and see what we can get. At 40 Hz CH B topped out at 402W at 0.1 % THD and was unable to produce much of anything more than that without going into distortion overdrive. That’s not bad, slightly worse than at 1 kHz, but it’s also a more demanding test. Still, 400W into 4 ohms is pretty decent. CH A did a bit worse, as I had to let distortion rise to 0.3% before I could hit 400W. Note that CH B hit 0.3% at 405W, so they are actually fairly closely matched. Either channel when driven even a couple of watts higher than that distorted well above 10% with only the smallest adjustment on the drive level. So this thing is great until it’s not, and then it poops out fast. This should be pretty obvious when listening, if it sounds good, you’re probable below this max limit, but if it’s starting to sound like crap, then no sense running it that hot, as you’re not gaining any significant volume, all you’re doing is increasing the distortion. Oddly enough the clip lights never came on during the 40 Hz tones, even when the amp was truly clipping. Might be a limitation in the clip detector circuit, that it cannot resolve the distortion as well at lower frequencies? Not sure, but best be mindful of the volume knob either way and for bass duties, don’t trust the clip light to tell you if you’re pushing the amp, or your poor speakers, too hard. As surely you’re speakers will pay the price in the end, not to mention your ears.

Also just to get an idea of the amplifier THD performance at a nominal power level or to compare with other amps, measurements are shown here at 5W into 4 ohms, one channel driven. CH B looks decent, the lowest noise we get 1 kHz is 0.014% but at 40 Hz it drops to an impressive 0.0045%. Well, it’s not at DAC/pre-amp impressive, but pretty good for a massive pro amplifier. CH A looks a little better at 1 kHz at 5W and measured 0.0045% and at 40 Hz measured 0.0047%. And those numbers actually held up all the way up to about 300W-350W or so where THD was still less than 0.01%. I mean there’s no better time to say this, but man, they just don’t build them like they used to.

The last measurements are just a compilation of THD measurements at varying power levels that I collected manually as I increased the power from 1W to max power. Since REW doesn’t really have an option to automate testing THD of amplifiers across output power, you basically just have to collect them one by one and make a spreadsheet of the results. So that’s what I have done here. Also I did the measurements in % THD as well as dB THD, or a sort of SINAD type measurement, if your more familiar with that measurement unit. Basically 0.1% is the same as -60 dB. For every 10x distortion up or down from there, you either add or subtract 20 dB to the SINAD.

Anyway, I think that’s about all I have to share on the Crest Audio Vs900 amplifier. I’ve modeled the input line-level stage of this amp using MultismLive if you want to play around with ways to make the amp more sensitive for use in home audio. The V/Vs series amps range from the 450 model all the way up to the 1500 model but the only difference in the pre-amp stage is how much gain they have it set to. You could easily make any of these amps able to drive full output with only 0.775V by changing R2 and R4 to 23.2k. I might take that on with this amp, because 1.0V is pretty high to have to drive this thing to max power. My Presonous Firestudio almost couldn’t do it, as the output in REW was set to -1.75 dBFS in order to get to the 450 watts. This amp will end up paired with a Yamaha RX-V1700 which is spec’d to be able to drive up to 2.0V out the subwoofer jack, so it might not be necessary, but still a little more up-front gain could be nice just to keep me from having to set the sub to +12 dB on the receiver and not have any room to go up without just turning up the overall volume.