DACs with High Voltage Output

I am interested in upgrading my DAC. However, the integrated I run and like a lot, the 845 based Bel Canto SETi 40, needs a high voltage output to sound its best (at least 4 v rms). I have been working on finding DACs that fit that parameter. I'd love to get some input from the Audiogon community. Additionally it needs to have a USB input. 
I'll give you a negative recommendation... don't do a PS audio Directstream DAC.  It has a lower than average output voltage.  
From the PDF Manual for the SETi 40:

"Input level for 37 watts output • 1.5 V rms"

Many DACs will be in the 2 - 3 Vrms range...for the SE analog outputs. 

"Peak power output • >70 watts"  Perhaps @almarg can provide the maximum Vrms needed for peak output, should it be a factor.

Also, is this amp related or / and/or speaker related? You haven't listed your speaker.
A 37 watt output at 8 ohms is a rail voltage of 17.2 volts RMS. The peak power is based on the RMS voltage x 1.414, which is 24.3 volts. Power is therefore 24.3 squared divided by 8, or 74 watts.

That’s why the spec says "> 70 watts". The 1.5 Vrms will give peak power.
Thanks for the mention, @david_ten , and for your good catch on those specs.

I agree with the comment by gs5556, and in fact I was suspicious of that 70 watt number as soon as I saw it. Class A amplifiers typically do not provide dynamic headroom anywhere close to a near doubling of output power relative to their continuous power rating. In fact they commonly provide essentially no dynamic headroom.

So consistent with gs5556’s comment the 70 watt number is presumably the product (multiplication) of peak voltage and peak current, rather than the product of rms voltage and rms current which is how continuous power ratings (37 watts in this case) are specified.

But, hypothetically speaking, if the amp were in fact able to put out 70 watts defined in the same manner as continuous power ratings are defined, since for a given load impedance power is proportional to the square of voltage, if 1.5 Vrms produces 37 watts the input voltage required to produce those 70 watts would be 1.5 x (square root (70/37) = 2.06 Vrms. Still low enough to be generated by most DACs.

On the other hand, though, those sensitivity numbers presumably apply when the amp’s volume control is at max, and the OP may be looking for a DAC providing greater than usual output voltage to lower the settings of the volume control that he uses. Although he should keep in mind that as has been said in a number of past threads here higher volume control settings (even close to max) more often than not tend to be sonically preferable to lower settings, if it makes any difference at all.

Best regards,

-- Al

+1 Benchmark DAC3 does up to 18 volts 
@almarg  Al, thank you for your response. You are terrific at explaining these things.

From the PDF Manual for the SETi 40:

"Input level for 37 watts output • 1.5 V rms"

Many DACs will be in the 2 - 3 Vrms range...for the SE analog outputs.

david_ten is correct. With these sort of input sensitivity any cdp, dac will work, they’re nearly all over 2v these days.

Cheers George
+1 Benchmark DAC3 has high voltage (studio level) output capabilities, and can be used as Preamp as well.
It can take USB input.

If I remember correctly, there are three adjustable output voltage settings (for XLR output) - 2V/4V/9.8V. (can be adjusted via internal jumper)

For RCA output it is fixed at 2V.
Sorry guys. Been busy and stepped away from the thread for a bit. Thanks for all of the various feedback. All helpful in different ways.

I do have a question that stems from a couple of the comments if anyone is still following. I have read the manual and 1.5 v rms is indeed stated to result in maximum output. However, I am not sure how that translates, as it is certainly not true that it translates into maximum decibels.

I am currently using a Burson DAC with a stepped attenuation that ranges from 0 - 10 v rms based on where the knob is set. The volume increases all the way up the voltage chain, as you would assume. My speakers are some fairly insensitive bookshelves (87db). Unless the Burson attenuator is set to a level that outputs approximately 4 v rms there is not enough headroom in the Bel Canto volume range to be comfortable. So that is where I am coming up with that number. What am I not understanding? Even though the manual states 1.5 v gets maximum output, increasing the voltage input from there continues to increase the volume, all the way to 10 v rms.

What is the specific model name of the Burson DAC?  And can you provide a link to a description and/or manual for it?


-- Al
Its the Burson HA160D. There is unfortunately not a manual online (that I could locate). I got the above cited info about the voltage as it relates to the attenutor from a review at 6moons:  http://www.6moons.com/audioreviews/burson5/1.html

First, when you said…

Even though the manual states 1.5 v gets maximum output, increasing the voltage input from there continues to increase the volume, all the way to 10 v rms…

…Is that with the volume control on the amp at max, or with it turned down significantly? I suspect you meant that the volume increases all the way to 10 vrms with the amp’s volume control positioned significantly lower than max. If the volume controls of the DAC and the amp were both at max and the amp was not severely clipping and distorting it would certainly point to the statement in the review about 10 volts being grossly incorrect.

In any event I think it might be a good idea to contact Burson and ask them if they can clarify or confirm the relevant statement in the review. (I did some further research, btw, which failed to turn up any information supporting the statement). For the benefit of others, that statement was:

If you want to use the HA160D as standalone DAC into a preamp/integrated, click 3 on the stepped attenuator equates to 1VRMS out, click 5 to 2V, fully open to 10V. This will vary just a bit with sources of course.

Also, as a point of information the attenuator is stated to have 24 steps. And in that regard I would add that the steps are almost certainly not linear (which they shouldn’t be), and become finer as the setting is increased. So your estimate of 4 volts might be off somewhat, with the actual number being somewhat lower. For example a change of 2 clicks, from position 3 to position 5, results in a change of 6 db (from 1 volt to 2 volts), that being an average of 3 db per click. While a change from position 5 to position 24 can be calculated to average about ¾ db per click, which I suspect means ½ db per click near the upper end of the range, and 1 db per click at some of the lower settings.

Also, I’m wondering if a miscommunication with the reviewer might have occurred, and the voltages that were stated should have been indicated as peak-to-peak, rather than as rms. For a sine wave (which is what such specs are based on), 10 volts peak-to-peak corresponds to about 3.5 volts rms, and the 4 volt number you referred to, if peak-to-peak, would correspond to about 1.4 volts rms.

And perhaps another contributing factor is some degree of inaccuracy in the sensitivity spec of the amp. Also, a minor contributing factor may be that some (although not most) recordings are engineered such that their peak levels are a few db below "full scale" (the maximum possible digital value), which would cause the DAC to output less voltage than it is capable of for a given setting of its attenuator.

Finally, I saw a reference somewhere to 10 db of gain the unit provides when it is used as a preamp, rather than as a standalone DAC. Perhaps that factored into a possible miscommunication with the reviewer. 10 volts rms less 10 db of gain would be about 3.2 volts rms.

That’s about all I can think of at this point. As I said, perhaps Burson can shed some further light.


-- Al

Al, this is great insight that I am not (yet) technically literate enough to have discerned on my own. I really appreciate your willingness to dig into this. I will contact Burson at the beginning of the week and see if I cannot get a fuller, or perhaps more accurate, picture of how their DAC outputs. 

As a follow up to your question, yes, the volume control on the Bel Canto is necessarily significantly below max when I run the attenuation on the Burson to full tilt.  

I will report back after talking with Burson.
I did talk with Burson. They say that the attenuation does indeed step the voltage output all the way from 0.5 to 10 vrms. That means the Bel Canto really needs at least 4 vrms to be happy.
Also, in response to david_ten: there is certainly a speaker component, as they are somewhat inefficient bookshelves (87 db - Rosso Fiorentino Fiesole). But I am in a small space and have heard them in other contexts with equivalent power (Matersound 845 based integrated) and there is plenty of headroom.