Great that you can still find the parts. I managed to find a surplus of those Murata DCDC converters on eBay (bag of 100). Handy to have around for any project needing +/-15v. Sometimes I wonder why we work so hard to make fancy supplies because these measure very clean as implemented in BTSB with an external LC filter. You can see how clean the noise floor is.
Doesn’t get much simpler than this (on BTSB Panel Mount):
On BTSB SMT/TH:
Doesn’t get much simpler than this (on BTSB Panel Mount):
On BTSB SMT/TH:
Last edited:
Thanks - I guess it isn’t able to fully drive an F4 then which needs 40Vpp - because it can’t handle a higher voltage power supply?
I also just wondered if it would work with 12.7v in for the input signal (I may, or may not, use all of the gain on BTSB just depend where it’s best to add gain - throughout the chain, or just add a whole slog of +20dB gain at the final stage of BTSB?).
Been watching this discussion and I would like to clarify some discrepancies I see in your original post/question. I apologize if I am overlooking something obvious.
8dB gain is 2.5X. I don’t know what “6.35” is in your original question. For now, I’m going to ignore it until you clarify.
So 2V RMS from your DAC (at 0dB fs, which means at maximum output) connected to a preamp stage of max gain of +8dB (2.5X), means the max output from the preamp will be (2V RMS * 2.5) = 5V RMS.
So 5V RMS is 7V peak or 14V peak to peak (Vpp).
There are multiple gain options for the BTSB, and they are all operated by manipulating different dip switches. Understand that even with those gain options, the BTSB is limited on its maximum output by the power supply voltage rails that you implement and on the input impedance of the next stage, which in this case you already know is 47K ohms (the F4). This is true for any circuit, whether BTSB or another preamp, etc…
A 47K ohm input impedance is not a hard load. The LME49724 (if used in the BTSB) will happily swing lots of volts into a 47K ohm load without impediment. How do I know this? Because the data sheet of the LME49724 shows that and it also shows that on DIFFICULT loads, such as 600 ohms, it does very, very well.
Now, in order to clip an F4, Nelson has already stated in his design doc, that you need 40Vpp, or 20Vp or 14.14V RMS.
In other words, if you were to inject 14.14V rms to the front end of an F4, you will obtain an output of 25 watts RMS at 0.5% distortion overall.
This makes sense right? P = V^2/R; where P is 25 watts and R is 8 ohms. Solve for V, you find that is 14.14V RMS. Remember, the F4 is a 0dB voltage buffer. It is a current buffer, without any overall gain (0dB). It has current gain not voltage gain.
Let’s go back to the original question/problem.
You have 5V RMS from your dac/preamp combo. That, I am going to assume is the max. And you need 14.14VRMS to clip the F4.
14.14V RMS/5V RMS = 2.828 times.
20 log (2.828) = +9 dB gain.
You need to set your BTSB for at least 9dB gain or the closest number which is 8dB gain. I would recommend you actually set it higher, at lets say, 14dB gain. What that means is that you won’t need to turn your volume as much on your preamp.
Remember, after your dac, the amount of max gain you have available is +8dB from your preamp, and +14dB from your BTSB/F4. So that’s a total of 22dB gain. 22dB gain is about 12.6X. If you take your dac output of 2V RMS multiply it by 12.6 times, you get 25V RMS. 25V RMS is comfortably above the 14.14V RMS that you need to clip the F4 which is a good thing. The BTSB will then be operating at low distortion levels when the F4 is being clipped. You certainly don’t want the BTSB to be clipping before that, right?
The way you can think of it is you have a dac with 2V RMS output connected to an “integrated“ amp with overall 22dB gain. That’s pretty normal right?!
You also need to run your BTSB with the highest voltage rails it will allow, which imho opinion after reading through the LME49724 data sheet and also Neurochrome’s measurements of his UB (which also uses the LME49724) is +/- 18V max. You can probably get by with +/- 15V supplies, given that the design sheet shows 52Vpp capability (20V RMS) at these supply voltages (ex. Figure 13 where RLoad is 2k ohms). The chip will also run cooler, which is a good thing for longevity.
I did a similar analysis a few months back here for another diyaudio member. I would recommend you peruse both links for your design purposes and come to your own conclusions.
Remember, all this analysis paralysis also depends on the sensitivity of your loudspeakers. I am going to assume you have speakers that are at least 85-86dB sensitivity (and preferably 90dB but less than 105dB). It is no accident that low sensitivity loudspeakers (84-87dB) when used with dacs that are 2V RMS, need more overall gain (i.e. 20 to 28dB)and that very high sensitivity loudspeakers, i.e. Zu, Klipsch (~ 100dB plus) can get by with overall gains that are much, much lower (i.e. 14 to 20dB).
If the dac output voltage is higher, you decrease the requirements of overall gain. Modern dacs have much higher gain, typically 4V RMS or more. 4V rms would result in overall gain requirements of 6dB less, for example. My Halo dac is 5.6V rms from the XLR/balanced outputs and 2.8V RMS from the RCA/unbalanced outputs. My speakers are 96dB sensitive, I don’t have a preamp and the amps I build are usually 14 to 20dB gain.
Hope this helps.
Best,
Anand.
Last edited:
Thank you for the very clear explanation and link.
6.35 is the gain for +8dB (source).
My question in post #361 was what is the max voltage allowable into BTSB. For example, DAC out would be ~2vrms, which we know is fine. But I could also use a tube buffer stage inbetween that would deliver say 12vrms - before any gain being added by BTSB. 12vrms would obviously reduce the need for addition gain.
So I would expect BTSB to be fine with 2 to 5 Vrms input, that’s a given. But would I over drive it if I give it some 12v Vrms in? Whats the max input voltage Vrms?
6.35 is the gain for +8dB (source).
My question in post #361 was what is the max voltage allowable into BTSB. For example, DAC out would be ~2vrms, which we know is fine. But I could also use a tube buffer stage inbetween that would deliver say 12vrms - before any gain being added by BTSB. 12vrms would obviously reduce the need for addition gain.
So I would expect BTSB to be fine with 2 to 5 Vrms input, that’s a given. But would I over drive it if I give it some 12v Vrms in? Whats the max input voltage Vrms?
@stretchneck
Let’s assume we use a +/- 15V supply (which is what the stock supply is
). Let’s also assume we take the data sheet results of 52V pp capability or ~18V rms max output [My 20V RMS statement above was with a lot of rounding - oops!]
The maximum input sensitivity for the BTSB will vary depending on what gain switch you implement.
If it is set to 0dB (Av=1) gain setting, then obviously the maximum input sensitivity will be 18V rms.
If it is set to 6dB (Av=2) gain setting, then the maximum input sensitivity will be 9V rms.
If it is set to 14dB (Av=5) gain setting, then the maximum input sensitivity will be what? Hint: It will be less than the 6dB setting. And about 5 times less than the 0dB (Av=1) setting ;-)
If it is set to 20dB (Av=10) gain setting, then the maximum input sensitivity will be what? Hint: It will be less than the 14dB setting. And about 10 times less than the 0dB (Av=1) setting ;-)
Gain structure is something all diy’ers and audiophiles must understand and is individualized for each audio system. All things being equal, and generally speaking - Excessive amounts of gain can raise overall noise levels especially in systems using higher sensitivity speakers. Minimizing noise will maximize the signal you hear. You do have to account for recordings that are at low levels too. And that’s why we have volume controls!
Remember we are dealing with line level voltages here. I don’t see how the acoustical power discussions (i.e. the JL Audio link) is relevant 🤨. A gain of 8dB when dealing with line level voltages is Av = 2.5 times.
The big advantage with the BTSB is that you have control on what gain settings you want to implement for whichever amplifier/preamp build you have.
Best,
Anand.
Let’s assume we use a +/- 15V supply (which is what the stock supply is
). Let’s also assume we take the data sheet results of 52V pp capability or ~18V rms max output [My 20V RMS statement above was with a lot of rounding - oops!]The maximum input sensitivity for the BTSB will vary depending on what gain switch you implement.
If it is set to 0dB (Av=1) gain setting, then obviously the maximum input sensitivity will be 18V rms.
If it is set to 6dB (Av=2) gain setting, then the maximum input sensitivity will be 9V rms.
If it is set to 14dB (Av=5) gain setting, then the maximum input sensitivity will be what? Hint: It will be less than the 6dB setting. And about 5 times less than the 0dB (Av=1) setting ;-)
If it is set to 20dB (Av=10) gain setting, then the maximum input sensitivity will be what? Hint: It will be less than the 14dB setting. And about 10 times less than the 0dB (Av=1) setting ;-)
Gain structure is something all diy’ers and audiophiles must understand and is individualized for each audio system. All things being equal, and generally speaking - Excessive amounts of gain can raise overall noise levels especially in systems using higher sensitivity speakers. Minimizing noise will maximize the signal you hear. You do have to account for recordings that are at low levels too. And that’s why we have volume controls!
Remember we are dealing with line level voltages here. I don’t see how the acoustical power discussions (i.e. the JL Audio link) is relevant 🤨. A gain of 8dB when dealing with line level voltages is Av = 2.5 times.
The big advantage with the BTSB is that you have control on what gain settings you want to implement for whichever amplifier/preamp build you have.
Best,
Anand.
Last edited:
I think stretchneck is asking when will he overdrive the input buffer op amps. So I looked up the opa1656 in Mouser and clicked on the datasheet link. It says that op amp can go up to 29.5 Vpp into a 2k load with a +/- 15 VDC PS. The protection diodes, D1 - D8 keep the max input around the rail voltages anyway. If your input is near that, I think you could up the PS to the board to +/- 18 VDC, unless someone sees an issue with that. The same applies if you are using the SE output through an opa1656.
@propitious
Thank you!
Yes, I was looking at the LME datasheet instead of the OPA1656 (since the OPA is ahead of the LME in this circuit and is the input buffer) - my bad!
Sorry for the massive confusion! I completely forgot the original schematic! Apologies!
The gain settings with the LME opamp and/or OPA1637 occur AFTER the OPA1656 so the OPA1656 is the primary impediment. The OPA1656 should not alter the gain at all (1:1 essentially). So 29.5Vpp is about 10Vrms. IF you had 12V rms from your source (dac, preamp, whatever), you could overdrive the input OPA1656 opamps. Raising the rail voltage may be okay, but I don’t know about the heat dissipation issues of the OPA1656/OPA1637/LME opamps at higher rail voltages than +/- 15V.
Just be careful with the volume control with your preamp. That would be an unusual situation. Perhaps with very inefficient speakers and a low gain amplifier.
That aside in my own build, I changed R33 to R36 from 4k99 to 2k49 so the dip switches would allow for 6dB lower overall gains:
-6 dB [all switches off]
0 dB [only switch 1 and 4 on]
+8dB [only switch 2 and 5 on] and
+14dB [only switch 3 and 6 on].
That might be a better set of gain settings for certain setups.
Best,
Anand.
Thank you!
Yes, I was looking at the LME datasheet instead of the OPA1656 (since the OPA is ahead of the LME in this circuit and is the input buffer) - my bad!
Sorry for the massive confusion! I completely forgot the original schematic! Apologies!
The gain settings with the LME opamp and/or OPA1637 occur AFTER the OPA1656 so the OPA1656 is the primary impediment. The OPA1656 should not alter the gain at all (1:1 essentially). So 29.5Vpp is about 10Vrms. IF you had 12V rms from your source (dac, preamp, whatever), you could overdrive the input OPA1656 opamps. Raising the rail voltage may be okay, but I don’t know about the heat dissipation issues of the OPA1656/OPA1637/LME opamps at higher rail voltages than +/- 15V.
Just be careful with the volume control with your preamp. That would be an unusual situation. Perhaps with very inefficient speakers and a low gain amplifier.
That aside in my own build, I changed R33 to R36 from 4k99 to 2k49 so the dip switches would allow for 6dB lower overall gains:
-6 dB [all switches off]
0 dB [only switch 1 and 4 on]
+8dB [only switch 2 and 5 on] and
+14dB [only switch 3 and 6 on].
That might be a better set of gain settings for certain setups.
Best,
Anand.
Last edited:
I was curious to see if the BTSB really could drive 52v output swing. My built in function generator on my Oscope only makes 4.9Vpp. So I connected that to the balanced +/- on the input and measured output on +/- (GND pin in the middle not used). Sure enough I got 49vpp output on Oscope and 17.45vrms on DMM. That is 49.3vrms. It works!
Here’s the setup:
Oscope shows 48.8Vpp:
(The glitch at the peak is artifact of the generator)
Here is DMM AC volts:
Power supply was 12v class 2 linear wall wart and onboard DCDC is +/-15v.
This is certainly more than enough to drive an F4 to clip at 25w.
Here’s the setup:
Oscope shows 48.8Vpp:
(The glitch at the peak is artifact of the generator)
Here is DMM AC volts:
Power supply was 12v class 2 linear wall wart and onboard DCDC is +/-15v.
This is certainly more than enough to drive an F4 to clip at 25w.
The BTSB Panel Mount version uses OPA1637 output balanced drivers. I am not sure what the max Vpp output is on that one. Datasheet does not say specifically.
@xrk971
In early January, Vunce and I were exchanging a bunch of pm’s and in one of the earlier ones I wrote:
So now we have most of the data, except for max distortion at the highest levels - which I’m least concerned with since that‘s usually not the levels we ordinarily listen to.
Best,
Anand.
In early January, Vunce and I were exchanging a bunch of pm’s and in one of the earlier ones I wrote:
So now we have most of the data, except for max distortion at the highest levels - which I’m least concerned with since that‘s usually not the levels we ordinarily listen to.
Best,
Anand.
Sounds promising, 31mA max means it cannot drive 600ohm load as well. But F4 is 47k load on input so shouldn’t be a problem.
It might be good to figure out a resistor swap to enable 28dB gain on position where 20dB is now. That makes it a standard gain amp. But many people will have a preamp that is going to drive 20dB stage just fine.
I should run same quick test on Panel Mount.
It might be good to figure out a resistor swap to enable 28dB gain on position where 20dB is now. That makes it a standard gain amp. But many people will have a preamp that is going to drive 20dB stage just fine.
I should run same quick test on Panel Mount.
@xrk971
Do you know if there is another option for U1 on the BTSB or SMT version that can automagically convert +12V dc to +/- 18V dc? Just thinking out loud. If not, on LME based implementations, one can just wire in a +/- 18V super low noise linear supply. +/- 18Vdc might give all the swing one desires from the LME although with +/- 15Vdc we are already at ~ 50V pp as you showed with your very cool portable o’scope!
Best,
Anand.
Do you know if there is another option for U1 on the BTSB or SMT version that can automagically convert +12V dc to +/- 18V dc? Just thinking out loud. If not, on LME based implementations, one can just wire in a +/- 18V super low noise linear supply. +/- 18Vdc might give all the swing one desires from the LME although with +/- 15Vdc we are already at ~ 50V pp as you showed with your very cool portable o’scope!
Best,
Anand.