Does Khozmo make a Balanced/Quad stepped attenuator with SMD resistors? I don't see it as an option on their site?
BTW, does anyone have an opinion of the Acoustic-Dimension 41-stepped series attenuator?
I like the fact it's shielded.
Last edited:
I'm looking for some cheaper dac to go with this amplifier and still get better signal from my PC with Realtek Alc892. I found something on ebay, any opinions? Or any better ideas? Thank you.
Link: CM6631A USB 2 0 to SPDIF Coaxial Optical 24bit 192kHz Sound Card DAC Board | eBay
Best regards, Žiga
Link: CM6631A USB 2 0 to SPDIF Coaxial Optical 24bit 192kHz Sound Card DAC Board | eBay
Best regards, Žiga
Speakers too
I'm doing the requisite 300 hours final break-in on a pair of Alpair 10peN that I got in this week, and built some temp. boxes for--got them hooked up yesterday to an SE and they are driven reasonably well (as loud as I'm supposed to run them until they hit the 500 hr. mark, based on the designer).
Sound on "simple" recordings was quite nice, but things began to get a bit thicker sounding on dense music; my sense was that the little SE was just loosing control. I'm waiting for the LPUHP boards that will power these eventually, but may take apart my BAL-BAL to see what improvements the extra drive will yield. (Plus the balanced out on my DAC is a good bit better than the SE out.)
I'm doing the requisite 300 hours final break-in on a pair of Alpair 10peN that I got in this week, and built some temp. boxes for--got them hooked up yesterday to an SE and they are driven reasonably well (as loud as I'm supposed to run them until they hit the 500 hr. mark, based on the designer).
Sound on "simple" recordings was quite nice, but things began to get a bit thicker sounding on dense music; my sense was that the little SE was just loosing control. I'm waiting for the LPUHP boards that will power these eventually, but may take apart my BAL-BAL to see what improvements the extra drive will yield. (Plus the balanced out on my DAC is a good bit better than the SE out.)
Got in an extra 4 pin XLR, so was able to use that to run the alpairs off the bal-bal. (I hate to unsolder and re solder lead connections on boards, so did not want to do this.)
The extra drive of the bal-bal and my balanced IV has made a real difference. This makes me even more anxious to get the LPUHP boards in! For once I even have all the parts at hand. But waitings fine--I still have weeks more break in on the speakers before they can be pushed.
The heat sinks get very hot to the touch--not so hot that I need to pull my finger away (though my wife joke that I have asbestos fingertips). I should dig out my temp. probe. What's the safe upper limit for the heatsinks to get? Considering what I put those chips through soldering them onto the board, I know they can take crazy heat.
The extra drive of the bal-bal and my balanced IV has made a real difference. This makes me even more anxious to get the LPUHP boards in! For once I even have all the parts at hand. But waitings fine--I still have weeks more break in on the speakers before they can be pushed.
The heat sinks get very hot to the touch--not so hot that I need to pull my finger away (though my wife joke that I have asbestos fingertips). I should dig out my temp. probe. What's the safe upper limit for the heatsinks to get? Considering what I put those chips through soldering them onto the board, I know they can take crazy heat.
I was going over the PSU V2 BOM in preperation of ordering parts when the boards are shipped. I am planning to feed the PSU with the Antek 25VA 12V toroid.
Looking at the BOM default parts I see that R29, R30, R52 and R55 are designed for 15V output (see below).
REF DES _ Item _________________________ Description
R29, R52 Voltage Set Resistor (for 15V output) RES 1.24M OHM 1/8W 1% 0805 SMD
R30, R55 Voltage Set Resistor (for 15V output) RES 105K OHM 1/8W .1% 0805 SMD
Do I need to only change R30, R55 for 12V output as indicated in the BOM and shown below?
What about R29 and R52 as implied above?
Desired Set Voltage: R30, R55:
18 ______________ 86,282
16 ______________ 97,919
15V _____________ 105,000
13 ______________ 122,753
12V _____________ 134,088
11 ______________ 147,730
10 ______________ 164,463
9 _______________185,469
8 _______________212,628
Looking at the TPS7A3301 datasheet (pg11) it seems for 12V that R29 and R52 should be changed to 1.5MΩ ?
Looking at the BOM default parts I see that R29, R30, R52 and R55 are designed for 15V output (see below).
REF DES _ Item _________________________ Description
R29, R52 Voltage Set Resistor (for 15V output) RES 1.24M OHM 1/8W 1% 0805 SMD
R30, R55 Voltage Set Resistor (for 15V output) RES 105K OHM 1/8W .1% 0805 SMD
Do I need to only change R30, R55 for 12V output as indicated in the BOM and shown below?
What about R29 and R52 as implied above?
Desired Set Voltage: R30, R55:
18 ______________ 86,282
16 ______________ 97,919
15V _____________ 105,000
13 ______________ 122,753
12V _____________ 134,088
11 ______________ 147,730
10 ______________ 164,463
9 _______________185,469
8 _______________212,628
Looking at the TPS7A3301 datasheet (pg11) it seems for 12V that R29 and R52 should be changed to 1.5MΩ ?
Last edited:
Any reason to use a 12V toroid Vs 15V toroid to feed the PSU-V2?
OPC: I was basing this on your suggestion to another poster to use the 12V Antek for an SE-SE.
I need to build three of these PSU-V2.
One each for all three Wire Amps: Bal-Bal, Bal-SE & SE-SE.
Frankly, I have no V preference. Don't know if I should sweat it and stick with 15V?
I'm no engineer
OPC: I was basing this on your suggestion to another poster to use the 12V Antek for an SE-SE.
I need to build three of these PSU-V2.
One each for all three Wire Amps: Bal-Bal, Bal-SE & SE-SE.
Frankly, I have no V preference. Don't know if I should sweat it and stick with 15V?
I'm no engineer
Last edited:
Hi Tribbs,
All the values in that table were calculated assuming that R29 and R52 remain as 1.24M ohm.
Only R30 and R55 change to the values indicated in the table.
Alternatively, if you want 12V, you can follow the table in the datasheet by changing all the resistors to the indicated values. Both are correct, they are just different ways of doing the same thing. I wanted users to be able to change voltages with a minimum amount of effort and cost, so I fixed the value of both R29 and R52.
Regards,
Owen
All the values in that table were calculated assuming that R29 and R52 remain as 1.24M ohm.
Only R30 and R55 change to the values indicated in the table.
Alternatively, if you want 12V, you can follow the table in the datasheet by changing all the resistors to the indicated values. Both are correct, they are just different ways of doing the same thing. I wanted users to be able to change voltages with a minimum amount of effort and cost, so I fixed the value of both R29 and R52.
Regards,
Owen
If you want 12V output, then use either a 10V or 12V toroid. If you use a 15V, you will just be dropping more voltage across the regs which will cause them to heat up unnecessarily.
If you want to run at 15V rails, then going to a 15V transformer might be necessary depending on the load and transformer ratings.
Ideally, you want to figure out how much voltage output you really need, and only build to that requirement. Running super high rails is of no benefit if you're only looking for 500mVRMS output. All it does is thermally stress all the components while wasting power.
Regards,
Owen
That's a tricky question
There are many different ways to go about it, and they may yield different outcomes. You can do it purely subjectively by listening to something as loud as you care to and working out the peak possible output level, or you can target some output level that makes you happy (like say the ability to hit 110dBSPL).
If I were you, I would start here:
http://www.diyaudio.com/forums/mult...much-voltage-power-do-your-speakers-need.html
The test was designed for loudspeakers, but is equally applicable to headphones. This will get you closer to a realistic target as it will be catered to your actual listening habits as opposed to aiming for some arbitrary level.
Don't be surprised if a few hundred millivolts is all you need...
Regards,
Owen
There are many different ways to go about it, and they may yield different outcomes. You can do it purely subjectively by listening to something as loud as you care to and working out the peak possible output level, or you can target some output level that makes you happy (like say the ability to hit 110dBSPL).
If I were you, I would start here:
http://www.diyaudio.com/forums/mult...much-voltage-power-do-your-speakers-need.html
The test was designed for loudspeakers, but is equally applicable to headphones. This will get you closer to a realistic target as it will be catered to your actual listening habits as opposed to aiming for some arbitrary level.
Don't be surprised if a few hundred millivolts is all you need...
Regards,
Owen
Rail Voltage
You just had to ruin my Friday evening listening session didn't cha'?
Well, here are my set of 120Hz test results at an absolute max volume level that I may ever listen to at most a couple of CD tracks. After having imbibed two different cans and three cans of beer...
0.5V = Grado 325i (32Ω, 98dB) circa 2003
2.3V = Sennheiser HD430 (600Ω, 94dB) circa 1980's
How do I correlate the above results to an appropriate rail voltage?
That's a tricky question
You just had to ruin my Friday evening listening session didn't cha'?
Well, here are my set of 120Hz test results at an absolute max volume level that I may ever listen to at most a couple of CD tracks. After having imbibed two different cans and three cans of beer...
0.5V = Grado 325i (32Ω, 98dB) circa 2003
2.3V = Sennheiser HD430 (600Ω, 94dB) circa 1980's
How do I correlate the above results to an appropriate rail voltage?
Hey Owen, dumb question. Can I achieve 15 VDC output from your PSU with 12VAC supply toroids?
I thought that it was possible as 12 VAC is equivalent to 12 x 1.41 DC after rectification..
I have my boards setup for default 15VDC and bought the 12 VAC transformers recommended in earlier post ><
Edit: I understand there's voltage losses in the rectification process (min 0.7 v) but is it still possible with your psu boards?
I thought that it was possible as 12 VAC is equivalent to 12 x 1.41 DC after rectification..
I have my boards setup for default 15VDC and bought the 12 VAC transformers recommended in earlier post ><
Edit: I understand there's voltage losses in the rectification process (min 0.7 v) but is it still possible with your psu boards?
Last edited:
The test assumes that 0dBfs can never be exceeded.You just had to ruin my Friday evening listening session didn't cha'?
Well, here are my set of 120Hz test results at an absolute max volume level that I may ever listen to at most a couple of CD tracks. After having imbibed two different cans and three cans of beer...
0.5V = Grado 325i (32Ω, 98dB) circa 2003
2.3V = Sennheiser HD430 (600Ω, 94dB) circa 1980's
How do I correlate the above results to an appropriate rail voltage?
The measurements you have given show the AC voltage at the output when the signal is @ -9dB ref 0dBfs. (or is it -12dB ref 0dBfs?)
The absolute maximum peak signal at the output is 4*2.3V = 9.2Vpk
You would need +-10Vdc to +-12Vdc to reproduce these 0dBfs signals without clipping. The maximum Sennheiser current is ~ 6mApk
(1.5times the 600r peak current).
The higher sensitivity Grado would need +-3Vdc to +-5Vdc to reproduce the 2Vpk signal @ 0dBfs. The maximum Grado current is ~ 94mApk.
Taking the worst case from both requirements you end up with an amplifier that can output at least 9.2Vpk and 100mApk.
Last edited:
Hi Aive,
This depends on quite a few factors like:
- What the mains voltage is
- How the transformer is rated (15VAC output at full load?)
- The regulation spec of the transformer
- The rectification diodes used (Schottky?)
- The current drawn by the amplifier
Running 15V rails with a 12V transformer puts you in the yellow zone where if some of the above things aren't ideal, you might dip out of regulation from time to time.
The dropout on those regs is quite low at about 290mV, so as long as you can maintain 15.5V or so on the high side, you should be alright.
If you have a multimeter, then this is very easy to check. Measure the DC value before the regulators at high line voltage, and low line voltage in your area (this varies with the time of day/weekends), and if it stays above 15.5V then you probably don't have anything to worry about.
If the transformer is 25VA or more, and is rated for 15VAC at full load, then it's likely that you will get more like 16-17VAC under light load like the one presented by the headphone amp. If this is the case, then you should have more than enough margin.
Regards,
Owen
You just had to ruin my Friday evening listening session didn't cha'?
Sorry about that! At least we all learned something
Besides, didn't you just have to listen a little louder than usual?
Well, here are my set of 120Hz test results at an absolute max volume level that I may ever listen to at most a couple of CD tracks. After having imbibed two different cans and three cans of beer...
0.5V = Grado 325i (32Ω, 98dB) circa 2003
2.3V = Sennheiser HD430 (600Ω, 94dB) circa 1980's
How do I correlate the above results to an appropriate rail voltage?
This seems pretty much spot on, and looks like some very valid numbers. In this case the test even accounted for environmental factors like how the consumption of alcohol often leads to louder listening levels!
Andrew perfectly summarized the results and how to interpret them. The amp running the Sennheisers will need roughly 10-12V rails, and the amp running the Grado headphones could do with much less. Even 5V rails would suffice for that amp.
For the BAL-BAL amp, you can run very low rails as it can swing twice the output voltage with the same given power rails as the SE output variants. that means you'd only need 6V rails for the BAL-BAL even with the Sennheiser headphones.
Regards,
Owen
Thank you kind sir! A great lesson.
So for the Bal-Bal, a 12V toroid and ±12V rails seems appropriate for me then.
Thank you Owen and AndrewT, for spending your time to teach all of us, with hang overs or not
A graph I noticed in the LME49600 spec sheet seems to show better numbers with higher rail voltages. Specifically:
Phase vs Frequency vs Supply Voltage Wide BW Mode (Pg 5, fig 7)
Through all this I've lost track of the reason (Other then perhaps for part availability/price.) why we may want to keep the rail voltage at a minimum or use any other rail voltage other then ±15V. To what benefit for example is a ±9V verses a ±15V rail? After all, the LME49600 spec sheet show most tests using Vs = ±15V
But, I can appreciate ratcheting down the rail to steer well clear of the Bal-Bal's OPA1632 absolute maximum ±16.5Vs rating.
Can we agree then that, at least for the Bal-Bal, a toroid with 12V rated output feeding the PSU-V2 set for ±12V rails is a good compromise? Although I am not clear on exactly what we are compromising, vis-à-vis ±15Vs.
An aside, I also noticed figure 23 on page 8:
VS = ±15V, RL = 32Ω, f = 1kHz
Blasting my Grado's yesterday at what I calculated to be about ¼Wrms @120Hz, and looking at the above referenced graph, would show THD+N through the roof! I think. But, who listens at that level anyway?
Last edited:
I would agree... running at +/-12VDC for the rails is a good compromise.
Reducing rail voltage really just reduces the thermal load on the buffers and op-amps. The LME49990 and OPA1632 don't have thermal pads, so the little SOIC package needs to dissipate all the heat associated with higher voltage rails. If you run the OPA1632 at +/-16 VDC it gets too hot to keep your fingers on.
I honestly don't think this is much of an issue, as that IC could likely survive for 50 years at +70C, but it does beg the question "why would you thermally stress the components if it's really not necessary?"
The phase and gain graphs referenced on the datasheet only really show significant differences up at 10-100MHz which is so far beyond the audio band that it's completely irrelevant for our purposes. The phase delta between +/-2.5V and +/-15V at 20kHz would not be measurable, much less audible.
As for distortion, at a 1/4W you're sitting at at the bottom of a very low trough
0.000029% is a pretty staggering number. Less than 0.5 parts per million distortion. With the LME49990 you'll do even better than that graph as the distortion and noise is dominated by the op-amp used to nest the buffer.
Unfortunately, the distortion of the Grados is going to spoil the party a bit at ~-60dB, but something like my favorite AH-D2000 cans from Denon keep up remarkably well for a transducer at closer to -90dB for the highest distortion component. I've seen many amplifiers do much worse than that!
Cheers,
Owen
Reducing rail voltage really just reduces the thermal load on the buffers and op-amps. The LME49990 and OPA1632 don't have thermal pads, so the little SOIC package needs to dissipate all the heat associated with higher voltage rails. If you run the OPA1632 at +/-16 VDC it gets too hot to keep your fingers on.
I honestly don't think this is much of an issue, as that IC could likely survive for 50 years at +70C, but it does beg the question "why would you thermally stress the components if it's really not necessary?"
The phase and gain graphs referenced on the datasheet only really show significant differences up at 10-100MHz which is so far beyond the audio band that it's completely irrelevant for our purposes. The phase delta between +/-2.5V and +/-15V at 20kHz would not be measurable, much less audible.
As for distortion, at a 1/4W you're sitting at at the bottom of a very low trough
0.000029% is a pretty staggering number. Less than 0.5 parts per million distortion. With the LME49990 you'll do even better than that graph as the distortion and noise is dominated by the op-amp used to nest the buffer.
Unfortunately, the distortion of the Grados is going to spoil the party a bit at ~-60dB, but something like my favorite AH-D2000 cans from Denon keep up remarkably well for a transducer at closer to -90dB for the highest distortion component. I've seen many amplifiers do much worse than that!
Cheers,
Owen
That's an M as in Mega-Hz, yeah
I didn't have my coffee yet.As for distortion, at a 1/4W you're sitting at at the bottom of a very low trough
Er, that should have been 5 Watts into the Sennheisers.
Remind me not to read my notes and graphs too early on a Saturday morning.
I should have bought those Denon's while I had the chance. If you had to what would you replace them with?
All the best!
Thanks for your feedback Owen.
I got 50 VA 12 VAC transformers - I think the capacity and light loading should result in high'ish regulation (I'm hoping). I'll measure the VDC after the rectifiers as you recommend and see how I go. Otherwise I'll have to put a mouser order in for 4 x surface mount resistors and pay $30 shipping lol ><
Given that you have a 50VA transformer, and you live in Australia, the land famous for its very high mains voltage, I'd bet money that you're well over 15.5VDC before the regs.
It's still worth checking, but I'd be surprised if it was an issue.
If it is a problem, let me know and I can send you some resistors for the cost of lettermail, which should only be a dollar or two.
Cheers,
Owen