ran_ph said:
Like this one:
An externally hosted image should be here but it was not working when we last tested it.
I dunno about the noise specs but I'm pretty sure the LM337/LM317's are generally regarded as being better than fixed voltage regs.......right?
Uh huh. That's the commonly held understanding. And indeed most of the published specs are "better". Then you decide the LM337 doesn't cut it and track down the LT1085s.�@Then I suppose you move into discrete pass transistors & outboard voltage reference chips before finally building your own regulator completely from scratch... which isn't necessarily a bad path to follow, but its a rather long and time consuming one.
Remember this project (the VSPS) was supposed to be simple, right?
Would two 7812s in the circuit shown be better than than the traditional 7812 7912 combination? Somthing tells me theres a hidden penalty there somewhere, but I cant see it at the moment.
the numbers game
bored on a hot day ...
Correct. Assuming the op-amp has about 80 dB of PSRR over the audio bandwidth, the 360 uV on the rails appears as 36nV on the output. Nanovolts.
For the VSPS of 40dB gain and a voltage noise of 1nV/rt Hz, the output noise from the opamp itself over the audio bandwidth is at least 14 uV. Microvolts. And thats for an extremely low noise op-amp neglecting current noise and the thermal noise of the cartridge itself. 100 uV is a better real world estimate.
Its fairly safe to say that in high-gain op-amp circuits the noise of the amplifier is going to dominate over the that of the power supply by something in the order of a factor of 100-1000.
At high frequencies, the PSRR of the opamp is poor as is the performance of the regulator. Diode switching noise, for example, may well be audible. However I think that's better dealt with by careful selection of diodes and use of snubbers if need be.
/R
bored on a hot day ...
Correct. Assuming the op-amp has about 80 dB of PSRR over the audio bandwidth, the 360 uV on the rails appears as 36nV on the output. Nanovolts.
For the VSPS of 40dB gain and a voltage noise of 1nV/rt Hz, the output noise from the opamp itself over the audio bandwidth is at least 14 uV. Microvolts. And thats for an extremely low noise op-amp neglecting current noise and the thermal noise of the cartridge itself. 100 uV is a better real world estimate.
Its fairly safe to say that in high-gain op-amp circuits the noise of the amplifier is going to dominate over the that of the power supply by something in the order of a factor of 100-1000.
At high frequencies, the PSRR of the opamp is poor as is the performance of the regulator. Diode switching noise, for example, may well be audible. However I think that's better dealt with by careful selection of diodes and use of snubbers if need be.
/R
Re: the numbers game
PSRR decreases with opamp gain. In the case of 40dB and above example you would end up with 3600nV of supply-induced noise at the output.
Your average opamp has perhaps 70dB of PSRR at 10kHz and less above.
The supply rail contains broadband noise, noise correlated to regulator instabilities (said Ben Duncan: 'just listen to that 317'), and more often than not a half-waved rectified and distorted copy of the signal itself. Or the adjacent channel.
In an audio circuit there is always need to fret much about the P.S.
rjm said:
PSRR decreases with opamp gain. In the case of 40dB and above example you would end up with 3600nV of supply-induced noise at the output.
Your average opamp has perhaps 70dB of PSRR at 10kHz and less above.
The supply rail contains broadband noise, noise correlated to regulator instabilities (said Ben Duncan: 'just listen to that 317'), and more often than not a half-waved rectified and distorted copy of the signal itself. Or the adjacent channel.
ran_ph said:
In an audio circuit there is always need to fret much about the P.S.
my 1st VSPS
thanks to rjm for releasing this design out on the net. 😉
i've wired up my first phonostage. i settled on this design as it looked simple to build, and should allow me to spin some LPs really soon (i just got a td150).
it's a dual mono design with separate regulators for each opamp. i have on hand a pair of opa627 from my decomissioned active preamp that i intend to use. also had BG NX 4.7uF caps from that same build.
i've used smd 1% resistors throughout and silver mica 1% caps for the RIAA section (all resistors are matched between channels).
initially wanted to put a DIP switch to select between 40db and 50db, but that would mean an 8 position DIP as a pair of resistor needs to be switched for each channel, thus 2 pairs for 40db/50db selection for a channel. the DIP switch would actually be much bigger than the opamp socket (which is the area of the VSPS RIAA itself... each DIP also measures about 1R on my DMM). so i went with a 40db design instead.
the onboard regs have 180uF/20v sanyo oscons at pre and post reg positions. psu used in actual listening test was 2 x 50va 15vdc regulated supply taken from one of my previous project... meaning that the vsps is on dual mono psu and double regulated on each channel.
test setup:
we ran thru a Sumiko BPS retipped with Shibata stylus, original 103, 103R, 103R retipped with Shibata stylus on a technics SP10 at my friend's place. (will post pic of the lineup later...)
initially when i powered on, i was quite worried as the VPSP made no sound at all. when a LP was played, were were greeted with music. no hum, no hiss, etc. just plain music!
well... my opinion is that the VSPS did not perform too badly vs the resident improved marantz M7 phonostage. VSPS being cleaner and the M7 warmer.
pics as follows.
top of VSPS, with cd in the background:
bottom of VSPS:
closeup showing smd resistors of RIAA section:
board top view:
bottom view:
close up of 47k resistor between rca + and - :
dangling VSPS with my dual mono 50va 15vdc supply:
the source that was used, technics sp10 tt:
i'm just happy that i can finally get a hammond enclosure to box up this thing. 🙂
cheers garbage
thanks to rjm for releasing this design out on the net. 😉
i've wired up my first phonostage. i settled on this design as it looked simple to build, and should allow me to spin some LPs really soon (i just got a td150).
it's a dual mono design with separate regulators for each opamp. i have on hand a pair of opa627 from my decomissioned active preamp that i intend to use. also had BG NX 4.7uF caps from that same build.
i've used smd 1% resistors throughout and silver mica 1% caps for the RIAA section (all resistors are matched between channels).
initially wanted to put a DIP switch to select between 40db and 50db, but that would mean an 8 position DIP as a pair of resistor needs to be switched for each channel, thus 2 pairs for 40db/50db selection for a channel. the DIP switch would actually be much bigger than the opamp socket (which is the area of the VSPS RIAA itself... each DIP also measures about 1R on my DMM). so i went with a 40db design instead.
the onboard regs have 180uF/20v sanyo oscons at pre and post reg positions. psu used in actual listening test was 2 x 50va 15vdc regulated supply taken from one of my previous project... meaning that the vsps is on dual mono psu and double regulated on each channel.
test setup:
we ran thru a Sumiko BPS retipped with Shibata stylus, original 103, 103R, 103R retipped with Shibata stylus on a technics SP10 at my friend's place. (will post pic of the lineup later...)
initially when i powered on, i was quite worried as the VPSP made no sound at all. when a LP was played, were were greeted with music. no hum, no hiss, etc. just plain music!
well... my opinion is that the VSPS did not perform too badly vs the resident improved marantz M7 phonostage. VSPS being cleaner and the M7 warmer.
pics as follows.
top of VSPS, with cd in the background:
An externally hosted image should be here but it was not working when we last tested it.
bottom of VSPS:
An externally hosted image should be here but it was not working when we last tested it.
closeup showing smd resistors of RIAA section:
An externally hosted image should be here but it was not working when we last tested it.
board top view:
An externally hosted image should be here but it was not working when we last tested it.
bottom view:
An externally hosted image should be here but it was not working when we last tested it.
close up of 47k resistor between rca + and - :
An externally hosted image should be here but it was not working when we last tested it.
dangling VSPS with my dual mono 50va 15vdc supply:
An externally hosted image should be here but it was not working when we last tested it.
the source that was used, technics sp10 tt:
An externally hosted image should be here but it was not working when we last tested it.
i'm just happy that i can finally get a hammond enclosure to box up this thing. 🙂
cheers garbage
Even just thinking about the effort required to solder those itty-bitty resistors in place makes my head hurt. I can barely see them at all in the photo.
Certainly, I think we have a winner in most-compact-VSPS division...
-rjm
Certainly, I think we have a winner in most-compact-VSPS division...
-rjm
not my sp10mk3...
thanks ran_ph.
alas, the sp10mk3 belongs to my friend. he volunteered his setup to test my vsps as my td150 has been stripped, awaiting a replinth when i have time.
forgot to mention that the denons are mc carts and we used his denon permalloy stepup transformer to boast the signal before going into the vsps.
below are the pics that i promised:
lineup of carts to be tested. sumiko BPS in background, i'm not sure which is the R.
the 103R with elliptical stylus:
ran_ph said:
thanks ran_ph.
alas, the sp10mk3 belongs to my friend. he volunteered his setup to test my vsps as my td150 has been stripped, awaiting a replinth when i have time.
forgot to mention that the denons are mc carts and we used his denon permalloy stepup transformer to boast the signal before going into the vsps.
below are the pics that i promised:
lineup of carts to be tested. sumiko BPS in background, i'm not sure which is the R.
An externally hosted image should be here but it was not working when we last tested it.
the 103R with elliptical stylus:
An externally hosted image should be here but it was not working when we last tested it.
Mine is about that size, too. It's all through hole parts though. It is JAM PACKED. I have some resistors on the back of the board. Garbage, your's is definitely neater than mine. I wasn't so sure mine would even work. I'll try and post some pictures of it later.
nobody special said:
😉
neither was i. i only did rudimentary voltage checks.
i will take up richard's suggestion to remove the jackets of the lytic caps next.
forgot to make a correction in my post...
the last pic is of the cart that we listened to last. it is a denon 103R elliptical, not shibata as indicated in post 187.
also the pics of the carts.
1st pix: bps garrott shibata in background, 103R garrott elliptical in black headshell, and the 103R stock in silver headshell
2nd pix: 103R garrott elliptical
cheers
garbage
Well, then you did more than I did! 😀 I just fired it up, and hoped it didn't fire up! Actually, one channel didn't work at first.. until I found I hadn't plugged in the RCA connector all the way. DOH!
I've done a lot of prototyping / soldering in the past 12 years or so at work, so I knew what to look for. I did have to touch up one area, but I caught it before I tried it out, so no problem.
I can't really get to the chip on my board with a probe to check it from the top. I am still amazed it worked, and was as quiet as it was. My grounding scheme must have worked well, because I have zero hum. There is a little noise only when the amplifier is cranked all the way up. At normal sane levels, it's dead quiet.
The noise I'm getting sounds like rectifier noise. I used a lot of capacitance in the raw supply. Do you think snubber caps across the diodes would help? What's a good value, .01uf?
hi steve
i'm not sure if snubber diodes helps. if i'm not mistaken, most people use 0.1uF. i tried that with my buf634/opa627 head amp, but found no audible difference. the rectifiers used for that application were 4 x mur860 (actually the same psu that i used during my testing of the vsps).
cheers
garbage
i'm not sure if snubber diodes helps. if i'm not mistaken, most people use 0.1uF. i tried that with my buf634/opa627 head amp, but found no audible difference. the rectifiers used for that application were 4 x mur860 (actually the same psu that i used during my testing of the vsps).
cheers
garbage
The diode noise does tend to get through and form the low level background hiss. A quick test is to try batteries and see how much lower the noise floor is.
If it bothers you, there are four options:
If it bothers you, there are four options:
- substitute a different diode for rectification
- experiment a little more with snubber circuits, including RC types
- lower the amount of capacitance in the power supply
- finally, use batteries as a permanent solution
[/list=1]
-rjm
I guess it (the switching noise) really doesn't bother me. It's at a very low level, and isn't audible at any sane listening level.
I set cartridge up a little better last night, and listened to a few more lp's. The grainy upper mids are subsiding. I broke out an unopened copy of Phil Keaggy's "Way back Home" LP. It's mostly acoustic guitar and vocals, with some strings. I have heard this on cd many times over, and I was hearing things in the LP I hadn't ever heard before! There is something going on with the inner dynamics that is really amazing. I've never heard it before on my system. Very cool. 😀
On the other hand, there is still a flat spot in the mid/upper mid that is a little annoying. I'm wondering if it isn't my AT cartridge, which is known to sound a little on the warm side. Could it be the RIAA network?
Overall, this project has made me consider getting a really good TT. I've been listening to lp's since I was 8 years old (When I got my first real system- marantz turntable and receiver) but this is the first time I've heard the real potential of the LP.
I set cartridge up a little better last night, and listened to a few more lp's. The grainy upper mids are subsiding. I broke out an unopened copy of Phil Keaggy's "Way back Home" LP. It's mostly acoustic guitar and vocals, with some strings. I have heard this on cd many times over, and I was hearing things in the LP I hadn't ever heard before! There is something going on with the inner dynamics that is really amazing. I've never heard it before on my system. Very cool. 😀
On the other hand, there is still a flat spot in the mid/upper mid that is a little annoying. I'm wondering if it isn't my AT cartridge, which is known to sound a little on the warm side. Could it be the RIAA network?
Overall, this project has made me consider getting a really good TT. I've been listening to lp's since I was 8 years old (When I got my first real system- marantz turntable and receiver) but this is the first time I've heard the real potential of the LP.
Its worth giving audio circuits at least 100h of break in before making any real conclusions.
In your case, though, I really do think you need to urgently attend to point 3, above. I suggest removing, if only temporarily, the 6800uF caps and 22ohm resistors from the power supply and giving it a listen.
/rjm
In your case, though, I really do think you need to urgently attend to point 3, above. I suggest removing, if only temporarily, the 6800uF caps and 22ohm resistors from the power supply and giving it a listen.
/rjm
Could you explain to me how too much capacitance is a bad thing? And, why not the RC section? My thinking was, as long as my voltage coming in to the regulators is sufficiently high at the max current being drawn, how would a little resistance hurt? It definitely knocks out some of the noise; it suppresses it at least 10x by my simulations. As long as the input voltage is in the regulator's range, and there is sufficient filtering in close proximity to the opamp, why would it make any difference what is supplying it, as long as it's clean? I could see the caps near the opamp being critical in value.
I'll give it a try when I get the chance. I would really like to hear what you think about what I've said. I wasn't saying that you're wrong- only that I don't see what the problem is. It wouldn't be the first time I was wrong 🙂
I'll give it a try when I get the chance. I would really like to hear what you think about what I've said. I wasn't saying that you're wrong- only that I don't see what the problem is. It wouldn't be the first time I was wrong 🙂
I could make a pretty good case for 6800uF being too high. I dont have time to run through the numbers, but the as you add capacitance the diodes and transformer have to pass more current in a shorter space of time to maintain the same steady state output current. Both components become progressively less happy.
If you are into simulations, it is interesting to calculate the conduction angle as a function of filter capacitance and output current.
Of course if you remove the caps, you can't leave the resistors in there or you raise the power supply impedance to the value of the series resistance.
I could continue and try to explain the virtues of low impedance, low capacitance power supplies, but it would be quicker really if you just gave it a shot and decided if you liked it or not.
/R
If you are into simulations, it is interesting to calculate the conduction angle as a function of filter capacitance and output current.
Of course if you remove the caps, you can't leave the resistors in there or you raise the power supply impedance to the value of the series resistance.
I could continue and try to explain the virtues of low impedance, low capacitance power supplies, but it would be quicker really if you just gave it a shot and decided if you liked it or not.
/R
I understand. I 'll give it a shot. If anyone can point me toward a calculation for the conduction angle, I'd like to see for myself what is going on.
- Status
- Not open for further replies.