Sjöström Audio DIY forum

Projects => Power supplies => Topic started by: Keljian on June 13, 2005, 04:54:06 PM

Title: JSR03 - information
Post by: Keljian on June 13, 2005, 04:54:06 PM
Quote
A few questions about the JSR03, if you don't mind
>
> 1. What changes need to be made (aside the obvious for
>R8 and R9) for
> having the regulator output at 28.5v? (I'm guessing the
>25v caps would
> need to be increased to 35v - I'm looking for
>substitutes at the
> moment) note input voltage will be about 34v dc

Since the LM329 is fixed the only way to change the
voltage is the change the feedback. Strive to have the
impedance of the feedback at 500 ohms, meening that if you
take
R8 and R9 in parallel the value should be 500 ohms.

Caps wihc need to be changed are: C12, C4 = 35 V.  C1 may
not be changed if you can garantee max 35-37 volts. If you
aren't sure use 40 or 50 volts.

> 2. Regarding substitute parts, are there any you'd
>reccomend as
> potential upgrades for the original design? As I'm
>trying to get the
> absolute most out of the circuit. (aside using the
>LM329) Money is not
> an issue for parts.

I have not investigated possible "upgrades". Remember that
the whole purpose of the this regulator is to let the
AD825 do the major work to in an artifical way create low
output impedance.

>
> 3. Are there any potential problems problems running the
>BC32716TA
> instead of the BC860C as the latter has been
>discontinued? same goes
> for the BC560 (replacement BC557B)

Any >100 MHz, Hfe >300, Uce >50 will do.


>
> 5. I plan to be having low current draw (~500mA) will I
>have any
> problems using 1/8w resistors throughout the design?

I recommend 0.6 W metal film 1% ("plain metal flim") as
minimum but if you want to use smaller resistors make sure
you calculate the power dissipation in each.



(Thought I'd share because there may be other builders who require the same information)

Thanks for your patience per-anders
Title: JSR03 - information
Post by: Sigurd Ruschkowski on June 13, 2005, 06:25:16 PM
For high end audio use, I would change all caps into polypropylene types,
use other resistors like the Caddock MK132s or any type of Vishay's Bulk Metal Foil Type.


Sigurd
Title: JSR03 - information
Post by: Keljian on June 14, 2005, 12:02:43 AM
Quote from: Sigurd RuschkowskiFor high end audio use, I would change all caps into polypropylene types,
use other resistors like the Caddock MK132s or any type of Vishay's Bulk Metal Foil Type.


Sigurd

I'm guessing you're talking about these:

http://www.vishay.com/docs/26022/mkp1839.pdf ?
http://www.inter-technical.com/datasheets/MKP_10.PDF ?

Any preference of brand? I'm guessing WIMA?

I am planning to use vishay metal film resistors

I plan to use nichicon low impendance caps for the alu ones (with the exception of C12)
Title: JSR03 - information
Post by: Keljian on June 14, 2005, 03:53:29 AM
Struggling to find an equivalent part for the BC860(as the part is discontinued), as mentioned it has to be
>100 MHz, Hfe >300, Uce >50

The closest I've come up with is the BC557 which has all those specs however the collector current  is -100mA (whereas the 860 is -500mA)

http://www.fairchildsemi.com/ds/BC/BC557.pdf

Would this part work?

The R8 and R9 I'm planning on using are 2100 and 681 respectively, in parallel the value is 514~ ohms, is this close enough to 500 for the circuit?
Title: JSR03 - information
Post by: peranders on June 14, 2005, 08:51:21 AM
BC557 will be fine. The B type is sufficient but C is better.

R8 and R9 will also be fine if you want gain of 4. Remember though that R8//R9 = R7

When it comes to choosing caps, polyester is minimum in performance and polypropulene will work fine if they fit. Remember though that you mustn't add anything more at the output than the C12. If you add some 100 nF you will get oscillations

Sigurg is an audiophile, right(?), so sure he recommends something exotic. I'd say test it and come back about the results. At the moment neither I nor Sigurd have tested the difference. One thing is for sure and that is it feels better with "good" parts.
Title: JSR03 - information
Post by: Keljian on June 14, 2005, 10:30:52 AM
Thanks again P-A

I've made an excel spreadsheet with values and mouser/digikey part numbers, if you're interested in using it. Otherwise I've changed T4 to a d44h11, and will be using all polyproplyene caps (vishay mentioned earlier) they're about 1mm longer which I'll just bend the leads a bit for. Should fit ok. I'm more worried about the alu caps as they're a bit bigger then spec, but I'll figure something out.
Title: JSR03 - information
Post by: peranders on June 14, 2005, 10:40:39 AM
Have you noticed that all electrolythic caps have different voltage ratings so if you want some other bigger models, take a closer look at needed voltages.
Title: JSR03 - information
Post by: Keljian on June 14, 2005, 11:44:18 AM
Quote from: perandersHave you noticed that all electrolythic caps have different voltage ratings so if you want some other bigger models, take a closer look at needed voltages.


yeah but I'll be running 33v or so from the rectifier, I'll be using what you suggested, 50v for c1, 35v for c4 and c12, 25v for c3,c6 and c8, hoping that I have enough board space

.
Title: JSR03 - information
Post by: Sigurd Ruschkowski on June 14, 2005, 11:45:20 AM
Well, yes, I would say that I am kind of DIY audiophile. I do focus much on the quality of components. It always pays off!

Now, the superregulator is not an ordinary power supply regulator. It is very complex. I have not tested if special components wil make a sound system sound better. The quality of the sound system and the quality of how well one can listen to music .
However, I do like to have very very good components, and polyestercaps are not my favorites.


Sigurd

Quote from: peranders

Sigurg is an audiophile, right(?), so sure he recommends something exotic. I'd say test it and come back about the results. At the moment neither I nor Sigurd have tested the difference. One thing is for sure and that is it feels better with "good" parts.
:)
Title: JSR03 - information
Post by: Keljian on June 14, 2005, 11:56:27 AM
Quote from: Sigurd Ruschkowski
Now, the superregulator is not an ordinary power supply regulator. It is very complex. I have not tested if special components wil make a sound system sound better. The quality of the sound system and the quality of how well one can listen to music .
However, I do like to have very very good components, and polyestercaps are not my favorites.


Sigurd

Well it won't make it sound any worse I'm sure, so it can't hurt and it's only another $10USD or so for it, so I figure why not, I'm going all out on this project.
Title: JSR03 - information
Post by: Sigurd Ruschkowski on June 14, 2005, 11:56:33 AM
Yes, the MKP1839 or similar ones in polyprop are good. WIMA sells the MKP10 and MKP4 are very good, as are RIFA's blue PEH450.
Any of these will do a good job! These are also cheap.

A better alternative is Mundorf's tin foil polpropylene's - medium expensive.

If you are lucky you might find some 100 nF polystyrene caps - these are huge, though!

For electrolytic caps I would definitely go for ELNAs Cerafine series or Nichicon's Fine Gold,, or why not the top of the line Black Gate NX series!

I would definitely NOT use standard metall film resistors in a high en audio PS! Dale's RN60D-F are a good compromise between cost and performance. The Caddock MK132s are very expensive, and the Vishay Bulk Metal FOils are even more expensive.



Sigurd

Quote from: Keljian
Quote from: Sigurd RuschkowskiFor high end audio use, I would change all caps into polypropylene types,
use other resistors like the Caddock MK132s or any type of Vishay's Bulk Metal Foil Type.


Sigurd

I'm guessing you're talking about these:

http://www.vishay.com/docs/26022/mkp1839.pdf ?
http://www.inter-technical.com/datasheets/MKP_10.PDF ?

Any preference of brand? I'm guessing WIMA?

I am planning to use vishay metal film resistors

I plan to use nichicon low impendance caps for the alu ones (with the exception of C12)
Title: JSR03 - information
Post by: Keljian on June 14, 2005, 12:07:12 PM
Ok, I won't use black gates, they're just simply not worth it, nichicon are great caps (in my experience) and I'll be using elna SILMIC IIs in the device it's powering.

the nichicons I'm intending on using for the power regulator will be  
C1: UHE1H101MPD ( http://www.nichicon-us.com/english/seihin/pdfs/e-he.pdf )

C3,C6,C7: UPW1E101MEH ( http://www.nichicon-us.com/english/seihin/pdfs/e-pw.pdf )

C4, C12: UVZ1V101MED ( http://www.nichicon-us.com/english/seihin/pdfs/e-vz.pdf )


I'd prefer to use the vishay polyprop caps because they're axial leaded and would be easier to fit to the board rather than having say 7 or so huge caps hanging off the board, more a convenience thing and I suppose it'd be a last .5 % thing to have a different brand, this is already audiophile gear.

I was planning to use Vishay/dale CCFF55 resistors (I can't find the values I want in any others aside the RD55 which are too small in wattage)



Quote from: Sigurd RuschkowskiYes, the MKP1839 or similar ones in polyprop are good. WIMA sells the MKP10 and MKP4 are very good, as are RIFA's blue PEH450.
Any of these will do a good job! These are also cheap.

A better alternative is Mundorf's tin foil polpropylene's - medium expensive.

If you are lucky you might find some 100 nF polystyrene caps - these are huge, though!

For electrolytic caps I would definitely go for ELNAs Cerafine series or Nichicon's Fine Gold,, or why not the top of the line Black Gate NX series!

I would definitely NOT use standard metall film resistors in a high en audio PS! Dale's RN60D-F are a good compromise between cost and performance. The Caddock MK132s are very expensive, and the Vishay Bulk Metal FOils are even more expensive.



Sigurd
Title: JSR03 - information
Post by: Sigurd Ruschkowski on June 14, 2005, 12:22:28 PM
Keljian,
would be very interesting to see how you assembled the regulators:
Could you post some photos of your superregulators when they are finished?



Sigurd
Title: JSR03 - information
Post by: Keljian on June 14, 2005, 12:49:40 PM
I'm only making one of the regulators, don't need dual rails.

in other news I've substituted for RN60D resistors, however the values aren't spot on, *shrugs* they should be ok (they're within 5% of the values specced out for)

As for pictures, sure, once it's done.

I'd like to get the nichicon muse caps but they're just too difficult to obtain here in Australia. I'm already listening to a dtpakiller amp with the ones mentioned above and they should be fine. Though I might go panasonic FM now that it's been mentioned and I am making a digikey order..

Just so you know P-A (and all else who cares) it's for a maxxed out M3 ( http://www.amb.org/audio/mmm/  ) which I'm planning on building

I built a simple snubber power supply built around an LM317 but decided that it was the major weakness in the amp, plus the caps are some unknown brand (saxxon?) low esr, used uf4004 diodes for the rectifier and such - it'd work but I decided the amp deserved better.

will give you more info soon...
Title: JSR03 - information
Post by: Sigurd Ruschkowski on June 14, 2005, 02:55:07 PM
If you have a hard time (like me) to find high end audio related components there are two companies that have been very good to buy from:

Borbely Audio in Germany
http://www.borbelyaudio.com/audiophile_components.asp

and

Michael Percy Audio in the US
http://www.percyaudio.com/


Sigurd
Title: JSR03 - information
Post by: Keljian on June 14, 2005, 03:13:48 PM
Quote from: perandershttp://www.amb.org/audio/mmm/ .... surely

yep.. that'd be the one.. I was in a hurry.


will 150uF be ok for C4 and C12? (I can't seem to find 100v/35v panasonic FM)
Title: JSR03 - information
Post by: peranders on June 14, 2005, 05:30:21 PM
The exact value isn't critical. Walter Jung used the odd value 120 uF! 47-220 uF is "normal" I think.
Title: JSR03 - information
Post by: Keljian on June 15, 2005, 10:11:15 AM
parts ordered, more info about this project here:

http://www6.head-fi.org/forums/showthread.php?t=123686
Title: JSR03 - information
Post by: Sigurd Ruschkowski on June 15, 2005, 10:29:45 AM
Indeed,
and
performance
is not the same as good sound :D



Sigurd

Quote from: peranders
Quote from: Sigurd RuschkowskiWell, yes, I would say that I am kind of DIY audiophile. I do focus much on the quality of components.
Quality isn't the same thing as performance.  :D
Title: JSR03 - information
Post by: Sigurd Ruschkowski on June 15, 2005, 10:34:24 AM
P-A,

I see on the head-fi forum that you prefer the LM329 before the LM431"
"The voltage reference given in P-A's stock part list is LM431, he says the LM329 is better, so based on that, that's what I'm going to use.".

I am curious - me being a datasheet reader - why do you prefer the LM329?



Sigurd
Title: JSR03 - information
Post by: Sigurd Ruschkowski on June 15, 2005, 10:41:22 AM
Interesting thread, Jazper!

"As for rectifier diodes, diyaudio people seem to like the MUR820, which is overkill but cheap enough to use, so I plan to use a reasonable clone of it."


I would advice against the clones of the MUR820 as I have read that ONSemi's have the best ones. As a matter of fact I would use the soft recovery ones, MSR, instead.


I would definitlively go for a dual mono power supply! and also for each mono channel a dual rectifier bridge setup. For lowest noise and for best stereo separation.

Check out how I did with P-A:s RFB02s in dual bridge setup:
http://www.sjostromaudio.com/phpBB2/viewtopic.php?t=12


As for wires, do try solid silver wire (max 0.5mm in diam) - for everything.



Sigurd

Quote from: Keljianparts ordered, more info about this project here:

http://www6.head-fi.org/forums/showthread.php?t=123686
[/url]
Title: JSR03 - information
Post by: peranders on June 15, 2005, 10:47:58 AM
LM329 has less noise and is more stable but can't be adjusted. LM431 hasn't got any noise data but it is higher, still in real life with LP-filter the regulator has pretty low noise, way better than needed in most cases.

So if you want more than 6.9 volt and a very temperature stable output voltage LM329 is OK. If you want to save same money and in the same time want voltages från 2.5 V and up and also want to be able to tune the voltage in an easy way LM431 is to prefer.
Title: JSR03 - information
Post by: Keljian on June 15, 2005, 12:46:23 PM
Quote from: Sigurd RuschkowskiInteresting thread, Jazper!

"As for rectifier diodes, diyaudio people seem to like the MUR820, which is overkill but cheap enough to use, so I plan to use a reasonable clone of it."


I would advice against the clones of the MUR820 as I have read that ONSemi's have the best ones. As a matter of fact I would use the soft recovery ones, MSR, instead.


I would definitlively go for a dual mono power supply! and also for each mono channel a dual rectifier bridge setup. For lowest noise and for best stereo separation.


As for wires, do try solid silver wire (max 0.5mm in diam) - for everything.



Sigurd


The M3 has a rail splitter chip in the design, it's designed to have single supply.

As for the MUR820, I'm not too fussed to be frank, it's going to have 8000uF of capacitance after it, if that doesn't smooth the wave nothing will. The models I'm using are of high enough quality, and it's difficult to get ONsemi stuff in australia (I imagine it's easy in europe). I've just made a bulk order from mouser, so unless you're feeling charitable and feel like sending me some really nice rectifier diodes... :)

I'm sticking to silver coated copper wire.. I think the difference will be marginal compared to pure silver
Title: JSR03 - information
Post by: peranders on June 15, 2005, 01:24:23 PM
The most important thing is that it must feel right besides pure technical matters.

Your amp is a low power amp so I feel that choice of diodes are less important  than if it would have been a regular power amp.
Title: JSR03 - information
Post by: Keljian on June 15, 2005, 01:41:11 PM
Quote from: perandersThe most important thing is that it must feel right besides pure technical matters.

Your amp is a low power amp so I feel that choice of diodes are less important  than if it would have been a regular power amp.

How do you mean, feel right? (just clarifying)
Title: JSR03 - information
Post by: Keljian on June 15, 2005, 03:14:19 PM
ok .. after doing more research I've decided to go ahead and buy some harris RHRP860s locally
Title: JSR03 - information
Post by: peranders on June 15, 2005, 03:16:03 PM
It must feel good. That's it! Not more, not less. It's important to like the results and chosen parts.
Title: JSR03 - information
Post by: Sigurd Ruschkowski on June 15, 2005, 03:29:32 PM
8000uF of capacitance  will surely give you almost zero 50/60 Hz ripple, but the reason for having the soft recover diodes or the Schokky diodes is to have as little high req noise coming through from the rectifier diodes to the amplifier. If you have seen photos/images of the hf noise from a diode you will understand what I mean.

One way to get rid of the hf noise is to use smaller caps in paralell with the 8000uF caps you are using. I sometimes use 3 different polyprop caps in paralell; each cap 1/100th of the previous larger one.

Another way to minimise hf noise is to use soft recovery diodes or Schottky diodes.



Sigurd

Quote from: Keljian
Quote from: Sigurd RuschkowskiInteresting thread, Jazper!

"As for rectifier diodes, diyaudio people seem to like the MUR820, which is overkill but cheap enough to use, so I plan to use a reasonable clone of it."


I would advice against the clones of the MUR820 as I have read that ONSemi's have the best ones. As a matter of fact I would use the soft recovery ones, MSR, instead.


I would definitlively go for a dual mono power supply! and also for each mono channel a dual rectifier bridge setup. For lowest noise and for best stereo separation.


As for wires, do try solid silver wire (max 0.5mm in diam) - for everything.



Sigurd


The M3 has a rail splitter chip in the design, it's designed to have single supply.

As for the MUR820, I'm not too fussed to be frank, it's going to have 8000uF of capacitance after it, if that doesn't smooth the wave nothing will. The models I'm using are of high enough quality, and it's difficult to get ONsemi stuff in australia (I imagine it's easy in europe). I've just made a bulk order from mouser, so unless you're feeling charitable and feel like sending me some really nice rectifier diodes... :)

I'm sticking to silver coated copper wire.. I think the difference will be marginal compared to pure silver
Title: JSR03 - information
Post by: Sigurd Ruschkowski on June 15, 2005, 03:33:55 PM
It is like eating: one wants to eat what one likes.

P-A, I agree with you totally: it has to feel right with the components and with the design! The feeling is what it is all about.



Sigurd


Quote from: perandersIt must feel good. That's it! Not more, not less. It's important to like the results and chosen parts.
Title: JSR03 - information
Post by: Keljian on June 16, 2005, 03:54:35 AM
Well the new diodes almost certainly should fix the HF noise if there is any.

I think I understand what you mean, and that is why I've chosen the parts I have for this, and the main motivation behind using your design P-A, because in doing so, I know I'm building a high quality setup.
Title: JSR03 - information
Post by: Keljian on June 18, 2005, 05:28:33 AM
just a small update, I'm planning on a case similar to this one :

http://www.briangt.com/gallery/nigc-srmcgee

building it from scratch using carbon fiber, blackwood, aluminum and a few other things.. should come out really nice when it's done
Title: JSR03 - information
Post by: Sigurd Ruschkowski on June 20, 2005, 08:48:55 AM
Looks good on the photos.

Carbon fibre is really looking great! I am using it on my modell cars and boats.  My next preamp will also be in hardwood of some kind. Probably with an inner box of Al or copper plate/foil for shielding.
As a matter of fact, carbon fibre is conductive :-)


Sigurd

Quote from: Keljianjust a small update, I'm planning on a case similar to this one :

http://www.briangt.com/gallery/nigc-srmcgee

building it from scratch using carbon fiber, blackwood, aluminum and a few other things.. should come out really nice when it's done
Title: JSR03 - information
Post by: Keljian on June 20, 2005, 12:00:39 PM
well there have been a few setbacks, one that I need some browndog adapters for the amplifier, and two that I forgot to order some caps. The JSR03 board came(thanks PA), as did the bulk of the parts, but some are still due. it looks like this will be going by the end of next week as opposed to this week.
Title: JSR03 - information
Post by: Keljian on June 21, 2005, 12:05:06 AM
well, the JSR03 is being a real pain, the prop caps don't  fit - so there's much lead bending (I'm going to have to use polyester for a few) all the other components are really cramped, and the RN60D resistors are just too big (trying to tombstone and solder them on both sides of the board, but struggling. here I am wishing I had bought the SMD version and parts for it.

I'll persist and see if I can get it working
Title: JSR03 - information
Post by: peranders on June 21, 2005, 07:03:52 AM
I think you are working backwards a bit. Check first what the part should look like, then deside what to buy. The design goal was to make a small board for "normal" parts, at least at normal size.

You can use the solder side also if you spacers high enough.
Title: JSR03 - information
Post by: Keljian on June 23, 2005, 08:46:49 AM
(http://i6.photobucket.com/albums/y219/Keljian/IMG_0538.jpg)

(http://i6.photobucket.com/albums/y219/Keljian/IMG_0540.jpg) (note I haven't cleaned the flux yet, hence the messyness


it's working, however the output voltage is wrong - 17.95v (stable) (I'm feeding it 29v for test purposes)

I need above 28v.

R7 = 510R
R8=  2.15k
R9= 680R

(mistake in typing, not in the circuit)


parts substitution:

T4 = d44h11 output transistor
T2,T5= BC557
IC2 = LM329
IC4= AD797

Note one of the pads lifted during desoldering hence the abnormal lead bending of c13

help please?
Title: JSR03 - information
Post by: peranders on June 23, 2005, 11:01:22 AM
Looks good but you should use D44H11 OR BD139, not both! This is an option for TO126 or TO220 transistors. The same thing applies for the other transistors, TO92 or SOT23 types.

Notice also C9, should not be mounted.

I adwise you read the circuit description so you will see what you should mount and what you should omit. There are some options in the design which you must be aware of.

If you won't use the SENSE function you must put some tin on J1 and J2.

R9 and R8 isn't right. The gain forumla is Vout = Vref*((R8/R9)+1)

I see that you have chosen AD797. I have not yet tested it with this opamp but from what I know the regulator gets a bit more sensitive against capacitive load. AD825 is much better in this sense. Still if you have at least 50 mm wire to the load it will probably work alright.

One way to check if the regulator work properly is to measure the voltage  at the inputs of the opamp. They should have EXACTLY the same voltage. One other way to find out is to load the regulator. The voltage drop should be less then 1 mV at the lowest connection and some more at the board edge where you have some millivolts.
Title: JSR03 - information
Post by: Keljian on June 23, 2005, 11:27:15 AM
Quote from: perandersLooks good but you should use D44H11 OR BD139, not both! This is an option for TO126 or TO220 transistors. The same thing applies for the other transistors, TO92 or SOT23 types.

Notice also C9, should not be mounted.

I adwise you read the circuit description so you will see what you should mount and what you should omit. There are some options in the design which you must be aware of.

If you won't use the SENSE function you must put some tin on J1 and J2.

ok

removed BD139 and  c9, jumpered R10, tinned J1 and J2, removed R11

voltage out is now 8.27v

still much too low
Title: JSR03 - information
Post by: peranders on June 23, 2005, 11:39:34 AM
If you measure the voltages at the opamp inputs and they are the same, you must change the gain. Have you put the used resistor values into the formula? Please note which resistor is R8 and which is R9.

If you have more resistance on R9 than R8 the gain will be less than 2.

R8 = 2.15k

R9 = 680R
Title: JSR03 - information
Post by: Keljian on June 23, 2005, 11:40:35 AM
Quote from: peranders

R9 and R8 isn't right. The gain forumla is Vout = Vref*((R8/R9)+1)

I see that you have chosen AD797. I have not yet tested it with this opamp but from what I know the regulator gets a bit more sensitive against capacitive load. AD825 is much better in this sense. Still if you have at least 50 mm wire to the load it will probably work alright.


right

Vref = 6.9 (due to the use of LM329

so
6.9 * ((2150/680) +1) = 28.68
Title: JSR03 - information
Post by: Keljian on June 23, 2005, 11:44:02 AM
R8 and R9 are correct ( Just measured)
Title: JSR03 - information
Post by: peranders on June 23, 2005, 11:50:02 AM
Do you also have 6.9 volts across R9? If not the regulator isn't working properly. Remember also 33 volts in.
Title: JSR03 - information
Post by: Keljian on June 23, 2005, 12:00:30 PM
Quote from: perandersDo you also have 6.9 volts across R9? If not the regulator isn't working properly. Remember also 33 volts in.

I have 1.69v across R9 (with 32.2 v in)
Title: JSR03 - information
Post by: peranders on June 23, 2005, 12:28:00 PM
How about pin 7 of the opamp? Pin 6?

Does the LED shine?

Do you have 6.8 volts across DZ1?

Do you have 0.7 volts between base-emitter of all transistors?

Remember that the pcb is 100% error free so check very carefully that each part is the right one and they are turned right also.
Title: JSR03 - information
Post by: Keljian on June 23, 2005, 01:34:33 PM
led lights up

6.74v across DZ1

0.31v on one of the opamp pins, 9.7v on the other
Title: JSR03 - information
Post by: Keljian on June 23, 2005, 01:47:33 PM
all the parts are in the right way around also
Title: JSR03 - information
Post by: peranders on June 23, 2005, 04:05:54 PM
If it's possible, can you take one or more close up pictures so I can see the parts you have used? (Max 1 MB please.

You should have the output voltage at pin 7. Have you really checked the schematic and also understood it? You have two ways of feeding the opamp. R12 should not be used, only R11.
Check the picture very carefully:
http://home5.swipnet.se/~w-50674/hifi_pics/hifi_100pr/jsr03r0_topview.jpg

Is the opamp turned right? The notch should be a the left. The small ring at "IC3" is pin 1.
Title: JSR03 - information
Post by: Keljian on June 24, 2005, 02:34:50 AM
Quote from: perandersIf it's possible, can you take one or more close up pictures so I can see the parts you have used? (Max 1 MB please.


(http://i6.photobucket.com/albums/y219/Keljian/JSR03.jpg)

that's the best I can do with my digital camera (I tried about 20x to get it that close)

pin 1 on the opamp is near the little circle near IC4

the tombstoned resistor is a 56k (RN60D)
Title: JSR03 - information
Post by: peranders on June 24, 2005, 07:39:30 AM
R10? Should be a wire = 0 ohms. Check my picture. I have a wire jumper, that beige thing.
Title: JSR03 - information
Post by: Keljian on June 24, 2005, 09:15:47 AM
Quote from: perandersR10? Should be a wire = 0 ohms. Check my picture. I have a wire jumper, that beige thing.

R10 is jumpered on the other side of the board
Title: JSR03 - information
Post by: Keljian on June 24, 2005, 09:24:14 AM
I'm thinking I may have fried the 557 transistors, by firing up the board completely populated, I'm going to change them out for BC560Cs and see how I go
Title: JSR03 - information
Post by: peranders on June 24, 2005, 09:32:24 AM
Don't change the transistors until you have measured that are broken, I'll doubt that. Do you have 0.6-0.7 volts between the base and the emitter?

Do you have connection between the feedback R8, R9 and the inverting input of the opamp? R10 = 0 Ohms
Title: JSR03 - information
Post by: Keljian on June 24, 2005, 09:41:36 AM
Quote from: perandersDon't change the transistors until you have measured that are broken, I'll doubt that. Do you have 0.6-0.7 volts between the base and the emitter?

yes


Quote
Do you have connection between the feedback R8, R9 and the inverting input of the opamp? R10 = 0 Ohms

yes.. all of the above is correct.
Title: JSR03 - information
Post by: peranders on June 24, 2005, 09:50:50 AM
If you have 0.7 volts that's good.

Do you have the same voltage at pin 2 as over R9? Meening do you really have a wire in R10? On the solderside?
Title: JSR03 - information
Post by: Keljian on June 24, 2005, 09:55:15 AM
Quote from: perandersIf you have 0.7 volts that's good.

Do you have the same voltage at pin 2 as over R9? Meening do you really have a wire in R10? On the solderside?

yes, I can post a picture if you like, but I assure you I've bridged it on the solder side.

I checked it with my multimeter, 0 resistance over R10.
Title: JSR03 - information
Post by: peranders on June 24, 2005, 10:01:36 AM
AD797 may not like too much differential voltage. I don't recommend it as you may have noticed, still you can fix this by applying two diodes across the inputs.

Would you please measure:

Pin 2, 3, 4, 6, 7

Vout

Vin

Voltage across T4
Title: JSR03 - information
Post by: Keljian on June 24, 2005, 10:09:05 AM
Quote from: perandersAD797 may not like too much differntial voltage. I don't recommend it as you may have noticed, still you can fix this by applying to diodes across the inputs.

Would you please measure:

Pin 2, 3, 4, 6, 7

Vout

Vin

Voltage across T4

2-6 = 6.45v

2-7 = .302v

3-5 = 1.1v

3-6 = 7.16

4-7 = 1.3v

Vin = 33.23v

Vout = 8.27

T4 = 2.57, and .574
Title: JSR03 - information
Post by: peranders on June 24, 2005, 10:15:24 AM
Could you please measure the opamp voltage with ground as reference.

Are you sure you have solder all IC pins with success?

Aha, 431, startup problems, I may know what your problem is. The problem may be the feedback of the LM431. I'll come back soon with a solution. The resistor values must be chosen with taste. The problem is when the 431 is inactive, then the ref voltage must be LESS than the feedback from the output. If not you will get a stable hang-up condition.
Title: JSR03 - information
Post by: Keljian on June 24, 2005, 11:35:56 AM
Quote from: perandersCould you please measure the opamp voltage with ground as reference.

Are you sure you have solder all IC pins with success?

Aha, 431, startup problems, I may know what your problem is. The problem may be the feedback of the LM431. I'll come back soon with a solution. The resistor values must be chosen with taste. The problem is when the 431 is inactive, then the ref voltage must be LESS than the feedback from the output. If not you will get a stable hang-up condition.

well, I'm using an LM329 not a 431.
Title: JSR03 - information
Post by: peranders on June 24, 2005, 12:11:43 PM
Ok, in that to rule out latch-up, set the gain to 2 or around. Then you would get 13.5-14.0 volts. If you get those voltages you may have a problem with the opamp and the common mode limits. If so, try to solder in two diodes across the inputs.

Will you measure the opamp voltages with respect to ground.
Title: JSR03 - information
Post by: Keljian on June 24, 2005, 12:40:07 PM
Quote from: perandersOk, in that to rule out latch-up, set the gain to 2 or around. Then you would get 13.5-14.0 volts. If you get those voltages you may have a problem with the opamp and the common mode limits. If so, try to solder in two diodes across the inputs.

Will you measure the opamp voltages with respect to ground.

so, you're saying change R8, R9 and R7 to set gain 2, while trying to keep R7 around 500 ohms, while trying to keep R8 and R9 in parallel = R7.

(which involves working out values for all of them and desoldering/resoldering (just making sure I hear you right)

measuring opamp voltages now..
Title: JSR03 - information
Post by: Keljian on June 24, 2005, 12:44:12 PM
Pin 2 = 1.7 v
Pin 3 = 0.988v
pin 5 = 2.1v
Pin6 = 8.16v
pin7=  1.407v

the opamp with respect to ground
Title: JSR03 - information
Post by: peranders on June 24, 2005, 01:19:04 PM
Pin 7 should have more or less the output voltage, check the schematic. Are you REALLY sure that you have 47 ohms or less? The AD797 takes 8-10 mA which mean that you will loose 300-500 mV.

If you want to test different gains, let R8 be and connect resistors in parallel with R8.
Title: JSR03 - information
Post by: Keljian on June 24, 2005, 01:23:32 PM
Quote from: perandersPin 7 should have more or less the output voltage, check the schematic. Are you REALLY sure that you have 47 ohms or less? The AD797 takes 8-10 mA which mean that you will loose 300-500 mV.

If you want to test different gains, let R8 be and connect resistors in parallel with R8.

R11 is 46.7 ohms exactly (just measured)
Title: JSR03 - information
Post by: Keljian on June 24, 2005, 01:30:02 PM
I mixed up the pins, I'm tired.. sorry

pin 8= 2.1
pin 7 = 8.16
pin 6 = 1.407
Title: JSR03 - information
Post by: Keljian on June 24, 2005, 01:33:55 PM
Quote from: Keljian
Quote from: perandersAD797 may not like too much differntial voltage. I don't recommend it as you may have noticed, still you can fix this by applying to diodes across the inputs.

Would you please measure:

Pin 2, 3, 4, 6, 7

Vout

Vin

Voltage across T4

2-6 = 6.45v

2-7 = .302v

3-5 = 1.1v

3-6 = 7.16

4-7 = 1.3v

Vin = 33.23v

Vout = 8.27

T4 = 2.57, and .574

will also mean I mixed up these, here are the fixed versions - sorry I had an exam today


2-7 = 6.45v

2-6 = .302v

3-8 = 1.1v

3-7 = 7.16

4-6 = 1.3v
Title: JSR03 - information
Post by: peranders on June 24, 2005, 01:35:18 PM
Could you please repeat:

Pin 2

Pin 3

Pin 6

Pin 7

Output

At LM329, cathode

Have you notice how LM329 is, NOT like the LM431.
http://home5.swipnet.se/~w-50674/hifi_pics/hifi_100pr/jsr03r0_lm329.jpg
Title: JSR03 - information
Post by: Keljian on June 24, 2005, 01:37:00 PM
Quote from: perandersCould you please repeat:

Pin 2

Pin 3

Pin 6

Pin 7

Output

At LM329, cathode

pin 2-3-6-7 relative to ground? or relative to each other?

working on it now
Title: JSR03 - information
Post by: Keljian on June 24, 2005, 01:43:12 PM
Vout = 8.29

pin 2 = 1.7v to ground

pin 3= 0.988v to ground

pin 7 = 8.16v to ground

pin 6= 1.4v to ground

anode of LM329 = 0.239 to ground

cathode of LM329= 0 to ground
Title: JSR03 - information
Post by: peranders on June 24, 2005, 01:43:23 PM
Check my picture. Notice also that you should not have R4-R6.

Of this you can see that if you want to replace parts you must also check very carefully in datasheets so you are sure that you know what you. I'll hope this was the problem, fingers crossed!

I will add LM329 as an option to LM431 so this has to be documented later on.
Title: JSR03 - information
Post by: Keljian on June 24, 2005, 01:49:25 PM
Quote from: perandersCheck my picture. Notice also that you should not have R4-R6.

.

so what should I do with r4 and R6?

remove them completely? replace with jumpers?
Title: JSR03 - information
Post by: Keljian on June 24, 2005, 02:14:23 PM
Quote from: Keljian
Quote from: perandersCheck my picture. Notice also that you should not have R4-R6.

.

so what should I do with r4 and R6?

remove them completely? replace with jumpers?

after just removing R4 and r6 there is no change

I'm guessing R5 has to be removed also?
Title: JSR03 - information
Post by: peranders on June 24, 2005, 02:36:40 PM
Have you checked the picture above and turned the LM329?

R4-R6 removed.
Title: JSR03 - information
Post by: Keljian on June 24, 2005, 03:17:41 PM
Quote from: perandersHave you checked the picture above and turned the LM329?

R4-R6 removed.

29.82 v after all that

on Vout
Title: JSR03 - information
Post by: Keljian on June 24, 2005, 03:28:54 PM
26.4 under load (m3)

input is 33.3v

target is 28.5v (approx)

26 is a touch low..
but it'll suffice so long as it's low noise

LED goes off under load
Title: JSR03 - information
Post by: peranders on June 24, 2005, 08:44:25 PM
Good that you succeeded after all.

From zero load to max the voltage should not drop more than a few millivolts.

Have you checked that you min voltage isn't under 34 volts. You must also include the ripple. I think the voltmeter should read 36-37 volt in your case, which is a bit much if you have a 35 V cap at the input.

The LED should be lit all the time, if not the power stage hasn't got enough base current. How much current are you taking out?

If the output voltage isn't 100% stable you won't get low noise either.
Title: JSR03 - information
Post by: Keljian on June 25, 2005, 01:46:49 AM
Quote from: perandersGood that you succeeded after all.

From zero load to max the voltage should not drop more than a few millivolts.

Have you checked that you min voltage isn't under 34 volts. You must also include the ripple. I think the voltmeter should read 36-37 volt in your case, which is a bit much if you have a 35 V cap at the input.

The LED should be lit all the time, if not the power stage hasn't got enough base current. How much current are you taking out?

If the output voltage isn't 100% stable you won't get low noise either.

C1 is 50v

voltage in is 33.5 approx

I'll check current soon
Title: JSR03 - information
Post by: Keljian on June 25, 2005, 05:16:48 AM
Quote from: Keljian
Quote from: perandersGood that you succeeded after all.

From zero load to max the voltage should not drop more than a few millivolts.

Have you checked that you min voltage isn't under 34 volts. You must also include the ripple. I think the voltmeter should read 36-37 volt in your case, which is a bit much if you have a 35 V cap at the input.

The LED should be lit all the time, if not the power stage hasn't got enough base current. How much current are you taking out?

If the output voltage isn't 100% stable you won't get low noise either.

C1 is 50v

voltage in is 33.5 approx

I'll check current soon

Well I tried with a 30v AC (500mA) transformer, the same thing happened.

I'm only driving an m3.. which (while the quinescent current draw has been set to 130mA on each fet) should be drawing less than 500mA

ok the M3 is drawing (measured) exactly 309mA

So it isn't the transformer. (which is capable of 24v AC/0.87A)

Also the transformer worked with the old power supply design wihout voltage sag.
Title: JSR03 - information
Post by: peranders on June 25, 2005, 07:35:45 AM
Do you have a tin blob at the SENSE pads at the solder side?

If you meassure in the sense hole, X6. Stable voltage?

Just wondering but what does Morsel and the rest of the design team say about recommend power supply? Do you follow their instructions?

Personally I think 29 volt stabilized with 24 VAC is too little. A rule of thumb is to have the same AC voltage as DC so in your case 26-28 V AC for 29 V DC.

If you want 300-350 mA out you must have a transformer with 500-700 mA AC. I would have chosen 28 VAC, 1 A.

How much smoothing do you have? Have much ripple voltage do you have at 300 mA out?
Title: JSR03 - information
Post by: Keljian on June 25, 2005, 07:51:36 AM
Quote from: perandersDo you have a tin blob at the SENSE pads at the solder side?

If you meassure in the sense hole, X6. Stable voltage?

Just wondering but what does Morsel and the rest of the design team say about recommend power supply? Do you follow their instructions?

Personally I think 29 volt stabilized with 24 VAC is too little. A rule of thumb is to have the same AC voltage as DC so in your case 26-28 V AC for 29 V DC.

If you want 300-350 mA out you must have a transformer with 500-700 mA AC. I would have chosen 28 VAC, 1 A.

How much smoothing do you have? Have much ripple voltage do you have at 300 mA out?

Sense pads are tinned

How do I measure ripple?

as for their instructions, they're based on the opamp voltages but 24-29v/300ma is ok

8000uF worth of smoothing caps.. should be plenty..

Surely the JSR couldn't be drawing 300mA on its own..

if it worked with the other power supply (which was a snubberised steps clone)
Title: JSR03 - information
Post by: Keljian on June 25, 2005, 08:01:12 AM
the transformer is rated to .87A

I'm only trying to draw 300ma..
Title: JSR03 - information
Post by: peranders on June 25, 2005, 08:33:11 AM
The tricky part is that the current in theory will get 1.4 times more on the AC side plus in your case 8000 uF is considered as rather heavy smoothing and this increases losses in the transformer.

A thumb of rule is if you have heavy smoothing you can take out 50% in watts of the VA rating.

In your case: 300 mA x 34 volts = 10.2 Watts => 20.4 VA is required

3300-4700 uF is a normal max especially if you have a super regulator. If you want to have less you must have an oscilloscope in order to judge a safe lower limit.

With 24 VAC I don't recommend more than 24-26 volts DC unless you can get an oscilloscpoe.

BTW: How hot gets the LM317 with the rather small heatsink? The heastsink shouldn't be hotter than 60-70 degrees C.
Title: JSR03 - information
Post by: Keljian on June 25, 2005, 09:15:55 AM
Quote from: perandersThe tricky part is that the current in theory will get 1.4 times more on the AC side plus in your case 8000 uF is considered as rather heavy smoothing and this increases losses in the transformer.

A thumb of rule is if you have heavy smoothing you can take out 50% in watts of the VA rating.

In your case: 300 mA x 34 volts = 10.2 Watts => 20.4 VA is required

3300-4700 uF is a normal max especially if you have a super regulator. If you want to have less you must have an oscilloscope in order to judge a safe lower limit.

With 24 VAC I don't recommend more than 24-26 volts DC unless you can get an oscilloscpoe.

BTW: How hot gets the LM317 with the rather small heatsink? The heastsink shouldn't be hotter than 60-70 degrees C.

The transformer is 20VA.

I had the same problem with only 4700uF worth of caps, the caps I have are low ESR panasonic FM

The problem is still occuring with 25.7v (changed R8 value)

25.7v unloaded

23v loaded.

(26 going into the jsr when loaded)
Title: JSR03 - information
Post by: peranders on June 25, 2005, 09:25:55 AM
Have you checked X6, voltage drop also there?

The LED is lit also?

You have 10 mA base current available for T3/T4 which means that at least 1 A is possible, that assumes current gain of 100. It may be possible that used used transistor has a but low current gain. datasheet says min 20-40 so maybe you should decrease R14 to 47 or even 27 ohms. This will create more base current.

If you have such problems you will see a difference in voltage between the inputs of the opamp. The inverting input, pin 2 should have exactly the same voltage as the LM329.
Title: JSR03 - information
Post by: Keljian on June 25, 2005, 09:28:09 AM
Quote from: perandersHave you checked X6, voltage drop also there?

The LED is lit also?

voltage drop is at x6 also

led is not lit

when I switch it on the led goes on for a second, then goes off. If I unplug the load while running it lights up again
Title: JSR03 - information
Post by: Keljian on June 25, 2005, 09:35:36 AM
Quote from: Keljianthe transformer is rated to .87A

I'm only trying to draw 300ma..

max is 870mA, if I divide by 1.4 I get 614mA.. which is more than enough..
Title: JSR03 - information
Post by: peranders on June 25, 2005, 09:48:11 AM
If the LED is not lit, then you may increase the base current => decreasing R14.

The current is calculated like this:

(Ledvoltage-0.7)/R14 => (1.7-0.7)/R14
Title: JSR03 - information
Post by: Keljian on June 25, 2005, 10:07:38 AM
Quote from: perandersIf the LED is not lit, then you may increase the base current => decreasing R14.

The current is calculated like this:

(Ledvoltage-0.7)/R14 => (1.7-0.7)/R14

changed R14 to 27 ohms..

no change in output.

LED is  still off
Title: JSR03 - information
Post by: peranders on June 25, 2005, 10:18:34 AM
How much voltage across the LED?

How much voltage across R14?

Do you have 1 volt over R14 when you have a small load?
Title: JSR03 - information
Post by: Keljian on June 25, 2005, 10:25:18 AM
Quote from: perandersHow much voltage across the LED?

How much voltage across R14?

Do you have 1 volt over R14 when you have a small load?


0 load

LED = 1.8v

R14 = 1.07v

300mA load

R14 = 0.1v

LED = 0.8v
Title: JSR03 - information
Post by: peranders on June 25, 2005, 10:48:37 AM
R13 may be decreased, down to 2k2 is OK but at least so you'll see some light from the LED.

I'm off now, see you next week.
Title: JSR03 - information
Post by: Keljian on June 25, 2005, 10:53:24 AM
Quote from: perandersR13 may be decreased, down to 2k2 is OK but at least so you'll see some light from the LED.

I'm off now, see you next week.

ok see you
Title: JSR03 - information
Post by: Keljian on June 25, 2005, 03:13:52 PM
I think you were right P-A..

I'll get a much more powerful transformer (160VA ) in the coming week...
Title: JSR03 - information
Post by: Keljian on June 27, 2005, 12:03:23 PM
well after changing the transformer, it works a charm...

thanks for all your help PA.. oh .. I did some testing and you really need to feed it about 8-9v more than you want out of it.
Title: JSR03 - information
Post by: Keljian on June 28, 2005, 03:09:53 PM
Quote from: Keljianwell after changing the transformer, it works a charm...

thanks for all your help PA.. oh .. I did some testing and you really need to feed it about 8-9v more than you want out of it.

famous last words..

it's not working so well anymore, must have tapped it by mistake or something, Voltage across R9 is 2.3v, voltage across DZ1 is 6.7v, Vout is 8.5v

I've changed the regulator for a new one, and changed T5, not that it's made any difference.. I'm guessing there's a problem with the feedback loop but I don't know what the problem is.

voltages across transistors = 0.7v

resistors seem in spec

Followed the schematic, the problem is the reference, was outputting 22v
Title: JSR03 - information
Post by: peranders on June 30, 2005, 09:46:10 PM
Did you have enough cooling?

How are you getting 22 volts from the reference if you feed it from 8.5 volts?

What did you do to get the regulator broken? Remember that it has limited short circuit protection. It's not totally fool proof.
Title: JSR03 - information
Post by: Keljian on July 01, 2005, 01:43:17 AM
welcome back

I've always fed the regulator with the same 33.5v or so

I killed it by shorting the back of the regulator (the lm317) to output ground. it is outputting 8.54v, r9 is 2.30 v or so.. I intend to change the LM329 today as the voltage across it relative to ground is 22v or so

using a thermal probe I measured the temperature when it was running properly the d44h11 was running at 70C and the LM317(with heatsink) was running at 60C (ambient approx 25C )
Title: JSR03 - information
Post by: Keljian on July 01, 2005, 02:24:58 AM
Quote from: Keljianwelcome back

I've always fed the regulator with the same 33.5v or so

I killed it by shorting the back of the regulator (the lm317) to output ground. it is outputting 8.54v, r9 is 2.30 v or so.. I intend to change the LM329 today as the voltage across it relative to ground is 22v or so

using a thermal probe I measured the temperature when it was running properly the d44h11 was running at 70C and the LM317(with heatsink) was running at 60C (ambient approx 25C )

well it's not LM329.. I just changed it.
the voltage is still 8.54 coming out of the JSR

I'll trace the circuit again tonight..
Title: JSR03 - information
Post by: peranders on July 01, 2005, 08:35:04 AM
As a permanent solution I think 70 deg C is a bit too much. Notice that the lifetime of the electrolythic caps will be shortened.
Title: JSR03 - information
Post by: Keljian on July 01, 2005, 11:48:54 AM
Quote from: perandersAs a permanent solution I think 70 deg C is a bit too much. Notice that the lifetime of the electrolythic caps will be shortened.

I'll be putting a bigger heatsink on it when I get some thermal compound, once it's functional. I have one sitting here, in the meantime I need to get it working again, and I have no idea where to start.
Title: JSR03 - information
Post by: Keljian on July 01, 2005, 11:59:11 AM
9.74v across the lm317

6.91v across the diode

2.317v across r9

1.310v across r7

1.227v across r1
1.282v across R2
Title: JSR03 - information
Post by: Keljian on July 01, 2005, 12:56:12 PM
well I replaced the d44h11, 29.7v out.. output definately isn't stable


8.15v-8.55v across R9

6.75v across DZ1

0.75v across the led

7.15v going out of the LM329 (shouldn't this be 6.9?)


getting closer..
LED is off
Title: JSR03 - information
Post by: Keljian on July 01, 2005, 02:29:53 PM
replaced transistors(t2, t5) , replaced diode.. output voltage back to 8.5

I'm on the verge of giving up with this supply

readings as in previous post

9.74v across the lm317

6.91v across the diode

2.317v across r9

1.310v across r7

1.227v across r1
1.282v across R2
Title: JSR03 - information
Post by: peranders on July 01, 2005, 07:49:24 PM
Quote from: KeljianI'm on the verge of giving up with this supply

This design is pretty easy to get working I have built a number of them and tehy all worked right away.

The only thing you have to do is work carefully and don't rush.

All things you have done so far are the result of that you have been eager to get it ready.

You see have my reference design looks like, plan the changes, write them down, do some calculations if necessary and if you want some help just ask.

I'll be back in a week.
Title: JSR03 - information
Post by: Keljian on July 02, 2005, 03:17:15 AM
Quote from: peranders
Quote from: KeljianI'm on the verge of giving up with this supply

This design is pretty easy to get working I have built a number of them and tehy all worked right away.

The only thing you have to do is work carefully and don't rush.

All things you have done so far are the result of that you have been eager to get it ready.

You see have my reference design looks like, plan the changes, write them down, do some calculations if necessary and if you want some help just ask.

I'll be back in a week.

Perhaps, but I've had enough of the design, and the board is on the verge of being unusuable after having so many things soldered/desoldered from it. I've moved to a simpler design(on protoboard) for now, I may revisit it later, but for the momen I'm using what works
Title: JSR03 - information
Post by: peranders on July 02, 2005, 07:38:46 AM
I can see your disappointment. The pcb can take 3-5 times of disoldering even if you have professional desoldering tools so I can imgaine how the pcb looks like by now.

Anyway, the regualtor is a classic and the topology works very good in various applications.

Feel free to come back if you want to upgrade your "simple" regulator.
Title: JSR03 - information
Post by: Keljian on July 02, 2005, 09:22:59 AM
Quote from: perandersI can see your disappointment. The pcb can take 3-5 times of disoldering even if you have professional desoldering tools so I can imgaine how the pcb looks like by now.

Anyway, the regualtor is a classic and the topology works very good in various applications.

Feel free to come back if you want to upgrade your "simple" regulator.

I'd love to upgrade it, but if I were to upgrade it I would go for boutique parts all the way through and your design doesn't support that

I have been looking at using another circuit from diyaudio to clean up the sound, but to be honest, I can't see the benefit, because I can't tell the difference between the "simple" regulator, and the more advanced jung.


To be completely honest, the simpler solution sounds BETTER than the jung.. The bass slam is much better, and the amp seems faster
http://www6.head-fi.org/forums/showthread.php?p=1469609#post1469609
(if you want details)

as for the state of the board, don't ask, there are solderpads that have come off, there was one track lifted, and misc other problems, it really is not in a good state.

BTW, thankyou for your tireless support, really you've been fantastic with all the help you gave me, and advice in general :)
Title: JSR03 - information
Post by: peranders on September 13, 2005, 12:56:45 PM
Quote from: KeljianTo be completely honest, the simpler solution sounds BETTER than the jung.. The bass slam is much better, and the amp seems faster
http://www6.head-fi.org/forums/showthread.php?p=1469609#post1469609
(if you want details)
Since you eventually got the Super regulator working, do you still think that LM317 is better? The answer doesn't have to be "no" just to please me.
Title: JSR03 - information
Post by: Keljian on September 13, 2005, 03:03:24 PM
Quote from: peranders
Quote from: KeljianTo be completely honest, the simpler solution sounds BETTER than the jung.. The bass slam is much better, and the amp seems faster
http://www6.head-fi.org/forums/showthread.php?p=1469609#post1469609
(if you want details)
Since you eventually got the Super regulator working, do you still think that LM317 is better? The answer doesn't have to be "no" just to please me.

Better is a relative term. I think with the 317 there was more bass slam, but on the other hand I think the sound is cleaner with the jsr. I've put some more caps on the output of the jsr and that seems to have alleviated the problem a bit, there is still a "hiss" from the headphones at high volume
Title: JSR03 - information
Post by: peranders on September 13, 2005, 09:12:33 PM
Do you have hiss even with a unconnected amp or with particular sources? Do you use a PC and a soundcard?

I think you should ask other M3 builders if they also have hiss so you can leave out that it is the M3 design. You may have oscillations. Have you checked with an oscilloscope how this hiss look like?
Title: JSR03 - information
Post by: Keljian on September 21, 2005, 02:01:50 AM
Quote from: perandersDo you have hiss even with a unconnected amp or with particular sources? Do you use a PC and a soundcard?

I think you should ask other M3 builders if they also have hiss so you can leave out that it is the M3 design. You may have oscillations. Have you checked with an oscilloscope how this hiss look like?

Turns out the hiss is due to the dac I have connected atthe moment (tried another source) Will work on changing the power supply to the dac in the near future
Title: JSR03 - information
Post by: peranders on September 21, 2005, 06:55:16 AM
Is it a commercial DAC or have you built it yourself?
Title: JSR03 - information
Post by: Keljian on September 21, 2005, 06:58:05 AM
Quote from: perandersIs it a commercial DAC or have you built it yourself?

it's a guzzler dac as per here: http://www.mellowparenting.demon.co.uk/guzzler/usbdac.html

At the moment it uses a maxim switching DC-DC converter, has 3000uF worth of capacitance on the power rails and is powered by the usb port on my computer.

I have a LM7806 kicking around and a few transformers around, so I am thinking of creating a simple power supply for it
Title: JSR03 - information
Post by: peranders on September 21, 2005, 10:45:43 AM
From what I can see the designer has no (or little) control over the emi emissions. What about the emission over 100 MHz?

This is the hard thing about designing DAC's, make them "pure". Put a radio near the DAC and see or hear what you'll get in garbage.
Title: JSR03 - information
Post by: Keljian on September 21, 2005, 10:49:55 AM
Quote from: perandersFrom what I can see the designer has no (or little) control over the emi emissions. What about the emission over 100 MHz?

This is the hard thing about designing DAC's, make them "pure". Put a radio near the DAC and see or hear what you'll get in garbadge.

No garbage - it could be due to the fact that the aluminum case is grounded.

I think the problem is the power.. will experiment in the coming weeks.
Title: JSR03 - information
Post by: peranders on September 21, 2005, 11:07:43 AM
If you don't have tested with a FM or AM radio do that! This is a very cheap spectrum analyzer. Turn on and off your dac then you'll see what it transmits. Also using a DC-DC converter together with high fidelity ciruits need extra attention. The converter transmits both in wires and in air.
Title: JSR03 - information
Post by: Keljian on September 22, 2005, 03:54:33 AM
Quote from: perandersIf you don't have tested with a FM or AM radio do that! This is a very cheap spectrum analyzer. Turn on and off your dac then you'll see what it transmits. Also using a DC-DC converter together with high fidelity ciruits need extra attention. The converter transmits both in wires and in air.

you're right...  I get a bit of static on the radio when I put it right on top of the dac
Title: JSR03 - information
Post by: peranders on October 29, 2005, 03:46:35 PM
Quote from: Keljian
Better is a relative term. I think with the 317 there was more bass slam, but on the other hand I think the sound is cleaner with the jsr. I've put some more caps on the output of the jsr and that seems to have alleviated the problem a bit, there is still a "hiss" from the headphones at high volume
How about the slam and cleanness? Is the LM317 better or do you prefer the Super Regulator? Have you done anything to reduce the EMI problem of your DAC?
Title: JSR03 - information
Post by: Keljian on October 29, 2005, 11:30:16 PM
the sound is a bit "thinner" overall

I haven't had time lately to work on the dac - I have my final exams for the year coming up in less than a weeks time and that is my primary focus
Title: JSR03 - information
Post by: peranders on October 30, 2005, 08:10:28 AM
I'll guess you mean thin sound with the super regulator? You are sure that all parts are working properly? It might be that you'll have a clearer sound which you aren't used to.

The regulator works still I'll believe? Have you modified it anything?
Title: JSR03 - information
Post by: Keljian on October 30, 2005, 08:18:05 AM
the regulator is still working

The only thing I've changed in the circuit is the opamp used, I am using a socketed (on a browndog adapter) Ti THS4601 opamp on the regulator.

I find that with this opamp the overall background noise of the amplifier is less than the original AD825 (which is strange).

Overall the regulator is working fine, the sound out of the amp is a bit "thinner" than I am used to. This can be obviously fixed by changing opamps or tweaking the bass boost. So far I like the ths4631D chips in the M3 the most with the SA5000 headphones.

In a few weeks I'll be ordering another board from you, I'll give the design another attempt.

I am concerned about the heatsinks, they do get very hot, but are still within acceptable limits. When I build the next one I think I will drill up another heatsink that will be much larger. May I suggest for the next version that you consider better spacing so one can use some bigger heatsinks for the regulator and power transistor? (T4)  - an example of one is the Aavid-Thermalloy 531102B00000 as used on the mosfets on the M3.
Title: JSR03 - information
Post by: peranders on October 30, 2005, 09:50:33 AM
With AD825 you'll get approx. 5 uV noise and with LM317 you'll get a lot more. If you really hear different noise levels, try to switch of you DAC. The DAC may interfere.

When it comes to heat, the LM317 get hottest so there is no need to separate the transistor from the regulator.
Title: JSR03 - information
Post by: Keljian on October 30, 2005, 09:56:34 AM
Quote from: perandersWith AD825 you'll get approx. 5 uV noise and with LM317 you'll get a lot more. If you really hear different noise levels, try to switch of you DAC. The DAC may interfere.

When it comes to heat, the LM317 get hottest so there is no need to separate the transistor from the regulator.

I agree.

However the noise issue is there when I use other sources also(I was using another source when I detected it) - so it leads me to believe that it is inherent of the opamp.
Title: JSR03 - information
Post by: peranders on October 30, 2005, 11:10:23 AM
Can you actually hear noise coming from the headphone amp when the volume is turned down?
Title: JSR03 - information
Post by: Keljian on October 30, 2005, 01:27:40 PM
Quote from: perandersCan you actually hear noise coming from the headphone amp when the volume is turned down?

Yes, this particular noise I could hear.

Remember I have particularly resolving headphones (sony MDR-SA5000)
Title: JSR03 - information
Post by: peranders on October 30, 2005, 03:28:09 PM
The headphone amp of yours has around 117-125 dB in S/N. If noise really bothers you the level in your ears must be 20-30 dB at least. This means that something isn't working properly.

If you have the input of the headphone amp short circuited, do you get any noise then? Is the noise changing if the signal sources are turned on and off?
Title: JSR03 - information
Post by: Keljian on October 30, 2005, 11:28:41 PM
problem solved, was a bad connection at one of the inputs