View Full Version : PSU for Caiman with Pics
Dave Hewitt
20-04-2010, 08:43
Martin,I think you run the possibility of having a shorted turn on the tranny in your psu with the earth cable connected to the mounting bolt,just connect to case,dont make a continuous circuit thro the centre of the tranny.
Regards Dave.
Hi Dave
That's a good point. I shall re-route the earthing locally. Look out for a new posting as 'PSU for Caiman' when I rework it tonight.
Martin - do you know what is the optimum input voltage for the Caiman? I see that it states 15v to 18v on the back panel.
I have been thinking about just that today. The newer Caiman PSU puts out 15.5V and I was going to try up to 18V but before I do that I shall have an internal inspection and ensure that there are no 16V (a common value) electrolytics on the PSU rails.
HighFidelityGuy
20-04-2010, 13:06
I have been thinking about just that today. The newer Caiman PSU puts out 15.5V and I was going to try up to 18V but before I do that I shall have an internal inspection and ensure that there are no 16V (a common value) eletrolytics on the PSU rails.
From what I remember the main reservoir cap in the Caiman and standard 7520 are 16V. There is however a diode just after the power input that causes a voltage drop of about 0.5V. There's also a resistor which drops the voltage a bit more but I'm I can't remember if that's before or after the reservoir cap. So I would have thought you'd be ok supplying the Caiman with up to 16V but I'd be very cautious above that as there's very little safety margin.
I've been wondering about swapping the diode in my Caiman for a Schottky to reduce the voltage drop. I've done this successfully in my 7520 and that seems fine but that's only getting 12V.
With the standard diode in place in my Caiman the 12v Murrata reg is getting about 14.5V on it's input which is below the recommended 15V input. So a Schottky should bring the supply closer to 15v.
14.5V is still above the dropout voltage for that reg but the spec sheet says it provides the cleanest regulation from 15v+. So it makes sense to me to try and achieve that for completeness.
So I guess there's scope for supplying a bit more voltage, swapping the diode or both. I'm not sure if either of those on there own would make an improvement or whether simply using a quality linear regulated supply would overshadow these. Who knows. I guess these are all things you could experiment with. I'd be interested to hear how you get on as these are all areas I'm interested in exploring. :)
Whats the diode in the circuit for? If you are supplying the unit with DC, I can only "assume" its for protection. In which case, its not really needed unless something goes wrong with the PSU and it outputs AC.
*** purely educated speculation as I don't have a unit to look at or the diagram
With respect to the input voltage to the reg being 14.5v and not 15v, I wouldn't worry about that at all! So long as the Murata can supply the required 12v constantly and cleanly (the cleaner the better!) all will be fine. I intend to fairly soon take a look at one of these dacs as I'm really interested to see what I can do with it. Given the rocky start that Stan and I got off to, I'm hoping that he doesn't mind me taking a look and making my own suggestions here in the open!
Yes, the diode on the DC input is mainly for protection.
None yet - all I could glean from Dave is that he is going to put a much larger energy store (i.e. reservoir capacitor) in it.
He seemed to be quick to mention it on the first page of this thread so I thought he maybe keen to give some circuit details etc
Anyway let us know how you find the STR goes with the Caiman, I never tried the TR's with the Caiman but did try them with the analogue stage for the TDA1541 based dac and Sabre dac, tbh I thought it softened dynamics and made the bass loose some tightness even after adding CRD's on the output to give it more of a constant load, it may go much better with the Caiman though, of course only way to know these things is try it:) Its reported these regs are an excellent match with Naim pre-amps
HighFidelityGuy
20-04-2010, 16:51
Yes, the diode on the DC input is mainly for protection.
That's what I thought too. My idea of swapping it for a Schottky was simply to reduce the negative effect it has on the DC voltage to a negligible amount. You could of course just replace the diode with a wire link but then you'd loose the protection. Personally I'd prefer to keep the protection to be on the safe side.
As I said before, according to the spec sheet the 12V reg will work fine at 14.5V but 15V input will make sure it's output is as clean and accurate as possible right up to it's full output current rating. So it may be worth trying to see if it makes a noticeable change to the sound. Yes it's probably a bit overkill but then if you're going to fit fancy regs you may as well try and get the best out of them I think. :)
I've redeployed my PSU project for the Caiman DAC with very promising results. Some pics first:
http://www.mtc.me.uk/images/PSU Project 3.jpg
http://www.mtc.me.uk/images/PSU Project 4.jpg
http://www.mtc.me.uk/images/PSU Project 5.jpg
This project uses the Superteddy circuit described here http://www.teddypardo.com/. It features a discrete voltage regulator (rather than the LM-317T used in the Teddy) followed by a filter gyrator circuit. In experiments I have found this circuit to feature good, but not great, voltage regulation, slow start-up, some drift but settling after about a minute, very fine control over output voltage, high current (2.5A) capability and very low noise indeed. This latter is what attracted me to the circuit in the first place.
In addition and not shown in the pic above, I added another Nichicon 2200uF Fine Gold electrolytic on the DC input to beef up the reserve charge. The transformer is a 9-0-9V AC toroid providing about 23V raw DC into the regulator. I have set the output to exactly 16V to prevent damage to the Caiman's electrolytics but provide as much voltage as possible into the Dexa 12V regulator in my modified Caiman.
For safety I have also connected the earth to the chassis rather than via the toroid mounting bolt as shown in the pic.
Feeding the Caiman, the PSU barely gets lukewarm so there is no heat dissipation issue.
Listening to the Caiman on CD source from this PSU (using my Blu-ray player as transport) has moved it up from very good indeed to touching on the performance of my Ayre SACD player. I am hearing a wealth of micro detail, very natural and clean treble, extended, deep and tight bass and excellent spatial placement. The background is deep and black. I haven't yet performed the additonal electrolytic mods to the Caiman so there is more performance yet to be had. Soundtracks from Sky HD are very impressive, especially programming on HD channels, again with that natural placement of fine detail distinguishing it from the standard 15V Caiman PSU.
I believe the low noise performance of this PSU is responsible for enhancing the DAC's sound, however the 16V feed may just be helping the Dexa regulator perform in its envelope too.
This is good!
Glad to hear it finally went well in the end with this psu.
so you get the good filtering of the STR and the low output impedance of the Dexa's
Thought I must be losing it just now, this posting seemed to 'grow'. Guess you're keeping the relevant stuff together, Leo. Thanks.
I'm still listening to it and getting ever more impressed. I'm watching '24' on Sky HD and it sounds marvellous - as if previously there was a faint harshness to the sound which has now gone.
It wasn't me Martin:scratch:
Huh? I created this as a new thread and suddenly it had a load of postings before my first one!
StanleyB
20-04-2010, 19:43
From what I remember the main reservoir cap in the Caiman and standard 7520 are 16V. There is however a diode just after the power input that causes a voltage drop of about 0.5V. There's also a resistor which drops the voltage a bit more but I'm I can't remember if that's before or after the reservoir cap.
The big resistor situated behind the 7805 regulator can be replaced with a wire link if you have the Murata regulator fitted.
One more mod I must make: to change the blue LED dropper resistor from 1k to 2k2 as it shines like a bloody searchlight at 20mA current :)
Stan: did I see you post somewhere that you had tried a linear PSU with the Caiman with favourable results, or did I dream it?
Huh? I created this as a new thread and suddenly it had a load of postings before my first one!
Yes, strange :mental: I wonder if your main post can be moved to the start
Was this the post from Stan you was looking for ;) http://theartofsound.net/forum/showthread.php?t=5893
That's the one, Leo. It seems that Stan found a big linear PSU to be beneficial with a modified Caiman. Matches my findings - I guess both the Murata and Dexa regulators respond rather well to clean noise-free volts in.
Hi guys (Martin and Leo),
Thought I must be losing it just now, this posting seemed to 'grow'. Guess you're keeping the relevant stuff together, Leo. Thanks.
It was me - I wanted the 'PSU for SL-1210 with pics' thread to stay on-topic, and so moved the Caiman-related posts here, when you started a separate thead :)
Marco.
Bit worried about exceeding 15.5v into the Caiman. The diode does reduce the voltage by 0.7v so the input cap would see around 14.8v leaving some margin of safety for a 16v rated cap. This ideally should be nearer 20% though to leave a little headroom.
If you feed 18v to the Caiman the cap will see 17.3v and I think that this would be quite a risk and you are relying on the manufacturer giving a generous factor of safety.
I suspect feeding 18v and taking the diode out could be the last straw.
Regards
Chris
I'm keeping to 16V exactly, putting about 15.4V or so across the main electrolytic which should be ok. I have now modified the Caiman with Elna Silmic electrolytics to replace many of the tantalums and smaller electrolytics, although not the big 10,000uF one as that is impossible to source in that size (it's a Beresford special).
See post here http://theartofsound.net/forum/showthread.php?p=118229#post118229
My thanks to Carson for a great post on the electrolytic mods.
Chris, if you are worried about a constant voltage on the cap being above the 16v rating for the std cap, replace it with something like a dumpy 35v 10000uF TSUP.
If not, you could use a conventional height TSUP cap in the psu and remove the 16v one in the dac. Its only the one pre the reg you need to be worried about.
Personally i'd try and fit something different in the dac itself
Covenant
21-04-2010, 18:50
Here is my Beresford home made power supply jobby working at last. It will probably get some Teddyreg upgrade work done soon.
Bet your wondering why there is a 2p piece stuck on my headphones.
Nice one Jerry. What's the regulator module in there?
StanleyB
21-04-2010, 19:15
Bet your wondering why there is a 2p piece stuck on my headphones.
Because you are short of a penny?
Covenant
21-04-2010, 19:27
Its this one:
http://kikitronic.blogspot.com/search/label/Power%20Regulator%20DC%200-24V
Puffin has had a look at it ( and fixed a problem), thinks it needs some noise supression.
The headphones are Superlux HD681's and have been modded to death. The 2p's were a suggestion for adding mass and therefore reducing vibration. Heavy aluminium rings will eventually replace the coins (I hope)
Chris, if you are worried about a constant voltage on the cap being above the 16v rating for the std cap, replace it with something like a dumpy 35v 10000uF TSUP.
If not, you could use a conventional height TSUP cap in the psu and remove the 16v one in the dac. Its only the one pre the reg you need to be worried about.
Personally i'd try and fit something different in the dac itself
Hi Ian,
Stuggling to find a higher rated cap that fits under the casing as it's a bit snug.
Been thinking about a linear multi-style power supply a bit more feeding regulated voltages straight into the Caiman. Think that 15v, 12v and 2 x 5v is needed. I can then strip out the diode and 16v cap. The Spower regs will be quite expensive though even though good.
Regards
Chris
you can get Panasonic TSUPs at 25v/ 10000uF that are 20mm high and 35mm diameter and 35v/ 10000uF 25mm high and 35mm diameter. Problem is, I can't find that physical size on Farnel or RS!
I'll see if I can source them! but i'm fairly sure that they will fit. Otherwise you could try something physically larger on its side??? Not ideal but would work if space permits
whats the height of the std cap? Can you fit a 30mm high cap in there? a 35mm would be good but I think that will be pushing it!!!
here we go again
21-04-2010, 22:43
whats the height of the std cap? Can you fit a 30mm high cap in there? a 35mm would be good but I think that will be pushing it!!!
From memory I think it's about 27mm - I looked at this quite a while ago now and managed to find a Panny TSUP that fitted OK.
A more elegant solution IMO is to remove the single squat 10,000u cap and replace it with 2 radial caps on their side - you'll notice on the silk screen that all the holes and tracks are already there. The research I did at the time gave me several more options - again from memory the smaller lytics can be up to 18mm dia and 25mm height - which means you can get a couple of 4700u/16v Panny FC's in there.
Actually if you are using a new linear psu you could probably get away with a lower total capacitance by using a pair of 3300u caps - with a 25V rating!! - I think I did actually buy a pair of suitable FC's and a couple of Ruby ZL's too - just never got them fitted :(
worth experimenting with I think but I reckon it's a better option.
Yes, the diode on the DC input is mainly for protection.
HI Leo, how are you doin! have you 'scoped the input and out put of the diode ;)
HighFidelityGuy
07-05-2010, 16:01
Thanks for the excellent write up Martin.
I've just ordered a SuperTeddyReg board to try this out for my self.
I've been thinking about the issue of the main reservoir cap being run close to it's 16V rating and I was wondering if it would be possible to simply remove it from the equation by replacing it with a higher voltage equivalent inside the new power supply itself? Is this likely to have any negative effects?
I guess doing this will put the cap on the opposite side of the input diode D1 and fuse but that could easily be resolved if it was an issue.
Thanks. :)
Hi Dave
I don't think you want to remove the big regulator from the circuit altogether, being close to the area of demand as it is. However, you could replace it with a smaller value higher voltage stubby cap and all would be well, the ST keeping it well topped up with clean DC. If anything, you want to ensure that there is plenty of reservoir on the input to the ST.
I am just playing on the built-in tolerances of a 16V cap by feeding it with 16V! No explosion yet...
HighFidelityGuy
08-05-2010, 13:09
Thanks Martin, that's a good point about keeping the cap close to the area of demand. I hadn't thought about it like that. :doh:
Ok, so like you say I could switch to a lower value higher voltage cap or as I think somebody else mentioned it would be possible to use two smaller caps. So I just need to work out which would allow the best balance.
Would using less capacitance be detrimental or wouldn't it really matter when using a big external linear supply? Also, I think I read somewhere that using two smaller caps can allow lower ESR than a single bigger cap. Is that something that could be beneficial in this application?
Many thanks. :)
I don't think a smaller value will be detrimental in this application, especially if you can use a good quality type, e.g. Elna Silmic, Black Gate, Sanyo Os-Con etc. If you can fit two smaller caps that can work too, but check that the ESR is really better than a single larger one.
HighFidelityGuy
09-05-2010, 18:21
Ok, thanks Martin. I'll see what I can find.
Another option I've thought of is laying one large cap on it's side and squeezing it in that way. This would mean extending the wires down to the board though. Do you think the extra wire length would cause any issues?
Also I've noticed that the Elna cap mod calls for a few 16V caps. Do you know if these 16V caps are at risk as well?
Apologies about all the questions. :)
Cheers.
Yes, a large reservoir lying down would be fine.
The 16V Elnas I think are all the other side of the 12V regulator, so no problem.
Just curious, how much did this little project cost all said and done? Trying to weigh by options on the exact same project. Where did you end up getting the toroid from?
how much did this little project cost all said and done?
It wasn't cheap, but that wasn't my goal. I wanted some fun (hadn't made an electronic project for a while) and was targetting lowest possible noise since this PSU was for a DAC.
SuperTeddy regulator $63 (Teddy Pardo)
Transformer, caps & other components $97 (Parts Connexion)
Case £11 (Maplin)
Total about £95.
Just running in the Superteddyreg :)
Did you get one too? What do you think?
Yes, Just the pcb and a set of j-fets, I already had most of the compents so it proved quite cheap to try.
Its certainly better than the original Teddyreg, only been on a couple of days so I'm giving it a little more time to run in, seems quite promising so far
Nice one. I'll be interested in your listening findings when you get it going with the Caiman.
Bass now seems a little warm, infact mids to lows seems to have taken on a slight syrup quality .
Could be down to the op-amp choice now though not being a good match so I'll roll in another . Think I'll stick the LM4562 back in see how it goes :)
I may also try the LM4562 again, not really decided between them and the AD826. It's down to presentation rather than one being better than the other.
For all its downsides I think LM4562 is possibly the best compromise IMO , for me anyway when using the STR
I really liked the AD826 before with the other PSU's but since the STR :scratch: Time for some fiddling I think
I can only describe the difference as the LM4562 presents large amounts of fine detail in a slightly flat stage while the AD826 presents the shape of the sound in a deeper stage. They are both fine op-amps and I would wish for a combination of both strengths.
HighFidelityGuy
22-06-2010, 22:47
My STR based PSU is still being run in and I'm also thinking about trying the AD826. I was wondering, have you both used a pair of AD826's and a pair of LM4562's, or have you tried combinations of the two?
Thanks.
I mainly just have one fitted for the line/ variable out because I have a few sources I use an external amp for headphone use
I upped the STR output to 15.6v (loaded) it was just over 14v before when loaded :doh: It sounds noticeably better now
Its worth trying both LM4562 and AD826 to see which suits you best tbh
I upped the STR output to 15.6v (loaded) it was just over 14v before when loaded
Aha! I'm running mine at exactly 16V as the dropper diode is still in circuit, bringing it down to about 15.4V at the reservoir cap. I think the STR is doing a fine job, the elimination of noise makes the Caiman initially sound a little smooth but that's to be expected. With the right op-amp the presentation is detailed, dynamic and punchy.
The lag on the STR's voltage adjustment takes some getting used to, doesn't it?
I was wondering, have you both used a pair of AD826's and a pair of LM4562's, or have you tried combinations of the two?
I swapped them in pairs but I don't use headphones so one of them has had no effect.
HighFidelityGuy
23-06-2010, 10:27
I swapped them in pairs but I don't use headphones so one of them has had no effect.
Aparently the opamps do have an effect on each other though. So perhaps you could try an LM4562 in the headphone socket and see if it makes any difference. :)
Unless I'm mistaken, the line output op-amp may affect the headphone op-amp but not the other way round? I only use (fixed) line output.
HighFidelityGuy
23-06-2010, 17:27
Unless I'm mistaken, the line output op-amp may affect the headphone op-amp but not the other way round? I only use (fixed) line output.
Ah, you're probably right. I knew there was some kind of interaction. :)
Its worth trying both LM4562 and AD826 to see which suits you best tbh
I've just gone back to the LM4562 as they are just a touch more detailed and less lush which suits my system better. Both are great op-amps.
Yes, these things are cheap to try too.
LM4562 doesn't like capactive loads, shouldn't be much of a concern in most set ups
Canetoad
22-09-2011, 09:46
I have built a PSU similar to MartinT to power my Caiman. 80va tx, hexfred rectifier, 2 x 4700 uF smoothing caps and STR. I have set the output voltage to 15.8v under load and now want to replace the large 16v 10000 uF cap in the Caiman with something which has a higher voltage limit. The choices I currently have lying around are:
33 uF 20v Oscon
2 x 2.2 uF 100v metallized polyester (as per MartinT's Caiman)
6800 uF 35v Panasonic
As a novice I haven't a clue which would be the best choice to use. I don't understand what differences the rating of a cap (uF) has on it's performance. I would appreciate some information/advice from those in the know. :)
Bernie - since the SuperTeddy is highly regulated and low noise, you don't need a large reservoir cap inside the Caiman so much as power supply decoupling. So let the ST do the regulating work and have the Caiman keep noise down to a minimum. I would therefore recommend the 2 x 2.2uF poly (or whatever you can get to fit) option rather than a large electrolytic as their high frequency response will be a lot better.
Canetoad
22-09-2011, 11:56
Cheers Martin,
What about the Oscon? Aren't they very low ESR? Isn't this a good thing in the power supply? :scratch:
I like Os-Cons and you could try it but it's still an electrolytic so I would expect the polys to be better in this position.
Canetoad
22-09-2011, 12:24
OK, I get it. :)
Regarding replacing the big smoothing cap, I did this:
http://theartofsound.net/forum/showpost.php?p=221168&postcount=42
I fitted a 22uf 20V Sanyo Os-Con (thanks Mark!) and it works fine and dandy.
No point in using a huge cap following a regulated PSU, all you want is low impedance & the Os-Con will give you that..
Powered by vBulletin® Version 4.2.3 Copyright © 2024 vBulletin Solutions, Inc. All rights reserved.