Stevenacnj said:
I agree with Chris, you need to finish your schemaic.
One big question for me is:
Does your toriod have split secondaries or a center tap?
There are some advantages of split seconds that can be realized by using two bridges.
Also, how you connect the caps to the bridge and the transformer can have an affect as done wrong you induce humm voltage on ground. You want to isolate the large currents that come from the center tap (dirty ground) from the system ground. That being after the caps ground. You want to make sure, very sure, you are connecting to the Cap ground in a way that minimizes any ripple current carrying resistance.
Draw that carefully or better yet, snap a picture and send it.
I am doing what I can tell you are doing, but don't have an earch ground
The amp chassis is an older one that didn't have that. The UcD dropped in nicely. At this point I float the power supply and let the RCA input set ground. Scary, but it works and is dead quiet.An old trick that helps with ground loops in single ended is to add a 100 ohm resistor in series with RCA ground and system ground, that may help as it minimizes the amount of current that can flow, vs a wire!
You still want to hook your -in directly to the RCA ground though.
Mike
Portlandmike said:
Hi Mike,
That old trick doesn't really apply when it comes to optimal, he's wanting best sounding quality for what he paid for.
That paper is available at their website as a free copy with a bit of clicking around. I found it hardly worth the effort. A few good points sure, but it doesnt' seem like the level of info that can solve the pin 1 problem universally.
Portlandmike said:
Here is one
www.campuspa.com/downloads/aes48-2005-f.pdf
Don't know if there's a more recent version.
classd4sure said:
Chris,
I agree, kind of, but he's not optimal. He's single ended. 😉
Point being, try it, if it solves a problem, great. Because it just identified the problem as a ground loop. If not, its in the box! perhaps very poor PS layout...
Balanced inputs really help. Its why I'm so high on the idea of moving the UcD input op amp to my preamp.
Best Regards,
Mike
Stevenacnj,
You might wanna just try a cheater plug on your amp to narrow down the source of the hum.
If it solves the problem, its a ground loop thru earth ground.
Your amp is now grounded thru the RCA ground. Kind of scary, if there is a fault, but it will likely sound great.
If this solves the hum then you may want to try what I believe Chris is getting at in that you want to remove all connections from your amps ground to chassis/earth ground.
If not, I think its likely in your PS wiring.
Regards,
Mike
You might wanna just try a cheater plug on your amp to narrow down the source of the hum.
If it solves the problem, its a ground loop thru earth ground.
Your amp is now grounded thru the RCA ground. Kind of scary, if there is a fault, but it will likely sound great.
If this solves the hum then you may want to try what I believe Chris is getting at in that you want to remove all connections from your amps ground to chassis/earth ground.
If not, I think its likely in your PS wiring.
Regards,
Mike
Hi,
Optimal for what he has 🙂
Actually my experience with the UCD was that single ended, configured in such a way that was ground loop free, or floating inputs, was good enough that little to no difference could be heard by going differential.
While in differential mode, ground loops still led to a great deal of hum.
All the same he hasn't gotten far enough with it yet to be in the troubleshooting stage. So at this point we should not plan on a ground lift since this is the time to be avoiding such a thing.
Are you seriously considering moving the differential input stage of the UCD into your pre-amp? Or did you mean to say you'd like to add differential output drivers so you can use true balanced inputs? I like the later idea.
Regards,
Chris
Optimal for what he has 🙂
Actually my experience with the UCD was that single ended, configured in such a way that was ground loop free, or floating inputs, was good enough that little to no difference could be heard by going differential.
While in differential mode, ground loops still led to a great deal of hum.
All the same he hasn't gotten far enough with it yet to be in the troubleshooting stage. So at this point we should not plan on a ground lift since this is the time to be avoiding such a thing.
Are you seriously considering moving the differential input stage of the UCD into your pre-amp? Or did you mean to say you'd like to add differential output drivers so you can use true balanced inputs? I like the later idea.
Regards,
Chris
Thanks for the help so far.
Just so you know I havent even built the amps yet. I have all the parts ready to go. I dont have any hum problems & never have with my system.
Just want to make certain that when I do build them I do it correctly.
Just so you know I havent even built the amps yet. I have all the parts ready to go. I dont have any hum problems & never have with my system.
Just want to make certain that when I do build them I do it correctly.
classd4sure said:
Hi Chris,
As for making my drive differential, my SACD777 is bizarre, and good. It has differential D/A's, then converts to single ended, then thru a bank of filters. For the SACD1, it then converted do balanced again. I haven't got the nerve yet, but I ponder taking differential out from the D/A I to V converters.
What I was really talking about is moving the op amp in the UcD to the preamp. I have yet to see any benifit of having it there in the amp. Enlighten me if you see any advantage. I tend to believe that less circuitry is better. I have a nice passive preamp, stepped divider from gold point, but from experience, its really best with at least a buffer to drive the cables. It gets more dynamic.
I've done that, but with discretes. I want to maintain DC coupling.
My thought is, if your going need those op amps in front of the UcD, and they are best ran from a seperate supply, it just seems logical to move them to the preamp. I'm not as keen on putting the signal through yet another pair (or three) op amps to make it differential. Also, since I'm doing monoblocks, I don't need two seperate supplies for each amp to supply the AD8620's.
Also, it doesn't seem like it could hurt to get those op amps away from the switching. And again, I then get my passive preamp active.
Look forward to your thoughts.
Mike
Stevenacnj said:
Stevan,
I'll ask again since I don't recall a response. Sorry if you answered. Does your toriod have two seperate windings on the secondary, or a center tap.
If you have two seperate windings, you may dig and find ClassD4sures nice picture of how to do a superior power supply with two bridges.
Good Luck and have fun.
Mike
Mike,
Yes sorry forgot to answer about the transformer.
They are 300va torroidal with only single primary & single secondary.
Yes sorry forgot to answer about the transformer.
They are 300va torroidal with only single primary & single secondary.
Mighty Mike: 🙂
Yeahhh! Do it, my man
And while you are there make it a high Vout, wich work best with passives (so no need to add noisy active circuitry).
(just forget who told you)
Good luck
Mauricio (the audio pusher) 😉
Yeahhh! Do it, my man
And while you are there make it a high Vout, wich work best with passives (so no need to add noisy active circuitry).
(just forget who told you)
Good luck
Mauricio (the audio pusher) 😉
Hi Mike,
I don't feel that relocating the input stage is the best idea. You'll end up with alot of parasitices in play that won't be compensated equally at all frequencies. Further I think you'd be giving up the usefulness of the instrumentation amp like setup altogether. May as well just remove them outright and drive the comparator directly, I think essentially that's what you'd be doing.
Regards,
Chris
I don't feel that relocating the input stage is the best idea. You'll end up with alot of parasitices in play that won't be compensated equally at all frequencies. Further I think you'd be giving up the usefulness of the instrumentation amp like setup altogether. May as well just remove them outright and drive the comparator directly, I think essentially that's what you'd be doing.
Regards,
Chris
Portlandmike said:
Mike,
fyi,
managed to locate these
Nichicon Muse Series 470uF 100V in 18 mm dia.
http://www.nichicon.co.jp/english/seihin/pdfs/e-fg.pdf
these should be better than stock I guess.
caps?
Out of the box,
The “audio grade” caps are not a good choice at all for several reasons. The main one is you pay extra for an inferior cap. The Nichicon HE series is a much better choice with an 820uf @ 100v part available in 18mm dia.
http://www.nichicon.co.jp/english/seihin/alm_mini/param_f.htm
Still not happy with the 2.4 amp ripple current rating on this one but probably adequate in this application as the main storage caps would do most of the work.
There are a lot of competing caps out there and the Panasonic FC is one of the better ones. Check it out.
Roger
Out of the box,
The “audio grade” caps are not a good choice at all for several reasons. The main one is you pay extra for an inferior cap. The Nichicon HE series is a much better choice with an 820uf @ 100v part available in 18mm dia.
http://www.nichicon.co.jp/english/seihin/alm_mini/param_f.htm
Still not happy with the 2.4 amp ripple current rating on this one but probably adequate in this application as the main storage caps would do most of the work.
There are a lot of competing caps out there and the Panasonic FC is one of the better ones. Check it out.
Roger
Roger,
Thanks for your suggestions. remarks about Nichicon noted (although I have had good experiences with Nichicon in the past)
Panasonic caps are indeed value for money and I have used them successfully in several audio PSU projects with excellent results. Its just that my experience with UcDs is fairly limited and I'm still coming to grips with the circuit topology.
While wading through the hot rodding thread, I recalled that there were some reservations about using low ESR caps as the PSU caps, while other had reported an increase in bass frequencies with a loss of mid range focus, therefore I'd excluded the FC range from my current short list.
Based on your comments, I will give the 680uF/100V FCswith a rated ripple current of 2.3A a shot, in the absence of anything better!
Regards,
Amit
Thanks for your suggestions. remarks about Nichicon noted (although I have had good experiences with Nichicon in the past)
Panasonic caps are indeed value for money and I have used them successfully in several audio PSU projects with excellent results. Its just that my experience with UcDs is fairly limited and I'm still coming to grips with the circuit topology.
While wading through the hot rodding thread, I recalled that there were some reservations about using low ESR caps as the PSU caps, while other had reported an increase in bass frequencies with a loss of mid range focus, therefore I'd excluded the FC range from my current short list.
Based on your comments, I will give the 680uF/100V FCswith a rated ripple current of 2.3A a shot, in the absence of anything better!
Regards,
Amit
Caps in question
Amit,
There are 2 decoupling caps next to the output devices rated at 10uf @ 100v that use their ESR to form a snubber. These are the only ones to not replace with Low ESR types. The ones we are talking about are isolated from the output devices with .05 ohm resistors so low ESR is desirable as this means less heating and/or higher ripple current ratings. It also means a lower impedance ground path for noise both incoming and outgoing. I don’t think their ripple rating is critical as they only need to supply current for a couple of microseconds till the main storage caps take over. Also the job is split with 2 caps on board so each has only half the work.
I have been using a 1200uf @ 63v low ESR type in this position to good effect. I can do this as my 60 volt rails are regulated and OV/OC protected.
I also have had good experience with the 105deg industrial type Nichicon’s but regularly replace the “audio types” with far lower ESR and higher temperature types. This has always been an improvement in measurements and sound quality for me.
Roger
Amit,
There are 2 decoupling caps next to the output devices rated at 10uf @ 100v that use their ESR to form a snubber. These are the only ones to not replace with Low ESR types. The ones we are talking about are isolated from the output devices with .05 ohm resistors so low ESR is desirable as this means less heating and/or higher ripple current ratings. It also means a lower impedance ground path for noise both incoming and outgoing. I don’t think their ripple rating is critical as they only need to supply current for a couple of microseconds till the main storage caps take over. Also the job is split with 2 caps on board so each has only half the work.
I have been using a 1200uf @ 63v low ESR type in this position to good effect. I can do this as my 60 volt rails are regulated and OV/OC protected.
I also have had good experience with the 105deg industrial type Nichicon’s but regularly replace the “audio types” with far lower ESR and higher temperature types. This has always been an improvement in measurements and sound quality for me.
Roger
Hi,
The local bulk storage caps in question (stock 470uF in case there's any doubt) are chosen strictly for their ripple current capability.
High temp for longer life expectancy is a desirable trait, along with low ESR. That really narrows down the choice, and I do believe leaves out the "audio" series as well in most cases.
What you're after is something designed more for low impedance at high frequency.
It was recommended they be ripple rated to handle at least 20% of the RMS output current. This was emphasized over actual capacitance value in question. This spec very well may be massaged with stastical music output at normal playing levels in mind, having to do with life expected lifespan.
I would think in most cases larger capacitance would lead to a slower attack, and fattened bass response.
An interesting tweak on the 700 may be to try different sizes based on an order of magnitude.
The local bulk storage caps in question (stock 470uF in case there's any doubt) are chosen strictly for their ripple current capability.
High temp for longer life expectancy is a desirable trait, along with low ESR. That really narrows down the choice, and I do believe leaves out the "audio" series as well in most cases.
What you're after is something designed more for low impedance at high frequency.
It was recommended they be ripple rated to handle at least 20% of the RMS output current. This was emphasized over actual capacitance value in question. This spec very well may be massaged with stastical music output at normal playing levels in mind, having to do with life expected lifespan.
I would think in most cases larger capacitance would lead to a slower attack, and fattened bass response.
An interesting tweak on the 700 may be to try different sizes based on an order of magnitude.
I wish Rubycon ZL/L caps was available in 63v, these are by far the best caps I've tried in decoupling positions bettering any grade of Blackgates in my opinion.
I've been trying a few different caps in a cdp I'm modding and the ZL/A are brilliant! I read somewhere its even advised to remove the plastic jacket to improve the sound which I thought was some kind of joke but I tried it anyway seeing as though it costs nothing😀 the results was actually very surprising of course its not recommended to do it if the caps are in close contact with any metal objects but this was not a problem in the cdp
I've been trying a few different caps in a cdp I'm modding and the ZL/A are brilliant! I read somewhere its even advised to remove the plastic jacket to improve the sound which I thought was some kind of joke but I tried it anyway seeing as though it costs nothing😀 the results was actually very surprising
of course its not recommended to do it if the caps are in close contact with any metal objects but this was not a problem in the cdp'Roger' that! 😀
That is indeed very useful information. Many thanks.
Chris thanks for your input too.
Keeping the above comments in balance, it does seem that the 680u/100V Panasonic FC would be one of the recommended choice.
It's rated ripple current is 2.3A and impedance is 0.036 ohms at 100KHz.(coudln't find the ESR value online)
It's very hard to get better specs in the commercially available cap range with the size restriction.
I'll keep everyone posted in case I come across any cap which can better the FC within the required specs.
That is indeed very useful information. Many thanks.
Chris thanks for your input too.
Keeping the above comments in balance, it does seem that the 680u/100V Panasonic FC would be one of the recommended choice.
It's rated ripple current is 2.3A and impedance is 0.036 ohms at 100KHz.(coudln't find the ESR value online)
It's very hard to get better specs in the commercially available cap range with the size restriction.
I'll keep everyone posted in case I come across any cap which can better the FC within the required specs.
Chris,classd4sure said:
I don't understand what parasitics your talking about. That statement is confusing to me. Could you talk more on that.
If you move the op amp to the preamp, you have the IA front end in the preamp. Then your driving balanced lines into what's left of the UcD, which is more like a diff amp as I recall, not the the comparator input directly. If that were so, it would be a concern with noise you might pick up. That diff amp, if it were the Bruno AES paper schematic, would have 1.8k resistors as input R's. Thus, extra cable "parasitics" wouldn' t be an issue. Extra cable capacitance "parasitics" could likely be as benificial as detremental, and the same with parasitic inductance.
On the other hand, operating anything to be construed as hi resolution, or accurate in the presents or proximity of a switching regulator is less than ideal, although possible in my experience.
LTC and Jim Williams had a product developed to try to remedy that.
The only way to know might be to try it, but if I'm not thinking of something, let me know!
Thanks
Mike
outofthebox said:
Chris said, "It was recommended they be ripple rated to handle at least 20% of the RMS output current."
Are those exact words, or translated I don't know. My understanding would be that the RMS output would be handled by the main supply caps and the local "470uF" would working at mostly handling the ripple current (its got HF ferrites isolating it from the mains to keep the HF hash on the board), which I think is about 2A peak to peak triangle wave. I'd guess the RMS ripple of that at under 1A. I forget the formula for tri wave RMS, but it might be .5 peak, or 500mA.
If your using minimal mains, 4700uF, then the 470uF is going to be seeing a bigger share of the ripple.
As for the 18mm Nichicons. Just know this, 18mm will likely fit tight. That is they will be very close or even in contact with the heat sink. So if your ruling out caps for ripple current, which is ultamately a cap life thing, you will be doing at least as much damage to life by having them in contact with the heat sink.
Also, be aware that the mounting screws will go right into the case if there long. Obvious stuff, but watch out.
Bottom line, with good mains bypassing, open up your options 🙂
Note: the higher the mains rails, the less ripple current due to load you have.
Mike
p.s. I'm about to change the stock caps. I've got the following and would like advice on which to try first.
470uF 100V FC
470uF 63V FC
470uF 100V M series panasonics.
Any thoughts?
They are stock UcD400 rev 6.1 except coupling caps are removed.
- Status
- Not open for further replies.