| ABO |
That's the question.
I did some listening with different types of input coupling caps. They all seem to degrade the sound. I tried:
- Intertechnik KPSn (tin foil) cap: good on trebles, but not in midrange, expensive
- Intertechnik Audyn cap (Polyprop): muddy! don't use these!
- Intertechnik MKP+ (low inductance polyprop): very good, perfect midrange, but slightly dull top-end and very expensive.
Then I think: why not use a output relay? If all, even the best, caps degrade the sound?
I currently have a DC coupled signal path straight from the DAC to the speakers. With very expensive drivers I think this is asking for trouble. If anything fails, my woofers will join the failing part.
So, what is the verdict on output relays? To quote Charles Hansen: 'Everything makes a difference.' Understood, but some things make more of a difference than others. So, caps or relay?
I know that there are kitsets available, but none seem to be of 'high-end' quality. I'm looking for a 'stand-alone'speaker protection kit that uses very good quality relays.
Thanks |
|
|
| ekaerin |
ABO,
If you don't want to do it from scratch you may want to get a cheap kit
with a nice topology (I assume it's the relay making up the cost) and
change the relay.
However, finding a relay that can handle both brute force and still deliver details needs some thinking.
Myself I'm not that "worried" about input caps, maybe I should be....
Anyhows, adding an input cap does not protect your speakers from an
amp failure. Guess which one is more likely to fail, your DAC or amp ?
/ Mattias |
|
|
| peranders |
ABO, before you do anything, why don't you test these caps together with an another person? I think you are influenced by opinions. Why do I say this?
I have designed an amp with BUF634 as output buffer. The pcb has an option for "medium speed" and "high speed" of this buffer. Some people here claim that they really can hear a significant difference. Myself and two more persons have tested this, jumper on...jumper off...jumper on... jumper off. I _wanted_ to hear a difference but I couldn't. Neither could the other persons. With this I wanted to say that despite my still rather good hearing the difference is VERY small but other people do really claim that they _can_ hear the difference. The thing is that those people never told us _how_ they could detect this difference.
With this I want to say, get second opinion before you are digging yourself into a too deep hole. |
|
|
| sam9 |
The value of the input cap rather than the brand is probably more important. The type has some importance as well. I came to this conclusion accidently in the course of some simulations followed by actually constructing an amp. If you look just at the roll of a 4.7uF metal film cap looks good enough for most circumstances. If your speakers can deliver really deep bass (20hz?) then 10uF metal film appears to do the job.
However, on further investigation, 10uF can be a source of noise and distortion (as comparred to DC coupling) well up t in to the spectrum - nearly to 1kHz - even though the frequency response is unaffected. To get noise and distortion down to the level of DC-coupled, it looked like you need a metal film cap of about 20uF. These are really big! For audio purposes the application is usually in speaker crossovers where the large size is less of a problem. On a PCB with limited space you will sometimes find a pair of oppossed (neg-to-neg) 47uF/25V tantalum caps, but these are expensive. I had very limited PCB space to wook with and found an even more atypical (and heretical) solution - a pair of opossed 100uF/25V aluminum electrolytics. They may not be as aptimal as a very large metal film cap, but they are highly effective and very inexpensive. If this bothers one's "audiophile concience", one could use Black Gates which will dent your wallet enough to make you notice.
The above is based only on my recent experience and should not be taken as authoritative - just an option to consider.
PS- there are also bi-polar electrolytics, both ordinary and BGs. |
|
|
| ABO |
P-A,
Of course my opinions are influenced, just like everybody else's. Just like you didn't hear a difference because you didn't expected it to exist.
OTOH, I realise this. I try my best to be as objective as possible. The differences between the caps is VERY clear. You'd have to be deaf not to hear the difference between DC coupling and a standard Audyn cap (or perhaps your system is not up to the job). With the other caps the difference is far more subtle and I wouldn't want to do a blind test. But A/B-ing a few times makes the difference clear: voices sound less palpable with KPsN and cymbals are less sharp with MKP+. The differences are there.
What I was really looking for are opinions about output relays. I have established sound degradation from input caps (to the best of ability- that's good enough for me). But I don't know about relays.
Ekaering is right. That's why relays would be the best solution: I would not need any coupling capacitor at all and not be afraid that any failure may wipe out my beloved woofer.
So let's have some opinions on relays!
.. and of course some concrete recommendations for kitsets or other easy to build solutions. |
|
|
| peranders |
| quote: | Originally posted by ABO
P-A, Of course my opinions are influenced, just like everybody else's. Just like you didn't hear a difference because you didn't expected it to exist. |
You are wrong here. I did actually _want_ to hear a difference.
If your problem is input caps, I think you will get even more problems with the relay contacts. Since I sell relays at work in know quite a deal about those but the knowledge is more "industrial" if you know what I mean?
The main problem is low voltage over the contacts and small currents...and powerful relays.
One contact material you should avoid is SilverCadmiumOxide AgCdO.
Avoid gold flashed contacts. They are only for protection and will wear out very fast.
Hardsilver is OK.
Goldplated (with thick gold) contacts are OK but _very_ rare.
A good relay, one of the best in it's category, is Elesta SGR282 but there is one problem. The delivery times :bawling:
We sell Elesta at work but we are changing brand because of too high prices (not the biggest problem) and rediculoius long delivery times. That's too sad because those Swiss made relays are really good. They even sound good, the click, like a Mercedes door!
But the state of the art comes from NAIS. They make really, really god relays. |
|
|
| Christer |
There are relays with double contacts, one contact of some
heavy duty material that can withstand quite strong arcs and
one with some good contact material, gold, beryllium or
whatever it is. It is designed such that the heavy duty
contact always makes before and breaks after the other
contact and thus protects it. Just don't ask
me where to get these relays, but I remember it has been
discussed before on the forum. |
|
|
| ABO |
I remember some sort of DC protection where a triac is used to shorten the powerlines. This would be a good solution. Nothing in the signal path and only a blown fuse in case of a mishap.
That's what I need!
BTW,
P-A,
Surely you know that such a test (opa 627/buf) doesn't mean anything. Everything is tied to something else and everything influences everything else.
For instance: what load did the BUF drive? Was it a 47 kohm amp input? Or did you use PMA's termination resistor? The latter would mean load resistance of 100 ohm, quite a difference with the former case. And such a difference may very well be audible.
I hate it when people draw generalized conclusions (nothing personal here). Usually this means misinforming other members.
ABo |
|
|
| Christer |
| quote: | Originally posted by ABO
I remember some sort of DC protection where a triac is used to shorten the powerlines. This would be a good solution. Nothing in the signal path and only a blown fuse in case of a mishap.
That's what I need!
|
Yes, it is called crowbar circuit, in case you want to search
for it. Slone has a schematic of one such circuit in his book,
for instance. |
|
|
| sam9 |
"between DC coupling and a standard Audyn "
What was the value of the cap? |
|
|
| ABO |
2,7 uF, giving corner frequency of 1,25 Hz with 47K input resistance.
I had no problems with the bass performance. My speakers go as low as 60 Hz:xeye:
But the midrange en top-end didn't sound good.
I used 3,3 uF KpSn and 4,7 uF MKP+ (stuff that I have lying around).
Let me know what you think |
|
|
| peranders |
| quote: | Originally posted by ABO
For instance: what load did the BUF drive? Was it a 47 kohm amp input? Or did you use PMA's termination resistor? The latter would mean load resistance of 100 ohm, quite a difference with the former case. And such a difference may very well be audible.
I hate it when people draw generalized conclusions (nothing personal here). Usually this means misinforming other members. |
My load was a 150 ohms headphone. |
|
|
| ekaerin |
ABO,
Firing of triacs to ground the output may also firing away your output
stage. I have done some testing on work, bigger stuff than HI-FI, but
it's amazing how stubborn even fast blow fuses are compared to silicon.
If you happen to have good margin design MOS-FET output stage
it is more likely this survive than a bipolar one (Now that's my own view).
Per Anders,
I'm all ears when it comes to your knoledge about relays.
For small signal it's easier but powerhandling+small signals, I'm lost.
I was planning to use Schrack relays for my next amp. You know thoose small orange guys. Is that a bad idea (I have some at work..)
btw, ever since your QRO got published (I was just a dreaming kid then) I have
been wondering, at least twice aweek. What does QRO stand for.
/ Mattias |
|
|
| sam9 |
OK. Relays. I use them. Simple "icecube" NO (normally open) type. They can be used to serve multiple purposes:
A- Delayed on and instant off to keep anu on-off noises out of the speakers. This may be aesthetic only, but I really appreciate the absense of thumps, bumps, crackles and fade outs.
B- DC protection for the speakers. I've never had this condition ocurr but it's nice to know it's there. DC at the speakers can happen due to a sorts of amp faiures independant from DC at the input. It is better if there is no failure, but things happen some times - imagine how you would feel to know that a $1 transistor failing detsroyed a pair of $5,000 speakers!
C-Thermal protection. Use a normally closed thermal switch attached to a heatsink or output device. If the temperature raises above the rating of the switch it opens, removes current from the relay coil and cuts the load from the amp. This is with class AB only. Class A, I think you would need something that cuts power to the rails of even AC power.
I'm convinced that good quality relays, of the appropriate rating do not affect the sound. The conacts are no better or worse than other mechanical connection such as binding posts. Eventually, they could accumulate wear but if you mount the relays in sockets, they are easil replaced every few years. It's certainly better than a fuse in any case.
I'm sure you know the following but in case other readers don't, DC protection via relay is a separate matter from over current (or short circuit) protection circuitry. That is to protect the amp from someone shorting the speaker terminals. |
|
|
| ABO |
I'm really only interested in saving my (expensive) speakers. In case of a shortcircuit I'll just replace the output stage, no worries here. I've never had that condition anyway.
Also, my amp doesn't thump on switch-on:smash: |
|
|
| peranders |
| quote: | Originally posted by ekaerin
Per Anders,
I'm all ears when it comes to your knoledge about relays.
For small signal it's easier but powerhandling+small signals, I'm lost.
I was planning to use Schrack relays for my next amp. You know thoose small orange guys. Is that a bad idea (I have some at work..)
btw, ever since your QRO got published (I was just a dreaming kid then) I have
been wondering, at least twice aweek. What does QRO stand for.
/ Mattias |
Triac's aren't my melody and it's very very very rare in poweramps. May I ask what kind of amp we are talking about?
At work we were importers of Schrack a couple of years back...They were good but Elesta is better.
In my QRO project I used Schrack with silver contacts but I don't know if it was AgNi or pure silver. I think it was "hardsilver", AgNi. They have worked good since 1989. I have had no problems at all.
QRO in morse code is "I increase the power"
I must also mention that those relays was controlled by an advanced DC protection circuit.... but very cheap... |
|
|
| peranders |
| quote: | Originally posted by sam9
I'm convinced that good quality relays, of the appropriate rating do not affect the sound. The conacts are no better or worse than other mechanical connection such as binding posts. |
Wrong wrong. The big thing is contact pressure, which relays have rather little of.
We have a big traction company which have 5 volts circuits plus microcontroller and relay contacts for a very critical part: the brakes of the train. The have had huge problems with DC, 5 volts via relay contacts.
Power relays are normally specified for 10-100 mA at least, often is the loudspeaker current less.
If you want to check good but expensive relays, check NAIS and their "vridankarreläer", "twist anchor relay", don't know the right word but they have a turning part, instead of linear moving part. |
|
|
| ChocoHolic |
Well, I never faced major issues with capacitors.
...even if they do affect the sound...
But relays often showed real poor properties... , but I often
used them because I prefered to protrect my speakers also
against failures which you cannot block by a cap at the input, such
as a blown outputstage or any open or shorted component that affects the operating points behind the input cap....
The idea with triac which shortens the rails (not the output!!!)
is quite old, but good.
I never dared to try this as it is less comfortable (it will also blow the fuse of the supply) and I had doubts about
the current ratings of the triac. The triac will have to shorten the
entire bank of supply caps... Did you ever see what happens, if you accidently shorten them with a screw driver?? :hot: Better don't try. And not at all without wearing protection glasses!!! :bigeyes:
OK, back to the triacs. In former times I did not know that you can get triacs for nearly any desired current. Yes, for really ANY current which audio people can imagine.
Today I would say that this protection might be the best solution, which I am aware of, because it does not touch the signal during normal operation.
It is less comfortable, because it will also blow the fuse of the power supply. But this is not an issue, because the protection typically is not acting very often.
Cheers
Markus |
|
|
| ChocoHolic |
If you want to try the option with the triac then I would
propose to use two triacs. One between positive rail and ground, the second between negative rail and ground.
In case of protection both should be turned on.
If you would only use one triac from positive to negative rail, then
it may happen that the stronger rail pulls the other one across ground.
This may cause heavy currents in the reverse diodes of your output stage and the loudspeaker...
Another reason for two triacs:
If you have fuses in both rails then one will blow first, while the second is still working. In this case the second rail would remain active....
Good luck
Markus
P.S.
Take care about the required currents to fire heavy triacs.... |
|
|
| jam |
Don't use either if you can get away with esp. with your amp.
Regards,
Jam |
|
|
| Christer |
Slone has one noteworthy point about crowbars, ie. triacs.
It is not uncommon that the triac also blows when it fires so
after fixing the amp after a failure, don't forget to check or
replace also the triac, or you will no longer have protection. |
|
|
| ChocoHolic |
...yes, the current load for the device which has to shorten the rails will be very high...
By the way: In the last posts I was saying triac, but thinking thyristor.
A thyristor would be sufficient as for the shortening the rails you will only need one direction for the current.
Bye
Markus |
|
|
| cunningham |
One soloution to the capacitor for small signal coupling might be to use a small audio transformer.
Just be sure that you match the impeadence (Z) correctly.
you can also add small caps or inductors to adjust the freq. responce. |
|
|
| thylantyr |
| quote: | Originally posted by Christer
There are relays with double contacts, one contact of some
heavy duty material that can withstand quite strong arcs and
one with some good contact material, gold, beryllium or
whatever it is. It is designed such that the heavy duty
contact always makes before and breaks after the other
contact and thus protects it. Just don't ask
me where to get these relays, but I remember it has been
discussed before on the forum. |
http://www.plitron.com/pages/Produc...imo/amplimo.htm |
|
|
| thylantyr |
I've seen pictures of the internals of name brand audiophile
amplifiers using speaker relays. :hehe:
After talking to a relay expert from a well known relay manufacturer, he recommended the mercury wet contact type
of relay. They form a clean contact on every swipe. The problem
with 'dry' relays is breaking that oxide layer that forms on
the relay contacts.
If you are paranoid about blowing up expensive speakers, use
fuses to protect the individual drivers. One fuse for the tweeter,
another for the midrange, another for the woofer, each will
have their own unique fuse rating. I fuse my tweeters all the
time and it works fine.
Speakers usually can take some DC voltage abuse, the
duration depends on power handling of the driver. The other
issue is the mechanical damage of the driver, higher DC
offset can probably do more harm than a lower voltage
from a mechanical point of view. :hehe:
If you don't want relays, capacitors, or crowbars shorting
the rails forcing a blown fuse situation as speaker
protection, then your choices narrow.
Why not just use a simple dc offset detector and have it
drive a relay connected to the main input power of the
amplifier to remove power when there is fault ? The detector
needs to run off it's own dedicated power supply because
it's the 'watch dog'.
Even if you remove AC input, the rails will take time to
drop in voltage and your speaker will still be subject to
DC for a short time. Your watch dog can drive another relay
that disconnects both rail voltages from the amplfiier section.
If you are clever, you can design a circuit to engage the
relay, force a high current situation to break the relay contact
oxide, then remove the high current and let the relay operate
as normal.
..... OR you can turn on your stereo system full blast for
a second during initial turn on -- to force high current <-- hehe |
|
|
| janneman |
| quote: | Originally posted by thylantyr
[snip]If you are clever, you can design a circuit to engage the
relay, force a high current situation to break the relay contact
oxide, then remove the high current and let the relay operate
as normal.
..... OR you can turn on your stereo system full blast for
a second during initial turn on -- to force high current <-- hehe |
Burn-in after all!
Jan Didden |
|
|
| Christer |
Thanks. I think somebody has once posted the name of a
european (dutch, I think) company that also sells that type
of relays.
| quote: | Originally posted by thylantyr
After talking to a relay expert from a well known relay manufacturer, he recommended the mercury wet contact type
of relay. They form a clean contact on every swipe. The problem
with 'dry' relays is breaking that oxide layer that forms on
the relay contacts.
|
Yes, and they have no minimum hold current either, I think.
Unfortunately, they are nowadays forbidden on this side of
the Atlantic.
| quote: |
If you are paranoid about blowing up expensive speakers, use
fuses to protect the individual drivers. One fuse for the tweeter,
another for the midrange, another for the woofer, each will
have their own unique fuse rating. I fuse my tweeters all the
time and it works fine.
|
Even the hardcore objectivist engineer Randy Slone advises
against speaker fuses. He claims they have thermal distorsion. |
|
|
| thylantyr |
Even the hardcore objectivist engineer Randy Slone advises
against speaker fuses. He claims they have thermal distorsion.
Another hardcore objectivist Rod Elliot says;
http://sound.westhost.com/project33.htm
Naturally, one can simply rely on fuses. Although these also have finite resistance it is small, and use of fast blow fuses can be quite effective.
We all know that audio system design is a game of compromise.
Sort your list of gremlins from top down. Fuses 'sonics' are
at the bottom of my list and protecting tweeters is on top.
I have revealing tweeters and I abuse my systems pretty
hard into high spl and I can't hear any 'extra' distortion
with an inline fuse. Maybe I use 'magic' fuses and don't even
realize it :clown: |
|
|
| Steven |
| quote: | Originally posted by Christer
Thanks. I think somebody has once posted the name of a
european (dutch, I think) company that also sells that type
of relays.
|
http://www.amplimo.nl/en-us/dept_56.html
See bottom of page. Actually it is the same as Plitron shows, I think. Cost is reasonable at Euro 6.21.
This is only for your information, I never used it myself.
Steven |
|
|
|
