|
|
Post by joethearachnid on Dec 19, 2014 21:41:53 GMT
Hey all, apparently this is a place that exists. So anyway, I've been very happy with my headphone rig for a while now, since the HD 600 and C.H.AMP combo treats me well in the majority of cases. I've been hunkering for some speakers though, and finally decided to get a pair of Wharfdale Diamond 220s for Christmas after a demo. I figured a Gainclone was really the only option for a high-quality, cheap and easy speaker amp build, so I built one and here it is, finished today:  I originally had a larger, matte silver switch with red LED ring but it turned out to be non-locking switch so I got this one from Maplin today. Drilling the 16mm hole required was... interesting.  The innards. I went with a 220VA toroid which might have been excessive, but the wording of the manual for this gainclone kit (chipamp.com) wasn't terribly clear on what power was needed depending on stereo or single PSU boards.  Sadly the back panel got a bit scuffed during drilling, especially with milling out that IEC hole. I still don't know a good way to go about doing it for a clean hole without a milling machine. The binding posts are from eBay, very chunky and well built, don't look bad either.  The amp in use with my mother's Mission Model 70 Mk IIs. It drives them much better than the ancient Creek integrated amp she had before, very clean sound and plenty of power. I'm loving it so far so I can't wait until the 25th so I can hear it with the Wharfdales. I also need to give the amp to my brother at some point so he can give it back to me, since he bought me the rather nice Modushop case. Couple of questions for people though (or Frans at least, hi Frans  ): I'm currently driving the amp straight out of my Fostex HP-A3 DAC since it has a reasonably high-voltage line output that is controllable with the volume knob. I don't have any of the rest of my rig with me to try preamping, but would either the Panda or C.H.AMP work well as a pre, or am I better off just using the Fostex? Would the sound be improved or just add an unecessary extra component in the chain? I'm also never sure I got one of the DC offsets on the Panda tamed very well (it would jump all over the place with temperature changes), presumably if amplified this would cause some problems? Also, whilst I do like the Modushop cases, if there's one thing I wish it's that there was an option for panels with no ventilation holes. They're certainly useful in high heat-output builds, but in a Gainclone where the heatsinks barely get warm, they just make the amp a whole lot louder by letting the transformer vibration echo out. The transformer is about as damped as can be (clamped down hard and middle filled with white-tack for good measure) so my current plan is to put a book or two on top of the amp, which vastly reduces the noise. Any ideas for covering the holes in a more permanent way without it looking unsightly? Glad to be here.  |
|
|
|
Post by jhelms on Dec 23, 2014 1:40:28 GMT
Amazing how good gainclones can sound. I have built a few myself over the years!
The diamond 200 series is not yet available in the US but I myself was planning on a pair of 250's (once I get to hear them) as we have a Wharfedale dealer here in Az. They look like a heck of a bargain! I still may change my mind and go with a set of 220's with active subs.
|
|
Crispy
very active
Madrigal music is playing - Voices can faintly be heard, "Please leave this patient undisturbed."
Posts: 779 
|
Post by Crispy on Dec 23, 2014 16:10:04 GMT
Also, whilst I do like the Modushop cases, if there's one thing I wish it's that there was an option for panels with no ventilation holes. They're certainly useful in high heat-output builds, but in a Gainclone where the heatsinks barely get warm, they just make the amp a whole lot louder by letting the transformer vibration echo out. The transformer is about as damped as can be (clamped down hard and middle filled with white-tack for good measure) so my current plan is to put a book or two on top of the amp, which vastly reduces the noise. Any ideas for covering the holes in a more permanent way without it looking unsightly? Glad to be here. Hi Joe, Great looking build  I had a think about your problem and the simplest and possibly the cheapest solution would be to damp the full case with PC sound deadening sheets like these: www.dabs.com/products/akasa-paxmate-plus--noise-reduction-matting-for-full-tower-pc-case-7M98.html?q=Akasa%20AK-PAX-2&src=16There is enough to do a full tower PC case, so if the noise is not damped after one application then put another layer on top until you reach a level where the sound does not intrude. If you then do need to use the case ventilated you can soon cut through the ventilation slots with a scalpel or craft knife through the sheets, or simply remove the sheets. I would even put some on the top of the toroidal and possible around it? Frans may know whether this is a good idea or not? |
|
solderdude
Administrator
measureutternutter
Posts: 4,886
|
Post by solderdude on Dec 23, 2014 19:03:01 GMT
Hi Joe,
If you can play the amp loud enough with just the HP-A3 then I would not use an extra pre-amp.
You could always change the gain of the gainclone should you need a better volume adjusting range.
The hum from the trafo may be because it is too close to the metal bottom.
When you unscrew the trafo from the bottom and lay it on a piece of foam is the hum in the enclosure gone ?
Usually toroids don't hum very loud, but some do.
You could lay a thin sheet of metal (black) on top of the top plate or buy another black top plate without ventilation holes.
Personally I would leave them open and if hum is the reason for you wanting to close the holes would look for ways to 'audibly shield' the trafo or replace the trafo instead.
I HATE humming trafo's !
|
|
|
|
Post by joethearachnid on Dec 24, 2014 1:43:05 GMT
I remember the trafo humming when I was testing the boards out of the case, just sitting on a thin MDF board on a carpet. It's currently on two rubber pads totalling maybe 6mm thick in the amp. I've always had some home with these (Vigortronix) trafos so I assumed it's normal. It's worse the larger the trafo; the C.H.AMP ones were near silent, the Panda one middling and this one easily the loudest. It's still not BAD so much, only really audible in the dead of night in the quiet English countryside. The amp was on a wooden table (pictured above) and was still noticeably humming, so I put in on the floor under the table and I couldn't hear it at all sitting maybe 1.5m away. This was with a few books on top of the amp covering the (top) vents though. In actual use the amp will likely be somewhere out of the way anyway so this is a realistic situation.
Also been working on a bit of hum from the Mission speakers, again not really noticeable in normal circumstances but worth considering since I'll be using the Wharfdales at pretty near-field. One thing I actually didn't do was ground the heatsinks properly which would likely help shield the amp circuits. I meant to do that but entirely forgot in the end. I also considered putting the trafo in a separate box which might have helped, but other builds I've seen have put the amp boards next to trafos with no complaints. I also found that swapping the crappy RCA cable I was using for one I made up specially (since the RCA connections are far apart it needs a wider Y-split) with shielded cable reduced the noise quite significantly. In fact with nothing on the inputs there's very little noise at all. I'll see what the noise level is when I get the Wharfdales.
I'm always a bit overzealous with noise on a new amp, usually after a while I stop caring so much.
|
|
solderdude
Administrator
measureutternutter
Posts: 4,886
|
Post by solderdude on Dec 24, 2014 16:04:33 GMT
When the enclosure is grounded the heatsinks probably also will be... when they are mounted on the bottom plate that is.
I would consider putting the trafo in a small and audibly damped enclosure in the cabinet or search for a more silent trafo.
When there is some hum on the speaker outputs sometimes finding another point to ground the enclosure helps.
|
|
|
|
Post by joethearachnid on Dec 29, 2014 2:54:57 GMT
Quick update:
I got the Wharfdales and they sound lovely. I'm still rather bad at analysing them since I don't have a lot of speaker experience, they're as yet not set up optimally and I haven't had everyone out of the house to play them loud. Very clear, very nice, articulated etc. I didn't really realise how little base slam you get comparatively with speakers until I put on my HD 25-1s the other day and was kind of blown sideways. I guess that's also partially a volume thing; being considerate about volume in home listening is something I haven't had for a while.
Anyway I think less amp noise is coming through on the Wharfdales than the Missions which is nice, though I did realise that the ground loop problem I've had with all my amps to date is due to my laptop PSU doing some odd things with its ground plane rather than anything else. Since my desktop that I now use mostly should have a sensible ground plane I should be fine without any messing around on the amp once I get it back to university. I'd still like to do some damping to stop transformer noise but right now I can't be bothered to crack the case open since I'm busy enjoying music.
Last thing I have to consider is possibly some kind of output switcher. I believe that I can split a source over a lot of amps without any degradation so current setup is the HP-A3 feeding two headamps through a splitter. However I'm unsure about using second splitter or replacing one of the headamps with the speaker amp since I'm unsure of the electronic specifics and it feels messy regardless. Therefore my options would be to swap the outputs over manually every time I want to use the speakers, buy or build an RCA switcher to switch outputs, or use one of my headphone amplifiers as a preamp. Sadly a lot of bought RCA switchers look rather grim and of lesser build quality...
|
|
solderdude
Administrator
measureutternutter
Posts: 4,886
|
Post by solderdude on Dec 29, 2014 8:46:27 GMT
There are interesting articles about hearing with the body.
We all have highly 'touch sensitive' cells (so you can feel light brushes/touches).
These aren't the same ones that make you feel pain when something happened.
These very sensitive cells appear to work along side hearing and is sensitive to low frequencies only.
Turns out they aid us greatly in the lows and the 'missing 6dB effect' with headphones.
This is why most people prefer lifted bass in HP's to compensate for impact.
The problem with splitting a line level signal is that when some amplifiers are 'off' the input transistors of these switched off devices acts as a diode limiter degrading the sound.
With some amps (those with a resistor in series with the input, which I always implement in my designs) this isn't a problem when the output R of the source is low.
So a splitter (or multiple ones) will depend on the amps in question.
A matter of listening....
Wharfdales often have a much lower sensitivity than Missions (depends greatly on the model though)
When the sensitivity is 5 dB lower, so will background noise be lower.
Most separation transformers do not break HF groundloops but they could help for audio range frequencies.
|
|
|
|
Post by joethearachnid on Dec 29, 2014 15:38:51 GMT
Interesting about the body stuff, I guess it's just the difference between the headphone and the speaker experience. The Wharfdales aren't exactly bass monsters anyway being bookshelf speakers but it's just a different effect to get used to. I had assumed that analogous to headphones, speakers would have drifted further into being overly sensitive as people downsized their systems, so I figured that the older Missions would be require a lot more welly. I guess in passive speakers you're always going to need an amp so there's no reason to mess around with sensitivities compared to headphones where low-power sources are now the norm.
I did some reasonable testing when I first installed the splitter in my rig (HP-A3 to a C.H.AMP and a Panda) and I couldn't detect anything in terms of loss of level or quality. My completely unscientific thought was that since the HP-A3 wasn't really 'driving' anything its power wasn't being used up and so there's no loss. Both the C.H.AMP and the Panda have inputs going straight into 50kohm pots, and from what you say the C.H.AMP has an input resistor and so does the Panda if I recall. I'm not sure if this is still the case with the LM3886 Gainclone, which seems to have a 1kohm input resistor and no volpot on my build. Generally amps that aren't being used are switched off, not sure if that makes a difference.
In terms of the ground loop it's only really a temporary problem until I can get back to my desktop. I'm currently using a Raspberry Pi as a controller for my DAC, which is not connected to mains ground anyway. I did look briefly at ground loop isolating units but they seem to degrade the sound quite a bit and I'm not that keen on transformers in the audio path anyway (though your zero NFB transformer amp looked very interesting Frans). I'll live for the time being I think.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Dec 30, 2014 0:49:55 GMT
I didn't really realise how little base slam you get comparatively with speakers until I put on my HD 25-1s the other day and was kind of blown sideways. I guess that's also partially a volume thing; Joe, I feel it's more a sensitivity/bass driver/cabinet size thing. Physics dictate that standmounts (generally) just wont 'slam' or go as low as floorstanders. At 86db/13cm/7litres,( Diamond 220) you're never going to 'move enough air' to give you the 'percieved slam' you get with a pair of Senn 25's. As I'm sure you know, and quite counter-intuitively, the bigger the speaker (bass-driver/cabinet volume), the easier it is to drive, the higher the sensitivity, and so the louder the volume you will achieve with a given amp power. And with larger bass-drivers within larger cabinet volumes (i.e. floorstanders rather than standmounters), the more air can be moved and the more 'slam' will be felt. Standmounts, while maybe not providing you with huge amounts of slam, (generally) have many other positive attributes: accurate imaging, less cabinet resonance, more ability to sonically 'disappear', less baffle interaction, and a generally more forgiving ease of placement. As ever, YMMV, Jeff |
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Dec 30, 2014 0:55:55 GMT
BTW, lovely build, you must be proud - you should be.
And nice to see you on our friendly little forum, always liked reading your posts at the other place,
Jeff
|
|
|
|
Post by joethearachnid on Dec 30, 2014 15:53:09 GMT
Jeff,
Thanks for your kind words! My builds are certainly a lot cleaner than they used to be but it's always a bit humbling when you see people using solid-core wire in perfect runs with right angles and supports and whatnot. Fortunately the slightly crummy pictures don't show all the blemishes I know are in there, but in the end nobody else is really going to see or notice them anyway. Clean internals are good for noise reduction etc. but in the end it's more just about selfish pride. Nothing wrong with that though!
I actually didn't know about the smaller speakers being harder to drive, that's pretty interesting information. The fact that small speakers can't output as much bass as large ones I was aware of, but then again it's not like I felt the Wharfdales were lacking, just that the HD 25s were so unexpected. I did look vaguely at subwoofers at one point, but for near-field listening I think the current bass quantity is fine, and good subwoofers seem to be a bit out of my price range for the most part. In the end the speakers are there as an alternative rather than a replacement for my headphones so if I want bass it's always available, albeit in another form. Also I feel a hefty sub might not go down well in a terraced house...
I'm very inexperienced in the whole speaker area so if you have any other advice please go ahead and educate me.
-JoetheArachnid
|
|
|
|
Post by joethearachnid on Jan 1, 2015 18:58:07 GMT
I've been considering maybe putting some attenuators on my HP-A3 when feeding the Gainclone to reduce the amplified noise, more as a thought experiment than anything I plan on doing particularly. As far as I can tell most of the noise is coming from the source itself, the noise volume doesn't change with moving the volume on the DAC, and I have plenty of gain. However, I'm getting some confusion when it comes to the required values. Using this site: leleivre.com/rf_pipad.html I've calculated that using a 750r resistor as R1 and a 3kr resistor for R2 would give an attenuation of about 6dB. This is going on the Gainclone having a Zo of 1kohm which I believe is correct for this version. However, that gives a 'Return Loss' of infinity, which I'm not entirely sure about. Any thoughts? |
|
solderdude
Administrator
measureutternutter
Posts: 4,886
|
Post by solderdude on Jan 2, 2015 9:08:42 GMT
Perhaps lowering the gain of the gain clone is more sensible.
This way you also lower the amplifiers noise (that should be low) and lower distortion and might even increase BW a bit.
The input is less sensitive as well and you get better tranvel on the A3 pot.
The attenuator circuit shown is for RF (Radio frequencies) and designed to lower voltages while keeping load and source side impedances the same.
For audio this isn't needed and you can leave out the R2 on the input (left) side.
If you want 6 dB attenuation just use 2 x 1k
This will lower the output voltage by half (-6dB), the amplifier will see a source with 500 Ohm output R (providing the A3 is close to 0 Ohm, and the A3 will 'see' a load of slightly lower tan 2k which is a normal opamp load.
|
|
): I'm currently driving the amp straight out of my Fostex HP-A3 DAC since it has a reasonably high-voltage line output that is controllable with the volume knob. I don't have any of the rest of my rig with me to try preamping, but would either the Panda or C.H.AMP work well as a pre, or am I better off just using the Fostex? Would the sound be improved or just add an unecessary extra component in the chain? I'm also never sure I got one of the DC offsets on the Panda tamed very well (it would jump all over the place with temperature changes), presumably if amplified this would cause some problems?
