| wintermute |
Well I haven't built it yet (but I am making progress) ;) hopefully it will eventuate a little quicker than my MTM's are......
I decided to start a thread specifically about this amp so when I have questions I'll dump them in, whether asking questions this way will work or not I guess I'll have to see ;)
OK so I have just whipped up a dodgy chassis for it (first try at doing my own chassis, and this is only a test amp, so don't be too harsh.... the way it is going it'll probably end up in the ugly amp thread anyway ;) ).... It's going to be relatively compact and I'm wondering about the orientation of the torroidal (its a 300VA one).
I've taken a pic of where I think I'll mount it ( the box is sitting on it's front in the pic)
The blue rectangles represent where the LM3886's (I'm doing it p2p) will be (the other option is to use a shortened version of the heat spreader that came with the sink and they will then be in roughly the same place but hanging down....
Now I read that torroids radiate pretty much in a narrow band straight through the center so I'm thinking that this orientation should be ok.... any cautions?
The main thing I don't like about this orientation is that it will not be symetrical ie the toroid is closer to one chip than the other, my original thought was to mount the torroid vertically on the front wall smack bang in the center.
I'm not worried about the asthetics as this amp will be used as a test bench amp :)
Tony. |
|
|
| inguz |
| If you're building P2P, you might want to leave yourself a *lot* of space around the chips. Repairs & adjustments are much trickier when you can't just lift up a PCB. So I'd consider putting the trafo on the base or behind the front panel. |
|
|
| wintermute |
Thanks inguz
behind the front panel will certainly leave a lot more room around the chips, and also has the advantage that then both sides can be removed.... so that is probably a better bet. I'm a little paranoid about transformer induced noise as I have had problems with my other amp with this.
The base (as a mount point) is a bit tricky in that it makes it harder to mount my PS board because the base actually has the smallest dimentions (because I decided to put the heat sink on top) and the ps board wont fit beside the toroid!
Tony. |
|
|
| wintermute |
OK another question, (bet no one saw that coming ;)
when implementing the signal ground star point in p2p what method do people use?
I'm thinking iether just solder all the leads together at one point (messy and possibly dangerous, due to possibility of shorts), or use some sort of anchor point and connect to that. How have you done it :)
the PS star point is taken care of already :)
Tony. |
|
|
| Ropie |
| quote: | Originally posted by wintermute
OK another question, (bet no one saw that coming ;)
when implementing the signal ground star point in p2p what method do people use?
|
I solder all the leads together on the appropriate phono-out ground tag. |
|
|
| wintermute |
Thanks Ropie, that's something I hadn't considered at all. I better start considering the layout of my rear panel too ;)
Tony. |
|
|
| wintermute |
ok still on the signal ground......how important is it that all leads terminate at the exact same point?? I've got 7 gnd connections to do, I'm thinking of making pin 7 on the chip the star earth point, though this could result in not all leads soldered to the same exact point..... bad idea??? I'm thinking this may minimise the length of the runs to the signal ground star...
Every time I look at the chip I'm thinking "what have I got myself into" ;) the pins are a bit closer together than what I'm used to for p2p work!! Tansistors with three pins can have their leads spread apart easily :xeye:
edit make that 6 ;) I'll run the cz gnd separately back to the ps gnd :) actually thinking about it, should I run the supply bypass cap gnds back to the PS gnd separately too?
I'm using brians NI schematic..... I was originally thinking everything except the spkr returns terminated in a single star which then runs back to the PS ground..... now I'm thinking R2, NFB cap, input gnd, pin 7, local star with link back to PS. each bypass cap pair, wire back to ps. Cz wire back to ps. mute cap wire back to ps. I've just re-read the datasheet, and my the answer to my first question is probably a big fat "don't it's a bad idea" ;)
Tony. |
|
|
| wintermute |
well I decided that I better start actually soldering something ;)
This is what I have so far...... taking my time, this is fiddly work!!
Tony. |
|
|
| wintermute |
Thnaks Vikash :) I hadn't seen either yours or Marks :), I've been trying to see as many as possible to give me inspiration :) I'm trying to get it as compact (around the chip) as I can without going crazy ;)
Decided to have three earth returns..... one is the PS bypass caps and the mute cap. One is the NFB, chip ground, and signal grounds, and the other is the output compensation, the speakers will also have their own runs.
I've made some more progress, but when I checked my camera I don't have the latest pic (at GF's place now), I've attached the latest pic I took which has the rest of the FB components on the chip, I now have the input resistors and signal ground return wire too. only think left to do (for the first channel) is to put the output comp resistor and cap on, and solder the bypass caps and mute circuit to the chip :)
please excuse the poor quality and angle of this pic ;)
Tony. |
|
|
| wintermute |
OK another status pic..... If I keep this up I might actually finish it ;)
this one with input resistors and signal earth wire attached. Its amazing what you see when you take a closeup ;) looks much more horible than it does to the naked eye ;) I think I need to stop shrinking the heat shrink with the soldering iron tip too :smash:
Tony. |
|
|
| wintermute |
OK not much to go for this channel now, just connection to input output and power......
Note that although the photo looks like pins are really close or touching, they have a safe distance ;)
Tony. |
|
|
| wintermute |
| and one from the back...... might be time to go back to the chassis ;) |
|
|
| wintermute |
well I decided to do the rear panel...... holes went well except one where I don't think I centre punched deep enough and the bit must have slipped.... so one input socket is a bit wonky.....
but the real problem was with the hole for the power plug.... I got half way and my cheap nibbling tool decided that it had had enough of cutting 2mm aluminium and went on strike.... me being foolish and not thinking about it a little more carefully decided that the tungsten carbide cutting bit along with the mild steel case from the PC power supply that I got the connector out of would be real easy..... use the pc case as a template and just run around it..... hmmmm bad idea :rolleyes: If I had have thought a bit longer I could have used my jig saw, or if I was really desperate chain drilled it, but I let speed seduce me and well now there is some extra ventilation around the power plug ;)
Only other problem is that my fuse needs a 15mm hole and I don't have any drill bits that big!!
Oh well soldier on.....
oh and by the way there are some extra holes and a rounded off corner because this was a scrap piece of aluminium I had.....
Tony. |
|
|
| Ropie |
| Some intricate soldering there! :up: |
|
|
| wintermute |
Thanks Ropie :) It's not as good as I would have liked, but that's mainly to do with the fact that I forgot to clean up all of the component leads.... still haven't learned after more than 20 years ;).... some weren't tinning properly so I ended up with some blobs due to having to add more solder to get it to tin :xeye: Hopefully the second channel will be better, if I clean up all the component leads.
Here is another pic. It's starting to look like an amplifier now!! though after my gaf with the power socket hole, I'm feeling a bit hesitant about doing the switch hole!!! it's only a small one though, 13mm X 18mm, so maybe I should try chain drilling and filing it.....
man I wish I had some decent metalworking tools ;) I have my battery operated drill, a jig saw and the dremel (and a now defunct nibbling tool).... a drill press would be very nice. I was reading Vikashes case thread earlier, came across it when I did a search on finishing aluminium. I'm so jealous of all your equipment Vikash!!! I've always wanted a lathe, and you have lots of other goodies too :)
anyway time for me to head for bed I think.
Tony. |
|
|
| Vikash |
| That lathe still hasn't been switched on ;) Table saw and drill press are dead handy for metal work, but none top the router as my fav tool overall! My £25 router is still king. |
|
|
| wintermute |
Come on Vikash, surely you can turn up some nice equipment spikes or something ;) To tell you the truth, I'm not sure what I would use a lathe for, I know there have been times I wanted one, but I think that was when I wanted to turn some special bolts for my Morris 1100.....
Haven't made any progress today, as I had to do my tax return, though I did just go out and buy some power cable, as the 12 gauge stuff I had was too heavy, and the thinner stuff (15A auto cable) was too stiff. So I got some nice 16 gauge super flexible OFC stuff..... might do a bit more this evening :)
Tony. |
|
|
| wintermute |
Come on Vikash, surely you can turn up some nice equipment spikes or something ;) To tell you the truth, I'm not sure what I would use a lathe for, I know there have been times I wanted one, but I think that was when I wanted to turn some special bolts for my Morris 1100.....
Haven't made any progress today, as I had to do my tax return, though I did just go out and buy some power cable, as the 12 gauge stuff I had was too heavy, and the thinner stuff (15A auto cable) was too stiff. So I got some nice 16 gauge super flexible OFC stuff..... might do a bit more this evening :)
Tony. |
|
|
| wintermute |
OK just attached the power leads..... I think this has been the hardest soldering job so far!!! I always dread soldering power leads on though :xeye:
once again the camera doesn't hide any flaws ;) and it also flattens the perspective making everything look like it is shorting together!! I think the closest gap between anything is at least 2mm, hopefully won't have any problems!
Tony. |
|
|
| wintermute |
and the rear side, this is the one that looks kinda nasty ;)
Tony. |
|
|
| wintermute |
| Almost ready for a smoke test ;) |
|
|
| wintermute |
Well it passed the smoke test ;)
I've just powered it up and checked the dc offset on the output. 0.4mV!! is that too good to be true??? hoping that there isn't anything wrong :xeye:
That is with an 8 ohm dummy load connected and no input connected.
Tony. |
|
|
| wintermute |
apparently the dc offset is real :D
Have just hooked up to the pc sound card, and an old midrange driver I found out the front during a council cleanup and its playing music :) obviously with such a lousy driver I can make no comments on SQ but it works!!!!
with no input connected there is quite a bit of hum, but with the input connected no hum and just a very low level hiss :)
very happy about that as my 100W mosfet amp I have been struggling with hum for years...
I guess I should check the dc offset again with the pc conneted and nothing playing :) OK that looks a little more normal :) 30mv that is with no dc blocking cap on the input.
Is dc offset normally measured with or without an input connected??
Tony. |
|
|
| inguz |
| With shorted input. |
|
|
| wintermute |
Thanks inguz :)
with shorted input the DC offset is 1.4mV :) can't complain about that! :cool:
also I hooked up one of my 3 ways (which has speaker protection) and it sounds very nice indeed :) I've only got 1000uF per rail per channel, but to me it isn't lacking bass at all.
I cranked it, and I think it was starting to clip on the transients, got a bit harsh sounding.... but the ps was staying pretty steady. dropped from 27.6V to about 26.8V (though I guess my DMM might have to slow a sampling rate to see what is really happening :)
all in all very happy, now I just have to do the other channel and finish up the chassis properly as it is a bit of a death trap as is ;)
also this heatsink is way overkill ;) played for 30 minutes on a bit higher than normal listening level and the chip didn't even get warm, and I don't even have any thermal grease on it yet! no switch on or turn off clicks or thumps either (especially surprising since at the moment the switch consists of pulling the plug out of the socket ;)
gotta run now, need to get ready to go to the GF's place.
Tony. |
|
|
| Ropie |
Good work!
It's nice to see someone doing a p2p gainclone, not just going for the kit option :)
I think big heatsinks do improve sound so Iwouldn't worry about the overkill ;) |
|
|
| Vikash |
big fat :up:
What schematic did you use? And what what was the grounding arrangement you settled on. As soon as those heatsinks arrive I'll be putting mine together and I still don't understand proper grounding practices :xeye: |
|
|
| wintermute |
| quote: | Originally posted by Ropie
Good work!
It's nice to see someone doing a p2p gainclone, not just going for the kit option :)
I think big heatsinks do improve sound so Iwouldn't worry about the overkill ;) |
Thanks Ropie :) I wanted to do a p2p one and was even more determined when I read someones post that said people don't do them for the LM3886 cause there are too many components ;)
I guess if it runs cool then it is possibly having a positive effect!
| quote: | Originally posted by Vikash
big fat :up:
What schematic did you use? And what what was the grounding arrangement you settled on. As soon as those heatsinks arrive I'll be putting mine together and I still don't understand proper grounding practices :xeye: |
Hi Vikash,
I used the BrianGT LM3886 one (which is basically pretty close to the datasheet), I think the differences are the inclusion of the 22k resistor on input, slight differences in the NFB resistors and the increase to 47uF of the NFB cap)... I had already decided to go with 47uF on the nfb cap and decided to go with Brians other changes as the kit gets a good rap :)
I have the star point in the middle of my power supply board. It is common to both channels, so I guess I won't know if it has worked properly till I make the other channel ;) ( the PS actually has separate rectifiers and cap banks for each channel but a single ground running down the middle...... did I post a pic of the PS earlier?? edit: don't have one, will have to take a better one with the bolt and bleed resistors installed... here is a post with the ps board ---> http://www.diyaudio.com/forums/show...5654#post745654 before installing star point bolt... this was testing one side just in case I stuffed something up I didn't want to blow all my diodes or caps ;)
First thing on the star point is the transformer (center tap)* return, followed by the speaker return. Then I have the local power ground return, then the compensation cap return and finally the signal ground return. I'll elaborate below on the various local star points.
On the chip, I connected the +ve and negative leads of the two filter caps and also connected the mute caps earth connection at that point this made one local power star. That has one lead running back to the ps central star point. (Braided with the +ve and -ve rail wires).
I have a separate wire running back from the ground lead of the output compensation cap to the main star. I decided to do this based on the information in the datasheet.
I decided to use pin 7 as the local star point for the signal ground. It has the ground connection from the 22k input resistor, the ground connection of the NFB cap, and the ground from the input rca. The thin black wire is the run back to the PS star point.
This is my first try at star grounding, so I hope it is all still working properly when I hook up the other channel!! Pay to wait and see I think ;)
Tony.
* I know it is not a center tap since it is a dual secondary toroidal, but I don't know the correct term in that instance. |
|
|
| Mick_F |
Let me also show some pictures of my just finished P2P LM3886 Chipamp. It is based on the schematic of the Gaincard (see this thread ) plus the extra parts (mute, ground pin) due to the other chip.
The amp is mounted on a aluminium bookend which I use as a sort of testing jig for different amp configurations.
Mick |
|
|
| wintermute |
Nice work Mick :) bit easier for you to change things on yours than on mine!! I made mine pretty much as a once complete don't touch it (except for the PS)!! Good to see I'm not the only one doing a p2p LM3886 ;) Have you got anything supporting those big caps?? I'm thinking I'll put some silicon between my local decoupling caps and the heatsink so that there is less strain on the chip pins.
Vikash, here is a closeup of the PS (best pic I could find on the camera) not at home right now.... I thought I had a better one, but maybe I didn't take it.
Tony. |
|
|
| Mick_F |
| quote: | Originally posted by wintermute
Nice work Mick :) bit easier for you to change things on yours than on mine!! I made mine pretty much as a once complete don't touch it (except for the PS)!! Good to see I'm not the only one doing a p2p LM3886 ;) Have you got anything supporting those big caps?? I'm thinking I'll put some silicon between my local decoupling caps and the heatsink so that there is less strain on the chip pins.
Tony. |
I will make some kind of support for a final version. Not because of stability - the solder junction keeps the caps well in place - but because they can vibrate now. For testing it is ok.
Mick |
|
|
| Vikash |
| I was thinking about this too. Perhaps once it's all fixed fill the whole thing with resin. Just because you can and the amps are cheap so no huge loss from not being able to mod it later on... |
|
|
| wintermute |
resin!!! that might be a bit further than what I wanted to go ;) certainly don't want to pot my transformer in :eek:
I think that (apart from the resistors) the chips were the cheapest part of this amp!!! I haven't really been keeping a track of the cost, but I rekon it would be approaching $150 Aus (though that's partly because the caps I had to buy in lots of 5, and the transformer was a bit more expensive than I originally intended probably getting close to half the cost).
Tony. |
|
|
| Mick_F |
| quote: | Originally posted by Vikash
big fat :up:
And what what was the grounding arrangement you settled on. As soon as those heatsinks arrive I'll be putting mine together and I still don't understand proper grounding practices :xeye: |
Vihash, I have just put a detailed documentation of the building procedure of my "Gaincard copy" on my website. Please have a look at
this site
A hopefully very clear description of the grounding scheme is included.
Comments are welcome.....
Mick |
|
|
| Vikash |
| Excellent pics and write up Mick. The only things you haven't mentioned are the DC offset, and what the amp sounds like ;) |
|
|
| Mick_F |
| quote: | Originally posted by Vikash
Excellent pics and write up Mick. The only things you haven't mentioned are the DC offset, and what the amp sounds like ;) |
The DC offset is rather high. 90 and 48 mV. But I tested with **** speakers and later on with my Dynaudio test speakers and found no movement of the cones when switching the amp on, or any other problems. So I decided not to take any actions to reduce the DC offset.
I like the sound very much as far as I can judge it from listening to my test system. I will let it run for a week or so before I move it to my main system.
Mick
Edit: I was not aware that c r a p is a "bad word". It was removed by the system, however. I'd just like to let you know that I did not test it with **** speakers. ;) |
|
|
| Vikash |
Yes and we don't take kindly to people spacing things out to get past the filter! ;)
I noticed you have an input connected in the pic showing your DC offset measurement. Just wondering if this was connected on the other end and potentially adding to your offset?
My memory escapes me, but does grounding have an effect on DC offset? |
|
|
| Mick_F |
| quote: | Originally posted by Vikash
Yes and we don't take kindly to people spacing things out to get past the filter! ;)
I noticed you have an input connected in the pic showing your DC offset measurement. Just wondering if this was connected on the other end and potentially adding to your offset?
My memory escapes me, but does grounding have an effect on DC offset? |
So it IS a bad word??? A REALLY bad word???
I have tried with and without a source connected. The result was identical, but this is understandable as I have a DC blocker on the input. The offset I measure is a matter of the chip itself and the parts connected. However, I just leave it as it is.....
;)
Mick |
|
|
| wintermute |
Better be careful Mick you might get thrown in the sin bin for BAD language ;) sometimes I think that the auto sensor is a bit sensitive, but then there is a very wide demographic of people here, so what seems tame to us could be very offensive to someone else.
I've been thinking I should do a proper writeup on mine too, I've kinda been doing it visually with this thread, but a bit more explanation would be good ;)
Did you use generic brand components in yours Mick, or more esoteric ones? Mines a bit of a mix, but nothing outrageous.
I have made very little progress today, I think I drilled 6 holes and chain drilled and filed the hole for the power switch (it's much better than when I attacked the power socket hole with the dremel!). Oh that was the other thing, I drilled the hole and mounted the power indicator neon :).... no blue LED's in this gainclone :whazzat: (I'm assuming it is neon as it runs direct of 240V, and the multimeter shows no continuity, jaycar sell it as an LED).
Tony. |
|
|
| Mick_F |
| quote: | Originally posted by wintermute
Did you use generic brand components in yours Mick, or more esoteric ones? Mines a bit of a mix, but nothing outrageous.
Tony. |
I did not take particular care in getting fancy parts. The polyester caps are Wima and the electrolytics Roederstein. I bought the resistors at a local shop and I dont think that they are something special.
Actually, as a physicist I find it very hard to believe that exotic resistors are better (whatever that means) than normal ones. As long as you dont use wirewounds, a resistor is a resistor and ohmic resistance is its dominating property. This might be different with capacitors, which from a physicists standpoint are a much more complex issue.
Mick |
|
|
| wintermute |
I know what you mean Mick, I get very skeptical about things unless I can see something that could possibly make a difference. I used 7c metal film resistors in mine, but for the decoupling caps and feedback caps I used nichicon FG, which are a little more expensive but not by a lot. The PS caps I used panasonic FC's because of their low ESR (and also the generaly good rap people have given them), and the decoupling cap bypasses I used WIMA polyesters (I really didn't think that polypropylene would make a difference here, and the polyesters are so much smaller).
I had been pretty sceptical about caps too (except for low ESR because I could see that could help in lowering the output impedance of the PS) but I read Walt Jungs choosing capacitors article, and well lets just say it has me thinking there is more to the claims that caps can make a big difference than one might first think. I used a polyproylene (nothing exotic) for the output compensation cap, because itis effectively in the signal path.
I thought about getting some Welwyn RC55 resistors for the feedback resistors, but mainly because they are highly accurate 0.1% tolerance, and highly stable with temp changes (which could be important when mounted so close to the chip). but I ended up using the 7c ones ;)
Tony. |
|
|
| wintermute |
OK,
after a weekend of doing nothing at all, I have today almost finished the chassis, and in the process knocked the amp of the overly cluttered desk and broken the ps board :bawling:
Oh well the project seemed to be going to well to be true ;) no project is complete without at least one disaster!
No major problem, will try to glue with some superglue, and resoder the broken tracks, even if I can't mend it, should not affect operation, unless I decide to add more caps.
The other channel is complete except for power supply, input and output connections so should get it done tomorrow hopefully. fingers crossed.
well I said I was going to post a pic of the ps board so here it is ;)
Tony.
note that the top track on the left has some cuts in it because I was using it for practice, before I did the real ones ;) |
|
|
| wintermute |
Well I have fixed the ps board, still haven't wired up the remaining channel, but I have finished the chassis :)
It is really rough, but considering that it is my first attempt at fabricating a chassis, I did it in a hurry, and I only had a battery operated drill and jigsaw to work with I think it turned out ok ;)
Kinda an industrial look.
Tony. |
|
|
| wintermute |
| and the obligatory rear view ;) |
|
|
| karma |
looks good for your first attempt at fabricating, better than my
first.:D |
|
|
| wintermute |
Thanks Karma,
truth be known it is strictly my second ;) I wasn't counting the first as it is mostly wood, but I guess I should have. It can be seen here in the ugly looking amp thread :)
It may compete well with your first effort :D
Tony. |
|
|
| Upupa Epops |
| Tony, only little notice : at the next time don't make square hole with round file.... ;) |
|
|
| wintermute |
They are special ventilation holes, improve the sonics you know ;) I used the dremel, really really bad idea!! the power switch and fuse holes went much better :D chain drilled and filed (with the correct ones :rolleyes: )
Tony. |
|
|
| Vikash |
| I think maybe you just need the correct bit/technique on the dremel. I used a scroll saw with a wood cutting blade on mine and went very slow ;) I find bodge-ups usually related to patience - a virtue I don't posess :xeye: |
|
|
| wintermute |
In this Vikash you are 100% correct :) I ran out of patience!! I put the tungsten carbide cutting bit in the dremel, cranked it up to 37,000 RPM and went for it..... not the right approach, even with a template!! just cut the template too ;)
My dad has a dremel scroll saw, and a table saw, only problem is he is 7 hours drive away.......
and just to show upupa that I have improved here is a pic of the other two holes, yeah I know it still looks like I used a round file on the rectangular hole (I didn't I swear) I just went too close with the drill on on the first hole ;) the oval hole for the fuse is the final one. As you can see I'm getting better :rolleyes: (though still not perfect).... at least with these ones you can't see into the case around them when the switch and fuse are installed ;)
Tony. |
|
|
| wintermute |
I've now almost completely finished my gainclone, the only thing remaining is to screw the terminal block to the base ;) (oh and perhaps connect the 0V to the case) chassis is earthed to the mains, but ps isn't.... need to make an earth buster first though, as I need it for use with my pc). hmmmm I wonder where I can fit an earth buster ;)
Tested the second channel and it measured 1.2mV dc offset with input shorted, which compares nicely with the 1.4mV of the other channel.
Haven't done a listening test yet as it is too late, will do one tomorrow, along with an RMAA test.
It's been a lot more work than I thought it would be, but has been fun :) I'm going to do a proper writeup of the whole process in gory detail but it might be a while before I get it up on the web ;)
I think for my next chassis I will go with something a little more conventional, this one was made to match the heatsink I had, and as you can see from the photo below it is pretty tight. Hopefully it will not have a hum problem due to the compactness, but the test of the first channel was promising.
Tony. |
|
|
| wintermute |
OK so I have run some objective tests.
1. 1000uF per rail per channel results in some pretty serious sag in the ps when pushed hard. Voltage drops from about 27.5V to 23.5V tried with various freqs from 20Hz to 4000Hz. all behaved about the same with regards to voltage drop.
2. ripple in the PS is quite high. no load it is about 130mv when pushed to just before clipping (into an 8 ohm load both channels) it increases to about 1.8V. With say a 1Khz sine wave the ripple is nice looking (even if big) at full power. with a 20 Hz sine wave at full power the ripple is very scary looking!!!! more capacitance would no doubt help here.... whether I decide to add any not sure, I'll wait and see how it sounds ;)
3. Clipping performance is well pretty ordinary really (had read this before, wanted to confirm for myself. picture attached, note the bottom channel is the output sine wave off the dummy load. ideally it should still look like a sine wave with flat tops and bottoms, as you can see once the amp starts clipping the waveform gets rather distorted. This was a 1Khz sine wave. top channel was ripple, I set the preamp to 0.1X so the ~300mv it is probably in fact 3V!!!! the transformer is a 300VA BTW.
4. I've done the RMAA test can be seen here ----> http://home.swiftdsl.com.au/~tonywww/rmaa/gc4.htm Not stella but not bad either :) I use a sound card preamp in between the amp and the sc. set to 0.1X the amp didn't have quite enough grunt to get into the yellow zone, set to 1X I couldn't get any decent results at all. so I bumped the gain up slightly on my SC to get into the ok zone for RMAA when running in 0.1X mode. Test of just the sound card preamp can be found here ----> http://home.swiftdsl.com.au/~tonyww...new_battery.htm
Note that I suspect some anomalies with the crosstalk results are due to my test leads...
5. I do need some thermal grease on the chips!! at least for very high output. the heat sink got quite warm (maybe 35 deg C), and the chips themselves too hot to touch (not blister inducing but hot). I was pusing it quite hard though :)
That's all for now :)
Tony. |
|
|
| javven |
Not trying to troll - but you did this without HSC?
The stuff the computer overclockers use is pretty decent, actually. I was planning on using it on my sinks. |
|
|
| wintermute |
Hi Javven,
hehehe no need to worry about me thinking you are trolling ;) my initial test of one channel playing music at a bit higher than normal level for 30 minutes resulted in no increase in temp on the heatsink, and a barely warm chip :) I wasn't sure if I was going to be pulling it apart again, so decided not to put the HSC in at that stage (it's messy stuff), and at that stage it didn't really seem necessary.
This latest pretty much worst case test shows that it would be a good idea afterall ;)
I have a tube of unick so I think I will put some on to be on the safe side :)
Just did some power tests and the results were interesting :) adjusted level on my pc until just before clipping.
20 Hz only managed 13.2 V so about 21W 10Khz managed 20.5V so about 52W! (into 8 Ohm dummy load, rail voltage about +-27.5V) no wonder people complain about lack of bass with low capacitance supplies!
Tony.
edit: note about ripple with 20Hz sine wave being scary, the level was too high and the sc preamp was clipping ;) not so bad afterall...
pic attached. top trace is the ripple, about 2.4V and the bottom one is the output off the dummy load. |
|
|
| wintermute |
It seems my multimeter gets quite inaccurate at higher ac freqs.... seems to be still ok at 5Khz but at 10Khz is extermely inaccurate....
just checked with true RTA and whilst the meter is reading 20.9V at 10Khz, True RTA is reading 14.4V I suspect that the 14.4V is actually more like it :)
so feel free to ignore my previous comments about the bass being effected ;)
Tony. |
|
|
| destroyer X |
So... as i am lazy to do those testings, i will stand waiting for your final results.
Congratulations and thank you.
regards,
Carlos |
|
|
| wintermute |
Hi Carlos,
You may want to sit down ;) It could be a while!!! I'm not really set up to do A/B comparisons.....
I finished the amp this afternoon. I found that without the 0V connected to the chassis it had some hum and buzzzzz I was concerned about earth loops with my PC so I made an earth loop buster based on the design here -----> http://sound.westhost.com/earthing.htm Seems I'm posting this link a lot lately ;)
Result I hear you ask??? Absolutely dead quiet!!!!! I'm so happy because my other amp has always been very noisy! I probably should clarify that you can only here the buzz in it if you are close to the speakers probably within 20 cm or so.
(The ironic thing is the noisy amp gets a better S/N ratio in RMAA than this one) though I did do the rmaa test without the earth loop buster so will have to retest.
Good news Vikash I think my grounding system was a complete success!!!! I'm particularly impressed with the quietness when you consider how tightly packed it all is!
To give an idea how quiet this amp is, with no input connected (not even shorted) and this amp doesn't have a volume pot so max volume, I have to put my ear physically against the speaker to hear anything. The tweeter has the slightest hiss and the mid a tiny amount of hum, the woofer I'm not sure whether I could hear any hum or not. With the tweeter if I moved even 2cm away I couldn't hear anything, same with the mid.
Ok so how does it sound??? Well I haven't listened to any music on it yet; I plugged it into my TV (lowe xelos) and used the volume control of the TV.... too much hassle to drag the pc downstairs or the speakers upstairs. My first impression was that the treble was more pronounced than my normal amp. sillibance was more noticeable (though I hadn't listened to the former for a day or two). Flicked channels listening to different peoples voices (always a good test I think) and nothing glaring or nasty, sounded nice. Kept thinking the treble was more pronounced, wondered if that was simply because I had read that about others impressions and my brain was making it so ;)
After a while I decided it was time to pack up, reinstalled the usual amp and had a listen.... silibance was still there, maybe not quite as pronounced.... was undecided. Since then I have looked at the rmaa results for my old amp, it starts rolling off (very slowly) at about 5Khz and is down about .4db at 20Khz. The GC is pretty much flat out to 20Khz before it starts to roll off. Now I'm pretty skeptical as to whether I can actually hear what may amount to maybe .2db difference at say 10Khz...... maybe I can who knows!!
If there are any big differences with the GC and my existing 100W mosfet (apart from power handling, and clipping performance, the GC is only managing about 20W with the current PS, and the mosfet will happily deliver 110W into the same load) then I suspect I'm not going to be able to hear it very well until I finish my morel MTM's.
If I feel up to the challenge I may cart the pc downstairs, gut the series 200 so I can splice into the power amp direct and do some listening comparisons (still not true A/B but close).... but in reality this amp was supposed to be built to allow me to test loudspeakers (and active crossovers) so I think maybe I should get on with the loudspeaker side and not get too bogged down in amp comparisons ;)
Maybe my series 200 is better than I thought (yeah I know I only listened to TV ;) ) maybe when I listen to music I will hear the difference, but at this stage (at least with my current speakers) I suspect not.......
Oh and another pic not much different to before, but it is finished ;) . Now has earthloop buster installed, and terminal block screwed down, oh and heat sink compound on the chips, not that you can see it!
I connected one side of the earth loop buster to the same point that the mains earth is connected to the chassis, and the other side to the point where the 0V line connects to the terminal block, where the transformer wires terminate. Decided this was the best place for saftey in case of a transformer meltdown, and certainly seems agreeable from a noise perspective :)
Tony. |
|
|
|
