Oh noes! Not another amp...

[hardnrg]

double-sided pcb turned out wicked!

 

any updates on the case then?

[markiemrboo]

I just stumbled on this thread and spent like 30 minutes reading lol, amazing stuff man that looks darn near professional keep it up

 

Cheers!

 

Very impressive. This just makes me feel stupid for not even attempting a CMOY amp. This amp is being used to power speakers, correct?

 

That's right Get yourself attempting the CMOY There's something very rewarding about creating something which makes sound

 

double-sided pcb turned out wicked!

 

any updates on the case then?

 

No updates on the case yet, sorry

 

I can now announce, after a few very stupid mistakes that had me puzzled forever, that it works fine and I have had it playing music nicely at low volumes , as the test heatsink is only attached to one of the chips.....and without any thermal paste.

 

All resistors are hand matched from 1% resistors down to about 0.00001% I did this using my yogurt pot LMS bridge, which is more accurate than my multimeter. This is actually quite important for equal current sharing between the chips, especially with the gain resistors.

 

Anyway, the mistakes were as follows:

 

Mistake #1: Originally I saw oscillation. I forgot to stick the bypass caps on the power pins on the back side of the board, and blamed this. The oscillation was still present, but only when the input was left open and not shorted to ground.

 

Mistake #2: I had the two small 47uF caps on the right hand side of the board the wrong way around (the PCB artwork was fine, it was my mistake when assembling). When I left it switched on for 2 seconds with the input shorted to check if the oscillation went, and it did, I heard BANG *very brief but noticable pause* BANG. The caps exploded, and cap juice sprayed up the edge of the heatsink

 

"Mistake" #3: The safety resistor from input ground to power ground was possibly a tad too large at 10ohms and went open circuit because of mistake #4. As a result each IC was operating at a gain of unity, where they are unstable. This resulted in the oscillation I was seeing. At the time I was unaware this resistor went open circuit, as it made no noise and didn't look physically damaged at all. I soldered a 0.47R in it's place and the oscillation went away and I measured 14mV DC offset. All looked fine until...

 

Mistake #4: This is the one that had me puzzled for quite some time, and was pretty much the cause of everything! Because everything looked fine with no signal, I connected my preamp up and played music with the volume right down. Immediately after switching on I heard a quiet fizzle noise which seemed to be coming from the dummy load resistor. It had gone open circuit! When I went back and tested the conditions that were previously fine, it was in oscillation again. The 0.47R safety resistor had also gone open circuit quietly. It was here that I realised the chips were operating at unity gain, which was why I was seeing oscillation. I couldn't figure out why they had gone open circuit though. I posted a long post on diyaudio and slept on it.

 

When I woke up, I looked over the PCB and schematic and it dawned on me that I had the oscilloscope test leads the wrong way around on the speaker terminals! I had the scope probe ground connected to the speaker +ve output, thus shorting the output to ground, which resulted in a lot of current flowing. This destroyed the safety resistors as the power ground was floating ( that'll teach me I guess ), and this then caused the oscillation because the chips gain had suddenly been set to unity. I replaced the 0.47R resistor, swapped the leads around, and finally saw what I was expecting.... the output was a nice clean amplified version of the input!

 

Here is it playing some quiet music in to a speaker, you can just about see it on the scope.

 

[Bleeble]

I was wondering what happened to your updates.

 

So you just learned all this by messing around on your own? Any good books or online resources you used?

 

I just ordered the parts for my Cmoy yesterday. Hopefully they'll be here tomorrow. Weee!

[markiemrboo]

I was wondering what happened to your updates.

 

So you just learned all this by messing around on your own? Any good books or online resources you used?

 

I just ordered the parts for my Cmoy yesterday. Hopefully they'll be here tomorrow. Weee!

 

always seems to be slow progress with things like this.

 

Strangely, yes. Lots of reading and messing about. Same way I learnt how to code... and my way around the BSD's... etc :thumbs-up: I have just started watching these though. Maybe I should go back to Uni and get a degree in electronics too.

 

Good luck with the CMOY though No doubt you will make a thread about it on here, so I will keep an eye out for that

[markiemrboo]

Right! As you probably all know by now, I have been quite ill recently. I'm still not better, but I have found myself able to actually physically do some stuff without dieing. You probably all know that I went back to the drawing board and totally redesigned this new amp to have three LM3886s per channel, in parallel, for increased power handling. I also ditched the regulated power supply, as I don't think it would have been able to cope and it was just extra fluff I couldn't be bothered with.

 

The end result was initially a 100mm x 100mm amplifier (mono) with all but the bridge rectifiers on the board. I tested this, it worked well after a couple of stupid mistakes and even without any kind of proper grounding. I was then looking over the PCB layout and realised it would have been so very easy to have the bridge rectifiers on the PCB without increasing the size, so I went back yet again and redid another set of PCBs. It is these PCBs that I will be speaking about right now.

 

The usual etching process was the same pain in the rear end it usually is, but ended up working well enough and produced two very usable boards to work with. Drilling was a bit of a pain, as it liked to lift the thinner traces... but on the second board I just said to hell with it and drilled with it flat against my (glass) desk. It drilled so much easier / better and I couldn't see any damage to my desk :thumbs-up: I need a workshop Here's the two finished PCBs side by side. You will notice I haven't bodged a 'tinning job' this time. I tried Maplins own brand PCB lacquer on the first revision, which seemed to get dirty quickly. I tried Servisol "Plastic Seal" Lacquer on these, which seem to work better! The two black parts between the large black caps are the bridge rectifiers. They are now "only" 8A types, one for the positive side and one for the negative side (16A?!) rather than two 35A types. More on how well this works later...

 

 

The next task was heatsinking. The holes needed to be redrilled as they were in different places to the (old) new amp. I covered the old holes up with some black electrical tape, which you can probably clearly see. Drilling was easier than I imagined, and I still used the LEGO "drill press" to keep it straight, because I am poor. It works well.

 

I isolated the front edge of the board where the LM3886s are standing (power traces run here) from the heatsink again using electrical tape, though the anodizing would probably prevent anything from happening anyway... I wanted to make sure. Ugly as it is, I isolated the boards from each other at the middle with a piece of electrical tape too. One day I might find some kind of H edging rubber / PVC that I can slide down the middle, which would look neater.

 

I use the metal tabbed packages for the LM3886s which also need isolating from the heatsink as the tab is internally connected to the -ve power pin. Previously I had used 'mica' sheet and AS Ceramique. This was messy and awkward. This time round I am using adhesive backed Sil-Pad 900-S (TO92-P) thermal pads, which falls a tiny bit short of the edges of the chips but is thick enough that those edges don't touch anyway. I must say, they are so much easier to work with. Paste is not needed, so it's just a case of sticking them on the heatsink in the right place and screwing the chips on! Nice!

 

 

 

The next thing I wanted to do was, of course, just briefly turn the things on and see if they will sit and idle without exploding. I'm using the light bulb trick, a bulb in series with the mains live, which should limit the current in case of a short etc. The caps are effectively a short circuit at power on time until they charge, so the light should flash bright and then fade out to nothing / almost nothing. I always get nervous at initial turn on time, but it worked great and flashed bright for a second then faded, and sat without exploding.

 

 

Logically then, the next step is playing some tones / music and seeing if it amplifies. I have no picture of this, but it did So next up was testing with a test speaker. I blew my last one up a long time ago, but recently remembered that I have four old satellite speakers from my old old old PC speaker set (CS FPS1000) which I could use! I wasn't sure what would happen if a speaker was plugged in, while the light bulb was still in circuit, and then powered on. Power up, however, was clean with no scary noises. Although you can't see it, the amp is on and the speaker is playing music quite nicely. I was actually quite impressed, as this (horrible) speaker was far less horrible on this amp. On the PC speaker set it used to distort horribly at quite low volumes. This sounded rather clean (though still poo) even when being fed boosted low frequency and at high volume!

 

 

I ran this for about half an hour playing music fairly loud, and nothing bad happened. The heatsink stayed pretty cool the whole time... warm, but certainly not too hot to touch. The rectifiers were both stone cold. It's a 6 ohm speaker, and I was sort of abusing it. Judging from light bulb brightness it was probably giving at least 20-30W in to the poor thing. Oh, a cool thing about the light bulb trick is that when bass hits the bulb lights brighter. My sister, mother and myself were quite mesmerized by that for a few minutes

 

Happy that the speaker hadn't exploded, I proceeded to test with my old Eltax Monitor IIIs. These are 4 ohm nominal speakers. I was again impressed with the initial testing. I have tied things to safety Earth now, and I could hear no humming or buzzing, just a faint hiss ( from less than 1cm away ). The amp was driving this speaker very well indeed. I seem to recall that with my Cambridge Audio A5 (and my current single LM3886 / regulated PSU) amp(s) at higher volumes the woofer seemed to bottle out and sounded terrible. I didn't seem to get that with this amp, even with the lower frequencies still boosted (30 and 60Hz), which sounded overly bassy but was a good test. I suppose this would be known as 'well controlled, tight bass'. I left it playing for a good half an hour and again rather loud, and the heatsink was still not at all too hot to touch. The rectifiers were still stone cold

 

I'm satisfied now. It's very neat, and it works with minimal wiring effort / attention to "proper" grounding (= more neatness) without any audible humming or buzzing noises! The case is coming.... and I can't wait

[suchuwato]

Awesome! Can't wait to see the finished thing

[markiemrboo]

Awesome! Can't wait to see the finished thing

 

my little sister loves it. She says she wants one I will probably build her a mini single chip version after.

 

I am now testing both channels with both of the old Eltax. It's still working very well The light bulb is telling me it's drawing somewhere near 100W (it's a 100W bulb and lighting up bright) Admittedly not much, but doesn't seem to be breaking much of a sweat. I love it!

 

[suchuwato]

How much would you charge to build one then?

[markiemrboo]

How much would you charge to build one then?

 

Well, I am not sure really I was considering getting these PCBs fabricated professionally. I don't think I would sell them with home made PCBs to be honest. It then depends on how many people would actually be interested / would buy PCBs. The price (per PCB) for getting them done professionally goes down a fair bit once you get to about 50 PCBs, but finding 25 people who want a stereo amplifier based on these PCBs.... I dunno if it will happen easily!

 

I expect that the PCBs would be about

Edited September 1, 2007 by markiemrboo

[markiemrboo]

I've tried this amp on my main speakers today. As much as I don't believe that amplifiers in their operating range can sound different, it does actually seem to sound different. It may be in my head because I am in a particularly happy mood etc, but for those who are curious here are my initial impressions. Purely subjective of course.

 

Bass lines seem to have become very easy to follow. They never really stood out before. Possibly what some might call tight and well controlled? It seems like it may also be going deeper, but with my bass lacking speakers, where I need to boost the bass quite a lot to get it acceptable, I may not be qualified to say

 

Treble and midrange are sounding really quite lovely. There seems to be a certain delicateness about them both that just makes it so nice to listen to. I seem to be noticing details in both the treble and midrange I hadn't noticed before. Airy echoes and such. It also seems that guitars are sounding nice and raw, but at the same time not screaming in my ears. I randomly put on The Verve's "The Drugs Don't Work", and the opening instantly caught my attention. The guitars twanging away just sounded so sweet. Now, I have listened to the track before obviously, but it never really caught my attention like that previously!

 

I haven't really noticed any differences in imaging / sound stage, or nothing that has yet stood out.

 

I have the house to myself, so I have tested how well it copes under load. This has also impressed me. I can turn it up all the way and it still sounds very clean, where previously in the last quarter turn it would usually start to become quite shouty and the bass would lose control (I never really ventured there). Obviously there are still limits, and the poor woofers can't cope at full volume with overly bassy tracks like Massive Attacks "Angel", heck.. even the X-fi + EQ seems to have trouble with tracks like that unless you lower the master volume. With sensible tracks though, it's very nice indeed.

 

I'm truly awful at describing sound, but I hope this has satisfied some curiosity.

[jammin]

Of course there are so many factors that can have an affect on how you hear a piece of music (and as you say, mood can be one of them).

 

I know over a reasonable amount of fiddling with different stuff (cables, isolation etc.), that things which logically shouldn't really make much difference can have a (subjective) noticeable affect on the sound.

 

It does sound like you might need to try some different speakers if you aren't getting the bass you would like from them. I'm a relative purist (not extreme) when it comes to sources, so I'd rather not fiddle with it to compensate for a perceived lacking on the equipment side. That's all money permitting though obviously, and you have to do the best you can with what you've got.