Class D amp module build, silly power...

[Reid Malenfant]

Well as i posted in the Class T thread earlier i received the pair of silly power class D modules in kit form, the only thing that was kind of put together was all the small diodes (1N4148, 2 x 5v6 zener diodes & 1 x 15V zener) were pre fitted to the PCBs but not soldered. They do look alike & the text on each component is tiny - so i hope they got it right :eek:

As i really want to get to hear what these things actually sound like i proceeeded to start assembling them in short shift :eyebrows: These things are tiny, it just doesn't make a lot of sense in all honesty... Each one is more than capable of over 200W RMS with a 8 ohm load on it's own or 400W RMS with a 4 ohm load :eek: They just need a big PSU :eyebrows:

If bridged this equates to 800W RMS 8 ohm load, they won't like 4 ohm bridged as the power mosfet (two in one die) won't be able to handle the current... But i tell you this is still mad as you are about to see, i have seen bigger 10W amplifiers :D

Ok, here is one of the PCBs as received with fitted diodes.

http://i939.photobucket.com/albums/ad240/speakermark/Class%20D%20amp%20modules/PCBwithfitteddiodes-1.jpg

Thanks to MartinT for a rather sensible suggestion of a blob of glue on the bottom of the control IC, ok i didn't go with that but it gave me the idea of holding it in place with good old selotape :eyebrows: Cheers Martin 4018 I soldered the leads at both ends on this side & then removed the selotape to do the opposite pins... Then did one at a time to allow the poor thing to cool down, it never felt warm - good news :cool:

http://i939.photobucket.com/albums/ad240/speakermark/Class%20D%20amp%20modules/HoldingtheIC-1.jpg

Now you really have no idea how to gauge scale here so here is a pic of the control IC & my shaky soldering next to a £1 coin, yep these things are kind of small :stalks:

http://i939.photobucket.com/albums/ad240/speakermark/Class%20D%20amp%20modules/ICsoldered-1.jpg

Where i am now before i had a phone call & vistit from a friend, nearly all resistors fitted. The second PCB is in the same state, so it looks like i'll be carefully initially testing one or both amplifiers this coming weekend, well i'm hoping at least lol..

http://i939.photobucket.com/albums/ad240/speakermark/Class%20D%20amp%20modules/Mostresistorsfitted-1.jpg

I'm fairly sure i have a nice big fat toroidal transformer (from a Pioneer 120W RMS 8 ohm channel) that should be ok with these PCBs, though i'll check as the voltage will be at the top of the range & possibly marginal. Initial testing will be done with a much lower supply voltage & with resistors in both + & - rails to restrict current incase something is up. Once all has been established as well with both amps they can be connected to the big supply :eyebrows:

More coming shortly...

[MartinT]

Very nice so far, Mark. Keep us informed :)

[MartinT]

Very nice so far, Mark. Keep us informed :)

[John]

looks like fun...enjoy

[biggzy]

Looking great, will keep an eye on this thread, maybe in 10 years time i will be at this stage of doing DIY hi-fi, that board looks so small for the amount of power it will push out, what case you putting these into Mark?

please keep post updated with pics as this not only looks great/cool but also learns us noobs a thing or two.

Will be nice seeing the final working product :)

Adam.

[Reid Malenfant]

Looking great, will keep an eye on this thread, maybe in 10 years time i will be at this stage of doing DIY hi-fi, that board looks so small for the amount of power it will push out, what case you putting these into Mark?
I guess one that is big enough for the power supply :eyebrows: Honestly the amps are tiny, while i'm sure that the heatsinking is also adequate for general musical listening i'll be fitting a nice fat right angled alloy bracket to the power mosfet & attaching this to a reasonably sized heatsink ;)

You see what they don't tell you is that conduction losses increase with the mosfet die temperature as the resistance of the mosfet (RDS on) increases. Keeping it cool will yield more power & higher reliability at the end of the day :)

At a guess both amplifier modules & PSU could be fitted into a case a little bigger than one of your Teac amplifiers :eek: No kidding...

[biggzy]

I guess one that is big enough for the power supply :eyebrows: Honestly the amps are tiny, while i'm sure that the heatsinking is also adequate for general musical listening i'll be fitting a nice fat right angled alloy bracket to the power mosfet & attaching this to a reasonably sized heatsink ;)

You see what they don't tell you is that conduction losses increase with the mosfet die temperature as the resistance of the mosfet (RDS on) increases. Keeping it cool will yield more power & higher reliability at the end of the day :)

At a guess both amplifier modules & PSU could be fitted into a case a little bigger than one of your Teac amplifiers :eek: No kidding...

That sounds good, really can't wait to see the end product now, that sort of power in such a small box is mad, now i don't understand why most the hi-fi companies still produce 43cm wide units when we could have nice small units with huge power and great sound.
how long do these kits take to build mate?

Adam.

[Reid Malenfant]

About 6 hours if you know what you are doing to build a pair, maybe less.. It's the control IC that is the pain to solder as it's tiny as you can see compared to that £1 coin :eek:

This is the first time i have tackled them as well, given practise (that i'm not likely to gain) 4 hours might be realistic to build the modules.

No idea what they sound like, but i'm hopeful which is why i decided to buy them ;) I just couldn't be doing with 20 or 40W, i want realistic volume levels so these caught my eye, as well as that looking at the specification it looks like a realistic power level to given the possible supply voltages. Can't say the same for some of the stuff i have seen out there :D

[Mark Grant]

Looks fun and they must be really small looking at the size of the pound coin.

Another tip to hold components in place such as the chip is PVC insulation tape as it peels off easier than sellotape or try a blob of blutac, allows easy repositioning before soldering.

[Butuz]

Looks like a real fun project!!

I do have one q though - if you are using 120w transformer how can you expect this to supply 200w?

Butuz

[Reid Malenfant]

Looks like a real fun project!!

I do have one q though - if you are using 120w transformer how can you expect this to supply 200w?
I think i said i had a transformer from a Pioneer amplifier that was rated 120W RMS per channel ;) That was an 8 ohm rating & it'd nearly double into 4 ohms which would be about 200W RMS 4 ohms per channel.

Seeing as this is a switching amp i'd expect it to swing even more output voltage as there would be vastly less circuitry between the output & voltage rails :eyebrows:

The transformer is rated at well over 500VA by the looks of things so it's more than man enough for the job :)

[Techno Commander]

Looking forward to seeing how these progress. :)

[Reid Malenfant]

Looking forward to seeing how these progress. :)
Oh they are already totally together :eyebrows: I had them done by the next day, simples...

Just got to dig out a low voltage +/- PSU now to test the things along with a couple of 2W 47ohm or 22ohm resistors in the power rails to check for faults.. If they smoke Houston has a problem :(

I'll get some more pics up soon but i keep forgetting to charge the camera battery, i'll do that now you have reminded me Andy :) Thanks ;)

[Reid Malenfant]

Ok, here are the pictures of the amp kits put together :) I reckon that with the standard heatsinks & a 4ohm load they'll be good for about 120W RMS per channel with the quoted 90% efficiency. Maybe a tad more, the problem is though is that as the heatsink & Mosfet get hotter it become less efficient due to higher Mosfet conduction losses :lolsign:

What i'll do eventually is remove the standard crappy heatsink & attach a thick aluminium angle to a bigger heatsink. The standard dual mosfet package is a fully insulated plastic package which won't help things in removing heat in all honesty but i reckon that a maximum of 25W dissipation shouldn't be a problem, which should indicate about 250W RMS continuous per channel capability again at 90% efficiency.

Music isn't continuous though so in reality these could make an awful amount of power :eek:

Pics....

http://i939.photobucket.com/albums/ad240/speakermark/Class%20D%20amp%20modules/Image1-1.jpg

http://i939.photobucket.com/albums/ad240/speakermark/Class%20D%20amp%20modules/Image2-1.jpg

http://i939.photobucket.com/albums/ad240/speakermark/Class%20D%20amp%20modules/Image3-1.jpg

http://i939.photobucket.com/albums/ad240/speakermark/Class%20D%20amp%20modules/Image4-1.jpg

http://i939.photobucket.com/albums/ad240/speakermark/Class%20D%20amp%20modules/Image5-1.jpg

http://i939.photobucket.com/albums/ad240/speakermark/Class%20D%20amp%20modules/Image6-1.jpg


Just need to test the things now... Now where was the last place i saw the PSU & capacitors that i used to power my MTM speakers back in the 90s :scratch:

:lolsign:

Back soon i hope with test pictures on a sillyscope :eyebrows: Power measurements etc... Then a bigger PSU & real power :mental:

[biggzy]

Looking good Mark, i still cant get over how small they are for the power they pack, they look even smaller than i thought they were in the pic with your hand.

You one of these blokes with huge hands? :lolsign:

Adam.

[Reid Malenfant]

:D

Pretty average for someone of 6' 2" height i think :eyebrows: Yeah they are small, crazy in certain respects, but i guess the proof of the pudding is in the eating :)

We shall see asap :cool:

[John]

Nice work, hope its worth the effort

[Reid Malenfant]

Nice work, hope its worth the effort
Hi John, yes it'd be nice to say that they sound good & frankly there is no reason i can think of why they shouldn't :) However, if they aren't much cop i have a cunning plan (once more no turnips :eyebrows:)...

If they aren't that good then they can drive the twin 12" Yamaha woofers in each enclosure that also be taking the Tannoy 7" DC units i bought a while back :) After all it'll only be in the upper mid & treble that they have a possibility of being a bit ropey, below 200HZ they'll be as good as anything else out there but much more efficient.

Either way it's a win - win situation ;)

Cheers for popping in m8 :)

[Butuz]

They look really cool!

Did you use heatsink transfer compound between the MOSFET thingys and the heatsink?? If not - then buy some from a computer shop it will gie much increased heat transfer and should help the temperature problem.

Butuz

[Reid Malenfant]

Hi fella, yes i did ;) You only need a smidge though :eyebrows:

I have already mentioned i'll be disposing of the standard heatsinks as they are too small imo. They are also too thin so the heat won't be spread evenly :rolleyes:

I'll bolt the semiconductors to an 8mm thick aluminium bracket & attach that to a heatsink. In this instance it's worth keeping the dual Mosfet as cool as possible as it increases efficiency.

[biggzy]

i'll be disposing of the standard heatsinks as they are too small

Just stick a big fan in it like this guy did....

http://img828.imageshack.us/img828/3475/computerfan1.jpg (http://img828.imageshack.us/i/computerfan1.jpg/)

That will sort it ;)

(yes that is a PC in the pic not an amp, thats all i could find).

Adam.

[MartinT]

I was going to suggest the same thing, well not exactly! I have found that 12V fans run at 5V are whisper quiet and you could direct one at the existing heatsinks and I doubt you would ever need more cooling. Would keep them a tad more linear, too.

[Butuz]

Hi fella, yes i did ;) You only need a smidge though :eyebrows:

Darn it I thought I was going to be able to suggest something useful then. Foiled again!!! Good work boss!

I've got some water cooling kit lying around - that would do the job methinks :lolsign:

Butuz

[HighFidelityGuy]

I like the look of these amp modules Mark. I've been thinking about getting something like this to power my subs as their stock circuitry is a rather badly built. I'll watch this project with interest.:popcorn:

[Reid Malenfant]

Just stick a big fan in it like this guy did....
:lolsign: I know what you mean, i'd rather trust a bigger heatsink though in all honesty :) i reckon i must have over 100KG of heatsinks here so i'm sure i'll find something to suit :eyebrows:

I was going to suggest the same thing, well not exactly! I have found that 12V fans run at 5V are whisper quiet and you could direct one at the existing heatsinks and I doubt you would ever need more cooling. Would keep them a tad more linear, too.
True Martin, i'm just not keen on fans. Yes i know my Mackie has a couple in it, but that won't last forever as it's going to see convection cooled heatsinks sooner rather than later & no doubt a new case as well, or as a minimum a severely modified one ;)

Darn it I thought I was going to be able to suggest something useful then. Foiled again!!! Good work boss!
Well at least you tried chap, the fact that i used to work designing, building & testing SMPS may well not have been known to you :scratch: As for the water cooling i'd lump that in with fans as potentially less reliable so another no no :eyebrows:

I like the look of these amp modules Mark. I've been thinking about getting something like this to power my subs as their stock circuitry is a rather badly built. I'll watch this project with interest.:popcorn:
Well i think i have found a suitable PSU to use to initially test these modules. I couldn't find what i was looking for but i found a Fisher integrated amp of sufficiently low power output that i can splice into the PSU to test them - maybe tomorrow or Monday :cool:

They'd probably be very good as a sub amp as they should have a high damping factor at low frequencies. They'd also be good if someone fancied making themselves a mains regenerator ;) All that would be needed is a low distortion wein bridge oscillator & a transformer on the output :)

[Magna Audio]

Interesting - I've looked at some of those fairly big power D-classers but got a 20W (haha) T-amp to see what all the fuss and criticism was about.
The T amp really gives about 6w / ch but I am using efficient horns and if not turned up too loud it sounds reasonable - Not going to put the tube amp out of commission for serious vol / listening (they do sound quite different) but casual listening sounds pretty good. Very detailed midrange / upper midrange and decent bass if not turned up loud - loud it all falls apart with a rather irritating mids / highs character and bass sort of disappears.
I also found that impedance of preamps is important. Sounded terrible with an AudioNote pre feeding but ok with TVC. The latter better than using the onboard vol pot!

It will be very interesting to hear what you find with your amps.

What speakers are you going to drive with them?

[Reid Malenfant]

What speakers are you going to drive with them?
Hi Steve, well from what i have seen these things have an output filter tailored for a 4 ohm impedance ;) If running 8 ohms there appears to be a bit of a peak in the response at high frequencies due to the components used in the output filter (12Db/octave)...

So, i'll test them & then give them a blast driving a pair of Celestion Ditton 66 that i happen to have kicking about in the other front room :eyebrows: I know what these sound like having owned them for at least 6 years so it'll be a reasonable reference.

Tbh i did think that some of the class T jobbies may well be overated power wise, however these modules will accept a +/- 50V DC PSU which suggests at least a 100W RMS amp given an 8 ohm load ;)

We shall see, i have an oscilloscope to look at the output & an HP distortion analyzer just to make things more interesting :)

[Jonboy]

I also found that impedance of preamps is important. Sounded terrible with an AudioNote pre feeding but ok with TVC. The latter better than using the onboard vol pot!

It will be very interesting to hear what you find with your amps.

What speakers are you going to drive with them?

Interesting, what AN pre were you using?, i must agree about the vol pot though, the original seems a bit veiled

[Magna Audio]

The AN pre was the pre/phono M9. Apparently more suitable for matching a valve power amp.
I've been thinking of bypassing the vol pot on my mini-T. Wonder if it would make much diff compared to it just being on max...

[Magna Audio]

yeah the power quoted is probably 20W but you would not want to hear them all;-), they apparently clip badly and therefore the usable power is quite low.
My speakers are efficient (horn) so I prob only use <1w for loud vols realistically and I find it sounds more together on low vols.



Hi Steve, well from what i have seen these things have an output filter tailored for a 4 ohm impedance ;) If running 8 ohms there appears to be a bit of a peak in the response at high frequencies due to the components used in the output filter (12Db/octave)...

So, i'll test them & then give them a blast driving a pair of Celestion Ditton 66 that i happen to have kicking about in the other front room :eyebrows: I know what these sound like having owned them for at least 6 years so it'll be a reasonable reference.

Tbh i did think that some of the class T jobbies may well be overated power wise, however these modules will accept a +/- 50V DC PSU which suggests at least a 100W RMS amp given an 8 ohm load ;)

We shall see, i have an oscilloscope to look at the output & an HP distortion analyzer just to make things more interesting :)

[Reid Malenfant]

Right, i have established that both modules are 100% working :) I'd have prefered to test them on a current limiting PSU but i don't have one handy, that will have to be remedied ;)

I simply put a resistor (47ohm) in each power line except for the ground connection & connected it up to the PSU of the Fisher amplifier i mentioned earlier. No smoke, LEDs on each module lit up. Low 1mV DC offset with approximately 0.1mV of AC noise on each output. To test if they amplified i fed them a bit of good old AC hum by injecting it via a screwdriver & finger :eyebrows:

Not very scientific but it did the job :cool: Now to get a decent PSU rigged up & some real testing :exactly:

E2A:- I have purchased a 500VA 0-35V 0-35V toroidal transformer which should do adequately for the PSU. The result should be at least 120W RMS with an 8 ohm load & hopefully 200W RMS into 4 ohm per channel. I have a nice pair of 33,000uf 100V computer grade electrolytics & i know i'll have a 35A bridge here somewhere, so effectively the PSU just needs putting together now :)

[MartinT]

Low 1mV DC offset with approximately 0.1mV of AC noise on each output.

That's probably good CMRR, Mark. Have you looked at the output on the 'scope, I'd like to know whether you can see any switching noise? What's the frequency of operation?

Not in the same league, I know, but I'd like to do the same with the little Mini-T amp I bought for my computer. It certainly sounds good but I'd like to get the scope on it.

[Reid Malenfant]

I'll find all this out when they are connected to a decent PSU Martin ;) I just wanted to make sure they both worked before setting up a load of kit:-

Function generator / frequency counter
Oscilloscope
4 & 8 ohm test load
AC voltmeter
Digital temperature guage

What you need to remember is that any T amp is going to vary the switching frequency depending on the output power ;) This is entirely what class T is about, at low output power the switching frequency is likely to be 1 - 1.5MHz to decrease distortion. At higher powers they go to as more conventional 300 - 400KHz to increase efficiency as switching losses would be too high at 1.5MHz.

These as far as i'm aware are just plain old class D with a fixed frequency of switching, but they could well be free running self oscillating units in which case the switching frequency might well vary.. Whatever i'll find out & let you know.

[Reid Malenfant]

E2A:- I have purchased a 500VA 0-35V 0-35V toroidal transformer which should do adequately for the PSU. The result should be at least 120W RMS with an 8 ohm load & hopefully 200W RMS into 4 ohm per channel. I have a nice pair of 33,000uf 100V computer grade electrolytics & i know i'll have a 35A bridge here somewhere, so effectively the PSU just needs putting together now :)
& here are the rather large electrolytics i was refering to :) Everything is pretty small except for these, the toroidal transformer i bought isn't that big either. If i'm going to fit these in a case i think i'll turn them sideways so as to make things as small as possible :eyebrows:

http://i939.photobucket.com/albums/ad240/speakermark/Class%20D%20amp%20modules/Capacitors-1.jpg

http://i939.photobucket.com/albums/ad240/speakermark/Class%20D%20amp%20modules/Caps2-1.jpg

These are a bit overkill but that's the way i like it. They were removed from a 500W RMS per channel power amp, think i'll use some of the heatsinking from the same amp to give the switching mosfets on the modules an easier time..

[MartinT]

With the peak charging current using those huge caps, you're going to have to use some seriously rated fast rectifiers. Do you have them yet?

[Reid Malenfant]

Hundreds :eyebrows: I have more components than a lot of repair shops Martin, i kid you not ;) For testing i'll use a standard 35A bridge, but when i come to build a proper stereo amp i'll use MUR3040PT diodes, probably 8 rather than 4 as that will give a 60A rating.

I have also contacted an ex work colleague about some different 63V rated Capacitors, 100V is totally OTT for a +/- 50V DC supply & i might be better off saving them for something more appropriate.

[MartinT]

I used to buy those bargain bags of components back when I was a teenager and built up boxes of good stuff and experimented a lot with amps and other circuits. My favourite project was light-beam communications using converted torches and opto sensors (made from standard can transistors with the tops sawn off and filled with liquid plastic to form a lens). My mate and I could get it working across a good sized field. Many of my ideas came from Elektor, ETI and Wireless World.

All that has gone now, disposed of over a series of house moves.

[Reid Malenfant]

:lolsign: I know where you are coming from Martin ;)

Unfortunately due to circumstances beyond my control (:rolleyes:) it'll be a while before i get these properly tested. I had a bit of a fit one day last week when the 230V - 0 35V 0 35V 500VA transformer arrived & it turned out to be 230 - 0 45V 0 45V :steam: I honestly couldn't believe it as the outside of the box had the bloody secondary voltages on it :mental:

Anyway, having calmed down a bit i came to realise this could actually be advantageous given a bit of ingenuity :eyebrows:

What i'm going to do is put some incredibly low Rds ON power mosfets between the bridge rectifiers & capacitors so i can open the circuit at the desired voltage on the storage capacitors. This means i'll be able to switch between 200W RMS into an 8 ohm load at high voltage & 200W RMS into 4 ohms (maybe a tad more) at a lower capacitor voltage. The circuit is pretty much sussed out in my head so it's just a case of waiting for the power mosfets (http://cgi.ebay.co.uk/ws/eBayISAPI.dll?ViewItem&item=180633918018&ssPageName=STRK:MEWAX:IT) to arrive from the USA. I'll be paralleling two of them up to lower losses as the pulsed current into the capacitors will no doubt be a bit high :eyebrows:

Sorry for the delay, back to it ASAP :)