|16th February 2004, 04:39 PM||#1|
Another A-75 comes to life
After a few fits, escaping magic smoke, miswiring and other assorted trials mostly related to working on it late at night, my A-75 is now singing sweetly. It's amazing that there is so much detail without sounding harsh. Just sweet! It is as much better than my Leach amp as that was better than my Yamaha DSP-A1 (Still temproraily doing preamp duty)
I can't wait to try the AX (Thanks, Chad!) once I collect the parts and find the time.
Pics and more details once I am done with my fatherly duties - and figure out how to insert pictures.
Thanks to Nelson and Norm for sharing the design.
|16th February 2004, 08:49 PM||#2|
Join Date: Sep 2003
Location: Paris - France
Another A-75 comes to life, wouahhh !!! I want to see the pictures....
where are the pictures.. just kidding.
|16th February 2004, 11:18 PM||#3|
Warning: Long post, no pictures yet.
An A-75 Builderís Odyssey
I remember thinking that Iíd like to build the A-75 back when the article was first published Ė yep, dating myself. I started collecting parts a couple of years ago. I had started my Leach amps and stumbled onto the Pass Labs DIY page (when it was still part of the commercial page.) Apex Jr. had a couple of massive toroids, 2 x 42@11 amps. Hmm. 75 Watts ought to be plenty, but since these are going to do double duty as HT amps why not bump the rails up to get more power? I started collecting parts before I found DIYaudio,
I used IRFD210/9210 for the input, figuring that Iíd get a little safety factor with the higher voltage rating, and lower input capacitance wouldnít hurt, either. I bought two sets of Old Colony boards (front end & PS). I planned to use one as a line stage, but when it came time to buy parts, hey you probably need to buy a bunch of IRFP240/9240 to get two matched sets, so why not buy enough to build two stereo A-75s and biamp?
I bought three sets of those forced air heat sinks at Apex Jr. Used one for my 4 channel Leach amp, planned to use the other two for the A-75s. I built up the front end boards. There are two spots labeled R24 (one being R81), but nothing a bit of trace following couldnít help.
At the time I finished the front end boards I didnít have a scope or Variac. I fired it up first with a 12 volt transformer to verify basic operation, then used a 30 volt transformer to run it at +/- 50V. Boy, are those CCS trimpots sensitive! If I were to do it again, Iíd go for multi turn types. After setting the offset and bias, I swapped in 12 volt zeners for the 9 volt reference I the power supply to bring the rails up to +/-65V. I figured that 84 VCT would give me 55 volt rails, so I wanted the 10 volt margin of the original design.
The IRFD210ís seemed pretty hot. I expected that the dissipation would be up around Ĺ watt (they are rated at 1 watt) but they were too hot to touch. As a preamp, it sounded very nice. It was motorboating at first, but moving the power leads cured that. I read here that I should add local bypassing for the power supplies, but didnít do anything about it yet.. When Alain Dupont was looking to start his A-75 he wanted more power and I told him about my experience with hot transistors and my plan to solder copper tabs on the drains of the inputs to cool them.
The project stagnated while I considered my options for a case and output stage. Work kept me away from it, for a while, too. I started drilling and tapping the heat sinks for the output FETs. Broke a tap and found the thread on tapping. Then I found jrsunís case group buy and bought two Ė never did like the idea of a fan. It was a bit smaller than I had planned, 8Ēw x 14Ēd x 10Ēh, but I figured that I could squeeze everything in there.
After much tinkering with the transformers, caps and boards, I finally figured out how to make it all fit. Iíd planned to add a soft start internally but there wasnít room. The CL-60 takes care of the inrush, so it is not needed. I used 6 15,000 uf/100 Volt caps from Apex Jr. per channel, 45,000 uf per rail. They are 2Ēd x 3Ē high. So they have a level to themselves. I also added a small 60VCT transformer for the regulated supply, figuring that a doubler running off 84VCT would put the regulator dissipation too high and threaten the capacitors.
Along the way I purchased a 25 MHz scope, a 7 amp Powerstat and a drill press. Projects are an excuse to buy tools, right? When I got the first case assembled the scope came in handy. Without bypassing the power leads I was getting a nice 140 KHz oscillation that went away with a mere .1 uf across the power leads to ground. To play it safe I added the 200-uf caps I had handy, too. Oops, got one backwards. Mr. Powerstat allowed me to keep it from popping,
Without the oscillation, the input fets were just comfortably warm with 64.8 V on the rails. Since Iíd had some trouble adjusting the bias and offset with a 1KHz input tone (my computer tone generator has a DC tick) I didnít try to use a tone on this board. Iíd listened to the other one, but not this one.
I hooked up the first outputs, and everything went smoothly. I used 6 Fairchild IRFP240/9240 per channel, wired point to point. I was a little surprised at how slowly the bias came up when I first cranked up the bias pot. After some fiddling, I got it stable at 1.2 Amps total and the DC offset down to .010V. Next channel. No problem there either. The bias really climbed when I laid the top on the case, so I spent some time getting it down.
This transformer gives me 60 volt rails when loaded to 300 watts. From jrsunís data, I figured Iíd be able to run about 150 watts idle dissipation to get a case temperature around 50C. I measured 40C C at that level, so I can take it up later.
I got a test tone CD, so once the bias was stable I put a signal to it. Hmm Ė one channel nothing and the other an attenuated signal. The phone rang, I was on the phone for a bit and POP! The channel that had a signal blew most of the output transistors. The 10 amp line fuse was not just blown, it was plated onto the glass.
Well, the channel that had no signal had a cold /broken joint on the wire from the RCA jack to the XLR. After disconnecting the outputs I verified that the front end boards were still working. At least they didnít burn up. They didnít amplify Ė the signal out was about 75% of the level going in at the output terminal. At the drive terminals there was a 7 volt signal for .2 volts in. Have you guessed my mistake yet?
I chose to start with a jumper for R81 and take full feedback from the output stage. Unfortunately, I put the jumper where the 75K R27 feedback resistor should have gone and the 75K in R81ís place. OK, now we know that the amp doesnít like being driven at unity gain. (The input network drops the signal level slightly, hence the attenuation observed.
Saturday was relatively warm for Pennsylvania in January. So I went for a bicycle ride to clear my head. I went further than I had planned, and the temperature dropped to 40 during the ride. Feeling better, but tired I began rebuilding.
I used the other main board, verifying that it amplified and otherwise worked. After matching output devices, I had put them back in the tube in Vgs order. I picked six 240s and six 9240s matched within .011V. After installing, I began rebiasing. The undamaged channel biased up fine. The second seemed to come along, although one end of the heat sink seemed hotter than the other. The chassis is too tight to check all the source resistor drops, so I just measured at one end Ė the cold one. I wanted to hear this thing, and after half an hour the bias seemed stable so I hooked it up.
I heard lots of detail, but the right (funky heat distribution) channel seemed grainy. Maybe itís this CD, as I got up to put another CD on, POP! FZZT! And the magic smoke came out again. It was nearly midnight (I donít normally stay up that late), so I disassembled it far enough to notice that two FETs on the hot end had melted somewhat. I had stuck the Vgs labels above each FET for reference (ďin caseĒ I blow one ) Oops. Yeah, the last 2 digits were close, but the 3.484 and 3.581-3.593 arenít a really good match, especially with 0R22 source resistors. Current hogs had gone wild Ė enough thermal mass to keep them alive for a while, but not long.
I considered increasing the source resistors, but didnít. The next morning I managed to get replace the blown FETs and get it reassembled for biasing after verifying that the current sharing was reasonable. I set it for 150W dissipation per channel and buttoned it up. Oddly, after tightening down the top and rear panel (3mm aluminum) the heat sinks are significantly cooler. I am not sure if this is due to the panels helping the heat sinks or the now more stagnant air causing the bias to drop. Perhaps Iíll open it up and run leads out to measure the bias in the sealed state.
Well, it finally keeps the smoke inside, where it belongs. I am glad I bought enough output devices to allow me to get several well matched sets. Assuming I pay attention on my second one, Iíll have enough left over to do a pair of low power Alephs or my AX.
The amp now sits in my living room, itís blue LED glowing (a little too) brightly. Detail without harshness, bass with authority, a blacker background. Played a few Zappa CDís in honor of Mr. Pass, very nice. I might be moving to Montana soonÖ I donít have any Goldfrapp, and my local CD store doesnít either. I have gone through a variety of music, enjoying all the newly found nuances. It just sounds better than it did before. Offset is <.015V, no noise or turn on thump. There is a slight delay while the output devices turn on. I have a slight turn off DC drift. when I biamp there will be a capacitor or relay protecting my tweeters.
I get a little more than 35 W RMS in class A at 1.2 amps bias Ė if that is what I still have. Plenty of reserve power for HT use, Pirates of the Caribbean coming up soon.
Pictures will have to wait until I get them developed and scanned in. No decent digital camera, yet. Iíll shoot up the roll and develop it tomorrow.
|17th February 2004, 02:08 PM||#4|
Join Date: Apr 2002
"ICs and Transistors are working with smoke!
If the smoke coming out, the part will not work anymore...."
Great story, nice to see, that with which kind of problems you can face, when you do not have right equipment.
And nice to see, how well you managed!
|17th February 2004, 02:48 PM||#5|
Join Date: Sep 2003
Location: Paris - France
>>I built up the front end boards. There are two spots labeled R24 (one
>>being R81), but nothing a bit of trace following couldnít help.
It's a pb that took me a full day following all traces, and relabelling the 2 wrong resistors. Thanks for your warning!
>> Boy, are those CCS trimpots sensitive! If I were to do it again, Iíd go for
>>multi turn types.
It reminds me the same pb, too sensitive, even with 2 multimeters to check the 1.00volt's it's relatively tricky...
>>The IRFD210ís seemed pretty hot. It was motorboating at first, but moving >>the power leads cured that. I read here that I should add local bypassing for >>the power supplies, but didnít do anything about it yet.. When Alain Dupont >>was looking to start his A-75 he wanted more power and I told him about my >>experience with hot transistors and my plan to solder copper tabs on the drains >>of the inputs to cool them.
Well, at this time the input fets are hot, but not too much, I'll check the temp
>>Broke a tap and found the thread on tapping.
It's a DIYer karma to break taps...
>>Ė never did like the idea of a fan.
In the original design the heatsinks used couldn't work without a fan,
I'll have to replace mine it's noisy, even working with 7 volts out of 15
>>Along the way I purchased a 25 MHz scope, a 7 amp Powerstat and a drill >>press. Projects are an excuse to buy tools, right? When I got the first case >>assembled the scope came in handy. Without bypassing the power leads I was >>getting a nice 140 KHz oscillation that went away with a mere .1 uf across the >>power leads to ground.
Got a welleman scope, usefull and handy.
When you talk about oscillation... I cured before having problems, I installed
4 * 100 micro farad at the arrival's of +- 50 volts power supply's on the front end
>>Since Iíd had some trouble adjusting the bias and offset with a 1KHz input >>tone (my computer tone generator has a DC tick) I didnít try to use a tone on >>this board. Iíd listened to the other one, but not this one.
My old Macintosh gave me a clean 1 KHz with the Audio Toolbox shareware
so I didn't have to buy or build a function generator yet...
I also used the Audio Toolbox to make a CD with 1KHz 0 Db sine, 10 KHz 0 Db sine and a 10 KHz 0 Db square, got so little ringing with the square, and looked so good compared to Stereophile ou Absolute Sound very sophisticated Hi end stuff
>>I got a test tone CD, so once the bias was stable I put a signal to it. Hmm Ė >>one channel nothing and the other an attenuated signal. The phone rang, I >>was on the phone for a bit and POP! The channel that had a signal blew most >>of the output transistors. The 10 amp line fuse was not just blown, it was >>plated onto the glass.
I too, blew the first set of 6 amp fuses with only CRC's hooked on each rectifier bridge ; so I installed 10 amps, worked fine...
>>I chose to start with a jumper for R81 and take full feedback from the output >>stage. Unfortunately, I put the jumper where the 75K R27 feedback resistor >>should have gone and the 75K in R81ís place.
Didn't put any thing in R81 infinite, I will try later, I wanted to listed to the amp as soon as possible
>>I heard lots of detail, but the right (funky heat distribution) channel seemed >>grainy. Maybe itís this CD, as I got up to put another CD on, POP! FZZT! And >>the magic smoke came out again.
It took me 3 days to set all right, The CCS trimpots are too much sensitive
so are the 2 pots for the bias, the final approach was done checking the temperature on each heatsink, and I installed 2 micro bananas on each last Fet
to be able to measure the bias without opening the case
>>I considered increasing the source resistors, but didnít.
I used 0.47 Ohms for the sources, instead of the 1 Ohms.
>>Well, it finally keeps the smoke inside, where it belongs. I am glad I bought >>enough output devices to allow me to get several well matched sets.
Gladly I blew only one IRF 232, I think it was Mr static !
>>The amp now sits in my living room, itís blue LED glowing (a little too) brightly.
Mine too, Danielle told me about they are too bright, "could you do something about this!"; I will replace the 2 * 10 KOhms with 2*15 KOhms soon!!!
>>Detail without harshness, bass with authority, a blacker background. Played a >>few Zappa CDís in honor of Mr. Pass, very nice. I might be moving to Montana >>soonÖ I donít have any Goldfrapp, and my local CD store doesnít either.
I have no Zappa, and not any Goldfrapp ; I must buy some to listen to that music
; but I used ZZTop's 2 albums and some of my best HDCD's
Rackmaninov, and The last vinyl of the "Pink Floyd" it's damned good ...
>> Offset is <.015V, no noise or turn on thump. There is a slight delay while the >>output devices turn on. I have a slight turn off DC drift.
About the same, Offset is <.02 drifting during the warmup, but <.01 after
>>I get a little more than 35 W RMS in class A at 1.2 amps bias Ė if that is what I >>still have. Plenty of reserve power for HT use, Pirates of the Caribbean coming >>up soon.
Very good work. I have to tell that 75 watts is plenty, I only use 1/5 th for current
high listening and I eventualy unplugged the Subwoofer, there are so much punch out of a A75. I just keep the sub for Moovies!!
>>Pictures will have to wait until I get them developed and scanned in. No decent >>digital camera, yet. Iíll shoot up the roll and develop it tomorrow.
We are waiting !!!
|17th February 2004, 11:08 PM||#10|
|Thread Tools||Search this Thread|
|Thread||Thread Starter||Forum||Replies||Last Post|
|for the life of me||CarlyBoy||Tubes / Valves||5||30th January 2008 10:06 AM|
|The DIY Life||DJ Exprice||Everything Else||11||25th December 2006 07:38 PM|
|Bridges for life !||mpmarino||Parts||4||11th December 2004 11:11 PM|
|Tube Life||Glowin Plates||Tubes / Valves||14||6th October 2004 07:01 PM|
|Life After the Oil Crash||roibm||The Lounge||141||31st July 2004 08:43 PM|
|New To Site?||Need Help?|