| norb |
Hello everybody,
completed the electronics of my A75 a few days ago (i will build the chassis subsequently) and i am extremely happy with it already.
Many thanks to Mr. Pass and Mr. Thagard for an excellent project. No problems at all, the amp worked perfect and stable right from the beginning and surprised me by having no turn on/off-thump. Nice add-on.
My A75 is a dual mono with self-designed and made pcb´s. Changes are the TO-220 IRF630/9630 on the output stage, 0R82 source resistors and Q11 mounted on the output stage heatsink for I found that it improves bias current regulation.
The amp is set up unbalanced (have to build a preamp yet), 100% feedback, folded cascode with R21/22 = 100R, and sounds very clear, has a wide stage and a powerful and precise low end. My only direct comparisons are Onkyo A7 and H&K 655Vxi (sorry for that) and the difference is stunningly large. With any well recorded music, it is like wiping a (thick) layer of grey dust from a window you thought to be pretty translucent already. Speakers are Isophon Europa.
I will test some other setups over the next weeks and hear what happens.
One thing i tested so far is thermal coupling of the input-fets for better offset regulation, so far coupling of only the two N´s and P´s of the differential pairs has decreased offset regulation. I expect coupling of all input fets to yield better results, maybe somebody can share his experience on this topic.
Building this amplifier is a huge reward (also for your local electricity provider) and enjoyment. Thanks again.
Norbert
For the photo: lower pic shows one channel: top pcb is front end power supply, lower pcb is front end, blue wires are speaker out, the red and black wires with lugs go to the capacitor power rails mounted on top (upper pic showing complete channel in crude test setup), the black rectangular bits are the source resistors, the square ones the mosfets. |
|
|
| MEGA-amp |
| Looks good so far!! What voltages are you running on the front end and output stage? |
|
|
| norb |
After warming up the front end voltage levels out at +/- 51V; output stage is +/-41 V with 120mA bias. Heatsinks are 45°C with fans at 800 rpm, so the Mosfets shoud feel comfortable.
I may switch to toroids with a few volts less, because the standard trafos do hum a bit and i will see how noisy they are once the chassis is completed. Luckily the fans are nearly silent. |
|
|
| moe29 |
I always think "serious amp" when i see A75 builds :)
nice job. |
|
|
| jupiterjune |
| quote: | | completed the electronics of my A75 a few days ago (i will build the chassis subsequently) and i am extremely happy with it already |
Well, that could be a mistake--
because you won't be able to listen to it while you are putting the electonics into the chasis. I had mine in an open unfinished chasis for over a year before I could bring myself to take it out of service.
Very nice work, BTW!
JJ |
|
|
| Blues |
| quote: | Originally posted by jupiterjune
Well, that could be a mistake--
because you won't be able to listen to it while you are putting the electonics into the chasis. I had mine in an open unfinished chasis for over a year before I could bring myself to take it out of service.
Very nice work, BTW!
JJ |
Advice to the knowledgeable in electronics...build the chassis first. The sweet spot can be addicting! |
|
|
| jacco vermeulen |
| quote: | Originally posted by norb
TO-220 IRF630/9630 |
Cheapest MOSFET output devices around, pleased to see another 630/9630 A-75 constructor
| quote: | | the capacitor power |
EPCOS B-41580 ? |
|
|
| jupiterjune |
Jacco-
| quote: | | Cheapest MOSFET output devices around, pleased to see another 630/9630 A-75 constructor |
I used 634's/9630's --does that count?
Is the metric equivalent of a #4 screw any easier to tap? tapping 48 #4 blind holes was an exercise in patience!!!!!
Norb--what input mosfet's did you use?
No experience here with thermally coupling them. Keep us posted as to how that works!
JJ |
|
|
| jacco vermeulen |
Tapping regular UNC, Metric or oldy Witworth makes no difference, blind tapping always is a pita.
same goes for tapping fine wire , no matter whether it's UNF, Metric fine or BSF.
If they were cheap they count, i've yet to build a more affordable output stage than the A-75's.
I recall paying 20 times more for a comparable 75 watt Pd, TO-220 RCA RFP type MOSFET for another design. |
|
|
| Vix |
| quote: | Originally posted by norb
I may switch to toroids with a few volts less, because the standard trafos do hum a bit and i will see how noisy they are once the chassis is completed. Luckily the fans are nearly silent. |
Oh, yes. They HUM. In any "normal" application this is not an issue, but when you listen at low volumes late at night this can be quite irritating. I have built pass amps (Zen V9 and F2) with standard EI transformers, and their HUM is the only flaw. I am also looking to replace them by torroid ones.
A-75 is a good choice, deserves a nice chasis
:) |
|
|
| norb |
I´m currently designing the chassis with CAD to have 2 and 3 mm metal plates laser cut in final shape with all the mounting holes, so i hope to complete the chassis on one weekend.
There are 46 tappings overall per channel, but using a box column drill with frequency control it wasn´t that hard, just broke one tap drill.
Caps are Epcos 41560 63V/22000µF, input fets are the original 110/9110. I couldn´t get 230/9231 at a reasonable price, and after comparing different mosfet datasheets to the 230/9231 came out with the 630/9630. I got them from one production lot and especially the 630 are like clones, max. bias current difference is 5mA.
Serious amp is quite true, a screwdriver didn´t survive shorting out one capacitor bank accidentially. Hopefully the amp will never do this to my speakers.....
Does anybody have a recommendation for a preamp for balanced operation ? I´m not sure about using the BalZenPre because of the A75´s low input impedance on the minus input.
norb |
|
|
| jupiterjune |
| quote: | | Does anybody have a recommendation for a preamp for balanced operation ? I´m not sure about using the BalZenPre because of the A75´s low input impedance on the minus input. |
I think the impedence on the positive input is supposed to be set low (to match the negative side) via a switch when using balanced inputs. Of course, you a could also increase the resistance of the resistor going to ground on the inputs to say 10Kohms or so. This will also increase the gain of the amp.
I don't think the BalZenPre would have any problems with the 3000 ohm input impedence anyway (I think 3K is stock -- thats what mine has). I used a BZP one with mine, and I didn't see any significant roll-off at 20Khz.
Just make sure you spruce it up a bit w SUSY feedback and maybe a mosfet CCS, and cascoding if you want to really get crazy.
My BalZenPre was slightly more detailed than my McIntosh preamp, and both are very musical. (I will be re-building the BZP!) |
|
|
| jacco vermeulen |
| quote: | Originally posted by norb
especially the 630 are like clones |
That's because they are, just a different package.
You can expect to pay $4 for an IRF230 in the US, up to $12 each in the EC zone.
Most TO3s overhere are SemeLab manufacture.
Think i got a few hundred of both IRF630 and IRF9630 for the A-75 for 40 cents each, which ensures affordable parallel matching.
What's the story with the black source resistors ? |
|
|
| norb |
The source resis are 3W 1% SMD metal film type SMR from company Isabellenhuette. I soldered them directly from each source pin to a circuit board which serves as a heat dissipator for the resis and as amp output rail.
jupiterjune, what is SUSY feedback ? Good to hear that the BZP works with the A75. Do you use balanced connection? |
|
|
| jacco vermeulen |
| quote: | Originally posted by norb
SMR Isabellenhuette |
Very nice, expensive stuff.
I take it they're soldered directly on the center copper output rail for cooling them up to the 3 watt max level ? |
|
|
| jupiterjune |
| quote: | | (I think 3K is stock -- thats what mine has) |
Actually, mine has 475 ohms in series with the mosfet gate, and then 1500 ohms to ground, just like the original schematic. I had to go back and look at my notes to refresh my memory.
| quote: | | I think the impedence on the positive input is supposed to be set low (to match the negative side) via a switch when using balanced inputs |
The switch actually grounds the negative input, not what I said above.....
SUSY feedback as it is often referred to here is running feedback from the outputs of a differential pair back to the inputs, symmetrically. Mr. Pass holds a patent on SUSY feedback, and for a real explanation, go to passlabs.com, go to the "other" tab and "Articles"--lots of other good stuff there too.
http://www.passlabs.com/articles.htm
I used an unbalanced input (I don't have any balanced sources) and the balanced output. My A-75 would get noisy after being run a while, and it gets worse with higher bias current. Using the balanced BZP outputs got rid of most of the noise.
http://www.diyaudio.com/forums/show...&threadid=50267
This thread has a good schematic for the preamp. I first cascoded mosfets in mine, then tried cascodes j-fets about a week after Nelson published the article on cascoding. (I think it was called "Cascode amp design). I don't think I was able to detect any difference between the j-fet and the mosfet version. |
|
|
| norb |
You´re right Jacco, but as they only have to dissipate 15mW they are bit over the top, but i liked a compact design of the output stage, so i paid the price (1,20 Euro each)
JJ, thanks very much for the hint on the enhanced BOSOZ, its exactly what i was looking for. That answers all questions so far and leaves me with the next piece of equipment to build.
I think that the noise on my A75 is largely induced from the variable unbalanced output of my CD-player, for it´s just a pot connected internally to the fixed output. The noise is most with the pot in mid position, as then the output impedance is highest. But I will check my amp to see if it has similar noise behaviour. |
|
|
|
