The Mini-A

Jump to Latest
#202
If I'm correct that's around 90mm x 60mm x 355mm. Compared to the size of this beautiful amplifier's heatsinks http://www.diyaudio.com/forums/showthread.php?postid=1496171#post1496171 you should be fine. My heatsink is 150mm^3 (a cube with 6" edge length) and houses both channels. The supply voltage is +/- 13V and after several hours of operation the heatsink's temperature is around 50°C (I can leave may hand on it for maybe ten seconds...). Do yourself a favour and try this amp, it's really sweet 😀

Dave
 
#204
Dave,

Thanks for your reply.
I've ordered the boards from Chipamp.com and hope I could make it.

The heatsinks are sitting around for a while and I'm going to order all the other parts soon.

The site you attached showed a interior pics and I saw couple components at the AC socket, I thought one is 10w resistor and the other is blue in color. Could you kindly tell me what are they? and what they are for?

My tube pre-amp with a big capacitor 4.7uf at the output. Is it safe to not putting a protector for the speakers. I have a feeling that not many people talking about that.

Thanks

Albert
 
#206
Which picture do you mean? If you refer to imix500's Mini A I think that is the soft start circuit he mentioned. Many people simply use a CL60 thermistor to connect the two primaries of the transformer to limit inrush current (see F4 Owner's manual / Service manual for a good example http://www.firstwatt.com/downloads.htm).

Speaker protection...I have never really considered using protective circuits...the preamps I've use so far were all AC-coupled (that is with a cap at the output) and at power-up I flip the switches approaching the speakers (like 1st CD, 2nd pre, last main amp) and at power-off away from the speakers (amp 1st, ...).

The Mini A has a little turn on thump and seemingly it gets bigger as the (optional) capacitor across D1 is increased. The turn off thump is a bit more hefty. I've read somewhere that this is due to asymmetric discharge of the PSU and that it could be tamed with the right choice of bleeder-resistors but I haven't done any experiments yet. My speakers are Fostex FE108EZ and they seem to be fine.

What circuit will you be building?

And be sure to check out this thread: http://www.diyaudio.com/forums/showthread.php?threadid=12923
 
#208
No, I think he means the grounding arrangement on my mini-A - thats the one linked in the post above.

What I've got is a star connection for the grounds (the copper bus between the loudspeaker terminals)which is then connected to electrical earth (at the IEC inlet- thin green wire) via a pair of diodes back to back 69 style, in parallel with a 15R 7W resistor and a low value cap, I think 0.01uF. Helps stop hum etc.

BTW, the heatsinks should be fine, mine are a good bit smaller than that and only get hand hot. IIRC my supply voltage is about 15V and the heatsinks get to ~50 deg C.

Any other questions about that build etc, fire away. Its a nice amp. Maybe not quite as sweet as a well built EL84 amp, but has a little more grunt too. I have this feeling that if you scaled it up a little (something your heatsinks will allow you to do) to 20 or 30WPC it might sound even better.

Fran
 
#209
If you want it to sound sweeter increase the bias. It need not be complicated--just lower the value of the Source resistors in the output stage. Given the low rail voltage of a stock Mini-A, you can run some pretty ferocious bias currents in the larger output cases (e.g TO-247) and still be able to manage the power dissipation fairly easily as long as you've got decent heatsinks.
A .22 Ohm Rsource will give you around 2.25A bias, which sounds intimidating until you realize that 15V rails lead to about 34W Pd...well within the DIY realm. Too hot? Use .33 or .39 Ohms.
Never underestimate the sonic benefits of higher bias. Just make sure you can manage the heat.

Grey
 
#211
Many thanks Grey for that suggestion. To be honest I bought some preassembled boards so never actually got into the biasing - I just put it all together! they are the brianGT boards, so I'm gonna pull them out over the weekend and have a look at the schematic and see which resistors you mean. The heatsinks get to about 45-50 degC as is, so from what I read they could stand another bit of bias anyway.

I'll find out whats in there now and report back...

Fran
 
#212
albertli said:
Grey,

What if I run this amp at +/- 21v rail, how big the heatsink I need for the 0.33 source resisitor?

Thanks

Albert

BTW I plan to mount just one pair of irfp240 per channel with the heatsink size 3 1/2" H, fin 2 1/2" D and 14"W
Click to expand...


I never try to estimate degrees/Watt for a heatsink based on dimensions people give over the web. There are just too many variables. If (as a lot of people do) the builder sits the heatsink directly on a table top, that inhibits air flow from beneath the heatsink and decreases efficiency. Note that I didn't say not to do it...I just said it decreases efficiency. I do it myself, but I take the decrease in convection current into account when I'm deciding what sort of heatsink to throw into the fray. Another factor that I can't control is when people use heatsinks with the fins running sideways because "that's the way Nelson does it." The same caveat applies--you have to derate the heatsink a bit. Your description omits the number of fins, which is a crucial factor in determining the surface area, which in turn determines how much heat the sink can dissipate. Too few fins and you don't add much surface area. Too many and you start to decrease the air flow and have to resort to using a fan. Etc.
Okay...so what do you do?
Well, after you've been at this for a while you can kinda estimate what you'd be able to do just by eyeballing the heatsink. Failing that--and something I still do if I think I'm going to run close to the limit--you clamp a flat-bottomed power resistor to the heatsink and run a known amount of power through it. Give it thirty minutes or an hour and stick a thermometer down between the fins, next to the main backplate. Better still, slide the tip of the thermometer into a screw hole. Don't found a religion on this method! Treat it as an estimate only! Some of the power will be lost through the body of the resistor and not through the heatsink. The footprint of the resistor will be different from a transistor. And other sundry et ceteras. However, it will give you a pretty good starting point.
Then the math part starts, so get out your calculator.
1) Assume about .5V across the Source resistor in a more-or-less normal Aleph (the Mini-A qualifies as 'normal'). This isn't going to be 100% accurate, but it's a good place to start.
2) .5V/.33 Ohms = 1.5A
3) 21V rails * 1.5A = 31.5W per output device
4) You say you're going to use a pair per heatsink, so that's 2 * 31.5W = 63W per heatsink
5) Decide how hot you will allow your heatsink to get. Nelson cooks on his. I'm more conservative in this area, preferring to be able to leave my hand on the heatsink indefinitely. I look for a maximum allowable temperature of about 110-115 degrees Fahrenheit. I also assume room temperature to be about 75 degrees. That means that my maximum allowable temperature rise is 115 degrees - 75 degrees = 40 degrees Fahrenheit.
6) 40 degrees/63W = .63 degrees/W
Compare this number to your heatsink estimate derived above. If your heatsink rating is less than .63. you're going to be okay. If it's above, then you're on shaky ground and you'll need to consider lowering the bias current, the rail voltage, or both to get the heat dissipation down.
When all else fails, think conservatively when it comes to heat.

Grey

P.S.: Don't have time to proof the post. If something's wonky, say so, and I'll fix it later.
 
#218
Could someone double check me and see if I'm correct:

the output source resistors to adjust for bias are R18 and 19 in the original grollins schematic (3W MO 0.47R) which are R27 and R28 on the brianGT mini-A boards?

Would I also need to change R20 and 21 in the grollins schema (R23 and 24 in brianGT board)?


Sorry for the newbie questions......:scratch:

Fran
 
#220
Could someone double check me and see if I'm correct:

the output source resistors to adjust for bias are R18 and 19 in the original grollins schematic (3W MO 0.47R) which are R27 and R28 on the brianGT mini-A boards?

Would I also need to change R20 and 21 in the grollins schema (R23 and 24 in brianGT board)?


Sorry for the newbie questions......

Fran
Click to expand...


Anyone? pretty please with a cherry on top!?
 
You must log in or register to reply here.
Top Bottom