fets for aleph P1.7

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Formerly "jh6you". R.I.P.
Joined 2006
potepuh said:
Now I am a bit confused. Why not matched?


A1.7 has output coupling caps, i.e. the output nodes see only AC.
So . . . we do not need to worry about the DC offset voltages at the outputs.
No one would blaim your matching MOSFETs.
No one would warn you about your unmatching MOSFETs, either.
Either way you could go unpon your free choice.
 
Let me try to convey 2 cents. Someone can certainly straiten me out, if I'm confused. The diffamp/balanced topology needs a + and - side doing exactly the same thing to perform at it's best. The current sources will not provide the same current without the source R's and Vth and bipolars Vbe and even Beta being exactly the same. The bias level of the gain transistors, I beleive, is set in an area of the curve, that would also complement each other's opposite action with signal present. Not operating with the same current, side for side, will also effect their gain and balanced behaviour. Yes the DC output is protected by caps for the load but the more ideal both sides are, the better the performance!
 
Dependant of your layout and budget you can use quite a few caps in this possition I guess. I usually by my parts from partsconnexion.com in Canada. They are quick and easy to deal with and don't charge a lot in postage. If you want the best you could check out Jensen caps, BG N caps or if yuo can fit the caps outside the board you could go for foil caps - but these will cost some if you want the more exotic types. Some good brands are: Mundorf, Jensen, Audio Note and AuriCap.

I followed Damians advice and use the BG N( 2x10uF 50v and 2x47uF 50v for each board). Try this link: www.partsconnexion.com
 
Formerly "jh6you". R.I.P.
Joined 2006
flg said:
Let me try to convey 2 cents. Someone can certainly straiten me out, if I'm confused. The diffamp/balanced topology needs a + and - side doing exactly the same thing to perform at it's best. The current sources will not provide the same current without the source R's and Vth and bipolars Vbe and even Beta being exactly the same. The bias level of the gain transistors, I beleive, is set in an area of the curve, that would also complement each other's opposite action with signal present. Not operating with the same current, side for side, will also effect their gain and balanced behaviour. Yes the DC output is protected by caps for the load but the more ideal both sides are, the better the performance!



You are right word by word scientifically. :)
 
Babowana, Scientifically :whazzat: not me. Sometimes I'm better than others. I beleive the idea is more important. Scientifically, would have included math and such, rather than adjectives.
And, potepuh. I forgot. The question of how many Transistors to buy to get a few good pair? I ussually buy 20-30 and that should produce 2 or 3, maybe 4 "good enough" matched pairs. Non- RoHs complant parts are cheaper now. At least in the US... You should try to get at least 1% or better yet, .1% R's when you need a match. However, a pile of 20-30 5% R/s and meter will do a good job too. For low values, maybe below a 100 ohms, use a 4 wire measuremet if you have a meter that does that.:D
 
OK maybe this can be the first helpful post in reply :smash:

I bought 20 '610s and got 8 that were matched as close as my
meter could measure (10 mV) for the amp and another pair the
same for the power supply. I also bout 10 '9610s and got those
matched within 10 mV too.

So my amps are all matched, every part, channel to channel within
10 mV.

You dont' need to match the BJTs.



W.;)
 
If you don't need to match the BJTs. Why then did you say: "So my amps are all matched, every part, channel to channel within
10 mV." Please don't take me wrong...
I stand on my original statements. The Vbe of the BJT as it corresponds to the current in the Source R's, set's the stage current.
Every little bit of matching, will contribute to the performance of the circuit. Some areas more than others, I'm sure. But, expecting gain from these devices, the BJT included, means matching likely is more important there. :apathic: :apathic: :apathic:
 
sorry for side question, but within range of the thread... i picked up 610's that don't have the hole for sink mounting. how essential will sinking be? i have some basic clip ons, but will have to stick some silpad in there as they are close together on veterans board and the clipon would span 2 of them - not an aesthetically appealling option.
 
Don't feel sorry, sounds close enough to the topic to me. As is being discussed. Keeping the addjacent transistors (matched or not) close together will preserve the matching and temperature tracking. I beleive the original circuit has the gain devices burning about .4 Watts ea. I would say this is not absolutely into the "need" catagory. But it's close. However, the less they raise in temperature, the less difference you are likely to encounter.
There are several heatsink styles that use clips and don't need the hole... Or even a built in clip mechanism...
There are also non electrically cunductive but purposely thermally conductive epoxies which would allow you to glue them to the heatsink...
In my circuit, I have them back to back with the heatsink between them
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.